I’m interested in assessing the performance of the publishing and cleanup tasks.
Is there a way to know when the Nextflow process have finished the publishing of all the required outputs? I assume that it starts as soon as the outputs are available.
Is there a way to know when Nextflow starts and ends the cleanup of the work folder once all tasks are finished?
I want to see how the number of threads in the head process affects performance in our typical workloads. According to Optimizing Nextflow for HPC and cloud a scale the head node uses as many threads as
the number of available CPUs
How Nextflow get the number of available CPUs? My particular running environment is a SLURM job. Does it honours the number of CPUs requested to SLURM? Or does it get as many CPUs as the machine have? Is there somewhere in the logs where I can see how many CPUs/threads the head node is using? In case it is taking more CPUs than allocated by SLURM, is there some flag/command line parameter/environment variable I can use to limit this number?
Ouch! I didn’t wanted to enter the realm of plugins, I’m not looking for new functionality, just information about current one. I was expecting to have all that information in the log files, but I was unable to find it, hence I was asking for it.