[mpich-discuss] MPICH configure
Zhou, Hui
zhouh at anl.gov
Thu May 7 13:28:55 CDT 2020
Hi Bruce,
Thanks for the details. Since you already instructed `sbatch` with `-N 2 -n 6` I thought you could simply run with `srun`, i.e. without further `-n`? I believe hydra also reads environment so you could also try `mpiexec` without `-n`. Did you also try `-N 1` and `-N 6`?
--
Hui Zhou
From: "Palmer, Bruce J" <Bruce.Palmer at pnnl.gov>
Date: Thursday, May 7, 2020 at 11:16 AM
To: "Zhou, Hui" <zhouh at anl.gov>, "discuss at mpich.org" <discuss at mpich.org>
Subject: Re: [mpich-discuss] MPICH configure
Hi Hui,
Sorry for the late reply, I keep getting pulled off on other projects. I’m actually running a test suite most of the time, so the job scripts look like
#!/bin/csh
#SBATCH -t 02:30:00
#SBATCH -A XGA
#SBATCH -p short,slurm,gpu
#SBATCH -N 2
#SBATCH -n 6
#SBATCH -o ./test.out
#SBATCH -e ./test.err
source /etc/profile.d/modules.csh
source ~/set_mpich
env | grep PATH
module list
#make check-ga MPIEXEC="mpirun -n 6 "
make check-ga MPIEXEC="srun -n 6 "
I’ve tried using mpirun, srun and mpiexec in the MPIEXEC variable. If I run a test standalone, then the job submission script is
#!/bin/csh
#SBATCH -t 02:30:00
#SBATCH -A XGA
#SBATCH -p short,slurm,gpu
#SBATCH -N 2
#SBATCH -n 6
#SBATCH -o ./test.out
#SBATCH -e ./test.err
source /etc/profile.d/modules.csh
source ~/set_mpich
env | grep PATH
module list
srun -n 6 test.x > test.out
Again, I’ve tried running with mpirun, srun, and mpiexec. The environment in the set_mpich file is
module purge
module load gcc/5.2.0
module load python/2.7.8
module load cmake/3.8.2
module load git
module load mkl
setenv CC gcc
setenv CFLAGS "-pthread"
setenv CXX g++
setenv CXXFLAGS "-pthread"
setenv FC gfortran
setenv FCFLAGS "-pthread"
unsetenv F90
unsetenv F90FLAGS
setenv PATH /people/d3g293/mpich/mpich-3.3.2/install/bin:${PATH}
setenv MANPATH /people/d3g293/mpich/mpich-3.3.2/install/share/man:${MANPATH}
setenv LD_LIBRARY_PATH /people/d3g293/mpich/mpich-3.3.2/install/lib:${LD_LIBRARY_PATH}
Bruce
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20200507/2be16177/attachment.html>
More information about the discuss
mailing list