The multiprocessing pool does not work expectively when importing some modules from pymatgen #3402
Labels
bug
needs investigation
Needs further research to pinpoint cause of issue
performance
Some functionality is too slow or regressed
pkg
Package health and distribution related stuff
HI! I was trying to use
CrystalNN
with a multiprocessing pool and found that the CPU usage was unexpected. I have given some small tests to figure out where the problems are and then found that when I import some modules from pymatgen, such asfrom pymatgen.analysis.local_env import CrystalNN
,from pymatgen.core import Structure
, the issue will occur.The following codes show minimal implementation to reproduce this issue.
The normal case.
The bugged case. Whatever change the number of processes, the total CPU usage is 200% (In the follow case, is 12.5 x 16 = 200%),
OS: Ubuntu 22.04.3 LTS
My Python(3.8.17) env:
The text was updated successfully, but these errors were encountered: