Merge branch 'develop' into warning

.github/workflows/test.yml

@@ -29,7 +29,7 @@ jobs:
 
       - name: Build
         run: |
-          cmake -B build -DBUILD_TESTING=ON -DENABLE_DEEPKS=ON -DENABLE_LIBXC=ON -DENABLE_LIBRI=ON -DENABLE_PAW=ON
+          cmake -B build -DBUILD_TESTING=ON -DENABLE_DEEPKS=ON -DENABLE_LIBXC=ON -DENABLE_LIBRI=ON -DENABLE_PAW=ON -DENABLE_GOOGLEBENCH=ON
           cmake --build build -j8
           cmake --install build
 

CMakeLists.txt

@@ -34,6 +34,7 @@ option(DEBUG_INFO "Print message for developers to debug." OFF)
 option(ENABLE_NATIVE_OPTIMIZATION "Enable compilation optimization for the native machine's CPU type" OFF)
 option(COMMIT_INFO "Print commit information in log" ON)
 option(ENABLE_FFT_TWO_CENTER "Enable FFT-based two-center integral method." ON)
+option(ENABLE_GOOGLEBENCH "Enable GOOGLE-benchmark usage." OFF)
 option(ENABLE_RAPIDJSON "Enable rapid-json usage." OFF)
 
 
@@ -428,8 +429,7 @@ endif()
 
 if(ENABLE_DEEPKS)
   # Torch uses outdated components to detech CUDA arch, causing failure on latest CUDA kits.
-  # See above for setting CMAKE_CUDA_ARCHITECTURES
-  set(TORCH_CUDA_ARCH_LIST CMAKE_CUDA_ARCHITECTURES)
+  # Set CMake variable TORCH_CUDA_ARCH_LIST in the form of "major.minor" if required.
   find_package(Torch REQUIRED)
   if(NOT Torch_VERSION VERSION_LESS "2.1.0")
     set_if_higher(CMAKE_CXX_STANDARD 17)
@@ -536,11 +536,8 @@ if(DEFINED Libxc_DIR)
   set(ENABLE_LIBXC ON)
 endif()
 if(ENABLE_LIBXC)
-  find_package(Libxc REQUIRED HINTS
-    ${Libxc_DIR}/share/cmake/Libxc
-    ${Libxc_DIR}/lib/cmake/Libxc
-    ${Libxc_DIR}/lib64/cmake/Libxc
-  )
+  # use `cmake/FindLibxc.cmake` to detect Libxc installation with `pkg-config`
+  find_package(Libxc REQUIRED)
   message(STATUS "Found Libxc: version " ${Libxc_VERSION})
   if(${Libxc_VERSION} VERSION_LESS 5.1.7)
     message(FATAL_ERROR "LibXC >= 5.1.7 is required.")
@@ -594,6 +591,25 @@ if(INFO)
   # modifications on blas_connector and lapack_connector
 endif()
 
+#  Add performance test in abacus
+IF (ENABLE_GOOGLEBENCH)
+  set(BUILD_TESTING ON)
+  find_package(benchmark HINTS ${BENCHMARK_DIR})
+  if(NOT ${benchmark_FOUND})
+    set(BENCHMARK_USE_BUNDLED_GTEST OFF)
+    include(FetchContent)
+    FetchContent_Declare(
+      benchmark
+      GIT_REPOSITORY https://github.com/google/benchmark.git
+      GIT_TAG "origin/main"
+      GIT_SHALLOW TRUE
+      GIT_PROGRESS TRUE 
+    )
+    set(BENCHMARK_ENABLE_TESTING OFF)
+    FetchContent_MakeAvailable(benchmark)
+  endif()
+endif()
+
 IF (BUILD_TESTING)
   set_if_higher(CMAKE_CXX_STANDARD 14) # Required in orbital
   include(CTest)
@@ -623,8 +639,14 @@ IF (BUILD_TESTING)
     endif()
 
     #dependencies & link library
-    target_link_libraries(${UT_TARGET} ${UT_LIBS}
-      Threads::Threads GTest::gtest_main GTest::gmock_main)
+    if(ENABLE_GOOGLEBENCH)
+      target_link_libraries(${UT_TARGET} ${UT_LIBS}
+        Threads::Threads GTest::gtest_main GTest::gmock_main benchmark::benchmark)
+      else()
+        target_link_libraries(${UT_TARGET} ${UT_LIBS}
+          Threads::Threads GTest::gtest_main GTest::gmock_main)
+    endif()
+
     if(USE_OPENMP)
       target_link_libraries(${UT_TARGET} OpenMP::OpenMP_CXX)
     endif()
@@ -634,6 +656,7 @@ IF (BUILD_TESTING)
       WORKING_DIRECTORY $<TARGET_FILE_DIR:${UT_TARGET}>
     )
   endfunction(AddTest)
+
 endif()
 
 add_subdirectory(source)

Dockerfile.cuda

@@ -2,7 +2,7 @@ FROM nvidia/cuda:12.2.0-devel-ubuntu22.04
 
 RUN apt update && apt install -y --no-install-recommends \
     libopenblas-openmp-dev liblapack-dev libscalapack-mpi-dev libelpa-dev libfftw3-dev libcereal-dev \
-    libxc-dev libgtest-dev libgmock-dev python3-numpy \
+    libxc-dev libgtest-dev libgmock-dev libbenchmark-dev python3-numpy \
     bc cmake git g++ make bc time sudo unzip vim wget
 
 ENV GIT_SSL_NO_VERIFY=true TERM=xterm-256color \

Dockerfile.gnu

@@ -1,7 +1,7 @@
 FROM ubuntu:22.04
 RUN apt update && apt install -y --no-install-recommends \
     libopenblas-openmp-dev liblapack-dev libscalapack-mpi-dev libelpa-dev libfftw3-dev libcereal-dev \
-    libxc-dev libgtest-dev libgmock-dev python3-numpy \
+    libxc-dev libgtest-dev libgmock-dev libbenchmark-dev python3-numpy \
     bc cmake git g++ make bc time sudo unzip vim wget gfortran
 
 ENV GIT_SSL_NO_VERIFY=true TERM=xterm-256color \

Dockerfile.intel

@@ -2,7 +2,7 @@ FROM ubuntu:22.04
 
 RUN apt-get update && apt-get install -y \
     bc cmake git gnupg gcc g++ python3-numpy sudo wget vim unzip \
-    libcereal-dev libxc-dev libgtest-dev libgmock-dev
+    libcereal-dev libxc-dev libgtest-dev libgmock-dev libbenchmark-dev
 
 # Following steps by https://software.intel.com/content/www/us/en/develop/documentation/installation-guide-for-intel-oneapi-toolkits-linux/top/installation/install-using-package-managers/apt.html .
 RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \

cmake/FindELPA.cmake

@@ -7,34 +7,50 @@
 #  ELPA_INCLUDE_DIR - Where to find ELPA headers.
 #
 
-find_path(ELPA_INCLUDE_DIR
+find_package(PkgConfig)
+
+find_path(ELPA_INCLUDE_DIRS
     elpa/elpa.h
     HINTS ${ELPA_DIR}
     PATH_SUFFIXES "include" "include/elpa"
     )
 if(USE_OPENMP)
-    find_library(ELPA_LIBRARY
-        NAMES elpa_openmp elpa
-        HINTS ${ELPA_DIR}
-        PATH_SUFFIXES "lib"
-        )
+    find_library(ELPA_LINK_LIBRARIES
+    NAMES elpa_openmp elpa
+    HINTS ${ELPA_DIR}
+    PATH_SUFFIXES "lib"
+    )
 else()
-    find_library(ELPA_LIBRARY
-        NAMES elpa
-        HINTS ${ELPA_DIR}
-        PATH_SUFFIXES "lib"
-        )
+    find_library(ELPA_LINK_LIBRARIES
+    NAMES elpa
+    HINTS ${ELPA_DIR}
+    PATH_SUFFIXES "lib"
+    )
+endif()
+
+if(NOT ELPA_INCLUDE_DIRS AND PKG_CONFIG_FOUND)
+  if(DEFINED ELPA_DIR)
+    string(APPEND CMAKE_PREFIX_PATH ";${ELPA_DIR}")
+  endif()
+  if(USE_OPENMP)
+    pkg_search_module(ELPA REQUIRED IMPORTED_TARGET GLOBAL elpa_openmp)
+  else()
+    pkg_search_module(ELPA REQUIRED IMPORTED_TARGET GLOBAL elpa)
+  endif()
+elseif(NOT PKG_CONFIG_FOUND)
+  message(
+    "ELPA : We need pkg-config to get all information about the elpa library")
 endif()
 
 # Handle the QUIET and REQUIRED arguments and
 # set ELPA_FOUND to TRUE if all variables are non-zero.
 include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(ELPA DEFAULT_MSG ELPA_LIBRARY ELPA_INCLUDE_DIR)
+find_package_handle_standard_args(ELPA DEFAULT_MSG ELPA_LINK_LIBRARIES ELPA_INCLUDE_DIRS)
 
 # Copy the results to the output variables and target.
 if(ELPA_FOUND)
-    set(ELPA_LIBRARIES ${ELPA_LIBRARY})
-    set(ELPA_INCLUDE_DIR ${ELPA_INCLUDE_DIR})
+    list(GET ELPA_LINK_LIBRARIES 0 ELPA_LIBRARY)
+    set(ELPA_INCLUDE_DIR ${ELPA_INCLUDE_DIRS})
 
     if(NOT TARGET ELPA::ELPA)
         add_library(ELPA::ELPA UNKNOWN IMPORTED)

cmake/FindLibxc.cmake

@@ -0,0 +1,36 @@
+include(FindPackageHandleStandardArgs)
+
+if(DEFINED Libxc_DIR)
+  string(APPEND CMAKE_PREFIX_PATH ";${Libxc_DIR}")
+endif()
+# Using CMake interface as default.
+# NO REQUIRED here, otherwhile it would throw error
+# with no LibXC found.
+find_package(Libxc HINTS
+    ${Libxc_DIR}/share/cmake/Libxc
+    ${Libxc_DIR}/lib/cmake/Libxc
+    ${Libxc_DIR}/lib64/cmake/Libxc
+  )
+if(NOT TARGET Libxc::xc)
+  find_package(PkgConfig REQUIRED)
+  pkg_search_module(Libxc REQUIRED IMPORTED_TARGET GLOBAL libxc)
+  find_package_handle_standard_args(Libxc DEFAULT_MSG Libxc_LINK_LIBRARIES Libxc_INCLUDE_DIRS)
+endif()
+
+
+# Copy the results to the output variables and target.
+# if find_package() above works, Libxc::xc would be present and
+# below would be skipped.
+if(Libxc_FOUND AND NOT TARGET Libxc::xc)
+	set(Libxc_LIBRARY ${Libxc_LINK_LIBRARIES})
+	set(Libxc_LIBRARIES ${Libxc_LIBRARY})
+	set(Libxc_INCLUDE_DIR ${Libxc_INCLUDE_DIRS})
+	add_library(Libxc::xc UNKNOWN IMPORTED)
+	set_target_properties(Libxc::xc PROPERTIES
+		IMPORTED_LOCATION "${Libxc_LIBRARY}"
+		INTERFACE_INCLUDE_DIRECTORIES "${Libxc_INCLUDE_DIR}")
+endif()
+
+set(CMAKE_REQUIRED_INCLUDES ${CMAKE_REQUIRED_INCLUDES} ${Libxc_INCLUDE_DIR})
+
+mark_as_advanced(Libxc_INCLUDE_DIR Libxc_LIBRARY)

docs/advanced/input_files/input-main.md

@@ -72,6 +72,7 @@
     - [mixing\_beta](#mixing_beta)
     - [mixing\_beta\_mag](#mixing_beta_mag)
     - [mixing\_ndim](#mixing_ndim)
+    - [mixing\_restart](#mixing_restart)
     - [mixing\_gg0](#mixing_gg0)
     - [mixing\_gg0\_mag](#mixing_gg0_mag)
     - [mixing\_gg0\_min](#mixing_gg0_min)
@@ -145,6 +146,8 @@
     - [out\_app\_flag](#out_app_flag)
     - [out\_ndigits](#out_ndigits)
     - [out\_interval](#out_interval)
+    - [band\_print\_num](#band_print_num)
+    - [bands\_to\_print](#bands_to_print)
     - [out\_element\_info](#out_element_info)
     - [restart\_save](#restart_save)
     - [restart\_load](#restart_load)
@@ -945,6 +948,8 @@ calculations.
   - **fixed**: fixed occupations (available for non-coductors only)
   - **gauss** or **gaussian**: Gaussian smearing method.
   - **mp**: methfessel-paxton smearing method; recommended for metals.
+  - **mp2**: 2-nd methfessel-paxton smearing method; recommended for metals.
+  - **mv** or **cold**: marzari-vanderbilt smearing method.
   - **fd**: Fermi-Dirac smearing method: $f=1/\{1+\exp[(E-\mu)/kT]\}$ and smearing_sigma below is the temperature $T$ (in Ry).
 - **Default**: gauss
 
@@ -1001,6 +1006,13 @@ We recommend the following options:
   For systems that are difficult to converge, one could try increasing the value of 'mixing_ndim' to enhance the stability of the self-consistent field (SCF) calculation.
 - **Default**: 8
 
+### mixing_restart
+
+- **Type**: Integer
+- **Description**: At `mixing_restart`-th iteration, SCF will restart by using output charge density from perivos iteration as input charge density directly, and start a new mixing. `mixing_restart=0|1` means SCF starts from scratch.
+
+- **Default**: 0
+
 ### mixing_gg0
 
 - **Type**: Real
@@ -1494,8 +1506,8 @@ These variables are used to control the output of properties.
 
 ### out_band
 
-- **Type**: Boolean
-- **Description**: Whether to output the band structure (in eV). For more information, refer to the [band.md](../elec_properties/band.md)
+- **Type**: Boolean Integer(optional)
+- **Description**: Whether to output the band structure (in eV), optionally output precision can be set by a second parameter, default is 8. For more information, refer to the [band.md](../elec_properties/band.md)
 - **Default**: False
 
 ### out_proj_band
@@ -1599,6 +1611,20 @@ These variables are used to control the output of properties.
 - **Description**: Control the interval for printing Mulliken population analysis, $r(R)$, $H(R)$, $S(R)$, $T(R)$, $dH(R)$, $H(k)$, $S(k)$ and $wfc(k)$ matrices during molecular dynamics calculations. Check input parameters [out_mul](#out_mul), [out_mat_r](#out_mat_r), [out_mat_hs2](#out_mat_hs2), [out_mat_t](#out_mat_t), [out_mat_dh](#out_mat_dh), [out_mat_hs](#out_mat_hs) and [out_wfc_lcao](#out_wfc_lcao) for more information, respectively.
 - **Default**: 1
 
+### band_print_num
+
+- **Type**: Integer
+- **Availability**: PW basis
+- **Description**: If you want to plot a partial charge density contributed from some chosen bands. `band_print_num` define the number of band list. The result can be found in "band*.cube".
+- **Default**: 0
+
+### bands_to_print
+
+- **Type**: vector
+- **Availability**: band_print_num > 0
+- **Description**: define which band you want to choose for partial charge density.
+- **Default**: []
+
 ### out_element_info
 
 - **Type**: Boolean
@@ -2776,9 +2802,9 @@ These variables are used to control berry phase and wannier90 interface paramete
 
 - **Type**: String
 - **Description**: the spin direction for the Wannier function calculation when nspin is set to 2
-  - "up": Calculate spin up for the Wannier function.
-  - "down": Calculate spin down for the Wannier function.
-- **Default**: "up"
+  - `up`: Calculate spin up for the Wannier function.
+  - `down`: Calculate spin down for the Wannier function.
+- **Default**: `up`
 
 ### out_wannier_mmn
 
@@ -2818,6 +2844,7 @@ These variables are used to control berry phase and wannier90 interface paramete
 - **Description**: write the "UNK.*" file in ASCII format or binary format.
   - 0: write the "UNK.*" file in binary format.
   - 1: write the "UNK.*" file in ASCII format (text file format).
+- **Default**: 1
 
 [back to top](#full-list-of-input-keywords)
 

docs/advanced/install.md

@@ -69,6 +69,16 @@ After building and installing, unit tests can be performed with `ctest`.
 
 To run a subset of unit test, use `ctest -R <test-match-pattern>` to perform tests with name matched by given pattern.
 
+## Build Performance Tests
+
+To build performance tests for ABACUS, define `ENABLE_GOOGLEBENCH` flag. You can also specify the path to a local installation of [Google Benchmark](https://github.com/google/benchmark.git) by setting `BENCHMARK_DIR` flags. If not found locally, the configuration process will try to download it automatically.
+
+```bash
+cmake -B build -DENABLE_GOOGLEBENCH=1
+```
+
+Google Benchmark requires Google Test to build and run the tests. When setting `ENABLE_GOOGLEBENCH` to ON, `BUILD_TESTING` is automatically enabled. After building and installing, performance tests can be executed with `ctest`.
+
 ## Build with CUDA support
 
 ### Extra prerequisites

docs/quick_start/easy_install.md

@@ -28,7 +28,7 @@ These requirements support the calculation of plane-wave basis in ABACUS. For LC
 Some of these packages can be installed with popular package management system, such as `apt` and `yum`:
 
 ```bash
-sudo apt update && sudo apt install -y libopenblas-openmp-dev liblapack-dev libscalapack-mpi-dev libelpa-dev libfftw3-dev libcereal-dev libxc-dev g++ make cmake bc git
+sudo apt update && sudo apt install -y libopenblas-openmp-dev liblapack-dev libscalapack-mpi-dev libelpa-dev libfftw3-dev libcereal-dev libxc-dev g++ make cmake bc git pkgconf
 ```
 
 > Installing ELPA by apt only matches requirements on Ubuntu 22.04. For earlier linux distributions, you should build ELPA from source.
@@ -111,12 +111,12 @@ Here, 'build' is the path for building ABACUS; and '-D' is used for setting up s
   - `LAPACK_DIR`: Path to OpenBLAS library `libopenblas.so`(including BLAS and LAPACK)
   - `SCALAPACK_DIR`: Path to ScaLAPACK library `libscalapack.so`
   - `ELPA_DIR`: Path to ELPA install directory; should be the folder containing 'include' and 'lib'.
-  > Note: If you install ELPA from source, please add a symlink to avoid the additional include file folder with version name: `ln -s elpa/include/elpa-2021.05.002/elpa elpa/include/elpa`. This is a known behavior of ELPA.
+  > Note: In ABACUS v3.5.1 or earlier, if you install ELPA from source , please add a symlink to avoid the additional include file folder with version name: `ln -s elpa/include/elpa-2021.05.002/elpa elpa/include/elpa` to help the build system find ELPA headers.
 
   - `FFTW3_DIR`: Path to FFTW3.
   - `CEREAL_INCLUDE_DIR`: Path to the parent folder of `cereal/cereal.hpp`. Will download from GitHub if absent.
   - `Libxc_DIR`: (Optional) Path to Libxc.
-  > Note: Building Libxc from source with Makefile does NOT support using it in CMake here. Please compile Libxc with CMake instead.
+  > Note: In ABACUS v3.5.1 or earlier, Libxc built from source with Makefile is NOT supported; please compile Libxc with CMake instead.
   - `LIBRI_DIR`: (Optional) Path to LibRI.
   - `LIBCOMM_DIR`: (Optional) Path to LibComm.
 
@@ -126,6 +126,7 @@ Here, 'build' is the path for building ABACUS; and '-D' is used for setting up s
   - `ENABLE_LIBRI=OFF`: [Enable LibRI](../advanced/install.md#add-libri-support) to suppport variety of functionals. If `LIBRI_DIR` and `LIBCOMM_DIR` is defined, `ENABLE_LIBRI` will set to 'ON'.
   - `USE_OPENMP=ON`: Enable OpenMP support. Building ABACUS without OpenMP is not fully tested yet.
   - `BUILD_TESTING=OFF`: [Build unit tests](../advanced/install.md#build-unit-tests).
+  - `ENABLE_GOOGLEBENCH=OFF`: [Build performance tests](../advanced/install.md#build-performance-tests)
   - `ENABLE_MPI=ON`: Enable MPI parallel compilation. If set to `OFF`, a serial version of ABACUS with PW basis only will be compiled. Currently serial version of ABACUS with LCAO basis is not supported yet, so `ENABLE_LCAO` will be automatically set to `OFF`.
   - `ENABLE_COVERAGE=OFF`: Build ABACUS executable supporting [coverage analysis](../CONTRIBUTING.md#generating-code-coverage-report). This feature has a drastic impact on performance.
   - `ENABLE_ASAN=OFF`: Build with Address Sanitizer. This feature would help detecting memory problems.

python/pyabacus/CMakeLists.txt

@@ -12,9 +12,14 @@ set(BASE_PATH "${PROJECT_SOURCE_DIR}/../../source/module_base")
 set(ABACUS_SOURCE_DIR "${PROJECT_SOURCE_DIR}/../../source")
 include_directories(${BASE_PATH} ${ABACUS_SOURCE_DIR})
 list(APPEND _sources
-    ${ABACUS_SOURCE_DIR}/module_basis/module_nao/numerical_radial.h
-    ${ABACUS_SOURCE_DIR}/module_basis/module_nao/numerical_radial.cpp
-    ${PROJECT_SOURCE_DIR}/src/py_numerical_radial.cpp)
+    #${ABACUS_SOURCE_DIR}/module_basis/module_nao/numerical_radial.h
+    #${ABACUS_SOURCE_DIR}/module_basis/module_nao/numerical_radial.cpp
+    ${ABACUS_SOURCE_DIR}/module_base/constants.h
+    ${ABACUS_SOURCE_DIR}/module_base/math_sphbes.h
+    ${ABACUS_SOURCE_DIR}/module_base/math_sphbes.cpp
+    ${PROJECT_SOURCE_DIR}/src/py_abacus.cpp
+    #${PROJECT_SOURCE_DIR}/src/py_numerical_radial.cpp
+    ${PROJECT_SOURCE_DIR}/src/py_math_base.cpp)
 python_add_library(_core MODULE ${_sources} WITH_SOABI)
 target_link_libraries(_core PRIVATE pybind11::headers)
 target_compile_definitions(_core PRIVATE VERSION_INFO=${PROJECT_VERSION})

python/pyabacus/src/py_abacus.cpp

@@ -0,0 +1,13 @@
+#include <pybind11/numpy.h>
+#include <pybind11/pybind11.h>
+
+namespace py = pybind11;
+
+void bind_numerical_radial(py::module& m);
+void bind_math_base(py::module& m);
+
+PYBIND11_MODULE(_core, m)
+{
+    // bind_numerical_radial(m);
+    bind_math_base(m);
+}