CP2K (perftools)
Issue #24
new
Santis + Cuda/6.5
Setup
module swap PrgEnv-cray PrgEnv-gnu
module load craype-accel-nvidia35+cuda65
module load fftw
export LD_LIBRARY_PATH=$CRAY_LD_LIBRARY_PATH:$LD_LIBRARY_PATH
module use /project/csstaff/proposals
module load perflite/622cuda
module load craype-hugepages8M
Compile
cd /apps/daint/sandbox/jgp/cp2k/svn/cp2k.notool+cuda65/
make ARCH=CRAY-XC30-gfortran-cuda VERSION=psmp
Run
cd /apps/daint/sandbox/jgp/cp2k/svn/cp2k.notool+cuda65/
cd exe/CRAY-XC30-gfortran-cuda/08CN
./sbatch.sh santis 10 ../cp2k.psmp 8 1 8 H2O-dft-ls_NREP2.inp
export CRAY_CUDA_MPS=1 # without perftool
export CRAY_CUDA_MPS=0 # with perftool
Not enough memory for NREP5 on Santis
Comments (7)
-
reporter -
reporter CP2K + PERFLITE/622
- export CRAY_CUDA_MPS=0
- aprun -n 8 -N 1 -d 8 -j 1 ../cp2k.psmp H2O-dft-ls_NREP2.inp
but issue with craype-hugepages8M
libhugetlbfs [nid00021:15894]: WARNING: Unable to verify address range 0x42e43000 - 0x43000000. Not empty? libhugetlbfscp2k.psmp: malloc.c:3091: sYSMALLOc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) - 1) * 2])) - __builtin_offsetof (struct malloc_chunk, fd)))) && old_size == 0) || ((unsigned long) (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize)) +((2 * (sizeof(size_t))) - 1)) & ~((2 * (sizeof(size_t))) - 1))) && ((old_top)->size & 0x1) && ((unsigned long)old_end & pagemask) == 0)' failed.
-
reporter - Issue with craype-hugepages8M still present
- Without craype-hugepages8M, ok:
-
reporter Santis + Cuda/6.5 + MPS=1
perftools-lite/6.2.3 supports CRAY_CUDA_MPS=1
-
reporter Daint
module swap PrgEnv-cray PrgEnv-gnu module load craype-accel-nvidia35 module load fftw module load craype-hugepages8M module use /project/csstaff/proposals module load perflite/625cuda cd /apps/daint/5.2.UP02/sandbox/jgp/cp2k/GNU482/cp2k-code-15721-trunk/cp2k/makefiles/ vim CRAY-XC30-gfortran-gpu.psmp make ARCH=CRAY-XC30-gfortran-gpu VERSION=psmp clean make ARCH=CRAY-XC30-gfortran-gpu VERSION=psmp -j8 cd ../exe/CRAY-XC30-gfortran-gpu export CRAY_CUDA_MPS=1 sbatch.sh daint 10 ./cp2k.psmp 8 1 8 H2O-dft-ls_NREP2.inp "" "" -Ausup
Currently Loaded Modulefiles: modules/3.2.10.3 nodestat/2.2-1.0502.53712.3.109.ari sdb/1.0-1.0502.55976.5.27.ari alps/5.2.1-2.0502.9041.11.6.ari lustre-cray_ari_s/2.5_3.0.101_0.31.1_1.0502.8394.15.1-1.0502.19897.18.2 udreg/2.3.2-1.0502.9275.1.12.ari ugni/5.0-1.0502.9685.4.24.ari gni-headers/3.0-1.0502.9684.5.2.ari dmapp/7.0.1-1.0502.11083.8.4.ari xpmem/0.1-2.0502.55507.3.2.ari hss-llm/7.2.0 Base-opts/1.0.2-1.0502.53325.1.2.ari craype-sandybridge craype-network-aries craype/2.4.0 slurm cray-mpich/7.2.2 ddt/5.0 gcc/4.8.2 totalview-support/1.1.4 totalview/8.11.0 cray-libsci/13.0.4 pmi/5.0.7-1.0000.10678.155.25.ari atp/1.8.2 PrgEnv-gnu/5.2.40 cray-libsci_acc/3.1.1 cudatoolkit/6.5.14-1.0502.9613.6.1 craype-accel-nvidia35 rca/1.0.0-2.0502.53711.3.127.ari perflite/625cuda fftw/3.3.4.3
-
reporter - sbatch.sh santis 10 ./cp2k.psmp 64 8 1 H2O-dft-ls_NREP2.inp
-
reporter - changed title to CP2K (perftools)
- Log in to comment
mpi+cuda test