Skip to content

Commit 1ee463b

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 c80282d commit 1ee463b

File tree

2 files changed

+688
-0
lines changed

2 files changed

+688
-0
lines changed

0 commit comments

Comments
 (0)