`request-cpus=16` is much slower than running on one CPU
Created by: adivijaykumar
Maybe the likelihood calculation is so fast that the communication overhead dominates? Maybe making it faster needs a HTC cluster and not a HPC cluster.
Created by: adivijaykumar
Maybe the likelihood calculation is so fast that the communication overhead dominates? Maybe making it faster needs a HTC cluster and not a HPC cluster.