CP2K 9.1 | Research Center for Computational Science

Webpage

https://www.cp2k.org/

Version

9.1 (9.1.0)

Build Environment

GCC 11.2.1 (gcc-toolset-11)
HPC-X 2.13.1 (Open MPI 4.1.5)

Files Required

cp2k-9.1.0.tar.gz
tc-install-libint.sh.diff

--- scripts/stage3/install_libint.sh.org 2022-01-26 09:34:28.922143775 +0900
+++ scripts/stage3/install_libint.sh 2022-01-26 09:41:44.440708192 +0900
@@ -90,6 +90,7 @@
# Fix bug in makefile for Fortran module
sed -i "s/\$(CXX) \$(CXXFLAGS)/\$(FC) \$(FCFLAGS)/g" fortran/Makefile
fi
+ sed -i 's/FCLIBS)/FCLIBS) -lstdc++/' fortran/Makefile

make -j $(get_nprocs) > make.log 2>&1
make install > install.log 2>&1

patch-openblas-makefile

--- scripts/stage2/install_openblas.sh.org 2023-01-17 10:19:31.000000000 +0900
+++ scripts/stage2/install_openblas.sh 2023-01-17 10:20:21.000000000 +0900
@@ -47,6 +47,7 @@
[ -d OpenBLAS-${openblas_ver} ] && rm -rf OpenBLAS-${openblas_ver}
tar -zxf ${openblas_pkg}
cd OpenBLAS-${openblas_ver}
+ sed -i -e "/NOTPARALLEL/s/$/ shared/" Makefile

# First attempt to make openblas using auto detected
# TARGET, if this fails, then make with forced

Build Procedure

#!/bin/sh

INSTDIR=/apl/cp2k/9.1

GITHUB_VERSION=9.1.0
VERSION=9.1.0
DBCSR_VERSION=v2.2.0

SOURCE_ROOT=/home/users/${USER}/Software/CP2K/${GITHUB_VERSION}

TARBALL=${SOURCE_ROOT}/cp2k-${VERSION}.tar.gz

TC_PATCH_3_1=${SOURCE_ROOT}/tc-install-libint.sh.diff
TC_PATCH_2_OB=${SOURCE_ROOT}/patch-openblas-makefile

PARALLEL=12

#---------------------------------------------------------------------------
umask 0022
export LANG=C
export LC_ALL=C
ulimit -s unlimited

module purge
module load gcc-toolset/11
module load openmpi/4.1.5-hpcx/gcc11

cd $INSTDIR
if [ -d cp2k-${VERSION} ]; then
mv cp2k-${VERSION} cp2k-erase
rm -rf cp2k-erase &
fi
tar zxf ${TARBALL}
sleep 5
mv cp2k-${VERSION}/* .
sleep 5
rm -rf cp2k-${VERSION}/{.dockerignore,.github}
rmdir cp2k-${VERSION}

cd ${INSTDIR}/tools/toolchain

# apply patches
patch -p0 < ${TC_PATCH_3_1}
patch -p0 < ${TC_PATCH_2_OB}

# ad hoc...
sed -i -e "s/https:.*/https:\/\/github.com\/spglib\/spglib\/archive\/refs\/tags\/v1.16.2.tar.gz \\\/" scripts/stage7/install_spglib.sh

export CC=gcc
export CXX=g++
export FC=gfortran
export MPICC=mpicc
export MPICXX=mpicxx
export MPIFC=mpif90

./install_cp2k_toolchain.sh --mpi-mode=openmpi \
--math-mode=openblas \
--with-gcc=system \
--with-cmake=system \
--with-openmpi=system \
--with-mpich=no \
--with-intelmpi=no \
--with-libxc=install \
--with-libint=install \
--with-fftw=install \
--with-acml=no \
--with-mkl=no \
--with-openblas=install \
--with-scalapack=install \
--with-libsmm=no \
--with-libxsmm=install \
--with-elpa=no \
--with-ptscotch=no \
--with-superlu=no \
--with-pexsi=no \
--with-quip=install \
--with-plumed=install \
--with-sirius=install \
--with-gsl=install \
--with-libvdwxc=install \
--with-spglib=install \
--with-hdf5=install \
--with-spfft=install \
--with-spla=install \
--with-cosma=no \
--with-libvori=install \
-j ${PARALLEL}

cp install/arch/local.psmp ../../arch/rccs.psmp

cd ${INSTDIR}
cd ${INSTDIR}/exts
rmdir dbcsr
git clone https://github.com/cp2k/dbcsr.git
cd dbcsr
git checkout refs/tags/${DBCSR_VERSION}
git submodule update --init --recursive
cd ${INSTDIR}

make -j ${PARALLEL} ARCH=rccs VERSION=psmp
make -j ${PARALLEL} ARCH=rccs VERSION=psmp libcp2k

Tests

Following scripts was used for testing.

#!/bin/sh
#PBS -l select=1:ncpus=16:mpiprocs=16:ompthreads=1
#PBS -l walltime=12:00:00

export LC_ALL=C
export LANG=""
export OMP_STACKSIZE=64M

module purge
module load gcc-toolset/11
module load openmpi/4.1.5-hpcx/gcc11

CP2K=/apl/cp2k/9.1

CP2K_ARCH=rccs
CP2K_VER=psmp
TIMEOUT=600
PARALLEL=16

ulimit -s unlimited

cd ${CP2K}/regtesting/${CP2K_ARCH}/${CP2K_VER}
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

# serial test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 1 \
-ompthreads 1 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi1_omp1.log
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

# omp test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 1 \
-ompthreads 2 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi1_omp2.log
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

# mpi test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 2 \
-ompthreads 1 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi2_omp1.log
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

# mpi/openmp test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 2 \
-ompthreads 2 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi2_omp2.log

# yet another mpi test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 8 \
-ompthreads 1 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi8_omp1.log
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

# yet another mpi/openmp test
../../../tools/regtesting/do_regtest \
-nobuild \
-arch ${CP2K_ARCH} \
-version ${CP2K_VER} \
-mpiranks 8 \
-ompthreads 2 \
-jobmaxtime ${TIMEOUT} \
-cp2kdir ../../../ \
-maxtasks ${PARALLEL} >& regtest_mpi8_omp2.log
rm -rf LAST-${CP2K_ARCH}-${CP2K_VER}

results

$ grep GREPME regtest_mpi*
regtest_mpi1_omp1.log:GREPME 0 0 3586 0 3586 X
regtest_mpi1_omp2.log:GREPME 0 0 3586 0 3586 X
regtest_mpi2_omp1.log:GREPME 0 0 3642 0 3642 X
regtest_mpi2_omp2.log:GREPME 0 0 3642 0 3642 X
regtest_mpi8_omp1.log:GREPME 11 16 3629 0 3656 X
regtest_mpi8_omp2.log:GREPME 11 17 3628 0 3656 X

Test results are available at /apl/cp2k/9.1/regtesting/rccs/psmp.

Notes

elpa doesn't work well. Simply disabled.
If openmpi is employed, superlu can't be enabled. pexsi and ptscotch (depends on superlu) are also disabled.
(intel mpi version is not intensively tested. At this time, there was no workaround available for the Intel MPI issue solved on Feb 3.)
(Jul 26, 2024) Runtime library of HPC-X 2.11 or 2.13.1 cause significant performance loss for calculation of H20-256 system with 128 MPI processes. This problem can be removed by switching to Open MPI 4.1.5, 4.1.6 or HPC-X 2.16 runtime libraries.

View PDF