I have a branching workflow where, if processes are freely allowed to run in parallel, they consume more memory than my system has available. I’m running on my local machine for now, but this will eventually be run on SLURM and other HPC environments.
I know I can specify how much memory to request, like this:
process processName {
maxRetries 3
memory {14.GB + 2.GB * task.attempt}
}
Is there a way to use this to limit the number of processes nextflow runs simultaneously to stay under a total memory cap?