mirror of https://gitlab.com/QEF/q-e.git
Updates: the combination PGI 12.9 + CUDA 5.0 has some issues on CRAY, default is Intel.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@9588 c92efa57-630b-4861-b058-cf58834340f0
This commit is contained in:
parent
8a232a9482
commit
8db11db601
|
@ -1,8 +1,5 @@
|
|||
Info by Filippo Spiga, Oct. 2012, valid for any version of QE after 5.
|
||||
|
||||
**************************************************************
|
||||
* THIS IS A DRAFT - EVALUATION IN STILL IN PROGRESS *
|
||||
**************************************************************
|
||||
|
||||
Machine name : TODI (Cray XKT) at CSCS, Lugano (CH)
|
||||
Machine spec : http://user.cscs.ch/hardware/todi_cray_xk7/index.html
|
||||
|
@ -13,11 +10,12 @@ Similar systems : TITAN (ORNL, USA)
|
|||
|
||||
... starting from a SVN checkout...
|
||||
|
||||
module switch PrgEnv-cray PrgEnv-pgi
|
||||
module switch PrgEnv-cray PrgEnv-intel
|
||||
module load cudatoolkit/5.0.33.103
|
||||
module unload atp hss-llm
|
||||
cd espresso
|
||||
cd GPU/
|
||||
./configure --enable-openmp --enable-cuda --with-gpu-arch=30 \
|
||||
./configure --enable-openmp --enable-cuda --with-gpu-arch=35 \
|
||||
--with-cuda-dir=${CRAY_CUDATOOLKIT_DIR} --disable-magma \
|
||||
--disable-profiling --enable-phigemm --enable-parallel \
|
||||
--with-scalapack ARCH=crayxt
|
||||
|
@ -27,19 +25,6 @@ make -f Makefile.gpu all-gpu
|
|||
|
||||
Executables will be located under "./bin"
|
||||
|
||||
...@todi1:~/espresso> ls -1 bin/*gpu*
|
||||
bin/cp-gpu.x
|
||||
bin/cppp-gpu.x
|
||||
bin/dynmat-gpu.x
|
||||
bin/generate_vdW_kernel_table-gpu.x
|
||||
bin/lambda-gpu.x
|
||||
bin/matdyn-gpu.x
|
||||
bin/neb-gpu.x
|
||||
bin/ph-gpu.x
|
||||
bin/pw-gpu.x
|
||||
bin/q2r-gpu.x
|
||||
bin/turbo_lanczos-gpu.x
|
||||
bin/wfdd-gpu.x
|
||||
|
||||
IMPORTANT NOTE (!): only pw-gpu.x, neb-gpu.x, ph-gpu.x use extensively
|
||||
the GPU card in multiple sections of the code. All
|
||||
|
@ -51,6 +36,11 @@ IMPORTANT NOTE : not all the codes are enough "big" or "computational
|
|||
missing in the above list do "make -f Makefile.gpu distclean"
|
||||
and follow the instruction in the file "README.CSCS_rosa"
|
||||
|
||||
IMPORTANT NOTE : CPU-only code supports PGI, GNU and Intel compilers. GPU+CPU
|
||||
code supports *ONLY* Intel compiler. A bug report has been
|
||||
filled to PGI and NVIDIA.
|
||||
|
||||
|
||||
|
||||
2. Good practices
|
||||
|
||||
|
@ -66,6 +56,7 @@ IMPORTANT NOTE : not all the codes are enough "big" or "computational
|
|||
mandatory to export the variable CRAY_CUDA_PROXY ("export CRAY_CUDA_PROXY=1")
|
||||
|
||||
|
||||
|
||||
3. Example scripts
|
||||
|
||||
#SBATCH --job-name="QE-BENCH-SPIGA"
|
||||
|
|
Loading…
Reference in New Issue