Core usage with fq2bam

Healthcare Parabricks
1

Is there any way to limit core usage with the fq2bam tool? We’re testing parabricks 4.0.0-1 with a Singularity container (converted from Docker) on an academic cluster. A typical GPU node might have 28-32 cores and 4 GPUs. fq2bam can be limited to use 2 GPUs with the “–num-gpus” flag. It would be very useful to limit fq2bam to say 14 or 16 CPU cores so that other compute jobs can execute on the remaining GPUs. The bwa step tends to use 18+ cores when it runs.

2

Hey @bgregor,

If you’re using Parabricks 4.0, you should be able to limit the cores in the Docker run command using --cpus or a similar flag. Does that solution work for you?

Thanks