Gromacs 2018.8 with GPU for LX (gnu)
Webpage
Version
2018.8
Build Environment
- Intel Parallel Studio XE 2018 update 4 (for Intel MPI)
- gcc 6.3.1 (devtoolset-6 Software Collections)
- cmake 3.8.2
Files Required
- gromacs-2018.8.tar.gz
- regressiontests-2018.8.tar.gz
- tests_CMakelists.patch
--- tests/CMakeLists.txt.org 2019-10-07 19:24:40.524863424 +0900
+++ tests/CMakeLists.txt 2019-10-07 19:25:13.736715189 +0900
@@ -125,6 +125,11 @@
endif()
#We should use MPIEXEC_NUMPROC_FLAG but gmxtest.pl doesn't let us pass it
endif()
+ if(GMX_THREAD_MPI)
+ set(GMX_TEST_NUMBER_PROCS 8 CACHE STRING "Number of processors used for testing")
+ mark_as_advanced(GMX_TEST_NUMBER_PROCS)
+ list(APPEND ARGS -nt ${GMX_TEST_NUMBER_PROCS})
+ endif()
if(GMX_BINARY_SUFFIX)
list(APPEND ARGS -suffix ${GMX_BINARY_SUFFIX})
endif()
(this patch changes number of cores used in tests to 8; just to avoid error in tests on 14-core CPU installed node. This patch is usually not necessary.)
- fftw-3.3.8.tar.gz (Because the build node "ccgpup" cannot access to fftw download site.)
Build Procedure
#!/bin/sh
VERSION=2018.8
INSTALL_PREFIX=/local/apl/lx/gromacs2018.8-gnu-CUDA
BASEDIR=/home/users/${USER}/Software/Gromacs/${VERSION}/
GROMACS_TARBALL=${BASEDIR}/gromacs-${VERSION}.tar.gz
REGRESSION_TARBALL=${BASEDIR}/regressiontests-${VERSION}.tar.gz
WORKDIR=/work/users/${USER}
REGRESSION_PATH=${WORKDIR}/regressiontests-${VERSION}
FFTW_VER=3.3.8
FFTW_PATH=${BASEDIR}/fftw-${FFTW_VER}.tar.gz
PATCH_TEST=${BASEDIR}/tests_CMakelists.patch
PARALLEL=12
#---------------------------------------------------------------------
umask 0022
module purge
module load scl/devtoolset-6
module load intel_parallelstudio/2017update8
module load cuda/9.1
module load cmake/3.8.2
cd ${WORKDIR}
if [ -d gromacs-${VERSION} ]; then
mv gromacs-${VERSION} gromacs_erase
rm -rf gromacs_erase &
fi
if [ -d regressiontests-${VERSION} ]; then
mv regressiontests-${VERSION} regressiontests_erase
rm -rf regressiontests_erase &
fi
tar xzf ${GROMACS_TARBALL}
tar xzf ${REGRESSION_TARBALL}
cd gromacs-${VERSION}
patch -p0 < ${PATCH_TEST}
# single precision, no MPI
mkdir rccs-s
cd rccs-s
cmake .. \
-DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX} \
-DCMAKE_VERBOSE_MAKEFILE=ON \
-DGMX_MPI=OFF \
-DGMX_GPU=ON \
-DGMX_DOUBLE=OFF \
-DGMX_THREAD_MPI=ON \
-DGMX_BUILD_OWN_FFTW=ON \
-DGMX_BUILD_OWN_FFTW_URL=${FFTW_PATH} \
-DREGRESSIONTEST_DOWNLOAD=OFF \
-DREGRESSIONTEST_PATH=${REGRESSION_PATH}
make -j${PARALLEL} && make check && make install
cd ..
# compiler setting for MPI versions
export CC=mpicc
export CXX=mpicxx
export F77=mpif90
export F90=mpif90
export FC=mpif90
# single precision, with MPI
mkdir rccs-mpi-s
cd rccs-mpi-s
cmake .. \
-DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX} \
-DCMAKE_VERBOSE_MAKEFILE=ON \
-DGMX_MPI=ON \
-DGMX_GPU=ON \
-DGMX_DOUBLE=OFF \
-DGMX_THREAD_MPI=OFF \
-DGMX_BUILD_OWN_FFTW=ON \
-DGMX_BUILD_OWN_FFTW_URL=${FFTW_PATH} \
-DREGRESSIONTEST_DOWNLOAD=OFF \
-DREGRESSIONTEST_PATH=${REGRESSION_PATH}
make -j${PARALLEL} && make check && make install
cd ..