Closed PaulHancock closed 4 years ago
Bonus! It appears that regular fitting also has this issue...
or rather the issue is that setting a limit on the number of workers in the queue doesn't automatically mean we have only that many cores in use. Some testing has shown to me that even when only two instances of the fitting are being run at any one time, all of my available compute resources will be consumed.
Maybe this is a problem with the pprocess
module (which I plan to discontinue using due to lack of support into python 3).
No longer an issue as of 2/3/2020. Not entirely sure what fixed this tbh.
When running in
--priorized
mode aegean doesn't respect the--cores
option. for exaple:but looking at the system monitor shows that all 8 of my cores are being used for the priorized fitting.