As of #602 , the default for locate_brightfield_ring() is processes=‘auto’ (use all available CPUs), while for batch() it’s processes=1 (no parallelization; see #499 ). It seems like it should be auto for both—a non-expert user would expect trackpy to just run as fast as possible on the available hardware. If the user is doing their own parallelization and wants to keep trackpy on one CPU, presumably they would be expert enough to change that option. And I’m sure there are edge cases in which multiprocessing slows down these computations, but it seems like those would be rare.
As of #602 , the default for
locate_brightfield_ring()
isprocesses=‘auto’
(use all available CPUs), while forbatch()
it’sprocesses=1
(no parallelization; see #499 ). It seems like it should beauto
for both—a non-expert user would expect trackpy to just run as fast as possible on the available hardware. If the user is doing their own parallelization and wants to keep trackpy on one CPU, presumably they would be expert enough to change that option. And I’m sure there are edge cases in which multiprocessing slows down these computations, but it seems like those would be rare.Any thoughts before I submit a PR?