Hello,
I have installed parabricks pipeline with singularity option, and I’m using an HPC that has nodes with one V100 card on them (per server). The pipeline worked on one node(i.e. one GPU). However, when I allocated two nodes to run on two GPUs I had the following error:
“Number of GPUs requested (2) is more than number of GPUs (1) in the system”
and the job configuration was:
#SBATCH --partition=gpu
#SBATCH --nodes=2
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=4
#SBATCH --mem=64000
So, Does parabricks support multi-host parallelism? If so how can I manage to solve the issue?
Thanks!