From 4c35b61f37b09b0ddcdd145a66e8f63ec192dbbb Mon Sep 17 00:00:00 2001 From: Andrea Valassi Date: Mon, 2 Sep 2024 17:41:10 +0200 Subject: [PATCH] [goodhel] ** COMPLETE GOODHEL ** rerun 30 tmad tests on itscrd90 - all as expected (failures in heft #833) STARTED AT Mon Sep 2 01:36:00 PM CEST 2024 (SM tests) ENDED(1) AT Mon Sep 2 05:30:03 PM CEST 2024 [Status=0] (BSM tests) ENDED(1) AT Mon Sep 2 05:39:56 PM CEST 2024 [Status=0] 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_d_inl0_hrd0.txt 1 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_m_inl0_hrd0.txt --- .../log_eemumu_mad_d_inl0_hrd0.txt | 226 ++++++++-------- .../log_eemumu_mad_f_inl0_hrd0.txt | 234 ++++++++--------- .../log_eemumu_mad_m_inl0_hrd0.txt | 222 ++++++++-------- .../log_ggtt_mad_d_inl0_hrd0.txt | 140 +++++----- .../log_ggtt_mad_f_inl0_hrd0.txt | 222 ++++++++-------- .../log_ggtt_mad_m_inl0_hrd0.txt | 220 ++++++++-------- .../log_ggttg_mad_d_inl0_hrd0.txt | 232 ++++++++--------- .../log_ggttg_mad_f_inl0_hrd0.txt | 230 ++++++++--------- .../log_ggttg_mad_m_inl0_hrd0.txt | 230 ++++++++--------- .../log_ggttgg_mad_d_inl0_hrd0.txt | 234 ++++++++--------- .../log_ggttgg_mad_f_inl0_hrd0.txt | 234 ++++++++--------- .../log_ggttgg_mad_m_inl0_hrd0.txt | 232 ++++++++--------- .../log_ggttggg_mad_d_inl0_hrd0.txt | 228 ++++++++-------- .../log_ggttggg_mad_f_inl0_hrd0.txt | 224 ++++++++-------- .../log_ggttggg_mad_m_inl0_hrd0.txt | 222 ++++++++-------- .../log_gqttq_mad_d_inl0_hrd0.txt | 244 +++++++++--------- .../log_gqttq_mad_f_inl0_hrd0.txt | 240 ++++++++--------- .../log_gqttq_mad_m_inl0_hrd0.txt | 240 ++++++++--------- .../log_heftggbb_mad_d_inl0_hrd0.txt | 178 ++++++------- .../log_heftggbb_mad_f_inl0_hrd0.txt | 32 +-- .../log_heftggbb_mad_m_inl0_hrd0.txt | 184 ++++++------- .../log_smeftggtttt_mad_d_inl0_hrd0.txt | 216 ++++++++-------- .../log_smeftggtttt_mad_f_inl0_hrd0.txt | 224 ++++++++-------- .../log_smeftggtttt_mad_m_inl0_hrd0.txt | 220 ++++++++-------- .../log_susyggt1t1_mad_d_inl0_hrd0.txt | 226 ++++++++-------- .../log_susyggt1t1_mad_f_inl0_hrd0.txt | 222 ++++++++-------- .../log_susyggt1t1_mad_m_inl0_hrd0.txt | 226 ++++++++-------- .../log_susyggtt_mad_d_inl0_hrd0.txt | 202 +++++++-------- .../log_susyggtt_mad_f_inl0_hrd0.txt | 210 +++++++-------- .../log_susyggtt_mad_m_inl0_hrd0.txt | 216 ++++++++-------- 30 files changed, 3205 insertions(+), 3205 deletions(-) diff --git a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt index 01107f564b..04b494ee87 100644 --- a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt @@ -3,8 +3,8 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/e make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cppnone +make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:42:55 +DATE: 2024-09-02_13:36:45 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 [UNWEIGHT] Wrote 3798 events (found 8192 events) - [COUNTERS] PROGRAM TOTAL : 0.6950s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6868s - [COUNTERS] Fortran MEs ( 1 ) : 0.0082s for 8192 events => throughput is 1.00E+06 events/s + [COUNTERS] PROGRAM TOTAL : 0.6905s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6826s + [COUNTERS] Fortran MEs ( 1 ) : 0.0079s for 8192 events => throughput is 1.04E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1770s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1693s - [COUNTERS] Fortran MEs ( 1 ) : 0.0077s for 8192 events => throughput is 1.07E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1835s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1761s + [COUNTERS] Fortran MEs ( 1 ) : 0.0074s for 8192 events => throughput is 1.10E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000766E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3730s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2895s - [COUNTERS] Fortran MEs ( 1 ) : 0.0835s for 90112 events => throughput is 1.08E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3412s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2668s + [COUNTERS] Fortran MEs ( 1 ) : 0.0744s for 81920 events => throughput is 1.10E+06 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661545E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1777s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1702s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0072s for 8192 events => throughput is 1.14E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743180E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1798s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1725s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0070s for 8192 events => throughput is 1.17E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715404661545E-002) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218279743180E-002) differ by less than 3E-14 (2.220446049250313e-16) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000753E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3648s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2879s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0766s for 90112 events => throughput is 1.18E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3355s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2647s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0706s for 81920 events => throughput is 1.16E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602020000753E-002) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967943114212E-002) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.167196e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.150952e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.165900e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.182097e+06 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1752s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1704s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0045s for 8192 events => throughput is 1.83E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1779s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1734s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0043s for 8192 events => throughput is 1.90E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715404661532E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218279743152E-002) differ by less than 3E-14 (0.0) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000753E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3353s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2887s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0463s for 90112 events => throughput is 1.94E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09162 [9.1623967943114198E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3092s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2664s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0425s for 81920 events => throughput is 1.93E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602020000753E-002) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967943114198E-002) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.918558e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.923354e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.023579e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.976799e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1786s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1750s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0033s for 8192 events => throughput is 2.48E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743180E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1826s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1789s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0034s for 8192 events => throughput is 2.43E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715404661532E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218279743180E-002) differ by less than 3E-14 (2.220446049250313e-16) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000739E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3295s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2928s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0363s for 90112 events => throughput is 2.48E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.2975s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2659s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0313s for 81920 events => throughput is 2.62E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602020000739E-002) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967943114212E-002) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.640473e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.565409e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.831088e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.669213e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1752s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1718s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0031s for 8192 events => throughput is 2.65E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743180E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1790s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1756s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0031s for 8192 events => throughput is 2.66E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715404661532E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218279743180E-002) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000739E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3209s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2867s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0338s for 90112 events => throughput is 2.66E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.2978s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2667s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0308s for 81920 events => throughput is 2.66E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602020000739E-002) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967943114212E-002) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.678759e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.683870e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.813366e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.831997e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1736s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1692s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0040s for 8192 events => throughput is 2.04E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09212 [9.2123218279743180E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1772s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1728s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0041s for 8192 events => throughput is 1.98E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715404661532E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218279743180E-002) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000739E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3322s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2913s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0405s for 90112 events => throughput is 2.22E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3084s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2693s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0388s for 81920 events => throughput is 2.11E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602020000739E-002) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967943114212E-002) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.108602e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.175244e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.253882e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.169307e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.6096s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6084s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.32E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.6068s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6057s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.44E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cuda (9.3382715404661532E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (9.2123218279743152E-002) and cuda (9.2123218279743152E-002) differ by less than 3E-14 (0.0) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000753E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.7166s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7111s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0049s for 90112 events => throughput is 1.84E+07 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114198E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.7009s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6958s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0045s for 81920 events => throughput is 1.81E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cuda (9.1515602020000753E-002) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (9.1623967943114212E-002) and cuda (9.1623967943114198E-002) differ by less than 3E-14 (1.1102230246251565e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.377977e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.444883e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.939853e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.931438e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.088090e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.099519e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.478718e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.504269e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.243737e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.131221e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.989285e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.064642e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.238682e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.115637e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.131222e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.137806e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt index 617aae1ec8..febb8fc1cd 100644 --- a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt @@ -1,22 +1,22 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum - make USEBUILDDIR=1 BACKEND=cuda + make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 +make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 + +make USEBUILDDIR=1 BACKEND=cpp512y +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' - -make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:43:11 +DATE: 2024-09-02_13:37:01 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 [UNWEIGHT] Wrote 3798 events (found 8192 events) - [COUNTERS] PROGRAM TOTAL : 0.7259s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7175s - [COUNTERS] Fortran MEs ( 1 ) : 0.0084s for 8192 events => throughput is 9.72E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.6968s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6891s + [COUNTERS] Fortran MEs ( 1 ) : 0.0078s for 8192 events => throughput is 1.05E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1878s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1797s - [COUNTERS] Fortran MEs ( 1 ) : 0.0081s for 8192 events => throughput is 1.01E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1814s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1736s + [COUNTERS] Fortran MEs ( 1 ) : 0.0078s for 8192 events => throughput is 1.05E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000766E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3875s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3018s - [COUNTERS] Fortran MEs ( 1 ) : 0.0857s for 90112 events => throughput is 1.05E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3434s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2688s + [COUNTERS] Fortran MEs ( 1 ) : 0.0745s for 81920 events => throughput is 1.10E+06 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382703205998396E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1866s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1794s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0070s for 8192 events => throughput is 1.17E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09212 [9.2123206245609227E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1820s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1747s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0071s for 8192 events => throughput is 1.16E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382703205998396E-002) differ by less than 4E-4 (1.306308462512007e-07) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123206245609227E-002) differ by less than 4E-4 (1.3063084580711148e-07) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515590123565249E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3784s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3020s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0761s for 90112 events => throughput is 1.18E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09162 [9.1623955989899517E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3325s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2658s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0665s for 81920 events => throughput is 1.23E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515590123565249E-002) differ by less than 4E-4 (1.2999352305698153e-07) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623955989899517E-002) differ by less than 4E-4 (1.304594743345433e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.232262e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.202276e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.234403e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.223143e+06 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382700723828302E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1808s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1776s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.88E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09212 [9.2123203796917311E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1774s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1744s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0029s for 8192 events => throughput is 2.87E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382700723828302E-002) differ by less than 4E-4 (1.5721146218172777e-07) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123203796917311E-002) differ by less than 4E-4 (1.5721146207070547e-07) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515587612890761E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3276s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2977s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0297s for 90112 events => throughput is 3.03E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623953457373258E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.2940s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2676s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0263s for 81920 events => throughput is 3.11E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515587612890761E-002) differ by less than 4E-4 (1.5742791048545257e-07) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623953457373258E-002) differ by less than 4E-4 (1.5809990860926604e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.119755e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.033812e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.282267e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.149920e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382700679354239E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1827s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1799s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0025s for 8192 events => throughput is 3.27E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123203753043101E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1908s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1877s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0029s for 8192 events => throughput is 2.83E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382700679354239E-002) differ by less than 4E-4 (1.576877179942926e-07) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123203753043101E-002) differ by less than 4E-4 (1.576877178832703e-07) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515587619408464E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3317s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3038s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0276s for 90112 events => throughput is 3.26E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623953469147923E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3087s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2830s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0255s for 81920 events => throughput is 3.22E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515587619408464E-002) differ by less than 4E-4 (1.573566908996682e-07) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623953469147923E-002) differ by less than 4E-4 (1.5797139785167502e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.481016e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.377135e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.570800e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.424358e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382700679354239E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1855s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1828s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0024s for 8192 events => throughput is 3.35E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123203753043101E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1825s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1798s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0025s for 8192 events => throughput is 3.22E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382700679354239E-002) differ by less than 4E-4 (1.576877179942926e-07) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123203753043101E-002) differ by less than 4E-4 (1.576877178832703e-07) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515587619408464E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3314s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3041s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0270s for 90112 events => throughput is 3.33E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09162 [9.1623953469147923E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.2945s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2707s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0237s for 81920 events => throughput is 3.45E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515587619408464E-002) differ by less than 4E-4 (1.573566908996682e-07) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623953469147923E-002) differ by less than 4E-4 (1.5797139785167502e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.644439e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.491135e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.697078e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.632724e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382704335459282E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1845s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1814s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0027s for 8192 events => throughput is 3.04E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09212 [9.2123207359836520E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1805s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1775s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.96E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382704335459282E-002) differ by less than 4E-4 (1.1853587900123586e-07) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123207359836520E-002) differ by less than 4E-4 (1.1853587877919125e-07) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515591296252558E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3372s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3079s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0290s for 90112 events => throughput is 3.10E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.09162 [9.1623957131729453E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3032s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2782s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0248s for 81920 events => throughput is 3.30E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515591296252558E-002) differ by less than 4E-4 (1.1717945325173673e-07) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623957131729453E-002) differ by less than 4E-4 (1.1799734289663633e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.387501e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.243117e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.616268e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.622259e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382706077425631E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.6084s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6073s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.48E+07 events/s + [XSECTION] Cross section = 0.09212 [9.2123209078308149E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.5992s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5982s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0005s for 8192 events => throughput is 1.54E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cuda (9.3382706077425631E-002) differ by less than 4E-4 (9.988182347875352e-08) +OK! xsec from fortran (9.2123218279743152E-002) and cuda (9.2123209078308149E-002) differ by less than 4E-4 (9.988182325670891e-08) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515592892887687E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.7292s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7238s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0048s for 90112 events => throughput is 1.86E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.09162 [9.1623958748796724E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.6932s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6884s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0043s for 81920 events => throughput is 1.92E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cuda (9.1515592892887687E-002) differ by less than 4E-4 (9.973286385633884e-08) +OK! xsec from fortran (9.1623967943114212E-002) and cuda (9.1623958748796724E-002) differ by less than 4E-4 (1.0034838804351409e-07) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.601368e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.784013e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.718163e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.692965e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.633474e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.612656e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.898384e+09 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.916341e+09 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.829286e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.651764e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.104797e+09 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.098672e+09 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.012752e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.923712e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.802072e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.800108e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt index e51bbf394d..383b979bbf 100644 --- a/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt @@ -3,8 +3,8 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/e make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cppnone +make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:43:26 +DATE: 2024-09-02_13:37:18 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 [UNWEIGHT] Wrote 3798 events (found 8192 events) - [COUNTERS] PROGRAM TOTAL : 0.6983s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6906s - [COUNTERS] Fortran MEs ( 1 ) : 0.0077s for 8192 events => throughput is 1.06E+06 events/s + [COUNTERS] PROGRAM TOTAL : 0.7114s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7036s + [COUNTERS] Fortran MEs ( 1 ) : 0.0078s for 8192 events => throughput is 1.06E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715404661532E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1791s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1711s - [COUNTERS] Fortran MEs ( 1 ) : 0.0080s for 8192 events => throughput is 1.02E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218279743152E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1799s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1721s + [COUNTERS] Fortran MEs ( 1 ) : 0.0077s for 8192 events => throughput is 1.06E+06 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_eemumu_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602020000766E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3694s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2869s - [COUNTERS] Fortran MEs ( 1 ) : 0.0825s for 90112 events => throughput is 1.09E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967943114212E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3431s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2683s + [COUNTERS] Fortran MEs ( 1 ) : 0.0748s for 81920 events => throughput is 1.10E+06 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715420701395E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1846s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1767s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0075s for 8192 events => throughput is 1.09E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218295566675E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1811s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1734s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0074s for 8192 events => throughput is 1.10E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715420701395E-002) differ by less than 2E-4 (1.7176482458580722e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218295566675E-002) differ by less than 2E-4 (1.7176482458580722e-10) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602033080859E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3660s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2865s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0792s for 90112 events => throughput is 1.14E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09162 [9.1623967959435296E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3395s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2663s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0729s for 81920 events => throughput is 1.12E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602033080859E-002) differ by less than 2E-4 (1.4292744765498355e-10) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967959435296E-002) differ by less than 2E-4 (1.781310654536128e-10) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.124575e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.139482e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.154252e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.145809e+06 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715420701354E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1757s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1709s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0044s for 8192 events => throughput is 1.88E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218295566661E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1767s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1721s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0043s for 8192 events => throughput is 1.89E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715420701354E-002) differ by less than 2E-4 (1.7176438049659737e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218295566661E-002) differ by less than 2E-4 (1.717646025412023e-10) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602033080859E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3336s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2878s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0455s for 90112 events => throughput is 1.98E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967959435282E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3070s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2659s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0409s for 81920 events => throughput is 2.00E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602033080859E-002) differ by less than 2E-4 (1.4292744765498355e-10) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967959435282E-002) differ by less than 2E-4 (1.781310654536128e-10) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.982594e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.001112e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.052848e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.050890e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715383664494E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1749s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1711s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0035s for 8192 events => throughput is 2.31E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218259029332E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1793s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1758s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0032s for 8192 events => throughput is 2.55E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715383664494E-002) differ by less than 2E-4 (2.2484925032983938e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218259029332E-002) differ by less than 2E-4 (2.2484902828523445e-10) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602022697845E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3282s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2920s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0358s for 90112 events => throughput is 2.51E+06 events/s + [XSECTION] Cross section = 0.09162 [9.1623967950679383E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.2992s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2669s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0321s for 81920 events => throughput is 2.55E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602022697845E-002) differ by less than 2E-4 (2.947131427788463e-11) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967950679383E-002) differ by less than 2E-4 (8.256750838597782e-11) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.552156e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.562076e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.649390e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.726106e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715383664494E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1744s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1708s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0033s for 8192 events => throughput is 2.47E+06 events/s + [XSECTION] Cross section = 0.09212 [9.2123218259029332E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1754s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1721s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0030s for 8192 events => throughput is 2.70E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715383664494E-002) differ by less than 2E-4 (2.2484925032983938e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218259029332E-002) differ by less than 2E-4 (2.2484902828523445e-10) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602022697845E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3217s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2876s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0337s for 90112 events => throughput is 2.67E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09162 [9.1623967950679383E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3001s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2683s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0316s for 81920 events => throughput is 2.60E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602022697845E-002) differ by less than 2E-4 (2.947131427788463e-11) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967950679383E-002) differ by less than 2E-4 (8.256750838597782e-11) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.650509e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.673365e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.719714e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.813324e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715383664494E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.1750s - [COUNTERS] Fortran Overhead ( 0 ) : 0.1712s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0035s for 8192 events => throughput is 2.36E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09212 [9.2123218259029332E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.1773s + [COUNTERS] Fortran Overhead ( 0 ) : 0.1731s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0039s for 8192 events => throughput is 2.08E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cpp (9.3382715383664494E-002) differ by less than 2E-4 (2.2484925032983938e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cpp (9.2123218259029332E-002) differ by less than 2E-4 (2.2484902828523445e-10) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_eemumu_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602022697845E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.3264s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2866s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0395s for 90112 events => throughput is 2.28E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.09162 [9.1623967950679383E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.3096s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2725s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0368s for 81920 events => throughput is 2.23E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cpp (9.1515602022697845E-002) differ by less than 2E-4 (2.947131427788463e-11) +OK! xsec from fortran (9.1623967943114212E-002) and cpp (9.1623967950679383E-002) differ by less than 2E-4 (8.256750838597782e-11) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.207219e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.134126e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.300574e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.274050e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09338 [9.3382715392009194E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1591 events (found 1595 events) - [COUNTERS] PROGRAM TOTAL : 0.5992s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5980s + [XSECTION] Cross section = 0.09212 [9.2123218267261497E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1608 events (found 1612 events) + [COUNTERS] PROGRAM TOTAL : 0.5994s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5983s [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.38E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.3382715404661532E-002) and cuda (9.3382715392009194E-002) differ by less than 2E-4 (1.3548906441229747e-10) +OK! xsec from fortran (9.2123218279743152E-002) and cuda (9.2123218267261497E-002) differ by less than 2E-4 (1.3548873134539008e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_eemumu_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.09152 [9.1515602021089631E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1782 events (found 1787 events) - [COUNTERS] PROGRAM TOTAL : 0.7158s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7101s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0050s for 90112 events => throughput is 1.80E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.09162 [9.1623967944572490E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 1656 events (found 1661 events) + [COUNTERS] PROGRAM TOTAL : 0.7017s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6967s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0045s for 81920 events => throughput is 1.84E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (9.1515602020000766E-002) and cuda (9.1515602021089631E-002) differ by less than 2E-4 (1.1898038110302878e-11) +OK! xsec from fortran (9.1623967943114212E-002) and cuda (9.1623967944572490E-002) differ by less than 2E-4 (1.591593523642132e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.054665e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.369377e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.970842e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.920435e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.242307e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.096525e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.491734e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.464853e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.221256e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.114601e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.104459e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.089166e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.208981e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.119279e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_EPEM_MUPMUM_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.160987e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.165995e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt index a7e32898b2..6cd61fa607 100644 --- a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx -make USEBUILDDIR=1 BACKEND=cuda - +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone + make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-09-02_11:54:30 +DATE: 2024-09-02_13:37:34 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx @@ -58,9 +58,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 2601 events (found 5405 events) - [COUNTERS] PROGRAM TOTAL : 0.8124s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7696s - [COUNTERS] Fortran MEs ( 1 ) : 0.0427s for 8192 events => throughput is 1.92E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.7891s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7477s + [COUNTERS] Fortran MEs ( 1 ) : 0.0415s for 8192 events => throughput is 1.98E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -83,9 +83,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4153s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3732s - [COUNTERS] Fortran MEs ( 1 ) : 0.0421s for 8192 events => throughput is 1.95E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4032s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3623s + [COUNTERS] Fortran MEs ( 1 ) : 0.0409s for 8192 events => throughput is 2.00E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -108,9 +108,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x10_fortran > /tmp/ava [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694841] fbridge_mode=0 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.6227s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1956s - [COUNTERS] Fortran MEs ( 1 ) : 0.4272s for 81920 events => throughput is 1.92E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.5704s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1536s + [COUNTERS] Fortran MEs ( 1 ) : 0.4168s for 81920 events => throughput is 1.97E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -133,9 +133,9 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4170s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3716s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0450s for 8192 events => throughput is 1.82E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4014s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3570s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0440s for 8192 events => throughput is 1.86E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -167,9 +167,9 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694841] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.6246s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1721s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4522s for 81920 events => throughput is 1.81E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.5730s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1324s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4403s for 81920 events => throughput is 1.86E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.847939e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.897948e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.849424e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.875264e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -211,10 +211,10 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3978s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3720s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0255s for 8192 events => throughput is 3.21E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3843s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3595s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0245s for 8192 events => throughput is 3.35E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -245,10 +245,10 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694841] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.4441s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1853s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2583s for 81920 events => throughput is 3.17E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 1.3745s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1270s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2471s for 81920 events => throughput is 3.31E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.236325e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.249563e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.295969e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.346265e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -289,9 +289,9 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3860s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3698s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0158s for 8192 events => throughput is 5.19E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3745s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3584s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0157s for 8192 events => throughput is 5.21E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -323,9 +323,9 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694870] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.3276s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1689s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1583s for 81920 events => throughput is 5.17E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.2784s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1253s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1528s for 81920 events => throughput is 5.36E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.286779e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.318391e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.280807e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.180789e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -367,10 +367,10 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3938s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3788s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0146s for 8192 events => throughput is 5.63E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3726s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3584s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0138s for 8192 events => throughput is 5.94E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -401,9 +401,9 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694870] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.3223s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1742s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1477s for 81920 events => throughput is 5.55E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.2698s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1284s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1411s for 81920 events => throughput is 5.81E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.813418e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.777377e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.836888e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.827458e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -445,9 +445,9 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3965s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3734s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0227s for 8192 events => throughput is 3.61E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3808s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3584s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0219s for 8192 events => throughput is 3.73E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -479,9 +479,9 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694870] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.4061s - [COUNTERS] Fortran Overhead ( 0 ) : 1.1760s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2298s for 81920 events => throughput is 3.57E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.3488s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1291s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2193s for 81920 events => throughput is 3.74E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.602524e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.693600e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.588687e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.602777e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -523,10 +523,10 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.06 [47.059717237420635] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8074s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8059s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.22E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [COUNTERS] PROGRAM TOTAL : 0.7923s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7910s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.28E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** @@ -557,9 +557,9 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] ChannelId = 1 [XSECTION] Cross section = 47.13 [47.127145120694848] fbridge_mode=1 [UNWEIGHT] Wrote 1595 events (found 1600 events) - [COUNTERS] PROGRAM TOTAL : 1.6076s - [COUNTERS] Fortran Overhead ( 0 ) : 1.6004s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0065s for 81920 events => throughput is 1.27E+07 events/s + [COUNTERS] PROGRAM TOTAL : 1.5670s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5604s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0059s for 81920 events => throughput is 1.40E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.857596e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.161438e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.662816e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.700686e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.030840e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.283924e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.084214e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.082297e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.019958e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.248997e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.155673e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.161485e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.017100e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.222301e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.050231e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.062951e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt index 420861126b..f5b9fa7963 100644 --- a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt @@ -2,12 +2,12 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/g make USEBUILDDIR=1 BACKEND=cuda - make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 + +make USEBUILDDIR=1 BACKEND=cppsse4 +make USEBUILDDIR=1 BACKEND=cppavx2 make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:44:09 +DATE: 2024-09-02_13:37:58 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184803756640] fbridge_mode=0 + [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 2601 events (found 5405 events) - [COUNTERS] PROGRAM TOTAL : 0.8019s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7604s - [COUNTERS] Fortran MEs ( 1 ) : 0.0415s for 8192 events => throughput is 1.97E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.7917s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7503s + [COUNTERS] Fortran MEs ( 1 ) : 0.0414s for 8192 events => throughput is 1.98E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184803756640] fbridge_mode=0 + [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4215s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3800s - [COUNTERS] Fortran MEs ( 1 ) : 0.0416s for 8192 events => throughput is 1.97E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4060s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3640s + [COUNTERS] Fortran MEs ( 1 ) : 0.0420s for 8192 events => throughput is 1.95E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x10_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105695279989114] fbridge_mode=0 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7567s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3058s - [COUNTERS] Fortran MEs ( 1 ) : 0.4510s for 90112 events => throughput is 2.00E+05 events/s + [XSECTION] Cross section = 47.13 [47.127145120694841] fbridge_mode=0 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.5820s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1651s + [COUNTERS] Fortran MEs ( 1 ) : 0.4168s for 81920 events => throughput is 1.97E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094179692708323] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059712130113020] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4203s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3790s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0410s for 8192 events => throughput is 2.00E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.4024s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3614s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0408s for 8192 events => throughput is 2.01E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094179692708323) differ by less than 4E-4 (1.0852822573959031e-07) +OK! xsec from fortran (47.059717237420642) and cpp (47.059712130113020) differ by less than 4E-4 (1.0852822585061261e-07) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105688388783328] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7678s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3093s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4582s for 90112 events => throughput is 1.97E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 47.13 [47.127138118246172] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.5380s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1251s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4127s for 81920 events => throughput is 1.99E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105688388783328) differ by less than 4E-4 (1.462924120732012e-07) +OK! xsec from fortran (47.127145120694841) and cpp (47.127138118246172) differ by less than 4E-4 (1.4858631158798374e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.984608e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.975953e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.996032e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.961561e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094175707109216] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059708147430918] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3923s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3751s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0170s for 8192 events => throughput is 4.83E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [COUNTERS] PROGRAM TOTAL : 0.3806s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3632s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0172s for 8192 events => throughput is 4.77E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094175707109216) differ by less than 4E-4 (1.9315861321533845e-07) +OK! xsec from fortran (47.059717237420642) and cpp (47.059708147430918) differ by less than 4E-4 (1.9315861332636075e-07) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105684583433771] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.4893s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3053s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1837s for 90112 events => throughput is 4.90E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 47.13 [47.127134290549655] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2940s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1276s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1661s for 81920 events => throughput is 4.93E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105684583433771) differ by less than 4E-4 (2.2707562807866566e-07) +OK! xsec from fortran (47.127145120694841) and cpp (47.127134290549655) differ by less than 4E-4 (2.2980694369678645e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.831484e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.704918e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.765454e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.777305e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094173726920275] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059706168691257] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3873s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3779s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0091s for 8192 events => throughput is 8.99E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3700s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3604s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0094s for 8192 events => throughput is 8.75E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094173726920275) differ by less than 4E-4 (2.3520603253945893e-07) +OK! xsec from fortran (47.059717237420642) and cpp (47.059706168691257) differ by less than 4E-4 (2.3520603253945893e-07) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105684037363524] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.4091s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3116s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0972s for 90112 events => throughput is 9.27E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 47.13 [47.127133760638543] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2154s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1282s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0870s for 81920 events => throughput is 9.42E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105684037363524) differ by less than 4E-4 (2.386680745258829e-07) +OK! xsec from fortran (47.127145120694841) and cpp (47.127133760638543) differ by less than 4E-4 (2.4105122997664097e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.995090e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.070716e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.148417e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.376553e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094173726920275] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059706168691257] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3894s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3807s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0085s for 8192 events => throughput is 9.68E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [COUNTERS] PROGRAM TOTAL : 0.3710s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3623s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0085s for 8192 events => throughput is 9.69E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094173726920275) differ by less than 4E-4 (2.3520603253945893e-07) +OK! xsec from fortran (47.059717237420642) and cpp (47.059706168691257) differ by less than 4E-4 (2.3520603253945893e-07) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105684037363524] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.3961s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3040s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0917s for 90112 events => throughput is 9.82E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 47.13 [47.127133760638543] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2292s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1447s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0844s for 81920 events => throughput is 9.71E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105684037363524) differ by less than 4E-4 (2.386680745258829e-07) +OK! xsec from fortran (47.127145120694841) and cpp (47.127133760638543) differ by less than 4E-4 (2.4105122997664097e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.994646e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.710053e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.882184e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.970031e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094178448427996] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059710886743353] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3945s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3828s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0112s for 8192 events => throughput is 7.31E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3748s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3619s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0126s for 8192 events => throughput is 6.51E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094178448427996) differ by less than 4E-4 (1.3494932904478674e-07) +OK! xsec from fortran (47.059717237420642) and cpp (47.059710886743353) differ by less than 4E-4 (1.3494932948887595e-07) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105688391432061] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.5017s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3657s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1356s for 90112 events => throughput is 6.64E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 47.13 [47.127138112830536] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2452s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1291s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1159s for 81920 events => throughput is 7.07E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105688391432061) differ by less than 4E-4 (1.462361824966507e-07) +OK! xsec from fortran (47.127145120694841) and cpp (47.127138112830536) differ by less than 4E-4 (1.487012269985044e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.837763e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.796358e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.925566e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.945516e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184162782994] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059716596916104] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8112s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8099s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.43E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [COUNTERS] PROGRAM TOTAL : 0.7878s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7866s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.47E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cuda (47.094184162782994) differ by less than 4E-4 (1.3610462645807786e-08) +OK! xsec from fortran (47.059717237420642) and cuda (47.059716596916104) differ by less than 4E-4 (1.3610462978874693e-08) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105694501043516] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7829s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7765s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0057s for 90112 events => throughput is 1.58E+07 events/s + [XSECTION] Cross section = 47.13 [47.127144127498006] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.5697s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5639s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0052s for 81920 events => throughput is 1.59E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cuda (47.105694501043516) differ by less than 4E-4 (1.6536123581545326e-08) +OK! xsec from fortran (47.127145120694841) and cuda (47.127144127498006) differ by less than 4E-4 (2.1074835543366532e-08) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.085941e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.257477e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.178660e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.170617e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.983696e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.984579e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.406286e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.467027e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.010543e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.930796e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.536473e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.529973e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.527299e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.472386e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.475317e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.480246e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt index 65f004f30e..eefec1e0a9 100644 --- a/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt @@ -1,7 +1,7 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx -make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:44:34 +DATE: 2024-09-02_13:38:23 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184803756640] fbridge_mode=0 + [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 2601 events (found 5405 events) - [COUNTERS] PROGRAM TOTAL : 0.8115s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7704s - [COUNTERS] Fortran MEs ( 1 ) : 0.0411s for 8192 events => throughput is 1.99E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.8029s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7616s + [COUNTERS] Fortran MEs ( 1 ) : 0.0413s for 8192 events => throughput is 1.98E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x1_fortran > /tmp/aval [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184803756640] fbridge_mode=0 + [XSECTION] Cross section = 47.06 [47.059717237420642] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4214s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3805s - [COUNTERS] Fortran MEs ( 1 ) : 0.0408s for 8192 events => throughput is 2.01E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4022s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3598s + [COUNTERS] Fortran MEs ( 1 ) : 0.0425s for 8192 events => throughput is 1.93E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggtt_x10_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105695279989114] fbridge_mode=0 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7670s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3128s - [COUNTERS] Fortran MEs ( 1 ) : 0.4542s for 90112 events => throughput is 1.98E+05 events/s + [XSECTION] Cross section = 47.13 [47.127145120694841] fbridge_mode=0 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.5642s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1527s + [COUNTERS] Fortran MEs ( 1 ) : 0.4114s for 81920 events => throughput is 1.99E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094186141863901] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059718574548555] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4222s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3775s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0442s for 8192 events => throughput is 1.85E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.4211s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3741s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0466s for 8192 events => throughput is 1.76E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094186141863901) differ by less than 2E-4 (2.8413428720952538e-08) +OK! xsec from fortran (47.059717237420642) and cpp (47.059718574548555) differ by less than 2E-4 (2.8413428498907933e-08) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105696630006634] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7889s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3008s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4877s for 90112 events => throughput is 1.85E+05 events/s + [XSECTION] Cross section = 47.13 [47.127146469161161] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.6234s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1651s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4579s for 81920 events => throughput is 1.79E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105696630006634) differ by less than 2E-4 (2.865932691165085e-08) +OK! xsec from fortran (47.127145120694841) and cpp (47.127146469161161) differ by less than 2E-4 (2.8613367453189653e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.863098e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.850969e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.876650e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.846303e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094186141863901] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059718574548555] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4042s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3795s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0242s for 8192 events => throughput is 3.38E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3816s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3573s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0240s for 8192 events => throughput is 3.42E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094186141863901) differ by less than 2E-4 (2.8413428720952538e-08) +OK! xsec from fortran (47.059717237420642) and cpp (47.059718574548555) differ by less than 2E-4 (2.8413428498907933e-08) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105696630006626] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.5750s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3065s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2680s for 90112 events => throughput is 3.36E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 47.13 [47.127146469161168] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.3673s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1259s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2411s for 81920 events => throughput is 3.40E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105696630006626) differ by less than 2E-4 (2.8659326689606246e-08) +OK! xsec from fortran (47.127145120694841) and cpp (47.127146469161168) differ by less than 2E-4 (2.8613367675234258e-08) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.334875e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.356677e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.372227e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.357238e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094186169585456] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059718602249824] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3946s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3794s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0147s for 8192 events => throughput is 5.56E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3738s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3585s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0149s for 8192 events => throughput is 5.49E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094186169585456) differ by less than 2E-4 (2.9002069412698006e-08) +OK! xsec from fortran (47.059717237420642) and cpp (47.059718602249824) differ by less than 2E-4 (2.9002069412698006e-08) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105696663215774] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.4696s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3034s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1658s for 90112 events => throughput is 5.44E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 47.13 [47.127146503336576] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2840s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1336s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1502s for 81920 events => throughput is 5.46E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105696663215774) differ by less than 2E-4 (2.9364318976377035e-08) +OK! xsec from fortran (47.127145120694841) and cpp (47.127146503336576) differ by less than 2E-4 (2.9338542262280498e-08) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.223051e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.338031e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.767945e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.437945e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094186169585456] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059718602249824] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4019s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3874s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0142s for 8192 events => throughput is 5.78E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3736s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3596s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0137s for 8192 events => throughput is 6.00E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094186169585456) differ by less than 2E-4 (2.9002069412698006e-08) +OK! xsec from fortran (47.059717237420642) and cpp (47.059718602249824) differ by less than 2E-4 (2.9002069412698006e-08) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105696663215774] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.4595s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3077s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1514s for 90112 events => throughput is 5.95E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 47.13 [47.127146503336576] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.2631s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1244s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1383s for 81920 events => throughput is 5.92E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105696663215774) differ by less than 2E-4 (2.9364318976377035e-08) +OK! xsec from fortran (47.127145120694841) and cpp (47.127146503336576) differ by less than 2E-4 (2.9338542262280498e-08) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.889622e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.908357e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.919078e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.978005e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x1_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094186169585456] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059718602249824] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4002s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3783s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0214s for 8192 events => throughput is 3.82E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.3827s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3610s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0214s for 8192 events => throughput is 3.83E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cpp (47.094186169585456) differ by less than 2E-4 (2.9002069412698006e-08) +OK! xsec from fortran (47.059717237420642) and cpp (47.059718602249824) differ by less than 2E-4 (2.9002069412698006e-08) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggtt_x10 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105696663215774] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.5451s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3093s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2354s for 90112 events => throughput is 3.83E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 47.13 [47.127146503336576] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.3509s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1375s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2130s for 81920 events => throughput is 3.85E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cpp (47.105696663215774) differ by less than 2E-4 (2.9364318976377035e-08) +OK! xsec from fortran (47.127145120694841) and cpp (47.127146503336576) differ by less than 2E-4 (2.9338542262280498e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.737875e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.695651e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.863403e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.857471e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.09 [47.094184798437830] fbridge_mode=1 + [XSECTION] Cross section = 47.06 [47.059717232105726] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8029s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8014s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.26E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [COUNTERS] PROGRAM TOTAL : 0.7907s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7893s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.27E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.094184803756640) and cuda (47.094184798437830) differ by less than 2E-4 (1.1293987967064822e-10) +OK! xsec from fortran (47.059717237420642) and cuda (47.059717232105726) differ by less than 2E-4 (1.1293987967064822e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggtt_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 47.11 [47.105695279068492] fbridge_mode=1 - [UNWEIGHT] Wrote 1744 events (found 1749 events) - [COUNTERS] PROGRAM TOTAL : 1.7390s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7315s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0066s for 90112 events => throughput is 1.36E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 47.13 [47.127145119233873] fbridge_mode=1 + [UNWEIGHT] Wrote 1595 events (found 1600 events) + [COUNTERS] PROGRAM TOTAL : 1.5635s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5569s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0059s for 81920 events => throughput is 1.40E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (47.105695279989114) and cuda (47.105695279068492) differ by less than 2E-4 (1.954369999168648e-11) +OK! xsec from fortran (47.127145120694841) and cuda (47.127145119233873) differ by less than 2E-4 (3.100053547200332e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.004360e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.103387e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.618155e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.564083e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.337805e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.263451e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.064726e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.059733e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.321717e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.271551e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.141622e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.134914e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.487761e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.219837e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.948699e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.039373e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt index c52a8af2f9..b1bed3a696 100644 --- a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt @@ -13,14 +13,14 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. -make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' +make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:45:01 +DATE: 2024-09-02_13:38:48 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 [UNWEIGHT] Wrote 365 events (found 1496 events) - [COUNTERS] PROGRAM TOTAL : 0.6887s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3666s - [COUNTERS] Fortran MEs ( 1 ) : 0.3221s for 8192 events => throughput is 2.54E+04 events/s + [COUNTERS] PROGRAM TOTAL : 0.6939s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3572s + [COUNTERS] Fortran MEs ( 1 ) : 0.3367s for 8192 events => throughput is 2.43E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6558s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3350s - [COUNTERS] Fortran MEs ( 1 ) : 0.3208s for 8192 events => throughput is 2.55E+04 events/s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6772s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3501s + [COUNTERS] Fortran MEs ( 1 ) : 0.3271s for 8192 events => throughput is 2.50E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.1103s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5412s - [COUNTERS] Fortran MEs ( 1 ) : 3.5692s for 90112 events => throughput is 2.52E+04 events/s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.7475s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5043s + [COUNTERS] Fortran MEs ( 1 ) : 3.2432s for 81920 events => throughput is 2.53E+04 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749110] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6762s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3380s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3370s for 8192 events => throughput is 2.43E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6813s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3421s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3381s for 8192 events => throughput is 2.42E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748607749110) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179356801529E-002) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717666E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.2687s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5495s - [COUNTERS] CudaCpp MEs ( 2 ) : 3.7180s for 90112 events => throughput is 2.42E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.07903 [7.9025940519967336E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.8497s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4519s + [COUNTERS] CudaCpp MEs ( 2 ) : 3.3966s for 81920 events => throughput is 2.41E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238481932717666E-002) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025940519967336E-002) differ by less than 3E-14 (3.3306690738754696e-16) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.517328e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.488051e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.477316e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.512862e+04 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607748863] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.5207s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3399s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1801s for 8192 events => throughput is 4.55E+04 events/s + [XSECTION] Cross section = 0.07512 [7.5121179356799669E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.5333s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3518s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1808s for 8192 events => throughput is 4.53E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748607748863) differ by less than 3E-14 (2.453592884421596e-14) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179356799669E-002) differ by less than 3E-14 (2.475797344914099e-14) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717666E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 3.4936s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5370s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.9559s for 90112 events => throughput is 4.61E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07903 [7.9025940519967128E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 3.2050s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4382s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.7661s for 81920 events => throughput is 4.64E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238481932717666E-002) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025940519967128E-002) differ by less than 3E-14 (2.9976021664879227e-15) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.723167e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.681252e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.710741e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.746383e+04 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749110] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4289s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3383s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0899s for 8192 events => throughput is 9.11E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4314s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3428s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0881s for 8192 events => throughput is 9.30E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748607749110) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179356801529E-002) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.5415s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5644s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.9765s for 90112 events => throughput is 9.23E+04 events/s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.3281s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4461s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.8814s for 81920 events => throughput is 9.29E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238481932717722E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025940519967364E-002) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.063994e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.142026e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.113779e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.308691e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749110] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4521s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3684s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0831s for 8192 events => throughput is 9.86E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4217s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3425s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0786s for 8192 events => throughput is 1.04E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748607749110) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179356801529E-002) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.4440s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5615s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8818s for 90112 events => throughput is 1.02E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.2605s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4663s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.7936s for 81920 events => throughput is 1.03E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238481932717722E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025940519967364E-002) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.056563e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.044274e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.066565e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.064470e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749110] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4498s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3385s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1106s for 8192 events => throughput is 7.41E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4529s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3414s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1108s for 8192 events => throughput is 7.39E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748607749110) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179356801529E-002) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.7606s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5479s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.2119s for 90112 events => throughput is 7.44E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.5464s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4476s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.0982s for 81920 events => throughput is 7.46E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238481932717722E-002) differ by less than 3E-14 (0.0) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025940519967364E-002) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.524660e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.357368e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.502357e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.248899e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.8444s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8355s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0059s for 8192 events => throughput is 1.38E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0029s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.7751s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7673s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0051s for 8192 events => throughput is 1.62E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cuda (0.10112748607749111) differ by less than 3E-14 (0.0) +OK! xsec from fortran (7.5121179356801529E-002) and cuda (7.5121179356801529E-002) differ by less than 3E-14 (0.0) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717736E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 1.9827s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9565s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0233s for 90112 events => throughput is 3.86E+06 events/s + [XSECTION] Cross section = 0.07903 [7.9025940519967378E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.8868s + [COUNTERS] Fortran Overhead ( 0 ) : 1.8634s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0206s for 81920 events => throughput is 3.98E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cuda (7.9238481932717736E-002) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9025940519967364E-002) and cuda (7.9025940519967378E-002) differ by less than 3E-14 (2.220446049250313e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.637288e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.644558e+06 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.243124e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.978425e+06 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.002014e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.930899e+06 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.239487e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.239083e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.002136e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.946055e+06 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.250655e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.249407e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.001900e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.934423e+06 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.746731e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.756574e+06 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt index b25cff31e4..d87931d536 100644 --- a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt @@ -5,9 +5,9 @@ make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone + make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 - make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:45:43 +DATE: 2024-09-02_13:39:27 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 [UNWEIGHT] Wrote 365 events (found 1496 events) - [COUNTERS] PROGRAM TOTAL : 0.6879s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3658s - [COUNTERS] Fortran MEs ( 1 ) : 0.3221s for 8192 events => throughput is 2.54E+04 events/s + [COUNTERS] PROGRAM TOTAL : 0.6955s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3620s + [COUNTERS] Fortran MEs ( 1 ) : 0.3335s for 8192 events => throughput is 2.46E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6575s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3322s - [COUNTERS] Fortran MEs ( 1 ) : 0.3252s for 8192 events => throughput is 2.52E+04 events/s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6713s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3472s + [COUNTERS] Fortran MEs ( 1 ) : 0.3240s for 8192 events => throughput is 2.53E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.0903s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5245s - [COUNTERS] Fortran MEs ( 1 ) : 3.5658s for 90112 events => throughput is 2.53E+04 events/s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.7560s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4961s + [COUNTERS] Fortran MEs ( 1 ) : 3.2599s for 81920 events => throughput is 2.51E+04 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112722616246457] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6630s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3346s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3273s for 8192 events => throughput is 2.50E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0010s + [XSECTION] Cross section = 0.07512 [7.5120986282454630E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6696s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3437s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3251s for 8192 events => throughput is 2.52E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112722616246457) differ by less than 4E-4 (2.570171934723753e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5120986282454630E-002) differ by less than 4E-4 (2.5701719349457974e-06) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238468293717765E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.1318s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5454s - [COUNTERS] CudaCpp MEs ( 2 ) : 3.5854s for 90112 events => throughput is 2.51E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0010s + [XSECTION] Cross section = 0.07903 [7.9025908787272253E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.7206s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4454s + [COUNTERS] CudaCpp MEs ( 2 ) : 3.2743s for 81920 events => throughput is 2.50E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238468293717765E-002) differ by less than 4E-4 (1.721259623721494e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025908787272253E-002) differ by less than 4E-4 (4.0154783231827196e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.562809e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.558507e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.549301e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.582610e+04 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112720694019242] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4414s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3412s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0997s for 8192 events => throughput is 8.22E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.07512 [7.5120972003450887E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4389s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3386s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0999s for 8192 events => throughput is 8.20E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112720694019242) differ by less than 4E-4 (2.760251535116609e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5120972003450887E-002) differ by less than 4E-4 (2.760251535116609e-06) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238454783817719E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.6571s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5548s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.1018s for 90112 events => throughput is 8.18E+04 events/s + [XSECTION] Cross section = 0.07903 [7.9025895068314325E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.4438s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4434s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.9999s for 81920 events => throughput is 8.19E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238454783817719E-002) differ by less than 4E-4 (3.4262266690454624e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025895068314325E-002) differ by less than 4E-4 (5.751485239180454e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.333170e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.373373e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.397937e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.412522e+04 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112721757974454] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.3825s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3366s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0455s for 8192 events => throughput is 1.80E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.07512 [7.5120979906897659E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.3864s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3416s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0445s for 8192 events => throughput is 1.84E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112721757974454) differ by less than 4E-4 (2.655042234289695e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5120979906897659E-002) differ by less than 4E-4 (2.655042234178673e-06) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238453732924513E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.0649s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5567s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5077s for 90112 events => throughput is 1.77E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.07903 [7.9025894930997334E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.9071s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4507s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4560s for 81920 events => throughput is 1.80E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238453732924513E-002) differ by less than 4E-4 (3.558850765195132e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025894930997334E-002) differ by less than 4E-4 (5.76886143077715e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.821951e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.849738e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.834362e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.841450e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112721757974454] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.3803s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3381s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0418s for 8192 events => throughput is 1.96E+05 events/s + [XSECTION] Cross section = 0.07512 [7.5120979906897659E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.3821s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3392s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0425s for 8192 events => throughput is 1.93E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112721757974454) differ by less than 4E-4 (2.655042234289695e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5120979906897659E-002) differ by less than 4E-4 (2.655042234178673e-06) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238453732924513E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.0303s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5712s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4587s for 90112 events => throughput is 1.96E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.07903 [7.9025894930997334E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.8680s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4482s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4194s for 81920 events => throughput is 1.95E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238453732924513E-002) differ by less than 4E-4 (3.558850765195132e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025894930997334E-002) differ by less than 4E-4 (5.76886143077715e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.018262e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.036217e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.019326e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.999731e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112723389095883] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.3929s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3375s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0549s for 8192 events => throughput is 1.49E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.07512 [7.5120992023461633E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.3941s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3409s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0528s for 8192 events => throughput is 1.55E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112723389095883) differ by less than 4E-4 (2.493748653908945e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5120992023461633E-002) differ by less than 4E-4 (2.4937486537979225e-06) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238464413054557E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.1189s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5295s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5889s for 90112 events => throughput is 1.53E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.07903 [7.9025905745768593E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.9806s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4395s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5408s for 81920 events => throughput is 1.51E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238464413054557E-002) differ by less than 4E-4 (2.2110043929046697e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025905745768593E-002) differ by less than 4E-4 (4.40035240800718e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.561264e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.535198e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.545662e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.533647e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112725654777677] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.7590s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7568s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0010s for 8192 events => throughput is 8.12E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.07512 [7.5121008853771057E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.7741s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7720s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0010s for 8192 events => throughput is 7.97E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cuda (0.10112725654777677) differ by less than 4E-4 (2.269706518509551e-06) +OK! xsec from fortran (7.5121179356801529E-002) and cuda (7.5121008853771057E-002) differ by less than 4E-4 (2.2697065186205734e-06) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238470908598507E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 1.9627s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9510s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0105s for 90112 events => throughput is 8.59E+06 events/s + [XSECTION] Cross section = 0.07903 [7.9025913315442994E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.8769s + [COUNTERS] Fortran Overhead ( 0 ) : 1.8663s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0096s for 81920 events => throughput is 8.56E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cuda (7.9238470908598507E-002) differ by less than 4E-4 (1.3912582552677577e-07) +OK! xsec from fortran (7.9025940519967364E-002) and cuda (7.9025913315442994E-002) differ by less than 4E-4 (3.44248030303973e-07) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.151184e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.169340e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.548948e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.551921e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.576425e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.555910e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.715469e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.635132e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.585156e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.547762e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.753005e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.786425e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.440113e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.424166e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.293588e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.285164e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt index b6592dfe65..e3d99bc79b 100644 --- a/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg - make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone + make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:46:20 +DATE: 2024-09-02_13:40:03 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttg.mad/SubProcesses/P1_gg_ttxg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 [UNWEIGHT] Wrote 365 events (found 1496 events) - [COUNTERS] PROGRAM TOTAL : 0.6929s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3702s - [COUNTERS] Fortran MEs ( 1 ) : 0.3227s for 8192 events => throughput is 2.54E+04 events/s + [COUNTERS] PROGRAM TOTAL : 0.6816s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3577s + [COUNTERS] Fortran MEs ( 1 ) : 0.3239s for 8192 events => throughput is 2.53E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748607749111] fbridge_mode=0 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6641s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3385s - [COUNTERS] Fortran MEs ( 1 ) : 0.3256s for 8192 events => throughput is 2.52E+04 events/s + [XSECTION] Cross section = 0.07512 [7.5121179356801529E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6705s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3440s + [COUNTERS] Fortran MEs ( 1 ) : 0.3265s for 8192 events => throughput is 2.51E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttg_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481932717722E-002] fbridge_mode=0 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.1698s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5570s - [COUNTERS] Fortran MEs ( 1 ) : 3.6128s for 90112 events => throughput is 2.49E+04 events/s + [XSECTION] Cross section = 0.07903 [7.9025940519967364E-002] fbridge_mode=0 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.7633s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5040s + [COUNTERS] Fortran MEs ( 1 ) : 3.2593s for 81920 events => throughput is 2.51E+04 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748700702684] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.6766s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3338s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3416s for 8192 events => throughput is 2.40E+04 events/s + [XSECTION] Cross section = 0.07512 [7.5121180047294511E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.6925s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3448s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3466s for 8192 events => throughput is 2.36E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748700702684) differ by less than 2E-4 (9.191721828116783e-09) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121180047294511E-002) differ by less than 2E-4 (9.191721828116783e-09) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238482679400354E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 5.3154s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5455s - [COUNTERS] CudaCpp MEs ( 2 ) : 3.7687s for 90112 events => throughput is 2.39E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.07903 [7.9025941264524710E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 4.9024s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4549s + [COUNTERS] CudaCpp MEs ( 2 ) : 3.4464s for 81920 events => throughput is 2.38E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238482679400354E-002) differ by less than 2E-4 (9.423232416594374e-09) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025941264524710E-002) differ by less than 2E-4 (9.421682767296602e-09) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.463950e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.409219e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.478616e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.461368e+04 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748702805033] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.5103s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3345s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1749s for 8192 events => throughput is 4.68E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07512 [7.5121180062911519E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.5166s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3414s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1745s for 8192 events => throughput is 4.69E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748702805033) differ by less than 2E-4 (9.399612865834683e-09) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121180062911519E-002) differ by less than 2E-4 (9.399612643790078e-09) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238482683055667E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 3.4746s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5384s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.9354s for 90112 events => throughput is 4.66E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07903 [7.9025941267115887E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 3.2225s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4623s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.7595s for 81920 events => throughput is 4.66E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238482683055667E-002) differ by less than 2E-4 (9.469362849401364e-09) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025941267115887E-002) differ by less than 2E-4 (9.454471872061276e-09) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.832626e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.750746e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.815562e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.820842e+04 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748681415580] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4266s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3394s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0865s for 8192 events => throughput is 9.47E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.07512 [7.5121179904022894E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4324s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3424s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0894s for 8192 events => throughput is 9.16E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748681415580) differ by less than 2E-4 (7.284514991212632e-09) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179904022894E-002) differ by less than 2E-4 (7.284514991212632e-09) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238482534347232E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.4911s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5269s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.9636s for 90112 events => throughput is 9.35E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.07903 [7.9025941113462506E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.3203s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4396s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.8802s for 81920 events => throughput is 9.31E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238482534347232E-002) differ by less than 2E-4 (7.592642958798024e-09) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025941113462506E-002) differ by less than 2E-4 (7.510130739518672e-09) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.435081e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.545867e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.477580e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.512958e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748681415580] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4142s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3362s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0774s for 8192 events => throughput is 1.06E+05 events/s + [XSECTION] Cross section = 0.07512 [7.5121179904022894E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4205s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3421s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0778s for 8192 events => throughput is 1.05E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748681415580) differ by less than 2E-4 (7.284514991212632e-09) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121179904022894E-002) differ by less than 2E-4 (7.284514991212632e-09) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238482534347232E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.3905s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5342s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8556s for 90112 events => throughput is 1.05E+05 events/s + [XSECTION] Cross section = 0.07903 [7.9025941113462506E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.2442s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4579s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.7857s for 81920 events => throughput is 1.04E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238482534347232E-002) differ by less than 2E-4 (7.592642958798024e-09) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025941113462506E-002) differ by less than 2E-4 (7.510130739518672e-09) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.087061e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.041867e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.088736e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.077655e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748700265108] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.4463s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3356s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1100s for 8192 events => throughput is 7.45E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07512 [7.5121180044044028E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.4719s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3570s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1143s for 8192 events => throughput is 7.17E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cpp (0.10112748700265108) differ by less than 2E-4 (9.148451995955043e-09) +OK! xsec from fortran (7.5121179356801529E-002) and cpp (7.5121180044044028E-002) differ by less than 2E-4 (9.148451995955043e-09) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttg_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238482666076374E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 2.7724s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5419s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.2297s for 90112 events => throughput is 7.33E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.07903 [7.9025941249627085E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 2.6023s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4569s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.1447s for 81920 events => throughput is 7.16E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cpp (7.9238482666076374E-002) differ by less than 2E-4 (9.255082034087536e-09) +OK! xsec from fortran (7.9025940519967364E-002) and cpp (7.9025941249627085E-002) differ by less than 2E-4 (9.233167119759855e-09) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.268797e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.219727e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.343356e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.314594e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.1011 [0.10112748601943165] fbridge_mode=1 - [UNWEIGHT] Wrote 386 events (found 1179 events) - [COUNTERS] PROGRAM TOTAL : 0.7682s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7592s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0060s for 8192 events => throughput is 1.36E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0029s + [XSECTION] Cross section = 0.07512 [7.5121179313672848E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 376 events (found 1385 events) + [COUNTERS] PROGRAM TOTAL : 0.7765s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7687s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0050s for 8192 events => throughput is 1.63E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.10112748607749111) and cuda (0.10112748601943165) differ by less than 2E-4 (5.74121417074025e-10) +OK! xsec from fortran (7.5121179356801529E-002) and cuda (7.5121179313672848E-002) differ by less than 2E-4 (5.74121417074025e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.07924 [7.9238481937154381E-002] fbridge_mode=1 - [UNWEIGHT] Wrote 1898 events (found 1903 events) - [COUNTERS] PROGRAM TOTAL : 1.9875s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9612s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0233s for 90112 events => throughput is 3.86E+06 events/s + [XSECTION] Cross section = 0.07903 [7.9025940522311808E-002] fbridge_mode=1 + [UNWEIGHT] Wrote 2328 events (found 2333 events) + [COUNTERS] PROGRAM TOTAL : 1.8912s + [COUNTERS] Fortran Overhead ( 0 ) : 1.8667s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0217s for 81920 events => throughput is 3.78E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0029s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9238481932717722E-002) and cuda (7.9238481937154381E-002) differ by less than 2E-4 (5.5991211667105745e-11) +OK! xsec from fortran (7.9025940519967364E-002) and cuda (7.9025940522311808E-002) differ by less than 2E-4 (2.966671353021866e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.654166e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.452858e+06 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.808330e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.877090e+06 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.001990e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.922383e+06 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.235577e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.232766e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.000218e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.905289e+06 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.245999e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.245744e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.996930e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.922069e+06 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.726284e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.730718e+06 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt index 9f965c04b5..2bffbaf5f3 100644 --- a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt @@ -1,11 +1,11 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg -make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone - make USEBUILDDIR=1 BACKEND=cppsse4 + make USEBUILDDIR=1 BACKEND=cppavx2 make USEBUILDDIR=1 BACKEND=cpp512y @@ -18,8 +18,8 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Nothing to be done for 'all'. -make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' +make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:47:02 +DATE: 2024-09-02_13:40:43 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 [UNWEIGHT] Wrote 11 events (found 187 events) - [COUNTERS] PROGRAM TOTAL : 4.5167s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2657s - [COUNTERS] Fortran MEs ( 1 ) : 4.2511s for 8192 events => throughput is 1.93E+03 events/s + [COUNTERS] PROGRAM TOTAL : 4.5090s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2517s + [COUNTERS] Fortran MEs ( 1 ) : 4.2572s for 8192 events => throughput is 1.92E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.4866s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2643s - [COUNTERS] Fortran MEs ( 1 ) : 4.2223s for 8192 events => throughput is 1.94E+03 events/s + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.4879s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2537s + [COUNTERS] Fortran MEs ( 1 ) : 4.2343s for 8192 events => throughput is 1.93E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099815] fbridge_mode=0 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 48.4461s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8443s - [COUNTERS] Fortran MEs ( 1 ) : 46.6018s for 90112 events => throughput is 1.93E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751484] fbridge_mode=0 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 44.1739s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6521s + [COUNTERS] Fortran MEs ( 1 ) : 42.5218s for 81920 events => throughput is 1.93E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222236] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.6404s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2618s - [COUNTERS] CudaCpp MEs ( 2 ) : 4.3690s for 8192 events => throughput is 1.88E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0096s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.6460s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2534s + [COUNTERS] CudaCpp MEs ( 2 ) : 4.3836s for 8192 events => throughput is 1.87E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0090s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556621222236) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099799] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 49.9380s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7954s - [COUNTERS] CudaCpp MEs ( 2 ) : 48.1336s for 90112 events => throughput is 1.87E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751481] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 45.6667s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6786s + [COUNTERS] CudaCpp MEs ( 2 ) : 43.9790s for 81920 events => throughput is 1.86E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0090s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083266099799) differ by less than 3E-14 (7.771561172376096e-16) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205240751481) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.926413e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.904888e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.935484e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.920830e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222236] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 2.6125s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2606s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.3472s for 8192 events => throughput is 3.49E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0046s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 2.5856s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2483s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.3326s for 8192 events => throughput is 3.51E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0047s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556621222236) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099785] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 27.5257s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8027s - [COUNTERS] CudaCpp MEs ( 2 ) : 25.7180s for 90112 events => throughput is 3.50E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0050s + [XSECTION] Cross section = 0.222 [0.22201205240751465] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 25.1460s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6806s + [COUNTERS] CudaCpp MEs ( 2 ) : 23.4608s for 81920 events => throughput is 3.49E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0045s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083266099785) differ by less than 3E-14 (1.3322676295501878e-15) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205240751465) differ by less than 3E-14 (8.881784197001252e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.649842e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.610646e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.636818e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.642851e+03 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222231] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.2653s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2598s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.0030s for 8192 events => throughput is 8.17E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0026s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.2689s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2506s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.0160s for 8192 events => throughput is 8.06E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0023s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556621222231) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099799] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 12.8598s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7908s - [COUNTERS] CudaCpp MEs ( 2 ) : 11.0665s for 90112 events => throughput is 8.14E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0025s + [XSECTION] Cross section = 0.222 [0.22201205240751481] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 11.7827s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6768s + [COUNTERS] CudaCpp MEs ( 2 ) : 10.1035s for 81920 events => throughput is 8.11E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0024s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083266099799) differ by less than 3E-14 (7.771561172376096e-16) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205240751481) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.344831e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.301783e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.416676e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.293607e+03 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222231] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.1673s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2599s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.9051s for 8192 events => throughput is 9.05E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0023s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.1596s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2529s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.9045s for 8192 events => throughput is 9.06E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0022s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556621222231) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099799] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 11.7872s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8132s - [COUNTERS] CudaCpp MEs ( 2 ) : 9.9717s for 90112 events => throughput is 9.04E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751481] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 10.7574s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6788s + [COUNTERS] CudaCpp MEs ( 2 ) : 9.0764s for 81920 events => throughput is 9.03E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0022s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083266099799) differ by less than 3E-14 (7.771561172376096e-16) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205240751481) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.472083e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.440086e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.534343e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.421946e+03 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222231] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.3936s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2589s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.1317s for 8192 events => throughput is 7.24E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0030s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.3924s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2531s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.1366s for 8192 events => throughput is 7.21E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556621222231) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099799] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 14.2691s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8171s - [COUNTERS] CudaCpp MEs ( 2 ) : 12.4493s for 90112 events => throughput is 7.24E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751481] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 13.0947s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6786s + [COUNTERS] CudaCpp MEs ( 2 ) : 11.4133s for 81920 events => throughput is 7.18E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083266099799) differ by less than 3E-14 (7.771561172376096e-16) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205240751481) differ by less than 3E-14 (1.1102230246251565e-16) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.935643e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.281195e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.348983e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.303696e+03 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222225] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.7693s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6983s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0363s for 8192 events => throughput is 2.26E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0347s + [XSECTION] Cross section = 0.1963 [0.19631538095144732] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.7566s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6857s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0365s for 8192 events => throughput is 2.25E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0345s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cuda (0.46320556621222225) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.19631538095144735) and cuda (0.19631538095144732) differ by less than 3E-14 (1.1102230246251565e-16) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099782] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 2.6062s - [COUNTERS] Fortran Overhead ( 0 ) : 2.2048s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3669s for 90112 events => throughput is 2.46E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0344s + [XSECTION] Cross section = 0.222 [0.22201205240751462] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 2.4784s + [COUNTERS] Fortran Overhead ( 0 ) : 2.1111s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3330s for 81920 events => throughput is 2.46E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0342s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cuda (0.22558083266099782) differ by less than 3E-14 (1.4432899320127035e-15) +OK! xsec from fortran (0.22201205240751484) and cuda (0.22201205240751462) differ by less than 3E-14 (9.992007221626409e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.290486e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.273048e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.506388e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.494353e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.134196e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.124513e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.177921e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.158303e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.129278e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.134303e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.155764e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.149706e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.126990e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.125829e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.446377e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.445941e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt index cd633f37c7..f38ccf4b33 100644 --- a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt @@ -1,8 +1,8 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg - make USEBUILDDIR=1 BACKEND=cuda + make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 @@ -13,8 +13,8 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' - make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' + make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:51:00 +DATE: 2024-09-02_13:44:28 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 [UNWEIGHT] Wrote 11 events (found 187 events) - [COUNTERS] PROGRAM TOTAL : 4.4959s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2635s - [COUNTERS] Fortran MEs ( 1 ) : 4.2323s for 8192 events => throughput is 1.94E+03 events/s + [COUNTERS] PROGRAM TOTAL : 4.5597s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2577s + [COUNTERS] Fortran MEs ( 1 ) : 4.3020s for 8192 events => throughput is 1.90E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.4788s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2631s - [COUNTERS] Fortran MEs ( 1 ) : 4.2156s for 8192 events => throughput is 1.94E+03 events/s + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.4997s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2475s + [COUNTERS] Fortran MEs ( 1 ) : 4.2522s for 8192 events => throughput is 1.93E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099815] fbridge_mode=0 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 48.4352s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8357s - [COUNTERS] Fortran MEs ( 1 ) : 46.5995s for 90112 events => throughput is 1.93E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751484] fbridge_mode=0 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 44.2773s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6472s + [COUNTERS] Fortran MEs ( 1 ) : 42.6301s for 81920 events => throughput is 1.92E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320716615478996] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.5354s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2660s - [COUNTERS] CudaCpp MEs ( 2 ) : 4.2605s for 8192 events => throughput is 1.92E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0089s + [XSECTION] Cross section = 0.1963 [0.19631605903771740] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.5291s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2530s + [COUNTERS] CudaCpp MEs ( 2 ) : 4.2677s for 8192 events => throughput is 1.92E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0084s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320716615478996) differ by less than 4E-4 (3.4540659359372228e-06) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631605903771740) differ by less than 4E-4 (3.4540659359372228e-06) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558162567940870] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 48.5468s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7982s - [COUNTERS] CudaCpp MEs ( 2 ) : 46.7401s for 90112 events => throughput is 1.93E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0085s + [XSECTION] Cross section = 0.222 [0.22201284973114477] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 44.5010s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6872s + [COUNTERS] CudaCpp MEs ( 2 ) : 42.8054s for 81920 events => throughput is 1.91E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0084s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558162567940870) differ by less than 4E-4 (3.5154512074697664e-06) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201284973114477) differ by less than 4E-4 (3.591352907506007e-06) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.996945e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.984642e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.982014e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.976486e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320708851010073] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.4573s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2634s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.1914s for 8192 events => throughput is 6.88E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0025s + [XSECTION] Cross section = 0.1963 [0.19631602613041260] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.4212s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2515s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.1673s for 8192 events => throughput is 7.02E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0023s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320708851010073) differ by less than 4E-4 (3.2864412462529913e-06) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631602613041260) differ by less than 4E-4 (3.2864412462529913e-06) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558157380141428] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 14.6570s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7854s - [COUNTERS] CudaCpp MEs ( 2 ) : 12.8693s for 90112 events => throughput is 7.00E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0024s + [XSECTION] Cross section = 0.222 [0.22201280004719651] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 13.5405s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6919s + [COUNTERS] CudaCpp MEs ( 2 ) : 11.8461s for 81920 events => throughput is 6.92E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0025s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558157380141428) differ by less than 4E-4 (3.2854760192435606e-06) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201280004719651) differ by less than 4E-4 (3.367563488287928e-06) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.255598e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.169863e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.246435e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.151481e+03 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320704806184321] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.7739s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2587s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5137s for 8192 events => throughput is 1.59E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s + [XSECTION] Cross section = 0.1963 [0.19631600898766727] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.7705s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2520s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5171s for 8192 events => throughput is 1.58E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320704806184321) differ by less than 4E-4 (3.199118769003917e-06) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631600898766727) differ by less than 4E-4 (3.199118769448006e-06) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558158459897135] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 7.4672s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7991s - [COUNTERS] CudaCpp MEs ( 2 ) : 5.6666s for 90112 events => throughput is 1.59E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s + [XSECTION] Cross section = 0.222 [0.22201280838919157] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 6.8281s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6751s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.1517s for 81920 events => throughput is 1.59E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0013s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558158459897135) differ by less than 4E-4 (3.333341597855366e-06) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201280838919157) differ by less than 4E-4 (3.4051380028987666e-06) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.606140e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.630834e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.576957e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.622899e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320704806184321] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.7680s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2709s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4957s for 8192 events => throughput is 1.65E+04 events/s + [XSECTION] Cross section = 0.1963 [0.19631600898766727] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.7112s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2524s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4575s for 8192 events => throughput is 1.79E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0013s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320704806184321) differ by less than 4E-4 (3.199118769003917e-06) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631600898766727) differ by less than 4E-4 (3.199118769448006e-06) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558158459897135] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 6.7809s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7804s - [COUNTERS] CudaCpp MEs ( 2 ) : 4.9992s for 90112 events => throughput is 1.80E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s + [XSECTION] Cross section = 0.222 [0.22201280838919157] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 6.2393s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6833s + [COUNTERS] CudaCpp MEs ( 2 ) : 4.5549s for 81920 events => throughput is 1.80E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558158459897135) differ by less than 4E-4 (3.333341597855366e-06) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201280838919157) differ by less than 4E-4 (3.4051380028987666e-06) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.849666e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.858866e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.858554e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.868238e+04 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320713685871445] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.8187s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2599s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5571s for 8192 events => throughput is 1.47E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0017s + [XSECTION] Cross section = 0.1963 [0.19631604662148008] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.8193s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2534s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5646s for 8192 events => throughput is 1.45E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320713685871445) differ by less than 4E-4 (3.390819555360025e-06) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631604662148008) differ by less than 4E-4 (3.390819555360025e-06) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558162184774774] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 7.9104s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7899s - [COUNTERS] CudaCpp MEs ( 2 ) : 6.1190s for 90112 events => throughput is 1.47E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s + [XSECTION] Cross section = 0.222 [0.22201284659931389] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 7.3597s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6753s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.6830s for 81920 events => throughput is 1.44E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0013s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558162184774774) differ by less than 4E-4 (3.4984654515568536e-06) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201284659931389) differ by less than 4E-4 (3.5772463271577237e-06) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.496224e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.475909e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.504281e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.472171e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320719394836651] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.7396s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6908s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0246s for 8192 events => throughput is 3.32E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0242s + [XSECTION] Cross section = 0.1963 [0.19631607081716690] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.7385s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6898s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0247s for 8192 events => throughput is 3.32E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0241s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cuda (0.46320719394836651) differ by less than 4E-4 (3.5140686183154912e-06) +OK! xsec from fortran (0.19631538095144735) and cuda (0.19631607081716690) differ by less than 4E-4 (3.514068618537536e-06) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558167135091578] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 2.4680s - [COUNTERS] Fortran Overhead ( 0 ) : 2.1917s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2521s for 90112 events => throughput is 3.57E+05 events/s + [XSECTION] Cross section = 0.222 [0.22201289508761571] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 2.3598s + [COUNTERS] Fortran Overhead ( 0 ) : 2.1049s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2308s for 81920 events => throughput is 3.55E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0241s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cuda (0.22558167135091578) differ by less than 4E-4 (3.7179130325526444e-06) +OK! xsec from fortran (0.22201205240751484) and cuda (0.22201289508761571) differ by less than 4E-4 (3.795650243887394e-06) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.382988e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.375421e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.717142e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.746565e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.139748e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.143274e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.304954e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.266413e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.085623e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.170287e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.300454e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.266584e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.130448e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.123293e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.397157e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.397306e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt index 27512be658..d8af7f3c37 100644 --- a/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg - make USEBUILDDIR=1 BACKEND=cuda - make USEBUILDDIR=1 BACKEND=cppnone + + make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -14,8 +14,8 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' -make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' +make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:54:10 +DATE: 2024-09-02_13:47:26 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 [UNWEIGHT] Wrote 11 events (found 187 events) - [COUNTERS] PROGRAM TOTAL : 4.4700s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2619s - [COUNTERS] Fortran MEs ( 1 ) : 4.2081s for 8192 events => throughput is 1.95E+03 events/s + [COUNTERS] PROGRAM TOTAL : 4.5326s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2489s + [COUNTERS] Fortran MEs ( 1 ) : 4.2837s for 8192 events => throughput is 1.91E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x1_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556621222242] fbridge_mode=0 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.4683s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2604s - [COUNTERS] Fortran MEs ( 1 ) : 4.2079s for 8192 events => throughput is 1.95E+03 events/s + [XSECTION] Cross section = 0.1963 [0.19631538095144735] fbridge_mode=0 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.4955s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2537s + [COUNTERS] Fortran MEs ( 1 ) : 4.2418s for 8192 events => throughput is 1.93E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttgg_x10_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083266099815] fbridge_mode=0 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 48.3196s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8366s - [COUNTERS] Fortran MEs ( 1 ) : 46.4830s for 90112 events => throughput is 1.94E+03 events/s + [XSECTION] Cross section = 0.222 [0.22201205240751484] fbridge_mode=0 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 44.1801s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6465s + [COUNTERS] Fortran MEs ( 1 ) : 42.5335s for 81920 events => throughput is 1.93E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556893412546] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 4.6760s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2586s - [COUNTERS] CudaCpp MEs ( 2 ) : 4.4088s for 8192 events => throughput is 1.86E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0086s + [XSECTION] Cross section = 0.1963 [0.19631538210504193] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 4.6933s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2498s + [COUNTERS] CudaCpp MEs ( 2 ) : 4.4342s for 8192 events => throughput is 1.85E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0093s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556893412546) differ by less than 2E-4 (5.876231279344779e-09) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538210504193) differ by less than 2E-4 (5.876231279344779e-09) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083370546855] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 50.5724s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8031s - [COUNTERS] CudaCpp MEs ( 2 ) : 48.7604s for 90112 events => throughput is 1.85E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0089s + [XSECTION] Cross section = 0.222 [0.22201205347145192] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 46.3174s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6875s + [COUNTERS] CudaCpp MEs ( 2 ) : 44.6212s for 81920 events => throughput is 1.84E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0087s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083370546855) differ by less than 2E-4 (4.630138050742971e-09) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205347145192) differ by less than 2E-4 (4.792249264085058e-09) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.909521e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.901554e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.899981e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.890855e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556780656974] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 2.5687s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2576s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.3063s for 8192 events => throughput is 3.55E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0048s + [XSECTION] Cross section = 0.1963 [0.19631538162716225] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 2.5947s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2518s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.3382s for 8192 events => throughput is 3.50E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0047s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556780656974) differ by less than 2E-4 (3.4419864736179306e-09) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538162716225) differ by less than 2E-4 (3.4419864736179306e-09) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083390630859] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 27.4318s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7915s - [COUNTERS] CudaCpp MEs ( 2 ) : 25.6356s for 90112 events => throughput is 3.52E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0047s + [XSECTION] Cross section = 0.222 [0.22201205360581136] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 25.2389s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6870s + [COUNTERS] CudaCpp MEs ( 2 ) : 23.5469s for 81920 events => throughput is 3.48E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0050s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083390630859) differ by less than 2E-4 (5.520462087460487e-09) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205360581136) differ by less than 2E-4 (5.397438940235588e-09) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.646364e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.381915e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.634455e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.267115e+03 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556770726795] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.2686s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2604s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.0056s for 8192 events => throughput is 8.15E+03 events/s + [XSECTION] Cross section = 0.1963 [0.19631538158507625] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.3332s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2636s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.0671s for 8192 events => throughput is 7.68E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0025s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556770726795) differ by less than 2E-4 (3.2276070704995163e-09) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538158507625) differ by less than 2E-4 (3.2276070704995163e-09) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083379720220] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 12.9032s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7920s - [COUNTERS] CudaCpp MEs ( 2 ) : 11.1088s for 90112 events => throughput is 8.11E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0024s + [XSECTION] Cross section = 0.222 [0.22201205346697014] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 12.8031s + [COUNTERS] Fortran Overhead ( 0 ) : 1.7945s + [COUNTERS] CudaCpp MEs ( 2 ) : 11.0061s for 81920 events => throughput is 7.44E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0025s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083379720220) differ by less than 2E-4 (5.036793426782538e-09) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205346697014) differ by less than 2E-4 (4.7720620788282986e-09) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.153831e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.801878e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.410165e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.585006e+03 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556770726795] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.1480s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2607s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8850s for 8192 events => throughput is 9.26E+03 events/s + [XSECTION] Cross section = 0.1963 [0.19631538158507625] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.2214s + [COUNTERS] Fortran Overhead ( 0 ) : 0.2666s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.9526s for 8192 events => throughput is 8.60E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0023s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556770726795) differ by less than 2E-4 (3.2276070704995163e-09) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538158507625) differ by less than 2E-4 (3.2276070704995163e-09) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083379720220] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 11.5478s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7830s - [COUNTERS] CudaCpp MEs ( 2 ) : 9.7625s for 90112 events => throughput is 9.23E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0024s + [XSECTION] Cross section = 0.222 [0.22201205346697014] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 11.1979s + [COUNTERS] Fortran Overhead ( 0 ) : 1.7760s + [COUNTERS] CudaCpp MEs ( 2 ) : 9.4196s for 81920 events => throughput is 8.70E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0023s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083379720220) differ by less than 2E-4 (5.036793426782538e-09) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205346697014) differ by less than 2E-4 (4.7720620788282986e-09) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.509937e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.010103e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.503575e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.994196e+03 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556770726795] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 1.3881s - [COUNTERS] Fortran Overhead ( 0 ) : 0.2592s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.1259s for 8192 events => throughput is 7.28E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0031s + [XSECTION] Cross section = 0.1963 [0.19631538158507625] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 1.8981s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6506s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.2445s for 8192 events => throughput is 6.58E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0030s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cpp (0.46320556770726795) differ by less than 2E-4 (3.2276070704995163e-09) +OK! xsec from fortran (0.19631538095144735) and cpp (0.19631538158507625) differ by less than 2E-4 (3.2276070704995163e-09) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttgg_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083379720220] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 14.4378s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7995s - [COUNTERS] CudaCpp MEs ( 2 ) : 12.6355s for 90112 events => throughput is 7.13E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s + [XSECTION] Cross section = 0.222 [0.22201205346697014] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 14.3768s + [COUNTERS] Fortran Overhead ( 0 ) : 1.8097s + [COUNTERS] CudaCpp MEs ( 2 ) : 12.5642s for 81920 events => throughput is 6.52E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0029s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cpp (0.22558083379720220) differ by less than 2E-4 (5.036793426782538e-09) +OK! xsec from fortran (0.22201205240751484) and cpp (0.22201205346697014) differ by less than 2E-4 (4.7720620788282986e-09) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.378664e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.504950e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.252552e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.496063e+03 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.4632 [0.46320556665261842] fbridge_mode=1 - [UNWEIGHT] Wrote 11 events (found 168 events) - [COUNTERS] PROGRAM TOTAL : 0.7612s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6909s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0360s for 8192 events => throughput is 2.27E+05 events/s + [XSECTION] Cross section = 0.1963 [0.19631538113809560] fbridge_mode=1 + [UNWEIGHT] Wrote 9 events (found 177 events) + [COUNTERS] PROGRAM TOTAL : 0.8243s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7536s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0364s for 8192 events => throughput is 2.25E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0343s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.46320556621222242) and cuda (0.46320556665261842) differ by less than 2E-4 (9.507570286615419e-10) +OK! xsec from fortran (0.19631538095144735) and cuda (0.19631538113809560) differ by less than 2E-4 (9.507572507061468e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttgg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 104 [XSECTION] ChannelId = 112 - [XSECTION] Cross section = 0.2256 [0.22558083224243403] fbridge_mode=1 - [UNWEIGHT] Wrote 18 events (found 294 events) - [COUNTERS] PROGRAM TOTAL : 2.5943s - [COUNTERS] Fortran Overhead ( 0 ) : 2.1940s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3660s for 90112 events => throughput is 2.46E+05 events/s + [XSECTION] Cross section = 0.222 [0.22201205203112193] fbridge_mode=1 + [UNWEIGHT] Wrote 14 events (found 319 events) + [COUNTERS] PROGRAM TOTAL : 2.6042s + [COUNTERS] Fortran Overhead ( 0 ) : 2.2358s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3340s for 81920 events => throughput is 2.45E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0343s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.22558083266099815) and cuda (0.22558083224243403) differ by less than 2E-4 (1.855495090907766e-09) +OK! xsec from fortran (0.22201205240751484) and cuda (0.22201205203112193) differ by less than 2E-4 (1.695371509846666e-09) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.292672e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.131083e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.513091e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.510243e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.132768e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.117451e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.151465e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.153045e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.134281e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.105050e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.177596e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.182850e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.130147e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.125282e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GG_TTXGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.451952e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.452383e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt index dab5f736a0..10576a2bb5 100644 --- a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt @@ -1,19 +1,19 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg - make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 -make USEBUILDDIR=1 BACKEND=cpp512y +make USEBUILDDIR=1 BACKEND=cppsse4 make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' +make USEBUILDDIR=1 BACKEND=cppavx2 make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' +make USEBUILDDIR=1 BACKEND=cpp512y + make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:59:36 +DATE: 2024-09-02_13:52:44 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 102.0811s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5116s - [COUNTERS] Fortran MEs ( 1 ) : 101.5694s for 8192 events => throughput is 8.07E+01 events/s + [COUNTERS] PROGRAM TOTAL : 102.9323s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5549s + [COUNTERS] Fortran MEs ( 1 ) : 102.3774s for 8192 events => throughput is 8.00E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 102.0739s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5163s - [COUNTERS] Fortran MEs ( 1 ) : 101.5576s for 8192 events => throughput is 8.07E+01 events/s + [COUNTERS] PROGRAM TOTAL : 101.0911s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5000s + [COUNTERS] Fortran MEs ( 1 ) : 100.5910s for 8192 events => throughput is 8.14E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x10_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086655967E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1120.7697s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3745s - [COUNTERS] Fortran MEs ( 1 ) : 1116.3951s for 90112 events => throughput is 8.07E+01 events/s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500328E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1015.2383s + [COUNTERS] Fortran Overhead ( 0 ) : 4.0774s + [COUNTERS] Fortran MEs ( 1 ) : 1011.1609s for 81920 events => throughput is 8.10E+01 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939193E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822387E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 122.6268s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5175s - [COUNTERS] CudaCpp MEs ( 2 ) : 121.9186s for 8192 events => throughput is 6.72E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1907s + [COUNTERS] PROGRAM TOTAL : 126.0231s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5348s + [COUNTERS] CudaCpp MEs ( 2 ) : 125.2893s for 8192 events => throughput is 6.54E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1990s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985227939193E-006) differ by less than 3E-14 (1.5543122344752192e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600451822387E-007) differ by less than 3E-14 (1.5543122344752192e-15) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656014E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1388.7153s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3988s - [COUNTERS] CudaCpp MEs ( 2 ) : 1384.1234s for 90112 events => throughput is 6.51E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1931s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500378E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1274.4706s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2092s + [COUNTERS] CudaCpp MEs ( 2 ) : 1270.0636s for 81920 events => throughput is 6.45E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1978s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993086656014E-007) differ by less than 3E-14 (1.9984014443252818e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711194500378E-007) differ by less than 3E-14 (2.220446049250313e-15) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.880201e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.207915e+01 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.389775e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.632250e+01 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939197E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822398E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 60.8180s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5182s - [COUNTERS] CudaCpp MEs ( 2 ) : 60.1993s for 8192 events => throughput is 1.36E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1005s + [COUNTERS] PROGRAM TOTAL : 62.2959s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5423s + [COUNTERS] CudaCpp MEs ( 2 ) : 61.6510s for 8192 events => throughput is 1.33E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1027s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985227939197E-006) differ by less than 3E-14 (1.7763568394002505e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600451822398E-007) differ by less than 3E-14 (1.9984014443252818e-15) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656017E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 663.6261s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4076s - [COUNTERS] CudaCpp MEs ( 2 ) : 659.1171s for 90112 events => throughput is 1.37E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1014s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500384E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 619.2424s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2423s + [COUNTERS] CudaCpp MEs ( 2 ) : 614.8976s for 81920 events => throughput is 1.33E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1025s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993086656017E-007) differ by less than 3E-14 (2.220446049250313e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711194500384E-007) differ by less than 3E-14 (2.4424906541753444e-15) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.603881e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.561112e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.607115e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.556168e+02 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939191E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822387E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 28.7968s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5160s - [COUNTERS] CudaCpp MEs ( 2 ) : 28.2344s for 8192 events => throughput is 2.90E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0464s + [COUNTERS] PROGRAM TOTAL : 30.0902s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5259s + [COUNTERS] CudaCpp MEs ( 2 ) : 29.5150s for 8192 events => throughput is 2.78E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0492s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985227939191E-006) differ by less than 3E-14 (1.3322676295501878e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600451822387E-007) differ by less than 3E-14 (1.5543122344752192e-15) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656014E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 314.6312s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4324s - [COUNTERS] CudaCpp MEs ( 2 ) : 310.1525s for 90112 events => throughput is 2.91E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0464s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500378E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 299.5931s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2235s + [COUNTERS] CudaCpp MEs ( 2 ) : 295.3217s for 81920 events => throughput is 2.77E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0479s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993086656014E-007) differ by less than 3E-14 (1.9984014443252818e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711194500378E-007) differ by less than 3E-14 (2.220446049250313e-15) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.378917e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.370752e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.496128e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.363530e+02 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939191E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822387E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 25.3254s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5203s - [COUNTERS] CudaCpp MEs ( 2 ) : 24.7644s for 8192 events => throughput is 3.31E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0408s + [COUNTERS] PROGRAM TOTAL : 26.2832s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5361s + [COUNTERS] CudaCpp MEs ( 2 ) : 25.7054s for 8192 events => throughput is 3.19E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0417s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985227939191E-006) differ by less than 3E-14 (1.3322676295501878e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600451822387E-007) differ by less than 3E-14 (1.5543122344752192e-15) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656014E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 277.9808s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4083s - [COUNTERS] CudaCpp MEs ( 2 ) : 273.5305s for 90112 events => throughput is 3.29E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0420s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500378E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 263.4475s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2272s + [COUNTERS] CudaCpp MEs ( 2 ) : 259.1792s for 81920 events => throughput is 3.16E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0411s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993086656014E-007) differ by less than 3E-14 (1.9984014443252818e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711194500378E-007) differ by less than 3E-14 (2.220446049250313e-15) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.986386e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.848358e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.006448e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.856069e+02 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939191E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822387E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 25.0869s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5172s - [COUNTERS] CudaCpp MEs ( 2 ) : 24.5238s for 8192 events => throughput is 3.34E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0459s + [COUNTERS] PROGRAM TOTAL : 26.3791s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5407s + [COUNTERS] CudaCpp MEs ( 2 ) : 25.7904s for 8192 events => throughput is 3.18E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0480s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985227939191E-006) differ by less than 3E-14 (1.3322676295501878e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600451822387E-007) differ by less than 3E-14 (1.5543122344752192e-15) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656014E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 271.0840s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3948s - [COUNTERS] CudaCpp MEs ( 2 ) : 266.6404s for 90112 events => throughput is 3.38E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0489s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500378E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 261.3331s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2548s + [COUNTERS] CudaCpp MEs ( 2 ) : 257.0305s for 81920 events => throughput is 3.19E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0478s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993086656014E-007) differ by less than 3E-14 (1.9984014443252818e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711194500378E-007) differ by less than 3E-14 (2.220446049250313e-15) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.641160e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.399527e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.622116e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.420593e+02 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939195E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822387E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 3.2426s - [COUNTERS] Fortran Overhead ( 0 ) : 1.0583s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.0970s for 8192 events => throughput is 7.47E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 1.0873s + [COUNTERS] PROGRAM TOTAL : 3.2000s + [COUNTERS] Fortran Overhead ( 0 ) : 1.0251s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.0956s for 8192 events => throughput is 7.48E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 1.0792s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cuda (1.2403985227939195E-006) differ by less than 3E-14 (1.7763568394002505e-15) +OK! xsec from fortran (2.4071600451822350E-007) and cuda (2.4071600451822387E-007) differ by less than 3E-14 (1.5543122344752192e-15) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086656006E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 17.9203s - [COUNTERS] Fortran Overhead ( 0 ) : 4.9107s - [COUNTERS] CudaCpp MEs ( 2 ) : 11.9249s for 90112 events => throughput is 7.56E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 1.0847s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500368E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 16.6383s + [COUNTERS] Fortran Overhead ( 0 ) : 4.7109s + [COUNTERS] CudaCpp MEs ( 2 ) : 10.8448s for 81920 events => throughput is 7.55E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 1.0826s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cuda (2.3322993086656006E-007) differ by less than 3E-14 (1.7763568394002505e-15) +OK! xsec from fortran (2.3367711194500328E-007) and cuda (2.3367711194500368E-007) differ by less than 3E-14 (1.7763568394002505e-15) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.521131e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.512970e+03 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.292650e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.331664e+03 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.241733e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.250493e+03 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.585186e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.569158e+03 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.235154e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.250388e+03 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.473644e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.457995e+03 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.236111e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.222642e+03 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.235762e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.244319e+03 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt index 4ffdbee10a..4f7020b92d 100644 --- a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt @@ -1,7 +1,7 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg -make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_22:23:03 +DATE: 2024-09-02_15:11:29 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 101.3873s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5075s - [COUNTERS] Fortran MEs ( 1 ) : 100.8798s for 8192 events => throughput is 8.12E+01 events/s + [COUNTERS] PROGRAM TOTAL : 103.2618s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5129s + [COUNTERS] Fortran MEs ( 1 ) : 102.7490s for 8192 events => throughput is 7.97E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 102.2416s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5135s - [COUNTERS] Fortran MEs ( 1 ) : 101.7281s for 8192 events => throughput is 8.05E+01 events/s + [COUNTERS] PROGRAM TOTAL : 103.1875s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5159s + [COUNTERS] Fortran MEs ( 1 ) : 102.6716s for 8192 events => throughput is 7.98E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x10_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086655967E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1114.7300s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3428s - [COUNTERS] Fortran MEs ( 1 ) : 1110.3872s for 90112 events => throughput is 8.12E+01 events/s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500328E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1032.7688s + [COUNTERS] Fortran Overhead ( 0 ) : 4.1788s + [COUNTERS] Fortran MEs ( 1 ) : 1028.5900s for 81920 events => throughput is 7.96E+01 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -132,16 +132,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405719945779552E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074966904940680E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 111.0089s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5100s - [COUNTERS] CudaCpp MEs ( 2 ) : 110.3187s for 8192 events => throughput is 7.43E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1802s + [COUNTERS] PROGRAM TOTAL : 115.8301s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5483s + [COUNTERS] CudaCpp MEs ( 2 ) : 115.0919s for 8192 events => throughput is 7.12E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1899s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2405719945779552E-006) differ by less than 4E-4 (0.00013985165319851944) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4074966904940680E-007) differ by less than 4E-4 (0.0001398516531987415) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -167,16 +167,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326290777570335E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1216.8479s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4035s - [COUNTERS] CudaCpp MEs ( 2 ) : 1212.2644s for 90112 events => throughput is 7.43E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1800s + [XSECTION] Cross section = 2.337e-07 [2.3371006874132279E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1147.4386s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2297s + [COUNTERS] CudaCpp MEs ( 2 ) : 1143.0190s for 81920 events => throughput is 7.17E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1898s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3326290777570335E-007) differ by less than 4E-4 (0.00014139226908471692) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3371006874132279E-007) differ by less than 4E-4 (0.00014103561981393398) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -185,12 +185,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.795452e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.473035e+01 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.783118e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.450466e+01 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -212,16 +212,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405716994349971E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074961177294896E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 27.4750s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5164s - [COUNTERS] CudaCpp MEs ( 2 ) : 26.9120s for 8192 events => throughput is 3.04E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0465s + [COUNTERS] PROGRAM TOTAL : 28.4356s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5378s + [COUNTERS] CudaCpp MEs ( 2 ) : 27.8502s for 8192 events => throughput is 2.94E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0477s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2405716994349971E-006) differ by less than 4E-4 (0.00013961371115600585) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4074961177294896E-007) differ by less than 4E-4 (0.00013961371115600585) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -247,16 +247,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326284885505778E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 300.8248s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4082s - [COUNTERS] CudaCpp MEs ( 2 ) : 296.3700s for 90112 events => throughput is 3.04E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0466s + [XSECTION] Cross section = 2.337e-07 [2.3371000958320542E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 281.6312s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2237s + [COUNTERS] CudaCpp MEs ( 2 ) : 277.3595s for 81920 events => throughput is 2.95E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0480s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3326284885505778E-007) differ by less than 4E-4 (0.0001411396400787801) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3371000958320542E-007) differ by less than 4E-4 (0.00014078245801796285) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -265,12 +265,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.485944e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.369110e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.470723e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.372147e+02 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -292,16 +292,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405716646933743E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074960503087015E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 14.5936s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5183s - [COUNTERS] CudaCpp MEs ( 2 ) : 14.0522s for 8192 events => throughput is 5.83E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0231s + [COUNTERS] PROGRAM TOTAL : 15.2024s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5385s + [COUNTERS] CudaCpp MEs ( 2 ) : 14.6399s for 8192 events => throughput is 5.60E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0240s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2405716646933743E-006) differ by less than 4E-4 (0.00013958570271999093) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4074960503087015E-007) differ by less than 4E-4 (0.00013958570271999093) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -327,16 +327,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326277033163402E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 158.5014s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4348s - [COUNTERS] CudaCpp MEs ( 2 ) : 154.0430s for 90112 events => throughput is 5.85E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0236s + [XSECTION] Cross section = 2.337e-07 [2.3370993018073023E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 150.2042s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2174s + [COUNTERS] CudaCpp MEs ( 2 ) : 145.9624s for 81920 events => throughput is 5.61E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0244s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3326277033163402E-007) differ by less than 4E-4 (0.00014080296191987252) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3370993018073023E-007) differ by less than 4E-4 (0.0001404426623292654) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -345,12 +345,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.991558e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.778707e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.952358e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.789736e+02 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -372,16 +372,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405716646933743E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074960503087015E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 12.8606s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5199s - [COUNTERS] CudaCpp MEs ( 2 ) : 12.3203s for 8192 events => throughput is 6.65E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0204s + [COUNTERS] PROGRAM TOTAL : 13.4508s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5354s + [COUNTERS] CudaCpp MEs ( 2 ) : 12.8941s for 8192 events => throughput is 6.35E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0213s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2405716646933743E-006) differ by less than 4E-4 (0.00013958570271999093) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4074960503087015E-007) differ by less than 4E-4 (0.00013958570271999093) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -407,16 +407,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326277033163402E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 139.5398s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3981s - [COUNTERS] CudaCpp MEs ( 2 ) : 135.1212s for 90112 events => throughput is 6.67E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0205s + [XSECTION] Cross section = 2.337e-07 [2.3370993018073023E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 133.5140s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2093s + [COUNTERS] CudaCpp MEs ( 2 ) : 129.2835s for 81920 events => throughput is 6.34E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0212s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3326277033163402E-007) differ by less than 4E-4 (0.00014080296191987252) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3370993018073023E-007) differ by less than 4E-4 (0.0001404426623292654) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -425,12 +425,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.890802e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.750012e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 8.069181e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.757000e+02 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -452,16 +452,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405719257109645E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074965568484170E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 12.8130s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5166s - [COUNTERS] CudaCpp MEs ( 2 ) : 12.2739s for 8192 events => throughput is 6.67E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0225s + [COUNTERS] PROGRAM TOTAL : 13.4533s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5289s + [COUNTERS] CudaCpp MEs ( 2 ) : 12.9002s for 8192 events => throughput is 6.35E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0241s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2405719257109645E-006) differ by less than 4E-4 (0.00013979613314640815) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4074965568484170E-007) differ by less than 4E-4 (0.00013979613314685224) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -487,16 +487,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326283665697276E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 139.5916s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4260s - [COUNTERS] CudaCpp MEs ( 2 ) : 135.1428s for 90112 events => throughput is 6.67E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0228s + [XSECTION] Cross section = 2.337e-07 [2.3370999458663008E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 134.4516s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2546s + [COUNTERS] CudaCpp MEs ( 2 ) : 130.1730s for 81920 events => throughput is 6.29E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0240s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3326283665697276E-007) differ by less than 4E-4 (0.00014108733939433016) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3370999458663008E-007) differ by less than 4E-4 (0.00014071828153428) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -505,12 +505,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.223008e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.807802e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.135239e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.829274e+02 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -531,16 +531,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.241e-06 [1.2405721007137020E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4074968964647489E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 2.1089s - [COUNTERS] Fortran Overhead ( 0 ) : 1.0215s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5332s for 8192 events => throughput is 1.54E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.5542s + [COUNTERS] PROGRAM TOTAL : 2.1230s + [COUNTERS] Fortran Overhead ( 0 ) : 1.0238s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5399s for 8192 events => throughput is 1.52E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.5593s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cuda (1.2405721007137020E-006) differ by less than 4E-4 (0.00013993721904270728) +OK! xsec from fortran (2.4071600451822350E-007) and cuda (2.4074968964647489E-007) differ by less than 4E-4 (0.00013993721904292933) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -565,16 +565,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.333e-07 [2.3326295421688232E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 11.2844s - [COUNTERS] Fortran Overhead ( 0 ) : 4.8851s - [COUNTERS] CudaCpp MEs ( 2 ) : 5.8421s for 90112 events => throughput is 1.54E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.5572s + [XSECTION] Cross section = 2.337e-07 [2.3371010890959812E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 10.6020s + [COUNTERS] Fortran Overhead ( 0 ) : 4.6997s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.3430s for 81920 events => throughput is 1.53E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.5593s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cuda (2.3326295421688232E-007) differ by less than 4E-4 (0.00014159139095037965) +OK! xsec from fortran (2.3367711194500328E-007) and cuda (2.3371010890959812E-007) differ by less than 4E-4 (0.00014120751630408712) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -583,42 +583,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.533878e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.501634e+04 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.547825e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.546925e+04 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.147653e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.147797e+04 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.124611e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.183522e+04 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.134315e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.178929e+04 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.131039e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.180671e+04 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.139642e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.127836e+04 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.021489e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.036770e+03 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt index e8248fddca..5cddfbf771 100644 --- a/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg + make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone - make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_23:26:17 +DATE: 2024-09-02_16:11:54 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_ttggg.mad/SubProcesses/P1_gg_ttxggg @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 103.0122s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5139s - [COUNTERS] Fortran MEs ( 1 ) : 102.4983s for 8192 events => throughput is 7.99E+01 events/s + [COUNTERS] PROGRAM TOTAL : 103.1687s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5146s + [COUNTERS] Fortran MEs ( 1 ) : 102.6540s for 8192 events => throughput is 7.98E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x1_fortran > /tmp/a [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985227939174E-006] fbridge_mode=0 + [XSECTION] Cross section = 2.407e-07 [2.4071600451822350E-007] fbridge_mode=0 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 101.2993s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5294s - [COUNTERS] Fortran MEs ( 1 ) : 100.7699s for 8192 events => throughput is 8.13E+01 events/s + [COUNTERS] PROGRAM TOTAL : 104.0554s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5127s + [COUNTERS] Fortran MEs ( 1 ) : 103.5426s for 8192 events => throughput is 7.91E+01 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_ggttggg_x10_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993086655967E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1118.7642s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3619s - [COUNTERS] Fortran MEs ( 1 ) : 1114.4022s for 90112 events => throughput is 8.09E+01 events/s + [XSECTION] Cross section = 2.337e-07 [2.3367711194500328E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1031.3685s + [COUNTERS] Fortran Overhead ( 0 ) : 4.1858s + [COUNTERS] Fortran MEs ( 1 ) : 1027.1827s for 81920 events => throughput is 7.98E+01 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985299359844E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600590423770E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 125.7885s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5193s - [COUNTERS] CudaCpp MEs ( 2 ) : 125.0621s for 8192 events => throughput is 6.55E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.2071s + [COUNTERS] PROGRAM TOTAL : 127.4942s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5294s + [COUNTERS] CudaCpp MEs ( 2 ) : 126.7512s for 8192 events => throughput is 6.46E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.2136s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985299359844E-006) differ by less than 2E-4 (5.7578810608305275e-09) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600590423770E-007) differ by less than 2E-4 (5.7578812828751325e-09) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993212353001E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 1322.8827s - [COUNTERS] Fortran Overhead ( 0 ) : 4.3903s - [COUNTERS] CudaCpp MEs ( 2 ) : 1318.2870s for 90112 events => throughput is 6.84E+01 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.2054s + [XSECTION] Cross section = 2.337e-07 [2.3367711321033882E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 1227.1172s + [COUNTERS] Fortran Overhead ( 0 ) : 4.1494s + [COUNTERS] CudaCpp MEs ( 2 ) : 1222.7563s for 81920 events => throughput is 6.70E+01 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.2114s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993212353001E-007) differ by less than 2E-4 (5.389404034161771e-09) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711321033882E-007) differ by less than 2E-4 (5.414888537558227e-09) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.761597e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.520606e+01 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.724704e+01 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.700013e+01 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985295828471E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600583570665E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 62.4510s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5155s - [COUNTERS] CudaCpp MEs ( 2 ) : 61.8333s for 8192 events => throughput is 1.32E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1022s + [COUNTERS] PROGRAM TOTAL : 64.7240s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5303s + [COUNTERS] CudaCpp MEs ( 2 ) : 64.0852s for 8192 events => throughput is 1.28E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1085s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985295828471E-006) differ by less than 2E-4 (5.473184350179849e-09) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600583570665E-007) differ by less than 2E-4 (5.473184572224454e-09) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993222645653E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 684.8121s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4198s - [COUNTERS] CudaCpp MEs ( 2 ) : 680.2921s for 90112 events => throughput is 1.32E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.1003s + [XSECTION] Cross section = 2.337e-07 [2.3367711328932527E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 642.6144s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2204s + [COUNTERS] CudaCpp MEs ( 2 ) : 638.2899s for 81920 events => throughput is 1.28E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.1041s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993222645653E-007) differ by less than 2E-4 (5.830713245558172e-09) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711328932527E-007) differ by less than 2E-4 (5.752903931011133e-09) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.589042e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.546098e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.588931e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.542306e+02 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985293629285E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600579302847E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 27.0092s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5181s - [COUNTERS] CudaCpp MEs ( 2 ) : 26.4459s for 8192 events => throughput is 3.10E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0452s + [COUNTERS] PROGRAM TOTAL : 28.4622s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5310s + [COUNTERS] CudaCpp MEs ( 2 ) : 27.8846s for 8192 events => throughput is 2.94E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0466s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985293629285E-006) differ by less than 2E-4 (5.29588750630694e-09) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600579302847E-007) differ by less than 2E-4 (5.29588795039615e-09) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993222447204E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 298.0409s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4173s - [COUNTERS] CudaCpp MEs ( 2 ) : 293.5790s for 90112 events => throughput is 3.07E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0445s + [XSECTION] Cross section = 2.337e-07 [2.3367711328428580E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 284.4500s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2319s + [COUNTERS] CudaCpp MEs ( 2 ) : 280.1700s for 81920 events => throughput is 2.92E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0481s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993222447204E-007) differ by less than 2E-4 (5.822204496297445e-09) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711328428580E-007) differ by less than 2E-4 (5.731338070802394e-09) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.648206e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.541087e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.625373e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.523087e+02 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985293629285E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600579302847E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 24.3540s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5168s - [COUNTERS] CudaCpp MEs ( 2 ) : 23.7936s for 8192 events => throughput is 3.44E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0436s + [COUNTERS] PROGRAM TOTAL : 24.8181s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5208s + [COUNTERS] CudaCpp MEs ( 2 ) : 24.2581s for 8192 events => throughput is 3.38E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0392s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985293629285E-006) differ by less than 2E-4 (5.29588750630694e-09) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600579302847E-007) differ by less than 2E-4 (5.29588795039615e-09) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993222447204E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 269.6777s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4164s - [COUNTERS] CudaCpp MEs ( 2 ) : 265.2234s for 90112 events => throughput is 3.40E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0378s + [XSECTION] Cross section = 2.337e-07 [2.3367711328428580E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 247.6863s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2186s + [COUNTERS] CudaCpp MEs ( 2 ) : 243.4277s for 81920 events => throughput is 3.37E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0401s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993222447204E-007) differ by less than 2E-4 (5.822204496297445e-09) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711328428580E-007) differ by less than 2E-4 (5.731338070802394e-09) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.285493e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.124584e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.289545e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.129705e+02 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985293629285E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600579302847E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 25.1227s - [COUNTERS] Fortran Overhead ( 0 ) : 0.5145s - [COUNTERS] CudaCpp MEs ( 2 ) : 24.5642s for 8192 events => throughput is 3.33E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0441s + [COUNTERS] PROGRAM TOTAL : 26.2398s + [COUNTERS] Fortran Overhead ( 0 ) : 0.5263s + [COUNTERS] CudaCpp MEs ( 2 ) : 25.6664s for 8192 events => throughput is 3.19E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0471s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cpp (1.2403985293629285E-006) differ by less than 2E-4 (5.29588750630694e-09) +OK! xsec from fortran (2.4071600451822350E-007) and cpp (2.4071600579302847E-007) differ by less than 2E-4 (5.29588795039615e-09) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_ggttggg_ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993222447204E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 274.1583s - [COUNTERS] Fortran Overhead ( 0 ) : 4.4200s - [COUNTERS] CudaCpp MEs ( 2 ) : 269.6946s for 90112 events => throughput is 3.34E+02 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0436s + [XSECTION] Cross section = 2.337e-07 [2.3367711328428580E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 264.0733s + [COUNTERS] Fortran Overhead ( 0 ) : 4.2200s + [COUNTERS] CudaCpp MEs ( 2 ) : 259.8066s for 81920 events => throughput is 3.15E+02 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0467s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cpp (2.3322993222447204E-007) differ by less than 2E-4 (5.822204496297445e-09) +OK! xsec from fortran (2.3367711194500328E-007) and cpp (2.3367711328428580E-007) differ by less than 2E-4 (5.731338070802394e-09) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.625912e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.470561e+02 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.662510e+02 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.465615e+02 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 1.24e-06 [1.2403985217419736E-006] fbridge_mode=1 + [XSECTION] Cross section = 2.407e-07 [2.4071600431407972E-007] fbridge_mode=1 [UNWEIGHT] Wrote 70 events (found 407 events) - [COUNTERS] PROGRAM TOTAL : 2.7717s - [COUNTERS] Fortran Overhead ( 0 ) : 1.0261s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8763s for 8192 events => throughput is 9.35E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.8694s + [COUNTERS] PROGRAM TOTAL : 2.7666s + [COUNTERS] Fortran Overhead ( 0 ) : 1.0164s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.8779s for 8192 events => throughput is 9.33E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.8722s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (1.2403985227939174E-006) and cuda (1.2403985217419736E-006) differ by less than 2E-4 (8.480691704448873e-10) +OK! xsec from fortran (2.4071600451822350E-007) and cuda (2.4071600431407972E-007) differ by less than 2E-4 (8.480689484002824e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_ggttggg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.332e-07 [2.3322993078576733E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 303 events (found 1531 events) - [COUNTERS] PROGRAM TOTAL : 15.2659s - [COUNTERS] Fortran Overhead ( 0 ) : 4.8943s - [COUNTERS] CudaCpp MEs ( 2 ) : 9.5013s for 90112 events => throughput is 9.48E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.8704s + [XSECTION] Cross section = 2.337e-07 [2.3367711183341013E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 301 events (found 2088 events) + [COUNTERS] PROGRAM TOTAL : 14.2343s + [COUNTERS] Fortran Overhead ( 0 ) : 4.7257s + [COUNTERS] CudaCpp MEs ( 2 ) : 8.6379s for 81920 events => throughput is 9.48E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.8707s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.3322993086655967E-007) and cuda (2.3322993078576733E-007) differ by less than 2E-4 (3.464063480507207e-10) +OK! xsec from fortran (2.3367711194500328E-007) and cuda (2.3367711183341013E-007) differ by less than 2E-4 (4.775527751021968e-10) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.434661e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.419745e+03 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.089765e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.085213e+04 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.112116e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.111411e+04 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 512 32 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.160890e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.162842e+04 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.108390e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.111775e+04 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 128 128 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.111312e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.114977e+04 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 --bridge *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.109990e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.111775e+04 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 2048 8 1 *** Process = SIGMA_SM_GG_TTXGGG_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.638783e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.647598e+03 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt index b877c26fea..e51f6dcf45 100644 --- a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt @@ -2,21 +2,21 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/g make USEBUILDDIR=1 BACKEND=cuda - +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 -make USEBUILDDIR=1 BACKEND=cpp512y -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make USEBUILDDIR=1 BACKEND=cppavx2 make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make USEBUILDDIR=1 BACKEND=cpp512y + make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:58:09 +DATE: 2024-09-02_13:51:17 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 [UNWEIGHT] Wrote 404 events (found 1817 events) - [COUNTERS] PROGRAM TOTAL : 0.4754s - [COUNTERS] Fortran Overhead ( 0 ) : 0.4051s - [COUNTERS] Fortran MEs ( 1 ) : 0.0703s for 8192 events => throughput is 1.16E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.5281s + [COUNTERS] Fortran Overhead ( 0 ) : 0.4455s + [COUNTERS] Fortran MEs ( 1 ) : 0.0826s for 8192 events => throughput is 9.91E+04 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4153s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3445s - [COUNTERS] Fortran MEs ( 1 ) : 0.0708s for 8192 events => throughput is 1.16E+05 events/s + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4672s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3894s + [COUNTERS] Fortran MEs ( 1 ) : 0.0778s for 8192 events => throughput is 1.05E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=0 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3303s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5573s - [COUNTERS] Fortran MEs ( 1 ) : 0.7730s for 90112 events => throughput is 1.17E+05 events/s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=0 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.2912s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5516s + [COUNTERS] Fortran MEs ( 1 ) : 0.7395s for 81920 events => throughput is 1.11E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263335] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4189s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3418s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0764s for 8192 events => throughput is 1.07E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.2014 [0.20138695898218106] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4583s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3780s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0796s for 8192 events => throughput is 1.03E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539351263335) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695898218106) differ by less than 3E-14 (2.220446049250313e-16) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3766s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5374s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8384s for 90112 events => throughput is 1.07E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.3140s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5331s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.7803s for 81920 events => throughput is 1.05E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686556561295) differ by less than 3E-14 (0.0) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856931551) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.104999e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.081936e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.080050e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.079877e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351262541] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3875s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3450s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0419s for 8192 events => throughput is 1.96E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695898217512] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4067s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3636s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0426s for 8192 events => throughput is 1.92E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539351262541) differ by less than 3E-14 (2.90878432451791e-14) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695898217512) differ by less than 3E-14 (2.919886554764162e-14) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561281] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.0024s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5394s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4624s for 90112 events => throughput is 1.95E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856931479] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.9390s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5140s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4245s for 81920 events => throughput is 1.93E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686556561281) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856931479) differ by less than 3E-14 (3.3306690738754696e-15) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.937885e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.933038e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.972484e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.952025e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263341] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3673s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3427s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0240s for 8192 events => throughput is 3.41E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695898218106] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3922s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3667s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0249s for 8192 events => throughput is 3.29E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539351263341) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695898218106) differ by less than 3E-14 (2.220446049250313e-16) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.8108s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5445s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2657s for 90112 events => throughput is 3.39E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.7515s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5045s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2466s for 81920 events => throughput is 3.32E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686556561295) differ by less than 3E-14 (0.0) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856931551) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.384861e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.301583e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.378583e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.357253e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263341] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3684s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3456s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0222s for 8192 events => throughput is 3.69E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695898218106] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3868s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3643s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0220s for 8192 events => throughput is 3.72E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539351263341) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695898218106) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.7798s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5417s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2375s for 90112 events => throughput is 3.79E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.7213s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5031s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2177s for 81920 events => throughput is 3.76E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686556561295) differ by less than 3E-14 (0.0) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856931551) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.465878e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.720593e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.626688e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.675025e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263341] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3809s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3477s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0325s for 8192 events => throughput is 2.52E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.2014 [0.20138695898218106] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3985s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3650s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0330s for 8192 events => throughput is 2.48E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539351263341) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695898218106) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.8986s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5431s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3549s for 90112 events => throughput is 2.54E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.8846s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5419s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3422s for 81920 events => throughput is 2.39E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686556561295) differ by less than 3E-14 (0.0) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856931551) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.412835e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.473567e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.491870e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.447823e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263363] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.7705s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7685s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0008s for 8192 events => throughput is 1.03E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.2014 [0.20138695898218123] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.7944s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7925s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0008s for 8192 events => throughput is 1.04E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cuda (0.27110539351263363) differ by less than 3E-14 (1.3322676295501878e-15) +OK! xsec from fortran (0.20138695898218101) and cuda (0.20138695898218123) differ by less than 3E-14 (1.1102230246251565e-15) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561304] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.9737s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9648s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0078s for 90112 events => throughput is 1.15E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.9625s + [COUNTERS] Fortran Overhead ( 0 ) : 1.9541s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0073s for 81920 events => throughput is 1.13E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cuda (0.21510686556561304) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.21333720856931551) and cuda (0.21333720856931551) differ by less than 3E-14 (0.0) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.555983e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.440836e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.037158e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.006534e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.629928e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.091165e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.566255e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.524798e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.636845e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.097519e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.850724e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.815713e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.619360e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.100001e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.790736e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.778307e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt index 8ac388b886..aa11fb1c97 100644 --- a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu -make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cppnone +make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -22,8 +22,8 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Nothing to be done for 'all'. -make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:58:38 +DATE: 2024-09-02_13:51:47 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 [UNWEIGHT] Wrote 404 events (found 1817 events) - [COUNTERS] PROGRAM TOTAL : 0.4756s - [COUNTERS] Fortran Overhead ( 0 ) : 0.4044s - [COUNTERS] Fortran MEs ( 1 ) : 0.0711s for 8192 events => throughput is 1.15E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4630s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3916s + [COUNTERS] Fortran MEs ( 1 ) : 0.0714s for 8192 events => throughput is 1.15E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4108s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3420s - [COUNTERS] Fortran MEs ( 1 ) : 0.0688s for 8192 events => throughput is 1.19E+05 events/s + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4280s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3577s + [COUNTERS] Fortran MEs ( 1 ) : 0.0704s for 8192 events => throughput is 1.16E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=0 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3245s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5525s - [COUNTERS] Fortran MEs ( 1 ) : 0.7719s for 90112 events => throughput is 1.17E+05 events/s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=0 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.1701s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4632s + [COUNTERS] Fortran MEs ( 1 ) : 0.7069s for 81920 events => throughput is 1.16E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110463158198617] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4137s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3419s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0712s for 8192 events => throughput is 1.15E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138639299234948] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4385s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3651s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0728s for 8192 events => throughput is 1.12E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110463158198617) differ by less than 4E-4 (2.8104591991429118e-06) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138639299234948) differ by less than 4E-4 (2.8104591994759787e-06) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686347932190] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3233s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5375s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.7851s for 90112 events => throughput is 1.15E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.2133 [0.21333714811285437] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.2500s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5200s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.7294s for 81920 events => throughput is 1.12E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686347932190) differ by less than 4E-4 (9.698858494111562e-09) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333714811285437) differ by less than 4E-4 (2.8338451385501173e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.154270e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.139895e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.117776e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.148826e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110459183868807] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3703s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3439s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0260s for 8192 events => throughput is 3.15E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2014 [0.20138636346958042] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3899s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3638s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0257s for 8192 events => throughput is 3.19E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110459183868807) differ by less than 4E-4 (2.9570564231695684e-06) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138636346958042) differ by less than 4E-4 (2.9570564231695684e-06) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510683073685827] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.8197s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5348s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2844s for 90112 events => throughput is 3.17E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2133 [0.21333711526833204] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.7600s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4993s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2603s for 81920 events => throughput is 3.15E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510683073685827) differ by less than 4E-4 (1.6191372875784538e-07) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333711526833204) differ by less than 4E-4 (4.3734041565990367e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.998738e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.026155e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.994620e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.997369e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110460727141733] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3581s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3447s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0130s for 8192 events => throughput is 6.29E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2014 [0.20138637493357370] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3818s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3681s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0134s for 8192 events => throughput is 6.11E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110460727141733) differ by less than 4E-4 (2.9001312211729413e-06) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138637493357370) differ by less than 4E-4 (2.9001312212839636e-06) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510682516942223] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.6873s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5442s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1427s for 90112 events => throughput is 6.31E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2133 [0.21333711119055648] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.6342s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5028s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1311s for 81920 events => throughput is 6.25E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510682516942223) differ by less than 4E-4 (1.8779591537398943e-07) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333711119055648) differ by less than 4E-4 (4.5645464141763625e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.110364e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.237494e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.231132e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.181800e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110460727141733] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3551s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3423s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0124s for 8192 events => throughput is 6.61E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2014 [0.20138637493357370] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3729s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3608s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0118s for 8192 events => throughput is 6.92E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110460727141733) differ by less than 4E-4 (2.9001312211729413e-06) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138637493357370) differ by less than 4E-4 (2.9001312212839636e-06) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510682516942223] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.6706s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5390s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1312s for 90112 events => throughput is 6.87E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.2133 [0.21333711119055648] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.6235s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5028s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1204s for 81920 events => throughput is 6.80E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510682516942223) differ by less than 4E-4 (1.8779591537398943e-07) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333711119055648) differ by less than 4E-4 (4.5645464141763625e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.737889e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.565147e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.863785e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.769501e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110464220032526] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3592s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3420s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0167s for 8192 events => throughput is 4.91E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.2014 [0.20138640088003848] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3936s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3752s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0180s for 8192 events => throughput is 4.56E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110464220032526) differ by less than 4E-4 (2.771292368253242e-06) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138640088003848) differ by less than 4E-4 (2.7712923684752866e-06) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510685471570221] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.7199s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5400s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1795s for 90112 events => throughput is 5.02E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.2133 [0.21333714041997137] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.6726s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5062s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1661s for 81920 events => throughput is 4.93E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510685471570221) differ by less than 4E-4 (5.043963013928732e-08) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333714041997137) differ by less than 4E-4 (3.1944424794261295e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.872478e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.726557e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.938459e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.148798e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110477321990667] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.7679s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7663s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.31E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s + [XSECTION] Cross section = 0.2014 [0.20138649820615484] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.7947s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7933s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.28E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cuda (0.27110477321990667) differ by less than 4E-4 (2.2880132283242816e-06) +OK! xsec from fortran (0.20138695898218101) and cuda (0.20138649820615484) differ by less than 4E-4 (2.2880132283242816e-06) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510689318513457] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.9690s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9617s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0063s for 90112 events => throughput is 1.43E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s + [XSECTION] Cross section = 0.2133 [0.21333718455724227] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.9374s + [COUNTERS] Fortran Overhead ( 0 ) : 1.9307s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0059s for 81920 events => throughput is 1.39E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cuda (0.21510689318513457) differ by less than 4E-4 (1.2839907048700638e-07) +OK! xsec from fortran (0.21333720856931551) and cuda (0.21333718455724227) differ by less than 4E-4 (1.1255454879233184e-07) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.567743e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.411327e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.424411e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.392276e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.006580e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.043137e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.460162e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.454815e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.113271e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.056295e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.506902e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.511895e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.545880e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.743351e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.393633e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.393872e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt index 25661e1063..2ca2f1bbb9 100644 --- a/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt @@ -2,18 +2,18 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/g make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cppnone -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 +make USEBUILDDIR=1 BACKEND=cppavx2 make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-08_20:59:06 +DATE: 2024-09-02_13:52:15 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gq_ttq.mad/SubProcesses/P1_gu_ttxu @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 [UNWEIGHT] Wrote 404 events (found 1817 events) - [COUNTERS] PROGRAM TOTAL : 0.4768s - [COUNTERS] Fortran Overhead ( 0 ) : 0.4060s - [COUNTERS] Fortran MEs ( 1 ) : 0.0709s for 8192 events => throughput is 1.16E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4594s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3893s + [COUNTERS] Fortran MEs ( 1 ) : 0.0701s for 8192 events => throughput is 1.17E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x1_fortran > /tmp/ava [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539351263330] fbridge_mode=0 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4179s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3473s - [COUNTERS] Fortran MEs ( 1 ) : 0.0706s for 8192 events => throughput is 1.16E+05 events/s + [XSECTION] Cross section = 0.2014 [0.20138695898218101] fbridge_mode=0 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4296s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3585s + [COUNTERS] Fortran MEs ( 1 ) : 0.0711s for 8192 events => throughput is 1.15E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_gqttq_x10_fortran > /tmp/av [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686556561295] fbridge_mode=0 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3258s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5517s - [COUNTERS] Fortran MEs ( 1 ) : 0.7741s for 90112 events => throughput is 1.16E+05 events/s + [XSECTION] Cross section = 0.2133 [0.21333720856931551] fbridge_mode=0 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.1670s + [COUNTERS] Fortran Overhead ( 0 ) : 1.4583s + [COUNTERS] Fortran MEs ( 1 ) : 0.7087s for 81920 events => throughput is 1.16E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539350666329] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.4207s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3437s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0763s for 8192 events => throughput is 1.07E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 0.2014 [0.20138695897774622] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4567s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3773s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0787s for 8192 events => throughput is 1.04E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539350666329) differ by less than 2E-4 (2.2020940626532592e-11) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695897774622) differ by less than 2E-4 (2.2021162671137517e-11) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686560103207] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 2.3663s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5373s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.8282s for 90112 events => throughput is 1.09E+05 events/s + [XSECTION] Cross section = 0.2133 [0.21333720859416538] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 2.2746s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5121s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.7619s for 81920 events => throughput is 1.08E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686560103207) differ by less than 2E-4 (1.646582870051816e-10) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720859416538) differ by less than 2E-4 (1.164817131638074e-10) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.091070e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.070667e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.097593e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.091686e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539350666335] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3890s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3472s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0412s for 8192 events => throughput is 1.99E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695897774625] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.4067s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3643s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0419s for 8192 events => throughput is 1.95E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539350666335) differ by less than 2E-4 (2.2020718581927667e-11) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695897774625) differ by less than 2E-4 (2.2021051648835055e-11) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686560103204] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.9944s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5398s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4540s for 90112 events => throughput is 1.98E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720859416538] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.9307s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5116s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4186s for 81920 events => throughput is 1.96E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686560103204) differ by less than 2E-4 (1.6465806496057667e-10) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720859416538) differ by less than 2E-4 (1.164817131638074e-10) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.922053e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.930753e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.990970e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.982190e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539330887440] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3734s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3492s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0237s for 8192 events => throughput is 3.46E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695883082144] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3993s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3743s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0245s for 8192 events => throughput is 3.34E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539330887440) differ by less than 2E-4 (7.515855715567454e-10) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695883082144) differ by less than 2E-4 (7.515856825790479e-10) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686557693198] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.8003s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5375s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2622s for 90112 events => throughput is 3.44E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856057634] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.7613s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5152s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2456s for 81920 events => throughput is 3.33E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686557693198) differ by less than 2E-4 (5.262057456434377e-11) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856057634) differ by less than 2E-4 (4.0964120984199326e-11) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.424784e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.386093e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.455227e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.375171e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539330887440] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3680s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3463s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0211s for 8192 events => throughput is 3.88E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2014 [0.20138695883082144] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3894s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3670s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0219s for 8192 events => throughput is 3.74E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539330887440) differ by less than 2E-4 (7.515855715567454e-10) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695883082144) differ by less than 2E-4 (7.515856825790479e-10) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686557693198] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.7822s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5448s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2369s for 90112 events => throughput is 3.80E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856057634] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.7499s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5317s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2177s for 81920 events => throughput is 3.76E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686557693198) differ by less than 2E-4 (5.262057456434377e-11) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856057634) differ by less than 2E-4 (4.0964120984199326e-11) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.843024e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.814116e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.890496e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.801075e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539330887440] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.3872s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3503s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0362s for 8192 events => throughput is 2.26E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s + [XSECTION] Cross section = 0.2014 [0.20138695883082144] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.3985s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3640s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0340s for 8192 events => throughput is 2.41E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cpp (0.27110539330887440) differ by less than 2E-4 (7.515855715567454e-10) +OK! xsec from fortran (0.20138695898218101) and cpp (0.20138695883082144) differ by less than 2E-4 (7.515856825790479e-10) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_gqttq_x1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686557693198] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.9147s - [COUNTERS] Fortran Overhead ( 0 ) : 1.5452s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3689s for 90112 events => throughput is 2.44E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.2133 [0.21333720856057634] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.8582s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5175s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3401s for 81920 events => throughput is 2.41E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cpp (0.21510686557693198) differ by less than 2E-4 (5.262057456434377e-11) +OK! xsec from fortran (0.21333720856931551) and cpp (0.21333720856057634) differ by less than 2E-4 (4.0964120984199326e-11) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.300565e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.246156e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.415614e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.263038e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2711 [0.27110539343558537] fbridge_mode=1 - [UNWEIGHT] Wrote 404 events (found 1228 events) - [COUNTERS] PROGRAM TOTAL : 0.7684s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7665s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0008s for 8192 events => throughput is 1.09E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.2014 [0.20138695892494698] fbridge_mode=1 + [UNWEIGHT] Wrote 406 events (found 1504 events) + [COUNTERS] PROGRAM TOTAL : 0.8045s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8026s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0008s for 8192 events => throughput is 1.02E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.27110539351263330) and cuda (0.27110539343558537) differ by less than 2E-4 (2.8419910869104115e-10) +OK! xsec from fortran (0.20138695898218101) and cuda (0.20138695892494698) differ by less than 2E-4 (2.841992197133436e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_gqttq_x [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 0.2151 [0.21510686553631395] fbridge_mode=1 - [UNWEIGHT] Wrote 1939 events (found 1944 events) - [COUNTERS] PROGRAM TOTAL : 1.9688s - [COUNTERS] Fortran Overhead ( 0 ) : 1.9599s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0078s for 90112 events => throughput is 1.15E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0012s + [XSECTION] Cross section = 0.2133 [0.21333720853130850] fbridge_mode=1 + [UNWEIGHT] Wrote 2313 events (found 2318 events) + [COUNTERS] PROGRAM TOTAL : 1.9758s + [COUNTERS] Fortran Overhead ( 0 ) : 1.9672s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0075s for 81920 events => throughput is 1.09E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.21510686556561295) and cuda (0.21510686553631395) differ by less than 2E-4 (1.3620671257541517e-10) +OK! xsec from fortran (0.21333720856931551) and cuda (0.21333720853130850) differ by less than 2E-4 (1.7815460218173484e-10) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.565914e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.411382e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.104681e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.087661e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.636309e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.093131e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.555697e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.540825e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.642280e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.091339e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.824016e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.832323e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.612307e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.090721e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SM_GU_TTXU_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.778614e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.785821e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_d_inl0_hrd0.txt index 9204db3db0..0034bc1410 100644 --- a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_d_inl0_hrd0.txt @@ -4,19 +4,19 @@ make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone - make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 + make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:48:38 +DATE: 2024-09-02_17:30:05 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx @@ -58,9 +58,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 3321 events (found 6423 events) - [COUNTERS] PROGRAM TOTAL : 0.9141s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8671s - [COUNTERS] Fortran MEs ( 1 ) : 0.0470s for 8192 events => throughput is 1.74E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.9388s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8901s + [COUNTERS] Fortran MEs ( 1 ) : 0.0488s for 8192 events => throughput is 1.68E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -83,9 +83,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4185s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3716s - [COUNTERS] Fortran MEs ( 1 ) : 0.0468s for 8192 events => throughput is 1.75E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4199s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3717s + [COUNTERS] Fortran MEs ( 1 ) : 0.0481s for 8192 events => throughput is 1.70E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377569] fbridge_mode=0 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.7982s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2863s - [COUNTERS] Fortran MEs ( 1 ) : 0.5119s for 90112 events => throughput is 1.76E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453966] fbridge_mode=0 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.6818s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1962s + [COUNTERS] Fortran MEs ( 1 ) : 0.4856s for 81920 events => throughput is 1.69E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -133,10 +133,10 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256148] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4199s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3695s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0500s for 8192 events => throughput is 1.64E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.4254s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3736s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0514s for 8192 events => throughput is 1.59E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377564] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.8165s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2690s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5470s for 90112 events => throughput is 1.65E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453962] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.7140s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1964s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5170s for 81920 events => throughput is 1.58E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895240377564) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918317453962) differ by less than 3E-14 (2.220446049250313e-16) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.683813e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.644779e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.668738e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.643205e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -211,9 +211,9 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256152] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4071s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3797s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0270s for 8192 events => throughput is 3.03E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4026s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3741s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0280s for 8192 events => throughput is 2.92E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377564] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.5672s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2711s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2957s for 90112 events => throughput is 3.05E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453962] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.4754s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1930s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2820s for 81920 events => throughput is 2.91E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895240377564) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918317453962) differ by less than 3E-14 (2.220446049250313e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.037815e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.974684e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.993910e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.000194e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -289,10 +289,10 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256232] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.3883s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3715s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0164s for 8192 events => throughput is 5.00E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.3910s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3737s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0169s for 8192 events => throughput is 4.84E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377489] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.4641s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2801s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1835s for 90112 events => throughput is 4.91E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453917] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.3637s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1935s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1699s for 81920 events => throughput is 4.82E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895240377489) differ by less than 3E-14 (3.885780586188048e-15) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918317453917) differ by less than 3E-14 (2.4424906541753444e-15) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.902798e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.834817e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.886099e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.916506e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -367,9 +367,9 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256232] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.3876s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3719s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0152s for 8192 events => throughput is 5.38E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3891s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3731s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0156s for 8192 events => throughput is 5.26E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377489] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.4216s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2567s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1645s for 90112 events => throughput is 5.48E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453917] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.3535s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1958s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1573s for 81920 events => throughput is 5.21E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895240377489) differ by less than 3E-14 (3.885780586188048e-15) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918317453917) differ by less than 3E-14 (2.4424906541753444e-15) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.361206e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.285933e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.494947e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.287634e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -445,9 +445,9 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256152] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.3960s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3733s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0222s for 8192 events => throughput is 3.68E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3969s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3728s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0237s for 8192 events => throughput is 3.46E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377560] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.5023s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2627s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2391s for 90112 events => throughput is 3.77E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 2.04 [2.0403918317453957] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.4324s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1973s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2346s for 81920 events => throughput is 3.49E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895240377560) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918317453957) differ by less than 3E-14 (4.440892098500626e-16) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.615246e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.482595e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.662708e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.536605e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -523,9 +523,9 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_heftggb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256165] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.7949s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7934s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.20E+07 events/s + [COUNTERS] PROGRAM TOTAL : 0.8060s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8046s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.25E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_heftggb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377573] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.7013s - [COUNTERS] Fortran Overhead ( 0 ) : 1.6935s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0069s for 90112 events => throughput is 1.30E+07 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453962] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.6270s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6197s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0066s for 81920 events => throughput is 1.25E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cuda (2.0434895240377573) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (2.0403918317453966) and cuda (2.0403918317453962) differ by less than 3E-14 (2.220446049250313e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.844829e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.980509e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.285195e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.341053e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.255268e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.005394e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.760215e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.736304e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.235451e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.008686e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.038893e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.045650e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.241445e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.986642e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.725782e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.688896e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_f_inl0_hrd0.txt index ae36851550..8f251e163d 100644 --- a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_f_inl0_hrd0.txt @@ -6,8 +6,8 @@ make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 +make USEBUILDDIR=1 BACKEND=cppavx2 make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:49:04 +DATE: 2024-09-02_17:30:31 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx @@ -58,9 +58,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 3321 events (found 6423 events) - [COUNTERS] PROGRAM TOTAL : 0.9394s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8922s - [COUNTERS] Fortran MEs ( 1 ) : 0.0473s for 8192 events => throughput is 1.73E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.9329s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8843s + [COUNTERS] Fortran MEs ( 1 ) : 0.0486s for 8192 events => throughput is 1.69E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -83,9 +83,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4203s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3728s - [COUNTERS] Fortran MEs ( 1 ) : 0.0475s for 8192 events => throughput is 1.72E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4210s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3721s + [COUNTERS] Fortran MEs ( 1 ) : 0.0489s for 8192 events => throughput is 1.67E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377569] fbridge_mode=0 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.7988s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2854s - [COUNTERS] Fortran MEs ( 1 ) : 0.5133s for 90112 events => throughput is 1.76E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453966] fbridge_mode=0 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.6945s + [COUNTERS] Fortran Overhead ( 0 ) : 1.2063s + [COUNTERS] Fortran MEs ( 1 ) : 0.4881s for 81920 events => throughput is 1.68E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -133,9 +133,9 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_heftggbb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162897355760356] fbridge_mode=1 [UNWEIGHT] Wrote 1620 events (found 1625 events) - [COUNTERS] PROGRAM TOTAL : 0.4180s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3713s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0463s for 8192 events => throughput is 1.77E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4282s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3801s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0477s for 8192 events => throughput is 1.72E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** diff --git a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_m_inl0_hrd0.txt index d90f539fcf..3c705f76f7 100644 --- a/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_heftggbb_mad/log_heftggbb_mad_m_inl0_hrd0.txt @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:49:10 +DATE: 2024-09-02_17:30:37 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/heft_gg_bb.mad/SubProcesses/P1_gg_bbx @@ -58,9 +58,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 3321 events (found 6423 events) - [COUNTERS] PROGRAM TOTAL : 0.9158s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8684s - [COUNTERS] Fortran MEs ( 1 ) : 0.0474s for 8192 events => throughput is 1.73E+05 events/s + [COUNTERS] PROGRAM TOTAL : 1.0294s + [COUNTERS] Fortran Overhead ( 0 ) : 0.9798s + [COUNTERS] Fortran MEs ( 1 ) : 0.0496s for 8192 events => throughput is 1.65E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -83,9 +83,9 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x1_fortran > /tmp/ [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955499256161] fbridge_mode=0 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4209s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3739s - [COUNTERS] Fortran MEs ( 1 ) : 0.0470s for 8192 events => throughput is 1.74E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4182s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3699s + [COUNTERS] Fortran MEs ( 1 ) : 0.0482s for 8192 events => throughput is 1.70E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_heftggbb_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895240377569] fbridge_mode=0 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.8008s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2889s - [COUNTERS] Fortran MEs ( 1 ) : 0.5118s for 90112 events => throughput is 1.76E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918317453966] fbridge_mode=0 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.6804s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1964s + [COUNTERS] Fortran MEs ( 1 ) : 0.4840s for 81920 events => throughput is 1.69E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -134,9 +134,9 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955975930954] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4229s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3736s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0488s for 8192 events => throughput is 1.68E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4218s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3704s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0509s for 8192 events => throughput is 1.61E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -167,16 +167,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895706383660] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.8077s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2621s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5452s for 90112 events => throughput is 1.65E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 2.04 [2.0403918784059574] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.7026s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1907s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5115s for 81920 events => throughput is 1.60E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895706383660) differ by less than 2E-4 (2.2804427679545825e-08) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918784059574) differ by less than 2E-4 (2.286843137078165e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -186,13 +186,13 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.584312e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.540691e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.572139e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.552846e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -216,9 +216,9 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955975930958] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4000s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3717s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0278s for 8192 events => throughput is 2.94E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4119s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3817s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0298s for 8192 events => throughput is 2.75E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -249,16 +249,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895706383669] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.6068s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3000s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3063s for 90112 events => throughput is 2.94E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918784059583] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.4698s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1876s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2818s for 81920 events => throughput is 2.91E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895706383669) differ by less than 2E-4 (2.2804428123635034e-08) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918784059583) differ by less than 2E-4 (2.286843181487086e-08) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -268,13 +268,13 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.801476e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.826497e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.739519e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.841997e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -298,10 +298,10 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955953696393] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4107s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3912s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0191s for 8192 events => throughput is 4.29E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.3893s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3717s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0172s for 8192 events => throughput is 4.77E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -331,16 +331,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895701245432] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.4541s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2695s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1842s for 90112 events => throughput is 4.89E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 2.04 [2.0403918777124197] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.3743s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1990s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1750s for 81920 events => throughput is 4.68E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895701245432) differ by less than 2E-4 (2.255298392483951e-08) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918777124197) differ by less than 2E-4 (2.252852726591925e-08) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -350,13 +350,13 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.846731e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.670334e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.806331e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.761155e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -380,10 +380,10 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955953696393] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.3903s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3744s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0155s for 8192 events => throughput is 5.29E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.3881s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3720s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0157s for 8192 events => throughput is 5.22E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -413,16 +413,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895701245432] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.4306s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2629s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1673s for 90112 events => throughput is 5.39E+05 events/s + [XSECTION] Cross section = 2.04 [2.0403918777124197] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.3542s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1948s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1590s for 81920 events => throughput is 5.15E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895701245432) differ by less than 2E-4 (2.255298392483951e-08) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918777124197) differ by less than 2E-4 (2.252852726591925e-08) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -432,13 +432,13 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.198253e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.171512e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.334338e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.214684e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -462,10 +462,10 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955953691082] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.4086s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3841s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0240s for 8192 events => throughput is 3.41E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.3978s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3731s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0243s for 8192 events => throughput is 3.37E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** @@ -495,16 +495,16 @@ INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895701243878] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.5232s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2714s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2514s for 90112 events => throughput is 3.58E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 2.04 [2.0403918777122336] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.4335s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1889s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2442s for 81920 events => throughput is 3.36E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cpp (2.0434895701243878) differ by less than 2E-4 (2.255290776354002e-08) +OK! xsec from fortran (2.0403918317453966) and cpp (2.0403918777122336) differ by less than 2E-4 (2.2528436005586627e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -514,13 +514,13 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.375382e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.189364e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** INFO: The following Floating Point Exceptions have been reported: FE_UNDERFLOW Process = SIGMA_HEFT_GG_BBX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.300552e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.186080e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -543,10 +543,10 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_heftggb [XSECTION] ChannelId = 1 [XSECTION] Cross section = 2.016 [2.0162955503257827] fbridge_mode=1 [UNWEIGHT] Wrote 1617 events (found 1622 events) - [COUNTERS] PROGRAM TOTAL : 0.7989s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7974s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.20E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [COUNTERS] PROGRAM TOTAL : 0.8027s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8013s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.24E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** @@ -575,16 +575,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_heftggb [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 2.043 [2.0434895242795732] fbridge_mode=1 - [UNWEIGHT] Wrote 1818 events (found 1823 events) - [COUNTERS] PROGRAM TOTAL : 1.6979s - [COUNTERS] Fortran Overhead ( 0 ) : 1.6904s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0066s for 90112 events => throughput is 1.36E+07 events/s + [XSECTION] Cross section = 2.04 [2.0403918319065637] fbridge_mode=1 + [UNWEIGHT] Wrote 1666 events (found 1671 events) + [COUNTERS] PROGRAM TOTAL : 1.6233s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6160s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0065s for 81920 events => throughput is 1.26E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (2.0434895240377569) and cuda (2.0434895242795732) differ by less than 2E-4 (1.183348974365117e-10) +OK! xsec from fortran (2.0403918317453966) and cuda (2.0403918319065637) differ by less than 2E-4 (7.898837139919124e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -593,42 +593,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.835154e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.973933e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.144694e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.942621e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.230105e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.996427e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.705062e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.683594e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.235322e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.987952e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.035545e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.041388e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.242431e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.999891e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_HEFT_GG_BBX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.754474e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.759020e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_d_inl0_hrd0.txt index 5562e4c07e..61e6d05bb8 100644 --- a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_d_inl0_hrd0.txt @@ -1,7 +1,7 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx -make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:52:08 +DATE: 2024-09-02_17:33:30 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 1041 events) - [COUNTERS] PROGRAM TOTAL : 2.5941s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3442s - [COUNTERS] Fortran MEs ( 1 ) : 2.2499s for 8192 events => throughput is 3.64E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6756s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3640s + [COUNTERS] Fortran MEs ( 1 ) : 2.3116s for 8192 events => throughput is 3.54E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.6220s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3462s - [COUNTERS] Fortran MEs ( 1 ) : 2.2759s for 8192 events => throughput is 3.60E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6584s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3398s + [COUNTERS] Fortran MEs ( 1 ) : 2.3186s for 8192 events => throughput is 3.53E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x10_fortran > / [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438230E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 26.7017s - [COUNTERS] Fortran Overhead ( 0 ) : 1.8086s - [COUNTERS] Fortran MEs ( 1 ) : 24.8931s for 90112 events => throughput is 3.62E+03 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056948E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 24.8055s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6447s + [COUNTERS] Fortran MEs ( 1 ) : 23.1608s for 81920 events => throughput is 3.54E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084454E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156295E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.7821s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3463s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.4305s for 8192 events => throughput is 3.37E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0052s + [COUNTERS] PROGRAM TOTAL : 2.8273s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3395s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.4825s for 8192 events => throughput is 3.30E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0053s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896697955084454E-007) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682706312156295E-007) differ by less than 3E-14 (2.220446049250313e-16) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438187E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 28.5017s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7808s - [COUNTERS] CudaCpp MEs ( 2 ) : 26.7158s for 90112 events => throughput is 3.37E+03 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056948E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 26.5217s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6260s + [COUNTERS] CudaCpp MEs ( 2 ) : 24.8905s for 81920 events => throughput is 3.29E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0052s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668083551438187E-007) differ by less than 3E-14 (5.551115123125783e-16) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681070783056948E-007) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.542884e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.451382e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.530103e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.447250e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084412E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156274E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 1.6103s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3441s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.2634s for 8192 events => throughput is 6.48E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s + [COUNTERS] PROGRAM TOTAL : 1.6462s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3375s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.3059s for 8192 events => throughput is 6.27E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896697955084412E-007) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682706312156274E-007) differ by less than 3E-14 (5.551115123125783e-16) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438230E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 15.9197s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7936s - [COUNTERS] CudaCpp MEs ( 2 ) : 14.1234s for 90112 events => throughput is 6.38E+03 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056969E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 14.6418s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6118s + [COUNTERS] CudaCpp MEs ( 2 ) : 13.0272s for 81920 events => throughput is 6.29E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668083551438230E-007) differ by less than 3E-14 (0.0) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681070783056969E-007) differ by less than 3E-14 (2.220446049250313e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.656588e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.514217e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.664988e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.507842e+03 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084454E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.9116s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3446s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5653s for 8192 events => throughput is 1.45E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0017s + [COUNTERS] PROGRAM TOTAL : 0.9162s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3371s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5775s for 8192 events => throughput is 1.42E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896697955084454E-007) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682706312156316E-007) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438198E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 8.0033s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7755s - [COUNTERS] CudaCpp MEs ( 2 ) : 6.2261s for 90112 events => throughput is 1.45E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0017s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 7.4616s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6284s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.8317s for 81920 events => throughput is 1.40E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668083551438198E-007) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681070783056980E-007) differ by less than 3E-14 (4.440892098500626e-16) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.485686e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.450229e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.488153e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.461068e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084454E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.8483s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3476s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4991s for 8192 events => throughput is 1.64E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s + [COUNTERS] PROGRAM TOTAL : 0.8526s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3373s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5139s for 8192 events => throughput is 1.59E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896697955084454E-007) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682706312156316E-007) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438198E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 7.2914s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7820s - [COUNTERS] CudaCpp MEs ( 2 ) : 5.5079s for 90112 events => throughput is 1.64E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 6.7535s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6160s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.1361s for 81920 events => throughput is 1.59E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668083551438198E-007) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681070783056980E-007) differ by less than 3E-14 (4.440892098500626e-16) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.693554e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.659613e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.678028e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.647914e+04 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084454E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.9859s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3430s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.6411s for 8192 events => throughput is 1.28E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0018s + [COUNTERS] PROGRAM TOTAL : 1.0156s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3378s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.6759s for 8192 events => throughput is 1.21E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0019s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896697955084454E-007) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682706312156316E-007) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438198E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 8.8930s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7934s - [COUNTERS] CudaCpp MEs ( 2 ) : 7.0976s for 90112 events => throughput is 1.27E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0021s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 8.4055s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6304s + [COUNTERS] CudaCpp MEs ( 2 ) : 6.7733s for 81920 events => throughput is 1.21E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0018s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668083551438198E-007) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681070783056980E-007) differ by less than 3E-14 (4.440892098500626e-16) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.269596e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.228529e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.304260e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.235984e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084454E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156284E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.8106s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7739s + [COUNTERS] PROGRAM TOTAL : 0.8087s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7721s [COUNTERS] CudaCpp MEs ( 2 ) : 0.0172s for 8192 events => throughput is 4.76E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0196s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0194s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cuda (7.9896697955084454E-007) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (7.9682706312156316E-007) and cuda (7.9682706312156284E-007) differ by less than 3E-14 (4.440892098500626e-16) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438198E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 2.4031s - [COUNTERS] Fortran Overhead ( 0 ) : 2.1951s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1884s for 90112 events => throughput is 4.78E+05 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056948E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 2.2463s + [COUNTERS] Fortran Overhead ( 0 ) : 2.0547s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1721s for 81920 events => throughput is 4.76E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0195s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cuda (7.6668083551438198E-007) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (7.6681070783056948E-007) and cuda (7.6681070783056948E-007) differ by less than 3E-14 (0.0) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.836004e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.796804e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.223426e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.217951e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.196129e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.187638e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.417377e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.416333e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.149870e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.185186e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.416796e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.411886e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.156718e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.118787e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.752894e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.762594e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_f_inl0_hrd0.txt index e6a1cba79b..259c794857 100644 --- a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_f_inl0_hrd0.txt @@ -3,11 +3,11 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/s make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 -make USEBUILDDIR=1 BACKEND=cppavx2 +make USEBUILDDIR=1 BACKEND=cppavx2 make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:54:32 +DATE: 2024-09-02_17:35:47 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 1041 events) - [COUNTERS] PROGRAM TOTAL : 2.6010s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3425s - [COUNTERS] Fortran MEs ( 1 ) : 2.2584s for 8192 events => throughput is 3.63E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6593s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3447s + [COUNTERS] Fortran MEs ( 1 ) : 2.3146s for 8192 events => throughput is 3.54E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.6135s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3438s - [COUNTERS] Fortran MEs ( 1 ) : 2.2696s for 8192 events => throughput is 3.61E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6478s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3369s + [COUNTERS] Fortran MEs ( 1 ) : 2.3109s for 8192 events => throughput is 3.54E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x10_fortran > / [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438230E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 26.5878s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7922s - [COUNTERS] Fortran MEs ( 1 ) : 24.7956s for 90112 events => throughput is 3.63E+03 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056948E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 24.7466s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6376s + [COUNTERS] Fortran MEs ( 1 ) : 23.1089s for 81920 events => throughput is 3.54E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896784952157763E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682793076220665E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.7487s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3437s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.4000s for 8192 events => throughput is 3.41E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0050s + [COUNTERS] PROGRAM TOTAL : 2.8062s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3409s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.4605s for 8192 events => throughput is 3.33E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0049s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896784952157763E-007) differ by less than 4E-4 (1.088869447052332e-06) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682793076220665E-007) differ by less than 4E-4 (1.0888694468302873e-06) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668138450782073E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 28.1446s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7932s - [COUNTERS] CudaCpp MEs ( 2 ) : 26.3466s for 90112 events => throughput is 3.42E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0048s + [XSECTION] Cross section = 7.668e-07 [7.6681142241402160E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 26.1400s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6292s + [COUNTERS] CudaCpp MEs ( 2 ) : 24.5059s for 81920 events => throughput is 3.34E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0049s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668138450782073E-007) differ by less than 4E-4 (7.160651642745819e-07) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681142241402160E-007) differ by less than 4E-4 (9.318902889354774e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.577022e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.468397e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.590866e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.478257e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896766542858863E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682774716228357E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 1.0076s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3437s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.6623s for 8192 events => throughput is 1.24E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s + [COUNTERS] PROGRAM TOTAL : 1.0224s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3406s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.6803s for 8192 events => throughput is 1.20E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896766542858863E-007) differ by less than 4E-4 (8.584556829838164e-07) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682774716228357E-007) differ by less than 4E-4 (8.584556825397271e-07) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668121906848987E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 9.0575s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7825s - [COUNTERS] CudaCpp MEs ( 2 ) : 7.2734s for 90112 events => throughput is 1.24E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s + [XSECTION] Cross section = 7.668e-07 [7.6681125499905415E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 8.4557s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6348s + [COUNTERS] CudaCpp MEs ( 2 ) : 6.8194s for 81920 events => throughput is 1.20E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668121906848987E-007) differ by less than 4E-4 (5.002787206720427e-07) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681125499905415E-007) differ by less than 4E-4 (7.135639592537046e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.265218e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.231805e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.265996e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.231714e+04 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896764408326359E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682772587412892E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.6296s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3461s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2826s for 8192 events => throughput is 2.90E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0010s + [COUNTERS] PROGRAM TOTAL : 0.6357s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3405s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2943s for 8192 events => throughput is 2.78E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896764408326359E-007) differ by less than 4E-4 (8.31739528805997e-07) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682772587412892E-007) differ by less than 4E-4 (8.317395285839524e-07) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668124799901306E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 4.9000s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7718s - [COUNTERS] CudaCpp MEs ( 2 ) : 3.1273s for 90112 events => throughput is 2.88E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0010s + [XSECTION] Cross section = 7.668e-07 [7.6681127768020636E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 4.5624s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6247s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.9368s for 81920 events => throughput is 2.79E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668124799901306E-007) differ by less than 4E-4 (5.380134884269694e-07) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681127768020636E-007) differ by less than 4E-4 (7.431425135262515e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.939784e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.856464e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.964350e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.861077e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896764408326359E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682772587412892E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.6110s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3506s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2595s for 8192 events => throughput is 3.16E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s + [COUNTERS] PROGRAM TOTAL : 0.6064s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3390s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2666s for 8192 events => throughput is 3.07E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896764408326359E-007) differ by less than 4E-4 (8.31739528805997e-07) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682772587412892E-007) differ by less than 4E-4 (8.317395285839524e-07) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668124799901306E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 4.6623s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7820s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.8794s for 90112 events => throughput is 3.13E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s + [XSECTION] Cross section = 7.668e-07 [7.6681127768020636E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 4.3063s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6337s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.6718s for 81920 events => throughput is 3.07E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668124799901306E-007) differ by less than 4E-4 (5.380134884269694e-07) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681127768020636E-007) differ by less than 4E-4 (7.431425135262515e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.263231e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.167100e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.247254e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.180552e+04 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896778056937195E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682786199467915E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.6684s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3460s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.3212s for 8192 events => throughput is 2.55E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0013s + [COUNTERS] PROGRAM TOTAL : 0.6836s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3412s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.3414s for 8192 events => throughput is 2.40E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0011s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896778056937195E-007) differ by less than 4E-4 (1.0025677505964836e-06) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682786199467915E-007) differ by less than 4E-4 (1.0025677501523944e-06) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668139178203571E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 5.3279s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7717s - [COUNTERS] CudaCpp MEs ( 2 ) : 3.5549s for 90112 events => throughput is 2.53E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0013s + [XSECTION] Cross section = 7.668e-07 [7.6681142187850746E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 5.0375s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6273s + [COUNTERS] CudaCpp MEs ( 2 ) : 3.4092s for 81920 events => throughput is 2.40E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0010s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668139178203571E-007) differ by less than 4E-4 (7.255530953820255e-07) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681142187850746E-007) differ by less than 4E-4 (9.311919235699406e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.589261e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.440305e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.602723e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.452995e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896802503195373E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682810580250350E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.8100s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7757s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0172s for 8192 events => throughput is 4.77E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0171s + [COUNTERS] PROGRAM TOTAL : 0.8112s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7772s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0171s for 8192 events => throughput is 4.79E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0169s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cuda (7.9896802503195373E-007) differ by less than 4E-4 (1.3085410737190273e-06) +OK! xsec from fortran (7.9682706312156316E-007) and cuda (7.9682810580250350E-007) differ by less than 4E-4 (1.308541073274938e-06) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668190930428073E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 2.3814s - [COUNTERS] Fortran Overhead ( 0 ) : 2.1945s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1700s for 90112 events => throughput is 5.30E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0169s + [XSECTION] Cross section = 7.668e-07 [7.6681166790588467E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 2.2266s + [COUNTERS] Fortran Overhead ( 0 ) : 2.0547s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1550s for 81920 events => throughput is 5.29E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0168s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cuda (7.6668190930428073E-007) differ by less than 4E-4 (1.400569635601201e-06) +OK! xsec from fortran (7.6681070783056948E-007) and cuda (7.6681166790588467E-007) differ by less than 4E-4 (1.2520369179203072e-06) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.860775e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.878729e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.139558e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.129979e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.304686e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.324277e+06 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.344126e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.349427e+06 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.335964e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.314580e+06 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.345203e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.348307e+06 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.314317e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.315260e+06 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.679665e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.681194e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_m_inl0_hrd0.txt index 7e343e91b1..92e7dbcbcb 100644 --- a/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_smeftggtttt_mad/log_smeftggtttt_mad_m_inl0_hrd0.txt @@ -1,22 +1,22 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx - make USEBUILDDIR=1 BACKEND=cuda - make USEBUILDDIR=1 BACKEND=cppnone + + + make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 - make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:56:30 +DATE: 2024-09-02_17:37:40 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/smeft_gg_tttt.mad/SubProcesses/P1_gg_ttxttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 1 events (found 1041 events) - [COUNTERS] PROGRAM TOTAL : 2.5870s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3434s - [COUNTERS] Fortran MEs ( 1 ) : 2.2435s for 8192 events => throughput is 3.65E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6439s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3382s + [COUNTERS] Fortran MEs ( 1 ) : 2.3058s for 8192 events => throughput is 3.55E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x1_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697955084444E-007] fbridge_mode=0 + [XSECTION] Cross section = 7.968e-07 [7.9682706312156316E-007] fbridge_mode=0 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.5935s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3429s - [COUNTERS] Fortran MEs ( 1 ) : 2.2507s for 8192 events => throughput is 3.64E+03 events/s + [COUNTERS] PROGRAM TOTAL : 2.6515s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3376s + [COUNTERS] Fortran MEs ( 1 ) : 2.3139s for 8192 events => throughput is 3.54E+03 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_smeftggtttt_x10_fortran > / [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551438230E-007] fbridge_mode=0 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 26.4482s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7926s - [COUNTERS] Fortran MEs ( 1 ) : 24.6556s for 90112 events => throughput is 3.65E+03 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681070783056948E-007] fbridge_mode=0 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 24.7420s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6387s + [COUNTERS] Fortran MEs ( 1 ) : 23.1033s for 81920 events => throughput is 3.55E+03 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896696375074447E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682704736378117E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 2.7899s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3466s - [COUNTERS] CudaCpp MEs ( 2 ) : 2.4385s for 8192 events => throughput is 3.36E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0049s + [COUNTERS] PROGRAM TOTAL : 2.8409s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3364s + [COUNTERS] CudaCpp MEs ( 2 ) : 2.4992s for 8192 events => throughput is 3.28E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0053s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896696375074447E-007) differ by less than 2E-4 (1.9775660775600556e-08) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682704736378117E-007) differ by less than 2E-4 (1.9775661108667464e-08) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668081976882373E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 28.6799s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7926s - [COUNTERS] CudaCpp MEs ( 2 ) : 26.8820s for 90112 events => throughput is 3.35E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0052s + [XSECTION] Cross section = 7.668e-07 [7.6681069211684093E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 26.5514s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6234s + [COUNTERS] CudaCpp MEs ( 2 ) : 24.9230s for 81920 events => throughput is 3.29E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0051s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668081976882373E-007) differ by less than 2E-4 (2.0537305522871918e-08) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681069211684093E-007) differ by less than 2E-4 (2.049231762057957e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.507267e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.426056e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.511786e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.421956e+03 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896696285825688E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682704647368412E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 1.5883s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3421s - [COUNTERS] CudaCpp MEs ( 2 ) : 1.2436s for 8192 events => throughput is 6.59E+03 events/s + [COUNTERS] PROGRAM TOTAL : 1.6184s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3346s + [COUNTERS] CudaCpp MEs ( 2 ) : 1.2810s for 8192 events => throughput is 6.39E+03 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896696285825688E-007) differ by less than 2E-4 (2.089271267102788e-08) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682704647368412E-007) differ by less than 2E-4 (2.0892712893072485e-08) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668081890954375E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 15.4498s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7701s - [COUNTERS] CudaCpp MEs ( 2 ) : 13.6770s for 90112 events => throughput is 6.59E+03 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0028s + [XSECTION] Cross section = 7.668e-07 [7.6681069124822548E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 14.4021s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6262s + [COUNTERS] CudaCpp MEs ( 2 ) : 12.7732s for 81920 events => throughput is 6.41E+03 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0027s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668081890954375E-007) differ by less than 2E-4 (2.1658084770059816e-08) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681069124822548E-007) differ by less than 2E-4 (2.162508139225139e-08) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.943689e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.774555e+03 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.925887e+03 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.706407e+03 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896696427369838E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682704788533451E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.9098s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3504s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.5576s for 8192 events => throughput is 1.47E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0018s + [COUNTERS] PROGRAM TOTAL : 0.9111s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3367s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5729s for 8192 events => throughput is 1.43E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896696427369838E-007) differ by less than 2E-4 (1.9121123240317672e-08) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682704788533451E-007) differ by less than 2E-4 (1.9121123462362277e-08) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668082030339872E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 7.9207s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7702s - [COUNTERS] CudaCpp MEs ( 2 ) : 6.1490s for 90112 events => throughput is 1.47E+04 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681069262191980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 7.3590s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6207s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.7368s for 81920 events => throughput is 1.43E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668082030339872E-007) differ by less than 2E-4 (1.984004671662376e-08) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681069262191980E-007) differ by less than 2E-4 (1.983364283919542e-08) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.518105e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.460837e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.514088e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.468820e+04 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896696427369838E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682704788533451E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.8334s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3445s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4873s for 8192 events => throughput is 1.68E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0016s + [COUNTERS] PROGRAM TOTAL : 0.8433s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3354s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.5066s for 8192 events => throughput is 1.62E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0014s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896696427369838E-007) differ by less than 2E-4 (1.9121123240317672e-08) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682704788533451E-007) differ by less than 2E-4 (1.9121123462362277e-08) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668082030339872E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 7.1725s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7642s - [COUNTERS] CudaCpp MEs ( 2 ) : 5.4067s for 90112 events => throughput is 1.67E+04 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681069262191980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 6.6898s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6253s + [COUNTERS] CudaCpp MEs ( 2 ) : 5.0631s for 81920 events => throughput is 1.62E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0015s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668082030339872E-007) differ by less than 2E-4 (1.984004671662376e-08) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681069262191980E-007) differ by less than 2E-4 (1.983364283919542e-08) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.710218e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.677348e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.722202e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.681994e+04 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896696427369838E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682704788533451E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.9928s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3430s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.6479s for 8192 events => throughput is 1.26E+04 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0019s + [COUNTERS] PROGRAM TOTAL : 1.0267s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3384s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.6865s for 8192 events => throughput is 1.19E+04 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0018s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cpp (7.9896696427369838E-007) differ by less than 2E-4 (1.9121123240317672e-08) +OK! xsec from fortran (7.9682706312156316E-007) and cpp (7.9682704788533451E-007) differ by less than 2E-4 (1.9121123462362277e-08) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_smeftggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668082030339872E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 9.0659s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7892s - [COUNTERS] CudaCpp MEs ( 2 ) : 7.2749s for 90112 events => throughput is 1.24E+04 events/s + [XSECTION] Cross section = 7.668e-07 [7.6681069262191980E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 8.4842s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6249s + [COUNTERS] CudaCpp MEs ( 2 ) : 6.8574s for 81920 events => throughput is 1.19E+04 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0018s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cpp (7.6668082030339872E-007) differ by less than 2E-4 (1.984004671662376e-08) +OK! xsec from fortran (7.6681070783056948E-007) and cpp (7.6681069262191980E-007) differ by less than 2E-4 (1.983364283919542e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.210214e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.213959e+04 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.254889e+04 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.215759e+04 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.99e-07 [7.9896697918297644E-007] fbridge_mode=1 + [XSECTION] Cross section = 7.968e-07 [7.9682706275468005E-007] fbridge_mode=1 [UNWEIGHT] Wrote 215 events (found 963 events) - [COUNTERS] PROGRAM TOTAL : 0.8127s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7760s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0173s for 8192 events => throughput is 4.75E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0195s + [COUNTERS] PROGRAM TOTAL : 0.8106s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7738s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0173s for 8192 events => throughput is 4.74E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0194s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.9896697955084444E-007) and cuda (7.9896697918297644E-007) differ by less than 2E-4 (4.6042958334879813e-10) +OK! xsec from fortran (7.9682706312156316E-007) and cuda (7.9682706275468005E-007) differ by less than 2E-4 (4.60430027438008e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_smeftgg [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 7.667e-07 [7.6668083551547592E-007] fbridge_mode=1 - [UNWEIGHT] Wrote 1700 events (found 1705 events) - [COUNTERS] PROGRAM TOTAL : 2.4045s - [COUNTERS] Fortran Overhead ( 0 ) : 2.1952s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1896s for 90112 events => throughput is 4.75E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0197s + [XSECTION] Cross section = 7.668e-07 [7.6681070778960167E-007] fbridge_mode=1 + [UNWEIGHT] Wrote 1545 events (found 1550 events) + [COUNTERS] PROGRAM TOTAL : 2.2570s + [COUNTERS] Fortran Overhead ( 0 ) : 2.0646s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1730s for 81920 events => throughput is 4.73E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0194s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (7.6668083551438230E-007) and cuda (7.6668083551547592E-007) differ by less than 2E-4 (1.4264145420384011e-12) +OK! xsec from fortran (7.6681070783056948E-007) and cuda (7.6681070778960167E-007) differ by less than 2E-4 (5.3426263413314246e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.814747e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.816690e+05 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.187533e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.181859e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.164029e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.105886e+05 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.389995e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.385772e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.128645e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.110892e+05 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.372948e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.368396e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.119403e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.152842e+05 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_SMEFTSIM_TOPU3L_MWSCHEME_UFO_GG_TTXTTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.750060e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.748881e+05 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_d_inl0_hrd0.txt index 0fe0851e40..5df3f7385d 100644 --- a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_d_inl0_hrd0.txt @@ -1,8 +1,8 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x -make USEBUILDDIR=1 BACKEND=cuda +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 @@ -13,10 +13,10 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:50:54 +DATE: 2024-09-02_17:32:19 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 [UNWEIGHT] Wrote 1767 events (found 4306 events) - [COUNTERS] PROGRAM TOTAL : 0.6580s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6494s - [COUNTERS] Fortran MEs ( 1 ) : 0.0085s for 8192 events => throughput is 9.58E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.6624s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6540s + [COUNTERS] Fortran MEs ( 1 ) : 0.0084s for 8192 events => throughput is 9.78E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3938s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3851s - [COUNTERS] Fortran MEs ( 1 ) : 0.0086s for 8192 events => throughput is 9.50E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3881s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3795s + [COUNTERS] Fortran MEs ( 1 ) : 0.0085s for 8192 events => throughput is 9.59E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x10_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384407] fbridge_mode=0 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4272s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3345s - [COUNTERS] Fortran MEs ( 1 ) : 0.0927s for 90112 events => throughput is 9.72E+05 events/s + [XSECTION] Cross section = 0.3111 [0.31114133745840050] fbridge_mode=0 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2866s + [COUNTERS] Fortran Overhead ( 0 ) : 1.2019s + [COUNTERS] Fortran MEs ( 1 ) : 0.0847s for 81920 events => throughput is 9.67E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869291] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3960s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3874s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0082s for 8192 events => throughput is 9.99E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3890s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3801s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0085s for 8192 events => throughput is 9.62E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156027201869291) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027251986106652) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384418] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4271s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3353s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0914s for 90112 events => throughput is 9.86E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3111 [0.31114133745840061] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2685s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1832s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0850s for 81920 events => throughput is 9.64E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098556244384418) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114133745840061) differ by less than 3E-14 (4.440892098500626e-16) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.006217e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.923026e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.022578e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.008067e+06 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869291] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3903s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3856s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0043s for 8192 events => throughput is 1.89E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3956s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3906s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0047s for 8192 events => throughput is 1.75E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156027201869291) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027251986106652) differ by less than 3E-14 (0.0) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384418] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3937s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3444s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0489s for 90112 events => throughput is 1.84E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3111 [0.31114133745840061] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2263s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1803s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0456s for 81920 events => throughput is 1.79E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098556244384418) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114133745840061) differ by less than 3E-14 (4.440892098500626e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.897485e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.903276e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.985824e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.984094e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869291] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3921s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3888s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.88E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3983s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3950s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0030s for 8192 events => throughput is 2.77E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156027201869291) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027251986106652) differ by less than 3E-14 (0.0) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384418] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3531s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3221s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0306s for 90112 events => throughput is 2.95E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3111 [0.31114133745840061] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2133s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1838s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0291s for 81920 events => throughput is 2.81E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098556244384418) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114133745840061) differ by less than 3E-14 (4.440892098500626e-16) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.126014e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.090193e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.364824e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.349498e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869291] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3883s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3854s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0026s for 8192 events => throughput is 3.20E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3915s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3883s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.95E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156027201869291) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027251986106652) differ by less than 3E-14 (0.0) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384418] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3635s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3336s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0295s for 90112 events => throughput is 3.05E+06 events/s + [XSECTION] Cross section = 0.3111 [0.31114133745840061] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2228s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1934s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0290s for 81920 events => throughput is 2.82E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098556244384418) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114133745840061) differ by less than 3E-14 (4.440892098500626e-16) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.285096e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.275977e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.423598e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.495764e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869291] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3910s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3874s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0031s for 8192 events => throughput is 2.63E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.3103 [0.31027251986106641] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3892s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3856s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0032s for 8192 events => throughput is 2.59E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156027201869291) differ by less than 3E-14 (3.3306690738754696e-16) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027251986106641) differ by less than 3E-14 (3.3306690738754696e-16) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384418] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3563s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3235s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0323s for 90112 events => throughput is 2.79E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.3111 [0.31114133745840061] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2191s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1869s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0318s for 81920 events => throughput is 2.58E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098556244384418) differ by less than 3E-14 (4.440892098500626e-16) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114133745840061) differ by less than 3E-14 (4.440892098500626e-16) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.866364e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.830901e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.134151e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.203659e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869280] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.8164s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8152s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.37E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s + [XSECTION] Cross section = 0.3103 [0.31027251986106624] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.8198s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8186s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.36E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cuda (0.31156027201869280) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (0.31027251986106652) and cuda (0.31027251986106624) differ by less than 3E-14 (8.881784197001252e-16) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384401] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.7576s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7518s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0052s for 90112 events => throughput is 1.72E+07 events/s + [XSECTION] Cross section = 0.3111 [0.31114133745840045] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.6279s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6221s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0052s for 81920 events => throughput is 1.58E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cuda (0.31098556244384401) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (0.31114133745840050) and cuda (0.31114133745840045) differ by less than 3E-14 (2.220446049250313e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.730366e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.124464e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.967481e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.019973e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.198830e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.898984e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.649618e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.606446e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.170218e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.920833e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.903772e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 8.046571e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.201664e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.924481e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.319844e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.306340e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_f_inl0_hrd0.txt index 5c4b04cd13..3b9b71911b 100644 --- a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_f_inl0_hrd0.txt @@ -1,10 +1,10 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x + make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone - make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:51:19 +DATE: 2024-09-02_17:32:43 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 [UNWEIGHT] Wrote 1767 events (found 4306 events) - [COUNTERS] PROGRAM TOTAL : 0.6497s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6414s - [COUNTERS] Fortran MEs ( 1 ) : 0.0083s for 8192 events => throughput is 9.86E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.6519s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6432s + [COUNTERS] Fortran MEs ( 1 ) : 0.0087s for 8192 events => throughput is 9.46E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.4039s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3951s - [COUNTERS] Fortran MEs ( 1 ) : 0.0089s for 8192 events => throughput is 9.25E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3894s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3809s + [COUNTERS] Fortran MEs ( 1 ) : 0.0086s for 8192 events => throughput is 9.58E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x10_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384407] fbridge_mode=0 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4878s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3911s - [COUNTERS] Fortran MEs ( 1 ) : 0.0967s for 90112 events => throughput is 9.32E+05 events/s + [XSECTION] Cross section = 0.3111 [0.31114133745840050] fbridge_mode=0 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2857s + [COUNTERS] Fortran Overhead ( 0 ) : 1.2011s + [COUNTERS] Fortran MEs ( 1 ) : 0.0847s for 81920 events => throughput is 9.68E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156021439979276] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3975s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3887s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0084s for 8192 events => throughput is 9.70E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027246248031887] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3955s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3869s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0084s for 8192 events => throughput is 9.78E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156021439979276) differ by less than 4E-4 (1.8493660913776466e-07) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027246248031887) differ by less than 4E-4 (1.8493660891572006e-07) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098550550786874] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4264s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3345s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0916s for 90112 events => throughput is 9.84E+05 events/s + [XSECTION] Cross section = 0.3111 [0.31114128050357165] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2673s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1831s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0839s for 81920 events => throughput is 9.77E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098550550786874) differ by less than 4E-4 (1.8308237492714596e-07) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114128050357165) differ by less than 4E-4 (1.830513082223817e-07) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.034265e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.001874e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.024334e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.021487e+06 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156021343761686] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3905s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3875s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0026s for 8192 events => throughput is 3.09E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.3103 [0.31027246152211979] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3922s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3891s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.89E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156021343761686) differ by less than 4E-4 (1.8802485879376718e-07) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027246152211979) differ by less than 4E-4 (1.8802485879376718e-07) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098550488814170] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3711s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3420s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0289s for 90112 events => throughput is 3.12E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.3111 [0.31114127986228485] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2262s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1976s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0284s for 81920 events => throughput is 2.89E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098550488814170) differ by less than 4E-4 (1.8507515886501125e-07) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114127986228485) differ by less than 4E-4 (1.8511238697449528e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.288372e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.204376e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.432097e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.350742e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156021516056748] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3889s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3868s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0018s for 8192 events => throughput is 4.52E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027246323794905] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3937s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3913s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0021s for 8192 events => throughput is 3.90E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156021516056748) differ by less than 4E-4 (1.8249478717091705e-07) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027246323794905) differ by less than 4E-4 (1.8249478717091705e-07) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098550596898289] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3432s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3229s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0200s for 90112 events => throughput is 4.50E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.3111 [0.31114128107070654] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2061s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1870s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0189s for 81920 events => throughput is 4.34E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098550596898289) differ by less than 4E-4 (1.815996238940798e-07) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114128107070654) differ by less than 4E-4 (1.8122855172109098e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.077269e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.013145e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.403997e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.273682e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156021516056748] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3869s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3848s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0018s for 8192 events => throughput is 4.55E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027246323794905] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3908s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3886s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0019s for 8192 events => throughput is 4.22E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156021516056748) differ by less than 4E-4 (1.8249478717091705e-07) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027246323794905) differ by less than 4E-4 (1.8249478717091705e-07) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098550596898289] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3387s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3197s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0187s for 90112 events => throughput is 4.81E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 0.3111 [0.31114128107070654] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2030s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1845s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0182s for 81920 events => throughput is 4.51E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098550596898289) differ by less than 4E-4 (1.815996238940798e-07) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114128107070654) differ by less than 4E-4 (1.8122855172109098e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.322495e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.010922e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.427973e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.674541e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156021917867366] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3878s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3853s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0022s for 8192 events => throughput is 3.78E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3103 [0.31027246723944757] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3851s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3826s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0022s for 8192 events => throughput is 3.75E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156021917867366) differ by less than 4E-4 (1.695980652582918e-07) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027246723944757) differ by less than 4E-4 (1.695980649252249e-07) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098551029624061] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3406s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3185s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0218s for 90112 events => throughput is 4.14E+06 events/s + [XSECTION] Cross section = 0.3111 [0.31114128532877433] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2042s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1827s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0212s for 81920 events => throughput is 3.86E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098551029624061) differ by less than 4E-4 (1.6768496602370675e-07) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114128532877433) differ by less than 4E-4 (1.675432348413608e-07) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.424607e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.376050e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.888963e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.816333e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156022290359153] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.8169s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8154s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.46E+07 events/s + [XSECTION] Cross section = 0.3103 [0.31027247094896931] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.8105s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8091s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.47E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cuda (0.31156022290359153) differ by less than 4E-4 (1.576423758198331e-07) +OK! xsec from fortran (0.31027251986106652) and cuda (0.31027247094896931) differ by less than 4E-4 (1.576423759308554e-07) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098551341908548] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.7464s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7407s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0049s for 90112 events => throughput is 1.85E+07 events/s + [XSECTION] Cross section = 0.3111 [0.31114128850887690] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.6177s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6121s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0048s for 81920 events => throughput is 1.70E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0009s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cuda (0.31098551341908548) differ by less than 4E-4 (1.5764319793998283e-07) +OK! xsec from fortran (0.31114133745840050) and cuda (0.31114128850887690) differ by less than 4E-4 (1.5732246960720175e-07) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.032627e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.477530e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.278657e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.308581e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.543019e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.017426e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.578539e+09 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.574954e+09 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.555176e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.016868e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.658200e+09 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.658719e+09 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.883073e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.341320e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.705532e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.698435e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_m_inl0_hrd0.txt index 62624c2c92..58b5780370 100644 --- a/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggt1t1_mad/log_susyggt1t1_mad_m_inl0_hrd0.txt @@ -1,9 +1,9 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x +make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 @@ -13,14 +13,14 @@ make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. -make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' +make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x' make[1]: Nothing to be done for 'all'. @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:51:44 +DATE: 2024-09-02_17:33:06 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_t1t1.mad/SubProcesses/P1_gg_t1t1x @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 [UNWEIGHT] Wrote 1767 events (found 4306 events) - [COUNTERS] PROGRAM TOTAL : 0.6493s - [COUNTERS] Fortran Overhead ( 0 ) : 0.6409s - [COUNTERS] Fortran MEs ( 1 ) : 0.0084s for 8192 events => throughput is 9.81E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.6574s + [COUNTERS] Fortran Overhead ( 0 ) : 0.6488s + [COUNTERS] Fortran MEs ( 1 ) : 0.0086s for 8192 events => throughput is 9.53E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x1_fortran > /tm [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027201869302] fbridge_mode=0 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3992s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3909s - [COUNTERS] Fortran MEs ( 1 ) : 0.0083s for 8192 events => throughput is 9.85E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027251986106652] fbridge_mode=0 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3901s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3814s + [COUNTERS] Fortran MEs ( 1 ) : 0.0087s for 8192 events => throughput is 9.38E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggt1t1_x10_fortran > /t [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556244384407] fbridge_mode=0 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4133s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3208s - [COUNTERS] Fortran MEs ( 1 ) : 0.0925s for 90112 events => throughput is 9.75E+05 events/s + [XSECTION] Cross section = 0.3111 [0.31114133745840050] fbridge_mode=0 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2899s + [COUNTERS] Fortran Overhead ( 0 ) : 1.2049s + [COUNTERS] Fortran MEs ( 1 ) : 0.0850s for 81920 events => throughput is 9.64E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156028014369008] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3950s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3864s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0082s for 8192 events => throughput is 9.94E+05 events/s + [XSECTION] Cross section = 0.3103 [0.31027252795248106] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.4039s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3949s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0086s for 8192 events => throughput is 9.50E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156028014369008) differ by less than 2E-4 (2.6078411874408403e-08) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027252795248106) differ by less than 2E-4 (2.6078411874408403e-08) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098557069460298] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.4087s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3177s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0906s for 90112 events => throughput is 9.95E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3111 [0.31114134568048296] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2677s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1810s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0863s for 81920 events => throughput is 9.49E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098557069460298) differ by less than 2E-4 (2.6531003172181045e-08) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114134568048296) differ by less than 2E-4 (2.6425554855435962e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.803386e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.851627e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.910254e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.931503e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156028014369008] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3923s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3874s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0045s for 8192 events => throughput is 1.82E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3103 [0.31027252795248106] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3942s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3894s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0044s for 8192 events => throughput is 1.85E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156028014369008) differ by less than 2E-4 (2.6078411874408403e-08) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027252795248106) differ by less than 2E-4 (2.6078411874408403e-08) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098557069460298] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3653s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3175s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0474s for 90112 events => throughput is 1.90E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 0.3111 [0.31114134568048296] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2289s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1834s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0452s for 81920 events => throughput is 1.81E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098557069460298) differ by less than 2E-4 (2.6531003172181045e-08) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114134568048296) differ by less than 2E-4 (2.6425554855435962e-08) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.964224e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.964933e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.028853e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.013309e+06 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156028097537258] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3954s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3923s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0027s for 8192 events => throughput is 3.03E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027252878072603] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3952s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3921s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0028s for 8192 events => throughput is 2.96E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156028097537258) differ by less than 2E-4 (2.8747823010988327e-08) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027252878072603) differ by less than 2E-4 (2.8747823010988327e-08) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098557141632605] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3415s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3131s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0280s for 90112 events => throughput is 3.22E+06 events/s + [XSECTION] Cross section = 0.3111 [0.31114134641025609] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2167s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1875s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0288s for 81920 events => throughput is 2.84E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098557141632605) differ by less than 2E-4 (2.8851763866910574e-08) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114134641025609) differ by less than 2E-4 (2.8771026228113783e-08) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.237365e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.261336e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.416021e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.333987e+06 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156028097537258] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3940s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3909s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0027s for 8192 events => throughput is 3.04E+06 events/s + [XSECTION] Cross section = 0.3103 [0.31027252878072603] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3884s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3853s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0027s for 8192 events => throughput is 3.06E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156028097537258) differ by less than 2E-4 (2.8747823010988327e-08) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027252878072603) differ by less than 2E-4 (2.8747823010988327e-08) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098557141632605] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3467s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3184s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0279s for 90112 events => throughput is 3.23E+06 events/s + [XSECTION] Cross section = 0.3111 [0.31114134641025609] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2131s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1847s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0280s for 81920 events => throughput is 2.93E+06 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098557141632605) differ by less than 2E-4 (2.8851763866910574e-08) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114134641025609) differ by less than 2E-4 (2.8771026228113783e-08) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.347126e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.333705e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.589308e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.525000e+06 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156028097537258] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.3978s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3942s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0031s for 8192 events => throughput is 2.66E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.3103 [0.31027252878072603] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.3824s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3787s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0032s for 8192 events => throughput is 2.53E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cpp (0.31156028097537258) differ by less than 2E-4 (2.8747823010988327e-08) +OK! xsec from fortran (0.31027251986106652) and cpp (0.31027252878072603) differ by less than 2E-4 (2.8747823010988327e-08) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggt1 [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098557141632605] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.3501s - [COUNTERS] Fortran Overhead ( 0 ) : 1.3186s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0311s for 90112 events => throughput is 2.90E+06 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 0.3111 [0.31114134641025609] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.2317s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1993s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0320s for 81920 events => throughput is 2.56E+06 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cpp (0.31098557141632605) differ by less than 2E-4 (2.8851763866910574e-08) +OK! xsec from fortran (0.31114133745840050) and cpp (0.31114134641025609) differ by less than 2E-4 (2.8771026228113783e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.904623e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.881195e+06 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.114835e+06 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.180424e+06 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.3116 [0.31156027194560187] fbridge_mode=1 - [UNWEIGHT] Wrote 1636 events (found 1641 events) - [COUNTERS] PROGRAM TOTAL : 0.8152s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8140s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.39E+07 events/s + [XSECTION] Cross section = 0.3103 [0.31027251978827741] fbridge_mode=1 + [UNWEIGHT] Wrote 1642 events (found 1647 events) + [COUNTERS] PROGRAM TOTAL : 0.8262s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8251s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.38E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31156027201869302) and cuda (0.31156027194560187) differ by less than 2E-4 (2.345971195083507e-10) +OK! xsec from fortran (0.31027251986106652) and cuda (0.31027251978827741) differ by less than 2E-4 (2.3459734155295564e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 2 [XSECTION] ChannelId = 3 - [XSECTION] Cross section = 0.311 [0.31098556243340819] fbridge_mode=1 - [UNWEIGHT] Wrote 1828 events (found 1833 events) - [COUNTERS] PROGRAM TOTAL : 1.7501s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7444s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0051s for 90112 events => throughput is 1.75E+07 events/s + [XSECTION] Cross section = 0.3111 [0.31114133742418421] fbridge_mode=1 + [UNWEIGHT] Wrote 1675 events (found 1680 events) + [COUNTERS] PROGRAM TOTAL : 1.6289s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6232s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0052s for 81920 events => throughput is 1.59E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0006s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (0.31098556244384407) and cuda (0.31098556243340819) differ by less than 2E-4 (3.3557379097715057e-11) +OK! xsec from fortran (0.31114133745840050) and cuda (0.31114133742418421) differ by less than 2E-4 (1.0997025512438086e-10) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.842332e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.165635e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.019027e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.988832e+08 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.214756e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.992251e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.517612e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.485904e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.171297e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.043072e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.740991e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.842260e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.214875e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.037657e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_T1T1X_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.310258e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.295725e+08 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_d_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_d_inl0_hrd0.txt index 6131633fdd..ea32ac964e 100644 --- a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_d_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_d_inl0_hrd0.txt @@ -1,13 +1,13 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx - make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 + make USEBUILDDIR=1 BACKEND=cppavx2 +make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:49:37 +DATE: 2024-09-02_17:31:03 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 2620 events (found 5403 events) - [COUNTERS] PROGRAM TOTAL : 0.8016s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7599s - [COUNTERS] Fortran MEs ( 1 ) : 0.0418s for 8192 events => throughput is 1.96E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.8304s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7882s + [COUNTERS] Fortran MEs ( 1 ) : 0.0421s for 8192 events => throughput is 1.94E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4173s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3758s - [COUNTERS] Fortran MEs ( 1 ) : 0.0415s for 8192 events => throughput is 1.97E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4182s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3759s + [COUNTERS] Fortran MEs ( 1 ) : 0.0422s for 8192 events => throughput is 1.94E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256456] fbridge_mode=0 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.6984s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2478s - [COUNTERS] Fortran MEs ( 1 ) : 0.4506s for 90112 events => throughput is 2.00E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646522] fbridge_mode=0 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6239s + [COUNTERS] Fortran Overhead ( 0 ) : 1.2011s + [COUNTERS] Fortran MEs ( 1 ) : 0.4228s for 81920 events => throughput is 1.94E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419863] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4145s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3702s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0438s for 8192 events => throughput is 1.87E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4290s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3833s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0453s for 8192 events => throughput is 1.81E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598860065419863) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (44.566218791890641) and cpp (44.566218791890641) differ by less than 3E-14 (0.0) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256471] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7366s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2536s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4825s for 90112 events => throughput is 1.87E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [XSECTION] Cross section = 44.6 [44.597935381646536] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6241s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1716s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4521s for 81920 events => throughput is 1.81E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577523870256471) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (44.597935381646522) and cpp (44.597935381646536) differ by less than 3E-14 (2.220446049250313e-16) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.880754e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.836967e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.882930e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.840854e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3960s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3713s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0243s for 8192 events => throughput is 3.37E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4099s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3836s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0260s for 8192 events => throughput is 3.15E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598860065419856) differ by less than 3E-14 (0.0) +OK! xsec from fortran (44.566218791890641) and cpp (44.566218791890641) differ by less than 3E-14 (0.0) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256471] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.5199s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2483s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2711s for 90112 events => throughput is 3.32E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646529] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.4343s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1791s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2548s for 81920 events => throughput is 3.21E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577523870256471) differ by less than 3E-14 (2.220446049250313e-16) +OK! xsec from fortran (44.597935381646522) and cpp (44.597935381646529) differ by less than 3E-14 (2.220446049250313e-16) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.302363e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.299644e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.365112e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.267845e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890648] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3924s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3765s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0155s for 8192 events => throughput is 5.28E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3993s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3831s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0157s for 8192 events => throughput is 5.21E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598860065419856) differ by less than 3E-14 (0.0) +OK! xsec from fortran (44.566218791890641) and cpp (44.566218791890648) differ by less than 3E-14 (2.220446049250313e-16) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256485] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4183s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2503s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1675s for 90112 events => throughput is 5.38E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646536] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3245s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1659s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1582s for 81920 events => throughput is 5.18E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577523870256485) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (44.597935381646522) and cpp (44.597935381646536) differ by less than 3E-14 (2.220446049250313e-16) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.278183e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.265606e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.374748e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.221261e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890648] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3894s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3754s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0136s for 8192 events => throughput is 6.02E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3924s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3778s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0143s for 8192 events => throughput is 5.75E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598860065419856) differ by less than 3E-14 (0.0) +OK! xsec from fortran (44.566218791890641) and cpp (44.566218791890648) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256485] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.3978s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2454s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1520s for 90112 events => throughput is 5.93E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646536] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3134s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1684s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1446s for 81920 events => throughput is 5.67E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577523870256485) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (44.597935381646522) and cpp (44.597935381646536) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.775498e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.766797e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.841522e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.821996e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4047s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3821s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0222s for 8192 events => throughput is 3.70E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4003s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3771s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0227s for 8192 events => throughput is 3.60E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598860065419856) differ by less than 3E-14 (0.0) +OK! xsec from fortran (44.566218791890641) and cpp (44.566218791890641) differ by less than 3E-14 (0.0) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_d_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256485] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4927s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2545s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2377s for 90112 events => throughput is 3.79E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646536] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.4098s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1779s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2315s for 81920 events => throughput is 3.54E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577523870256485) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (44.597935381646522) and cpp (44.597935381646536) differ by less than 3E-14 (2.220446049250313e-16) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.798876e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.543383e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:DBL+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.612840e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.589793e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419849] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218791890627] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8126s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8111s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.24E+07 events/s + [COUNTERS] PROGRAM TOTAL : 0.8137s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8123s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.27E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cuda (44.598860065419849) differ by less than 3E-14 (1.1102230246251565e-16) +OK! xsec from fortran (44.566218791890641) and cuda (44.566218791890627) differ by less than 3E-14 (3.3306690738754696e-16) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_d_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256485] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.6862s - [COUNTERS] Fortran Overhead ( 0 ) : 1.6788s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0066s for 90112 events => throughput is 1.37E+07 events/s + [XSECTION] Cross section = 44.6 [44.597935381646536] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6094s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6022s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0065s for 81920 events => throughput is 1.27E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cuda (44.577523870256485) differ by less than 3E-14 (6.661338147750939e-16) +OK! xsec from fortran (44.597935381646522) and cuda (44.597935381646536) differ by less than 3E-14 (2.220446049250313e-16) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.869432e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.015336e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.714086e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.672777e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.311155e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.051079e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.083882e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.075118e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.322734e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.073477e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.159310e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.160443e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.296675e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.067237e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:DBL+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.098537e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.049333e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_f_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_f_inl0_hrd0.txt index 58b86df658..9081a79133 100644 --- a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_f_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_f_inl0_hrd0.txt @@ -1,13 +1,13 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx + make USEBUILDDIR=1 BACKEND=cuda -make USEBUILDDIR=1 BACKEND=cppnone +make USEBUILDDIR=1 BACKEND=cppnone make USEBUILDDIR=1 BACKEND=cppsse4 - - make USEBUILDDIR=1 BACKEND=cppavx2 + make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:50:03 +DATE: 2024-09-02_17:31:29 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 2620 events (found 5403 events) - [COUNTERS] PROGRAM TOTAL : 0.8051s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7635s - [COUNTERS] Fortran MEs ( 1 ) : 0.0416s for 8192 events => throughput is 1.97E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.8104s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7680s + [COUNTERS] Fortran MEs ( 1 ) : 0.0424s for 8192 events => throughput is 1.93E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4148s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3740s - [COUNTERS] Fortran MEs ( 1 ) : 0.0408s for 8192 events => throughput is 2.01E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4161s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3732s + [COUNTERS] Fortran MEs ( 1 ) : 0.0428s for 8192 events => throughput is 1.91E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256456] fbridge_mode=0 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7188s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2615s - [COUNTERS] Fortran MEs ( 1 ) : 0.4573s for 90112 events => throughput is 1.97E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646522] fbridge_mode=0 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6122s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1903s + [COUNTERS] Fortran MEs ( 1 ) : 0.4219s for 81920 events => throughput is 1.94E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598853620719339] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566212351906906] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4164s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3751s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0410s for 8192 events => throughput is 2.00E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4208s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3782s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0423s for 8192 events => throughput is 1.94E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598853620719339) differ by less than 4E-4 (1.4450370500185272e-07) +OK! xsec from fortran (44.566218791890641) and cpp (44.566212351906906) differ by less than 4E-4 (1.4450370500185272e-07) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577522280119403] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7041s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2499s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4538s for 90112 events => throughput is 1.99E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 44.6 [44.597933631628678] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.5846s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1630s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4213s for 81920 events => throughput is 1.94E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577522280119403) differ by less than 4E-4 (3.567127371262302e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597933631628678) differ by less than 4E-4 (3.923988478593543e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.004528e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.969672e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.989674e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.950364e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598849697851406] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566208431910063] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3933s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3758s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0172s for 8192 events => throughput is 4.76E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3920s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3748s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0170s for 8192 events => throughput is 4.82E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598849697851406) differ by less than 4E-4 (2.3246263325393812e-07) +OK! xsec from fortran (44.566218791890641) and cpp (44.566208431910063) differ by less than 4E-4 (2.3246263336496042e-07) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577518590213366] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4571s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2702s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1866s for 90112 events => throughput is 4.83E+05 events/s + [XSECTION] Cross section = 44.6 [44.597929889999619] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3408s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1673s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1732s for 81920 events => throughput is 4.73E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577518590213366) differ by less than 4E-4 (1.1844630731783212e-07) +OK! xsec from fortran (44.597935381646522) and cpp (44.597929889999619) differ by less than 4E-4 (1.231367967591268e-07) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.766493e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.677596e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 4.711541e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 4.693658e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598850036412124] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566208770222985] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3932s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3838s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0091s for 8192 events => throughput is 8.99E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3871s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3778s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0091s for 8192 events => throughput is 9.05E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598850036412124) differ by less than 4E-4 (2.2487139172966408e-07) +OK! xsec from fortran (44.566218791890641) and cpp (44.566208770222985) differ by less than 4E-4 (2.248713919517087e-07) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577518612400254] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.3456s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2495s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0958s for 90112 events => throughput is 9.40E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 44.6 [44.597929987884214] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.2541s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1614s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0924s for 81920 events => throughput is 8.86E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577518612400254) differ by less than 4E-4 (1.1794859255953583e-07) +OK! xsec from fortran (44.597935381646522) and cpp (44.597929987884214) differ by less than 4E-4 (1.2094197321399491e-07) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.204759e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.029763e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.210555e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.217191e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598850036412124] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566208770222985] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3855s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3769s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0083s for 8192 events => throughput is 9.85E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3919s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3831s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0085s for 8192 events => throughput is 9.59E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598850036412124) differ by less than 4E-4 (2.2487139172966408e-07) +OK! xsec from fortran (44.566218791890641) and cpp (44.566208770222985) differ by less than 4E-4 (2.248713919517087e-07) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577518612400254] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.3394s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2483s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0908s for 90112 events => throughput is 9.92E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s + [XSECTION] Cross section = 44.6 [44.597929987884214] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.2468s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1597s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0869s for 81920 events => throughput is 9.43E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0002s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577518612400254) differ by less than 4E-4 (1.1794859255953583e-07) +OK! xsec from fortran (44.597935381646522) and cpp (44.597929987884214) differ by less than 4E-4 (1.2094197321399491e-07) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.706656e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.715473e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 9.233766e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 9.884609e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598854350242270] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566213080895913] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3868s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3748s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0116s for 8192 events => throughput is 7.03E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [COUNTERS] PROGRAM TOTAL : 0.3956s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3830s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0123s for 8192 events => throughput is 6.64E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598854350242270) differ by less than 4E-4 (1.2814627048385319e-07) +OK! xsec from fortran (44.566218791890641) and cpp (44.566213080895913) differ by less than 4E-4 (1.2814627048385319e-07) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_f_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577522751628507] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.3825s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2565s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1256s for 90112 events => throughput is 7.17E+05 events/s + [XSECTION] Cross section = 44.6 [44.597934154796604] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.2815s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1611s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1201s for 81920 events => throughput is 6.82E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577522751628507) differ by less than 4E-4 (2.5093990219104967e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597934154796604) differ by less than 4E-4 (2.7509119093771517e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.942843e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.815732e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:FLT+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 6.910825e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 6.862385e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598870301426373] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566229020405565] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8091s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8078s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.43E+07 events/s + [COUNTERS] PROGRAM TOTAL : 0.8071s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8059s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0006s for 8192 events => throughput is 1.44E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cuda (44.598870301426373) differ by less than 4E-4 (2.2951273881410827e-07) +OK! xsec from fortran (44.566218791890641) and cuda (44.566229020405565) differ by less than 4E-4 (2.2951273859206367e-07) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_f_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577527268256027] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7098s - [COUNTERS] Fortran Overhead ( 0 ) : 1.7033s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0058s for 90112 events => throughput is 1.56E+07 events/s + [XSECTION] Cross section = 44.6 [44.597939917911440] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6011s + [COUNTERS] Fortran Overhead ( 0 ) : 1.5947s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0057s for 81920 events => throughput is 1.44E+07 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cuda (44.577527268256027) differ by less than 4E-4 (7.622674558227516e-08) +OK! xsec from fortran (44.597935381646522) and cuda (44.597939917911440) differ by less than 4E-4 (1.0171468423081365e-07) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.705094e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.913519e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.269887e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.219529e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.888199e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.576413e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.391800e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.411458e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.898622e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.568060e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.539526e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.538066e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.473018e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.173665e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:FLT+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 7.495430e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 7.445975e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) *** diff --git a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_m_inl0_hrd0.txt b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_m_inl0_hrd0.txt index 75d0c77429..873ea741da 100644 --- a/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_m_inl0_hrd0.txt +++ b/epochX/cudacpp/tmad/logs_susyggtt_mad/log_susyggtt_mad_m_inl0_hrd0.txt @@ -1,29 +1,29 @@ Working directory (build): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx -make USEBUILDDIR=1 BACKEND=cuda - +make USEBUILDDIR=1 BACKEND=cuda make USEBUILDDIR=1 BACKEND=cppnone -make USEBUILDDIR=1 BACKEND=cppsse4 +make USEBUILDDIR=1 BACKEND=cppsse4 make USEBUILDDIR=1 BACKEND=cppavx2 + make USEBUILDDIR=1 BACKEND=cpp512y make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' +make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make USEBUILDDIR=1 BACKEND=cpp512z make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' -make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Nothing to be done for 'all'. -make[1]: Nothing to be done for 'all'. -make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Nothing to be done for 'all'. +make[1]: Nothing to be done for 'all'. +make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' make[1]: Nothing to be done for 'all'. make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx' @@ -32,7 +32,7 @@ make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/ OMP_NUM_THREADS= -DATE: 2024-08-09_00:50:28 +DATE: 2024-09-02_17:31:53 On itscrd90.cern.ch [CPU: Intel(R) Xeon(R) Silver 4216 CPU] [GPU: 1x Tesla V100S-PCIE-32GB]: Working directory (run): /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/susy_gg_tt.mad/SubProcesses/P1_gg_ttx @@ -56,11 +56,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 2620 events (found 5403 events) - [COUNTERS] PROGRAM TOTAL : 0.8208s - [COUNTERS] Fortran Overhead ( 0 ) : 0.7796s - [COUNTERS] Fortran MEs ( 1 ) : 0.0412s for 8192 events => throughput is 1.99E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.8126s + [COUNTERS] Fortran Overhead ( 0 ) : 0.7701s + [COUNTERS] Fortran MEs ( 1 ) : 0.0425s for 8192 events => throughput is 1.93E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x1 (create events.lhe) *** -------------------- @@ -81,11 +81,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x1_fortran > /tmp/ [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860065419856] fbridge_mode=0 + [XSECTION] Cross section = 44.57 [44.566218791890641] fbridge_mode=0 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.4160s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3749s - [COUNTERS] Fortran MEs ( 1 ) : 0.0411s for 8192 events => throughput is 1.99E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4184s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3764s + [COUNTERS] Fortran MEs ( 1 ) : 0.0421s for 8192 events => throughput is 1.95E+05 events/s *** (1) EXECUTE MADEVENT_FORTRAN x10 (create events.lhe) *** -------------------- @@ -106,11 +106,11 @@ Executing ' ./madevent_fortran < /tmp/avalassi/input_susyggtt_x10_fortran > /tmp [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523870256456] fbridge_mode=0 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7104s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2559s - [COUNTERS] Fortran MEs ( 1 ) : 0.4544s for 90112 events => throughput is 1.98E+05 events/s + [XSECTION] Cross section = 44.6 [44.597935381646522] fbridge_mode=0 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6162s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1932s + [COUNTERS] Fortran MEs ( 1 ) : 0.4230s for 81920 events => throughput is 1.94E+05 events/s *** (2-none) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -131,16 +131,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598861353577519] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566220079105520] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) [COUNTERS] PROGRAM TOTAL : 0.4204s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3749s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0450s for 8192 events => throughput is 1.82E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3745s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0455s for 8192 events => throughput is 1.80E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598861353577519) differ by less than 2E-4 (2.888319694527297e-08) +OK! xsec from fortran (44.566218791890641) and cpp (44.566220079105520) differ by less than 2E-4 (2.888319694527297e-08) *** (2-none) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -165,16 +165,16 @@ Executing ' ./build.none_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577525144126803] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.7448s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2577s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.4867s for 90112 events => throughput is 1.85E+05 events/s + [XSECTION] Cross section = 44.6 [44.597936659329918] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6171s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1633s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.4535s for 81920 events => throughput is 1.81E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-none) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577525144126803) differ by less than 2E-4 (2.8576516486467085e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597936659329918) differ by less than 2E-4 (2.8648935890274174e-08) *** (2-none) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -183,12 +183,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.873127e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.841628e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.907422e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.858275e+05 ) sec^-1 *** (2-sse4) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -209,16 +209,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598861353577519] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566220079105520] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3960s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3712s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0244s for 8192 events => throughput is 3.36E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.4108s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3856s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0248s for 8192 events => throughput is 3.30E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598861353577519) differ by less than 2E-4 (2.888319694527297e-08) +OK! xsec from fortran (44.566218791890641) and cpp (44.566220079105520) differ by less than 2E-4 (2.888319694527297e-08) *** (2-sse4) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -243,16 +243,16 @@ Executing ' ./build.sse4_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577525144126810] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.5269s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2579s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2686s for 90112 events => throughput is 3.35E+05 events/s + [XSECTION] Cross section = 44.6 [44.597936659329932] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.4159s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1670s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2486s for 81920 events => throughput is 3.30E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-sse4) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577525144126810) differ by less than 2E-4 (2.857651670851169e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597936659329932) differ by less than 2E-4 (2.8648936334363384e-08) *** (2-sse4) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -261,12 +261,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.333942e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.275966e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/sse4+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.376975e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.354346e+05 ) sec^-1 *** (2-avx2) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -287,16 +287,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598861344883289] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566220070417643] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3926s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3769s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0153s for 8192 events => throughput is 5.37E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3982s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3821s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0157s for 8192 events => throughput is 5.21E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598861344883289) differ by less than 2E-4 (2.868825421664667e-08) +OK! xsec from fortran (44.566218791890641) and cpp (44.566220070417643) differ by less than 2E-4 (2.8688253994602064e-08) *** (2-avx2) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -321,16 +321,16 @@ Executing ' ./build.avx2_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577525178109212] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4173s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2508s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1662s for 90112 events => throughput is 5.42E+05 events/s + [XSECTION] Cross section = 44.6 [44.597936687258517] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3200s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1650s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1546s for 81920 events => throughput is 5.30E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-avx2) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577525178109212) differ by less than 2E-4 (2.9338838025694258e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597936687258517) differ by less than 2E-4 (2.9275166735232006e-08) *** (2-avx2) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -339,12 +339,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.335642e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.328517e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/avx2+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.330908e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.356875e+05 ) sec^-1 *** (2-512y) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -365,16 +365,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598861344883289] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566220070417643] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3897s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3750s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0143s for 8192 events => throughput is 5.74E+05 events/s + [COUNTERS] PROGRAM TOTAL : 0.3918s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3773s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0142s for 8192 events => throughput is 5.78E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512y) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598861344883289) differ by less than 2E-4 (2.868825421664667e-08) +OK! xsec from fortran (44.566218791890641) and cpp (44.566220070417643) differ by less than 2E-4 (2.8688253994602064e-08) *** (2-512y) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -399,16 +399,16 @@ Executing ' ./build.512y_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577525178109212] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4068s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2528s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.1536s for 90112 events => throughput is 5.87E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s + [XSECTION] Cross section = 44.6 [44.597936687258517] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3038s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1610s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.1426s for 81920 events => throughput is 5.75E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0003s *** (2-512y) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577525178109212) differ by less than 2E-4 (2.9338838025694258e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597936687258517) differ by less than 2E-4 (2.9275166735232006e-08) *** (2-512y) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -417,12 +417,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.855366e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.901217e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512y+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.947430e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.915938e+05 ) sec^-1 *** (2-512z) EXECUTE MADEVENT_CPP x1 (create events.lhe) *** -------------------- @@ -443,16 +443,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598861344883289] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566220070417643] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.3995s - [COUNTERS] Fortran Overhead ( 0 ) : 0.3772s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0219s for 8192 events => throughput is 3.75E+05 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0005s + [COUNTERS] PROGRAM TOTAL : 0.4010s + [COUNTERS] Fortran Overhead ( 0 ) : 0.3784s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0222s for 8192 events => throughput is 3.69E+05 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cpp (44.598861344883289) differ by less than 2E-4 (2.868825421664667e-08) +OK! xsec from fortran (44.566218791890641) and cpp (44.566220070417643) differ by less than 2E-4 (2.8688253994602064e-08) *** (2-512z) Compare MADEVENT_CPP x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -477,16 +477,16 @@ Executing ' ./build.512z_m_inl0_hrd0/madevent_cpp < /tmp/avalassi/input_susyggtt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577525178109212] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.4943s - [COUNTERS] Fortran Overhead ( 0 ) : 1.2580s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.2358s for 90112 events => throughput is 3.82E+05 events/s + [XSECTION] Cross section = 44.6 [44.597936687258517] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.3925s + [COUNTERS] Fortran Overhead ( 0 ) : 1.1694s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.2226s for 81920 events => throughput is 3.68E+05 events/s [COUNTERS] CudaCpp HEL ( 3 ) : 0.0004s *** (2-512z) Compare MADEVENT_CPP x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cpp (44.577525178109212) differ by less than 2E-4 (2.9338838025694258e-08) +OK! xsec from fortran (44.597935381646522) and cpp (44.597936687258517) differ by less than 2E-4 (2.9275166735232006e-08) *** (2-512z) Compare MADEVENT_CPP x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -495,12 +495,12 @@ OK! events.lhe.cpp.10 and events.lhe.ref.10 are identical *** EXECUTE CHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+BRDHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.733262e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.637908e+05 ) sec^-1 *** EXECUTE CHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CPP [gcc 11.3.1] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CPP:MIX+CXS:CURHST+RMBHST+MESHST/512z+CXVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.702855e+05 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.712874e+05 ) sec^-1 *** (3-cuda) EXECUTE MADEVENT_CUDA x1 (create events.lhe) *** -------------------- @@ -521,16 +521,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.6 [44.598860056955807] fbridge_mode=1 + [XSECTION] Cross section = 44.57 [44.566218783432781] fbridge_mode=1 [UNWEIGHT] Wrote 1603 events (found 1608 events) - [COUNTERS] PROGRAM TOTAL : 0.8053s - [COUNTERS] Fortran Overhead ( 0 ) : 0.8039s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.21E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [COUNTERS] PROGRAM TOTAL : 0.8133s + [COUNTERS] Fortran Overhead ( 0 ) : 0.8119s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0007s for 8192 events => throughput is 1.26E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x1 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.598860065419856) and cuda (44.598860056955807) differ by less than 2E-4 (1.8978174587402918e-10) +OK! xsec from fortran (44.566218791890641) and cuda (44.566218783432781) differ by less than 2E-4 (1.8978185689633165e-10) *** (3-cuda) Compare MADEVENT_CUDA x1 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -555,16 +555,16 @@ Executing ' ./build.cuda_m_inl0_hrd0/madevent_cuda < /tmp/avalassi/input_susyggt [XSECTION] MultiChannel = TRUE [XSECTION] Configuration = 1 [XSECTION] ChannelId = 1 - [XSECTION] Cross section = 44.58 [44.577523872560512] fbridge_mode=1 - [UNWEIGHT] Wrote 1743 events (found 1748 events) - [COUNTERS] PROGRAM TOTAL : 1.6927s - [COUNTERS] Fortran Overhead ( 0 ) : 1.6853s - [COUNTERS] CudaCpp MEs ( 2 ) : 0.0066s for 90112 events => throughput is 1.36E+07 events/s - [COUNTERS] CudaCpp HEL ( 3 ) : 0.0008s + [XSECTION] Cross section = 44.6 [44.597935383529340] fbridge_mode=1 + [UNWEIGHT] Wrote 1594 events (found 1599 events) + [COUNTERS] PROGRAM TOTAL : 1.6131s + [COUNTERS] Fortran Overhead ( 0 ) : 1.6059s + [COUNTERS] CudaCpp MEs ( 2 ) : 0.0064s for 81920 events => throughput is 1.28E+07 events/s + [COUNTERS] CudaCpp HEL ( 3 ) : 0.0007s *** (3-cuda) Compare MADEVENT_CUDA x10 xsec to MADEVENT_FORTRAN xsec *** -OK! xsec from fortran (44.577523870256456) and cuda (44.577523872560512) differ by less than 2E-4 (5.168643291142416e-11) +OK! xsec from fortran (44.597935381646522) and cuda (44.597935383529340) differ by less than 2E-4 (4.221756277900113e-11) *** (3-cuda) Compare MADEVENT_CUDA x10 events.lhe to MADEVENT_FORTRAN events.lhe reference (including colors and helicities) *** @@ -573,42 +573,42 @@ OK! events.lhe.cuda.10 and events.lhe.ref.10 are identical *** EXECUTE GCHECK(8192) -p 256 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.871837e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 2.000352e+07 ) sec^-1 *** EXECUTE GCHECK(8192) -p 256 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 5.622666e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 5.577798e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.299743e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.097038e+07 ) sec^-1 *** EXECUTE GCHECK(MAX) -p 16384 32 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.055606e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.063106e+08 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.302003e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.100716e+07 ) sec^-1 *** EXECUTE GCHECK(MAX128THR) -p 4096 128 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 1.140289e+08 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 1.138806e+08 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 --bridge *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURHST+RMBHST+BRDDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 3.319830e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.118534e+07 ) sec^-1 *** EXECUTE GCHECK(MAX8THR) -p 65536 8 1 *** Process = SIGMA_MSSM_SLHA2_GG_TTX_CUDA [nvcc 12.0.140 (gcc 11.3.1)] [inlineHel=0] [hardcodePARAM=0] Workflow summary = CUD:MIX+THX:CURDEV+RMBDEV+MESDEV/none+NAVBRK -EvtsPerSec[MECalcOnly] (3a) = ( 2.983678e+07 ) sec^-1 +EvtsPerSec[MECalcOnly] (3a) = ( 3.038098e+07 ) sec^-1 *** (3-hip) WARNING! SKIP MADEVENT_HIP (hip is not supported on this node) ***