Boian Petkantchin dc3258c617 [mlir][mesh] Add all-slice operation (#81218) ...

This op is the inverse of all-gather. It is useful to have an explicit
concise representation instead of having a blob of slicing logic.

Add lowering for the op that slices from the tensor based on the
in-group process index.

Make resharding generate an all-slice instead of inserting the slicing
logic directly.

2024-02-15 13:03:58 -08:00

2.8 KiB

Raw Blame History

View Raw