grasp
grasp copied to clipboard
running rangular_mpi and rmchdf_mpi very slow
when running not very large CSFs, the rangluar_mpi is very slow. CPU% <40, MEM% only 0.1. rmchdf_mpi 1 of 20 is about 100 cpu%, , 3 of 20 are about 6%. The others are only about 1%. So, how to improve! Increase MEM or change supercomputer?