Skip to content

Latest commit

 

History

History
212 lines (178 loc) · 5.04 KB

File metadata and controls

212 lines (178 loc) · 5.04 KB

AMBER24_RC4_test on NVIDIA GeForce GTX 1070 Ti 8GB with Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz

OS System: CentOS Linux release 7.6.1810 with CMake 3.27.6, CUDA 11.8, GCC 7.3, and OpenMPI 4.1.6 operated by Lmod
Computer System (Frontend node): 2 x NVIDIA GeForce GTX 1070 Ti 8GB with Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz

Load modules

ml cmake
ml openmpi4
ml cuda/11.8
ml cudnn/11.8-8.9.2

Extract AMBER

tar xvfj AmberTools24_rc4.tar.bz2
tar xvfj Amber24_rc4.tar.bz2 

Compile Serial CPU

Complie and install

cd amber24_src
cd build
cmake .. -DCMAKE_INSTALL_PREFIX=/md0/apps2/amber24 -DCOMPILER=GNU -DMPI=FALSE -DCUDA=FALSE -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j8

Test

cd /md0/apps2/amber24
source amber.sh
make test.serial

Compile for Serial GPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/md0/apps2/amber24 -DCOMPILER=GNU -DMPI=FALSE -DCUDA=TRUE -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-11.8 -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j8

Test

cd /md0/apps2/amber24
source amber.sh
make test.cuda.serial

Compile for Parallel CPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/md0/apps2/amber24 -DCOMPILER=GNU -DCUDA=FALSE -DMPI=TRUE -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test for 2 MPI threads

cd /md0/apps2/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 2'
make test.parallel

Test for 4 MPI threads

cd /md0/apps2/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 4'
make test.parallel

Compile for Parallel GPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/md0/apps2/amber24 -DCOMPILER=GNU -DMPI=TRUE -DCUDA=TRUE -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-11.8 -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test for 2 MPI threads

cd /md0/apps2/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 2'
make test.cuda.parallel

AMBER24_RC4_test on Tesla V100 SXM2 32GB with Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz

OS System: CentOS Linux release 7.6.1810 with CMake 3.20.1, CUDA 11.1.1, GCC 10.3.0, and OpenMPI 4.1.1 operated by Lmod Computer System: 4 x Tesla V100 SXM2 32GB with Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz

Load modules

ml CMake/3.20.1-GCCcore-10.3.0
ml OpenMPI/4.1.1-GCC-10.3.0
ml CUDAcore/11.1.1

Extract AMBER

tar xvfj AmberTools24_rc4.tar.bz2
tar xvfj Amber24_rc4.tar.bz2 

Compile Serial CPU

Complie and install

cd amber24_src
cd build
cmake .. -DCMAKE_INSTALL_PREFIX=/ist/users/bunditb/bunditb_bak2/apps/amber24 -DCOMPILER=GNU -DMPI=FALSE -DCUDA=FALSE -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test

sinteractive -p gpu-cluster --mem=32GB --cpus-per-task=1 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
make test.serial

Compile for Serial GPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/ist/users/bunditb/bunditb_bak2/apps/amber24 -DCOMPILER=GNU -DMPI=FALSE -DCUDA=TRUE -DCUDA_TOOLKIT_ROOT_DIR=/ist/apps/modules/software/CUDAcore/11.1.1 -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test

sinteractive -p gpu-cluster --nodes=1 --gres=gpu:1 --mem=32GB --cpus-per-gpu=1 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
make test.cuda.serial

Compile for Parallel CPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/ist/users/bunditb/bunditb_bak2/apps/amber24 -DCOMPILER=GNU -DCUDA=FALSE -DMPI=TRUE -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test for 2 MPI threads

sinteractive -p gpu-cluster --mem=32GB --ntasks=2 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 2'
make test.parallel

Test for 4 MPI threads

sinteractive -p gpu-cluster --mem=32GB --ntasks=4 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 4'
make test.parallel

Compile for Parallel GPU

Complie and install

cd amber24_src
cd build
make clean
cmake .. -DCMAKE_INSTALL_PREFIX=/ist/users/bunditb/bunditb_bak2/apps/amber24 -DCOMPILER=GNU -DMPI=TRUE -DCUDA=TRUE -DCUDA_TOOLKIT_ROOT_DIR=/ist/apps/modules/software/CUDAcore/11.1.1 -DINSTALL_TESTS=TRUE -DDOWNLOAD_MINICONDA=TRUE
make install -j16

Test for 2 MPI threads

sinteractive -p gpu-cluster --nodes=1 --gres=gpu:1 --mem=32GB --ntasks=2 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 2'
make test.cuda.parallel

Test for 4 MPI threads

sinteractive -p gpu-cluster --nodes=1 --gres=gpu:1 --mem=32GB --ntasks=4 --time=1-0:0:0
cd /ist/users/bunditb/bunditb_bak2/apps/amber24
source amber.sh
export DO_PARALLEL='mpirun -np 4'
make test.cuda.parallel