[CP2K:9400] Re: running cuda-enabled cp2k on multiple nodes with aprun/HPC

[Ada Sedova]

Thanks, Andreas.

Here are the various interactive requests:
qsub -I -A stf006 -l walltime=00:40:00 -l nodes=1
qsub -I -A stf006 -l walltime=00:40:00 -l nodes=2
qsub -I -A stf006 -l walltime=00:40:00 -l nodes=3
qsub -I -A stf006 -l walltime=00:40:00 -l nodes=4

Here is an example of the environment set-up in the work dir aprun call:

module load cudatoolkit
module load fftw
module swap PrgEnv-pgi/5.2.82 PrgEnv-gnu
export CRAY_CUDA_PROXY=1
aprun -n 3
/ccs/home/adaa/cp2k-titan-gnu4.9.3/cp2k/exe/titan_gnu_cuda/cp2k.psmp -i
H2O-32.inp -o  H2O-32.out &

(Also with various other values for -n, and also with/without export
OMP_NUM_THREADS=8, as in the CUDA profiling part of the website)

The ldd paths to the executable are identical within a qsub on the work
node and on the compile node, except the hexadecimal addresses.

By the way, I built 4.1 on Eos, also a Cray but with no GPUs, and with psmp
but no CUDA, and there are no problems running on multiple nodes.

Thanks again,

Ada


On Wed, Sep 13, 2017 at 11:12 AM, Andreas Glöss <andreas... at gmail.com>
wrote:

> Dear Ada,
>
> Can you please post the aprun-line and the requested resources for the
> interactive reservation. Also check that 'ldd [your_executable]' shows the
> same library paths on the compile-node and the compute-node.
>
> Best regards,
> Andreas
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "cp2k" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/
> topic/cp2k/zlt69l2xaqc/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> cp2k+uns... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at https://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170913/3ac0844c/attachment.htm>