Skip to content

Commit 6c36441

Browse files
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 fd7e46b commit 6c36441

File tree

2 files changed

+904
-0
lines changed

2 files changed

+904
-0
lines changed

0 commit comments

Comments
 (0)