search

aelzenaar / bella

New computational package for small-rank matrix groups
BSD 3-Clause "New" or "Revised" License
2 stars 1 forks source link

Utilise multithreading #6

Open aelzenaar opened 11 months ago

aelzenaar commented 11 months ago

Use the power of dask and/or pandas to parallelise limit set computations.

aelzenaar commented 11 months ago

Change *cayley_graph_mc to do every count in parallel and change coloured_limit_set_mc back to using the free_ variant, c.f. d4d6608

aelzenaar commented 11 months ago

Currently blocked by https://github.com/mpmath/mpmath/issues/380 unless I can come up with a workaround for mpmath matrices not being pickleable. Perhaps converting to numpy arrays (but maybe this is too slow for the speed boost we would get)... we already need to pass the whole class GroupCache into each child process so even if I get it working it might be too much of a memory drain anyway.