Command-line option for cpu count

[glennhickey]

Right now run_wga_gpu uses $(nproc) threads. This makes it difficult to use from workflows such as toil or cactus. Can you please add an option to specifiy the threads on the command line? A similar option to specify the number of GPUs would also be helpful. Thanks.

[glennhickey]

I usually run via WDL/Terra on the cloud where it makes sense to give SegAlign all cores on the node. But not being able to specify the CPU count can cause trouble on other environments Toil supports. Ex : ComparativeGenomicsToolkit/cactus#709

@gsneha26 Could you please add a command-line option to run_segalign as well as segalign_repeat_masker to specify the number of cpus?

I remember you mentioning taskset but from what I've (just) read, I don't think this will work in general. The reason is that it requires specific CPUs to be specified, and not the number of CPUs. This distinction is important, as it's not within Toil's scope to know which CPUs are free or not free -- it only cares about the total counts.

[richard-burhans]

If anyone else is having this issue, my workaround is to set the OMP_THREAD_LIMIT environment variable before running SegAlign as documented here