Closed privefl closed 2 years ago
The confusing thing would be that some functions use OpenMP instead of {foreach} to parallelize, therefore this won't work for all functions.
I think I'll forget about this, unless there is a strong argument for it.
E.g. if using
ncores = NULL
, use higher-level registration for parallel processing?Could for example use
to run one thing on many nodes on a Slurm cluster.