Dave May <[email protected]> writes: > I argue it does matter as I've seen runs on 32k cores where a huge amount > of time is spent in those global reductions. I can provide an > implementation which uses a sub comm (PCSemiRedundant) if someone thinks > doing reductions on less cores is beneficial.
It doesn't matter much on Blue Gene, but is a big deal on older Crays. Aires seems to be in between. The default GAMG configuration doesn't do any reductions in the coarse grid, so the issue is moot. If an iterative coarse solver was used, I think we would be more motivated to put the coarse problem on a subcomm.
pgpb8IWKSKaES.pgp
Description: PGP signature
