Closed AlyShmahell closed 4 years ago
As it turns out, the workers were being added but not used, removing with pmp.Pool(pmp.cpu_count()) as pool:
actually yielded better results for (bigger datasets/more computationally expensive partial_process function), the reason for which is the removal of the unused workers overhead, therefore I'm closing this issue.
Hm..interesting. Thanks for the posting.
Hello,
I have the following minimal working example:
As you can see, I'm wrapping my partition_all with pool, but I'm not using any pool functionality. Yet when I do a keyboard interrupt I get the following:
Why are there pool workers running? and are they all (in this case both) assigned to the partial_process function?