.github: Bump linux gpu max limit to 100 (#65831)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65831
Was noticing scaling issues last night due to the lack of
linux.8xlarge.nvidia.gpu machines, seems as though that even at max
capacity we were still about ~50 queued workflows behind, this should
close that gap.
Also since these run the longest types of tests these are the most
likely to overlap with scale messages being processed while available
runners are still maxed out
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>
Test Plan: Imported from OSS
Reviewed By: malfet
Differential Revision: D31275892
Pulled By: seemethere
fbshipit-source-id: b22ceda115b70d7bdd9c4bc207b55ffab50381ef
