Skip to content

Commit 1fce448

Browse files
arsenmarsenm
committed
AMDGPU: Add cost model tests for minimumnum/maximumnum
The f16 cases in particular look broken since every vector size has the same reported cost.
1 parent 4ced29b commit 1fce448

File tree

2 files changed

+904
-0
lines changed

2 files changed

+904
-0
lines changed

0 commit comments

Comments
 (0)