[llvm] [AMDGPU] Fix GFX11 WMMA intrinsic lowering regression for compute kernels (PR #164036)

[Shilei Tian via llvm-commits]

================
@@ -0,0 +1,74 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -mattr=+wavefrontsize32 -verify-machineinstrs < %s | FileCheck %s --check-prefix=GFX11-W32
+
+; Test GFX11 WMMA with amdgpu_kernel (compute) calling convention
+; This test is critical to prevent regression of compute kernel WMMA support
+
+declare <8 x float> @llvm.amdgcn.wmma.f32.16x16x16.f16(<16 x half>, <16 x half>, <8 x float>)
+declare <8 x float> @llvm.amdgcn.wmma.f32.16x16x16.bf16(<16 x i16>, <16 x i16>, <8 x float>)
+declare <8 x i32> @llvm.amdgcn.wmma.i32.16x16x16.iu8(i1, <4 x i32>, i1, <4 x i32>, <8 x i32>, i1)
+declare <8 x i32> @llvm.amdgcn.wmma.i32.16x16x16.iu4(i1, <2 x i32>, i1, <2 x i32>, <8 x i32>, i1)
+
+; GFX11-W32-LABEL: test_wmma_f32_16x16x16_f16_kernel:
----------------
shiltian wrote:

These check lines don't look like auto generated. Can you run it again?

https://github.com/llvm/llvm-project/pull/164036