I am having a hanging problem using bcast that I'm trying to track
down. I added the above, hoping to get some debugging output,
but didn't see any. I am using slurm in batch mode.
slurm -n 20 -b slurm_script.sh
and slurm_script.sh contains:
mpirun -np $SLURM_NPROCS --mca coll_tuned_base_verbose 1 /u/chiuk/cfd/main
By the way, I also tried:
mpirun -np $SLURM_NPROCS --mca coll ^tuned /u/chiuk/cfd/main
but got the same hang. This is on odin.cs.indiana.edu.
Thanks.
|