[goodhel+june24] move to generated processes (except gg_tt.mad) from …

…the latest upstream/master for easier merging git checkout upstream/master $(git ls-tree --name-only upstream/master *.mad | grep -v ^gg_tt.mad)
valassi · Sep 3, 2024 · f2f3fc2 · f2f3fc2
1 parent 585f94b
commit f2f3fc2
Show file tree

Hide file tree

Showing 1,016 changed files with 313,174 additions and 88,205 deletions.
diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/me5_configuration.txt b/epochX/cudacpp/ee_mumu.mad/Cards/me5_configuration.txt
@@ -116,7 +116,6 @@
 # cluster_type = condor
 # cluster_queue = madgraph
 # cluster_size = 150 
-# cluster_walltime =  # time in minute for slurm and second for condor (not supported for other scheduller)
 
 #! Path to a node directory to avoid direct writing on the central disk
 #!  Note that condor clusters avoid direct writing by default (therefore this

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/param_card.dat b/epochX/cudacpp/ee_mumu.mad/Cards/param_card.dat
@@ -1,5 +1,5 @@
 ######################################################################
-## PARAM_CARD AUTOMATICALLY GENERATED BY MG5 FOLLOWING UFO MODEL   ####
+## PARAM_CARD AUTOMATICALY GENERATED BY MG5 FOLLOWING UFO MODEL   ####
 ######################################################################
 ##                                                                  ##
 ##  Width set on Auto will be computed following the information    ##

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/param_card_default.dat b/epochX/cudacpp/ee_mumu.mad/Cards/param_card_default.dat
@@ -1,5 +1,5 @@
 ######################################################################
-## PARAM_CARD AUTOMATICALLY GENERATED BY MG5 FOLLOWING UFO MODEL   ####
+## PARAM_CARD AUTOMATICALY GENERATED BY MG5 FOLLOWING UFO MODEL   ####
 ######################################################################
 ##                                                                  ##
 ##  Width set on Auto will be computed following the information    ##

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/proc_card_mg5.dat b/epochX/cudacpp/ee_mumu.mad/Cards/proc_card_mg5.dat
@@ -8,7 +8,7 @@
 #*                *                       *                 *
 #*                                                          *
 #*                                                          *
-#*         VERSION 3.6.0_lo_vect         2024-06-17         *
+#*         VERSION 3.5.3_lo_vect         2023-12-23         *
 [1;31m#*                                                          *[1;0m
 [1;31m#*          WARNING: UNKNOWN DEVELOPMENT VERSION.           *[1;0m
 [1;31m#*            WARNING: DO NOT USE FOR PRODUCTION            *[1;0m

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/reweight_card_default.dat b/epochX/cudacpp/ee_mumu.mad/Cards/reweight_card_default.dat
@@ -19,16 +19,8 @@
 
 change mode NLO    # Define type of Reweighting. For LO sample this command
                    # has no effect since only "LO" mode is allowed.
-
-#uncomment if you do not want to overwrite the reweight file of Sudakov in rw_me
-#change rwgt_dir /PATH_MG5_BRANCH/NAME_FOLDER
-
-#uncomment if you want to use Sudakov Reweight
-#change include_sudakov True
-
 
 launch
-
 #  SPECIFY A PATH OR USE THE SET COMMAND LIKE THIS:
 # set sminputs 1 130 # modify 1/alpha_EW
 

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/run_card.dat b/epochX/cudacpp/ee_mumu.mad/Cards/run_card.dat
@@ -100,7 +100,8 @@
    -O3 -ffast-math -fbounds-check = global_flag ! build flags for all Fortran code (for a fair comparison to cudacpp; default is -O)
    --fast-math  = aloha_flag ! fortran optimization flag for aloha function. Suggestions: '-ffast-math'
    -O3 = matrix_flag ! fortran optimization flag for matrix.f function. Suggestions: '-O3'
-   16384 = vector_size ! size of fortran arrays allocated in the multi-event API for SIMD/GPU (VECSIZE_MEMMAX)
+   32 = vector_size ! size of fortran arrays allocated in the multi-event API for SIMD/GPU (VECSIZE_MEMMAX)
+   512 = nb_warp ! total number of warp/frontwave
 
 #*********************************************************************
 # Customization (custom cuts/scale/bias/...)                         *
@@ -144,7 +145,7 @@
 # Maximum and minimum absolute rapidity (for max, -1 means no cut)   *
 #*********************************************************************
  2.5  = etal    ! max rap for the charged leptons 
- 0.0  = etalmin ! min rap for the charged leptons
+ 0.0  = etalmin ! main rap for the charged leptons
  {} = eta_min_pdg ! rap cut for other particles (use pdg code). Applied on particle and anti-particle
  {} = eta_max_pdg ! rap cut for other particles (syntax e.g. {6: 2.5, 23: 5})
 #*********************************************************************
@@ -198,7 +199,6 @@
 systematics = systematics_program ! none, systematics [python], SysCalc [depreceted, C++]
 ['--mur=0.5,1,2', '--muf=0.5,1,2', '--pdf=errorset'] = systematics_arguments ! see: https://cp3.irmp.ucl.ac.be/projects/madgraph/wiki/Systematics#Systematicspythonmodule
 
-
 #***********************************************************************
 # SIMD/GPU configuration for the CUDACPP plugin
 #************************************************************************

diff --git a/epochX/cudacpp/ee_mumu.mad/Cards/run_card_default.dat b/epochX/cudacpp/ee_mumu.mad/Cards/run_card_default.dat
@@ -101,6 +101,7 @@
    --fast-math  = aloha_flag ! fortran optimization flag for aloha function. Suggestions: '-ffast-math'
    -O3 = matrix_flag ! fortran optimization flag for matrix.f function. Suggestions: '-O3'
    16 = vector_size ! size of fortran arrays allocated in the multi-event API for SIMD/GPU (VECSIZE_MEMMAX)
+   1 = nb_warp ! total number of warp/frontwave
 
 #*********************************************************************
 # Customization (custom cuts/scale/bias/...)                         *
@@ -144,7 +145,7 @@
 # Maximum and minimum absolute rapidity (for max, -1 means no cut)   *
 #*********************************************************************
  2.5  = etal    ! max rap for the charged leptons 
- 0.0  = etalmin ! min rap for the charged leptons
+ 0.0  = etalmin ! main rap for the charged leptons
  {} = eta_min_pdg ! rap cut for other particles (use pdg code). Applied on particle and anti-particle
  {} = eta_max_pdg ! rap cut for other particles (syntax e.g. {6: 2.5, 23: 5})
 #*********************************************************************
@@ -198,7 +199,6 @@
 systematics = systematics_program ! none, systematics [python], SysCalc [depreceted, C++]
 ['--mur=0.5,1,2', '--muf=0.5,1,2', '--pdf=errorset'] = systematics_arguments ! see: https://cp3.irmp.ucl.ac.be/projects/madgraph/wiki/Systematics#Systematicspythonmodule
 
-
 #***********************************************************************
 # SIMD/GPU configuration for the CUDACPP plugin
 #************************************************************************

diff --git a/epochX/cudacpp/ee_mumu.mad/MGMEVersion.txt b/epochX/cudacpp/ee_mumu.mad/MGMEVersion.txt
@@ -1 +1 @@
-3.6.0_lo_vect
+3.5.3_lo_vect
diff --git a/epochX/cudacpp/ee_mumu.mad/Source/DHELAS/aloha_functions.f b/epochX/cudacpp/ee_mumu.mad/Source/DHELAS/aloha_functions.f
@@ -351,7 +351,7 @@ subroutine oxxxxx(p,fmass,nhel,nsf , fo)
             fo(6) = ip     * sqm(abs(im))
          else
 
-c            pp = min(p(0),dsqrt(p(1)**2+p(2)**2+p(3)**2))
+            pp = min(p(0),dsqrt(p(1)**2+p(2)**2+p(3)**2))
             sf(1) = dble(1+nsf+(1-nsf)*nh)*rHalf
             sf(2) = dble(1+nsf-(1-nsf)*nh)*rHalf
             omega(1) = dsqrt(p(0)+pp)
@@ -2054,7 +2054,7 @@ subroutine CombineAmp(nb, ihels, iwfcts, W1, Wall, Amp)
       enddo
       return
       end
-
+      
       subroutine CombineAmpS(nb, ihels, iwfcts, W1, Wall, Amp)
 
       integer nb ! size of the vectors

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/MODEL/couplings.f b/epochX/cudacpp/ee_mumu.mad/Source/MODEL/couplings.f
@@ -18,16 +18,10 @@ SUBROUTINE COUP()
       READLHA = .TRUE.
       INCLUDE 'intparam_definition.inc'
       CALL COUP1()
-      IF (UPDATELOOP) THEN
-
-        CALL COUP2()
-
-      ENDIF
-
 C     
 couplings needed to be evaluated points by points
 C     
-      CALL COUP3(1)
+      CALL COUP2(1)
 
       RETURN
       END
@@ -71,7 +65,7 @@ SUBROUTINE UPDATE_AS_PARAM(VECID)
 couplings needed to be evaluated points by points
 C     
       ALL_G(VECID) = G
-      CALL COUP3(VECID)
+      CALL COUP2(VECID)
 
       RETURN
       END

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/MODEL/couplings2.f b/epochX/cudacpp/ee_mumu.mad/Source/MODEL/couplings2.f
@@ -2,10 +2,10 @@
 c      written by the UFO converter
 ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc
 
-      SUBROUTINE COUP2( )
+      SUBROUTINE COUP2( VECID)
 
       IMPLICIT NONE
-
+      INTEGER VECID
       INCLUDE 'model_functions.inc'
 
       DOUBLE PRECISION PI, ZERO

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/MODEL/makeinc.inc b/epochX/cudacpp/ee_mumu.mad/Source/MODEL/makeinc.inc
@@ -2,4 +2,4 @@
 #      written by the UFO converter
 #############################################################################
 
-MODEL = couplings.o lha_read.o printout.o rw_para.o model_functions.o couplings1.o couplings2.o couplings3.o
+MODEL = couplings.o lha_read.o printout.o rw_para.o model_functions.o couplings1.o couplings2.o
diff --git a/epochX/cudacpp/ee_mumu.mad/Source/dsample.f b/epochX/cudacpp/ee_mumu.mad/Source/dsample.f
@@ -37,8 +37,12 @@ subroutine sample_full(ndim,ncall,itmax,itmin,dsig,ninvar,nconfigs,VECSIZE_USED)
       integer jmax,i,j,ipole
       integer itmax_adjust
 
-      integer imirror, iproc, iconf
-      integer ivec ! position of the event in the vector (max is VECSIZE_MEMMAX, loops go over VECSIZE_USED)
+c      integer imirror, iproc, iconf
+      integer imirror_vec(NB_WARP), iproc, ICONF_VEC(NB_WARP) 
+      integer ivec              ! position of the event in the vector (max is VECSIZE_MEMMAX, loops go over VECSIZE_USED)
+      integer ilock             !  position of the event in the current warp (max is WARP_SIZE)
+      integer iwarp               ! position of the current warp (max is NB_WARP)
+c     NOTE THAT IVEC = (IWARP-1)*NB_WARP + ILOCK      
 
 c
 c     External
@@ -161,6 +165,8 @@ subroutine sample_full(ndim,ncall,itmax,itmin,dsig,ninvar,nconfigs,VECSIZE_USED)
       ievent = 0
       iter = 1
       ivec = 0
+      ilock = 0
+      iwarp = 1
       do while(iter .le. itmax)
 c
 c     Get integration point
@@ -174,6 +180,11 @@ subroutine sample_full(ndim,ncall,itmax,itmin,dsig,ninvar,nconfigs,VECSIZE_USED)
             CUTSPASSED=.FALSE.
             if (passcuts(p,VECSIZE_USED)) then
                ivec=ivec+1
+               ilock = ilock+1
+               if (ilock.gt.WARP_SIZE)then
+                  ilock = 1
+                  iwarp = iwarp +1
+               endif
 c              write(*,*) 'pass_point ivec is ', ivec
                all_p(:,ivec) = p(:)
                all_wgt(ivec) = wgt
@@ -187,25 +198,35 @@ subroutine sample_full(ndim,ncall,itmax,itmin,dsig,ninvar,nconfigs,VECSIZE_USED)
 c               bckp(i) = fx
 c               write(*,*) i, all_wgt(i), fx, all_wgt(i)*fx
 c               all_wgt(i) = all_wgt(i)*fx
-               if (ivec.lt.VECSIZE_USED)then
+               if (ilock.ne.WARP_SIZE)then
                   cycle
                endif
-               ivec=0
+
                if (VECSIZE_USED.le.1) then
                   all_fx(1) = dsig(all_p, all_wgt,0)
                else
-               do i=1, VECSIZE_USED
+c                 Here "i" is the position in the full grid of the event                  
+                  do i=(iwarp-1)*WARP_SIZE+1, iwarp*warp_size
+
 c                 need to restore common block                  
                   xbk(:) = all_xbk(:, i)
                   cm_rap = all_cm_rap(i)
                   q2fact(:) = all_q2fact(:,i)
                   CUTSDONE=.TRUE.
                   CUTSPASSED=.TRUE.
-                  call prepare_grouping_choice(all_p(1,i), all_wgt(i), i.eq.1)
+                  call prepare_grouping_choice(all_p(1,i), all_wgt(i),i.eq.(iwarp-1)*WARP_SIZE+1)
                enddo
-               call select_grouping(imirror, iproc, iconf, all_wgt, VECSIZE_USED)
+               call select_grouping(imirror_vec(iwarp), iproc, iconf_vec(iwarp), all_wgt, iwarp)
+               if (ivec.lt.VECSIZE_USED)then
+                  cycle
+               endif
+c              reset variable for the next grid               
+               ivec = 0
+               ilock = 0
+               iwarp =1
+
                call dsig_vec(all_p, all_wgt, all_xbk, all_q2fact, all_cm_rap,
-     &                          iconf, iproc, imirror, all_fx,VECSIZE_USED)
+     &                          iconf_vec, iproc, imirror_vec, all_fx,VECSIZE_USED)
 
                 do i=1, VECSIZE_USED
 c                 need to restore common block                  

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/genps.inc b/epochX/cudacpp/ee_mumu.mad/Source/genps.inc
@@ -29,8 +29,9 @@ c      parameter (max_host=99,maxplace=199,maxpoints=100,maxans=50)
 c*************************************************************************
 c	Parameters for helicity sums in matrixN.f
 c*************************************************************************
-c      REAL*8 LIMHEL
-c      PARAMETER(LIMHEL=1e-8) -> pass in the run_card.dat
+      REAL*8 LIMHEL
+c     PARAMETER(LIMHEL=1e-8) ! ME threshold for helicity filtering (Fortran default)
+      PARAMETER(LIMHEL=0) ! ME threshold for helicity filtering (force Fortran to mimic cudacpp, see #419)
       INTEGER MAXTRIES
       PARAMETER(MAXTRIES=25)
 C     To pass the helicity configuration chosen by the DiscreteSampler to 

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/run.inc b/epochX/cudacpp/ee_mumu.mad/Source/run.inc
@@ -106,7 +106,4 @@ c
        double precision tmin_for_channel
        integer sde_strat ! 1 means standard single diagram enhancement strategy,
 c                          2 means approximation by the denominator of the propagator
-       common/TO_CHANNEL_STRAT/tmin_for_channel, sde_strat
-c
-       double precision limhel
-       common/to_limhel/limhel
+       common/TO_CHANNEL_STRAT/tmin_for_channel, sde_strat
diff --git a/epochX/cudacpp/ee_mumu.mad/Source/run_card.inc b/epochX/cudacpp/ee_mumu.mad/Source/run_card.inc
@@ -88,8 +88,6 @@
 
       DSQRT_SHAT = 0.000000000000000D+00
 
-      LIMHEL = 0.000000000000000D+00
-
       PTJ = 2.000000000000000D+01
 
       PTB = 0.000000000000000D+00

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/setrun.f b/epochX/cudacpp/ee_mumu.mad/Source/setrun.f
@@ -162,21 +162,9 @@ subroutine setrun
 C       Fill common block for Les Houches init info
       do i=1,2
         if(lpp(i).eq.1.or.lpp(i).eq.2) then
-          if (nb_proton(i).eq.1.and.nb_neutron(i).eq.0) then
-              idbmup(i)=2212
-          elseif (nb_proton(i).eq.0.and.nb_neutron(i).eq.1) then
-              idbmup(i)=2112
-          else
-              idbmup(i) = 1000000000 + (nb_proton(i)+nb_neutron(i))*10
-     $                               + nb_proton(i)*10000
-          endif
+          idbmup(i)=2212
         elseif(lpp(i).eq.-1.or.lpp(i).eq.-2) then
-          if (nb_proton(i).eq.1.and.nb_neutron(i).eq.0) then
-              idbmup(i)=-2212
-          else
-              idbmup(i) = -1*(1000000000 + (nb_proton(i)+nb_neutron(i))*10
-     $                                    + nb_proton(i)*10000)
-          endif
+          idbmup(i)=-2212
         elseif(lpp(i).eq.3) then
           idbmup(i)=11
         elseif(lpp(i).eq.-3) then

diff --git a/epochX/cudacpp/ee_mumu.mad/Source/vector.inc b/epochX/cudacpp/ee_mumu.mad/Source/vector.inc
@@ -27,5 +27,9 @@ C     NB: THIS FILE CANNOT CONTAIN #ifdef DIRECTIVES
 C     BECAUSE IT DOES NOT GO THROUGH THE CPP PREPROCESSOR
 C     (see https://github.com/madgraph5/madgraph4gpu/issues/458).
 C     
+      INTEGER WARP_SIZE
+      PARAMETER (WARP_SIZE=32)
+      INTEGER NB_WARP
+      PARAMETER (NB_WARP=512)
       INTEGER VECSIZE_MEMMAX
       PARAMETER (VECSIZE_MEMMAX=16384)