1 files changed, 20 insertions, 15 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/fmin3.ll b/llvm/test/CodeGen/AMDGPU/fmin3.ll
index fa93fbcfb91..48d0eedba5b 100644
--- a/llvm/test/CodeGen/AMDGPU/fmin3.ll
+++ b/llvm/test/CodeGen/AMDGPU/fmin3.ll
@@ -95,22 +95,26 @@ define amdgpu_kernel void @test_fmin3_olt_1_f16(half addrspace(1)* %out, half ad
 ; SI-NEXT: v_min3_f32
 ; SI-NEXT: v_min3_f32
 
-; VI: v_min_f16_sdwa v4, v0, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
-; VI: v_min_f16_e32 v0, v0, v1
-; VI: v_min_f16_sdwa v1, v2, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
-; VI: v_min_f16_e32 v0, v2, v0
-; VI: v_min_f16_sdwa v1, v1, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_1
-; VI: v_min_f16_e32 v0, v0, v3
-; VI: v_or_b32_e32 v0, v0, v1
-
-; GFX9: v_pk_min_f16
-; GFX9: v_pk_min_f16
-; GFX9: v_pk_min_f16
-define <2 x half> @no_fmin3_v2f16(<2 x half> %a, <2 x half> %b, <2 x half> %c, <2 x half> %d) {
+; VI: s_waitcnt
+; VI-NEXT: v_min_f16_sdwa v4, v0, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
+; VI-NEXT: v_min_f16_e32 v0, v0, v1
+; VI-NEXT: v_min_f16_sdwa v1, v2, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
+; VI-NEXT: v_min_f16_e32 v0, v2, v0
+; VI-NEXT: v_min_f16_sdwa v1, v1, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_1
+; VI-NEXT: v_min_f16_e32 v0, v0, v3
+; VI-NEXT: v_or_b32_sdwa v0, v0, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_0 src1_sel:DWORD
+; VI-NEXT: s_setpc_b64
+
+; GFX9: s_waitcnt
+; GFX9-NEXT: v_pk_min_f16 v0, v0, v1
+; GFX9-NEXT: v_pk_min_f16 v0, v2, v0
+; GFX9-NEXT: v_pk_min_f16 v0, v0, v3
+; GFX9-NEXT: s_setpc_b64
+define <2 x half> @no_fmin3_v2f16(<2 x half> %a, <2 x half> %b, <2 x half> %c, <2 x half> %d) #2 {
 entry:
-  %min = tail call fast <2 x half> @llvm.minnum.v2f16(<2 x half> %a, <2 x half> %b)
-  %min1 = tail call fast <2 x half> @llvm.minnum.v2f16(<2 x half> %c, <2 x half> %min)
-  %res = tail call fast <2 x half> @llvm.minnum.v2f16(<2 x half> %min1, <2 x half> %d)
+  %min = call <2 x half> @llvm.minnum.v2f16(<2 x half> %a, <2 x half> %b)
+  %min1 = call <2 x half> @llvm.minnum.v2f16(<2 x half> %c, <2 x half> %min)
+  %res = call <2 x half> @llvm.minnum.v2f16(<2 x half> %min1, <2 x half> %d)
   ret <2 x half> %res
 }
 
@@ -121,3 +125,4 @@ declare <2 x half> @llvm.minnum.v2f16(<2 x half>, <2 x half>)
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone speculatable }
+attributes #2 = { nounwind "no-nans-fp-math"="true" }