Skip to content

Commit 72e085d

Browse files
committed
AMDGPU: Update llvm.sqrt.bf16.ll
1 parent 463f200 commit 72e085d

File tree

1 file changed

+2
-0
lines changed

1 file changed

+2
-0
lines changed

llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -66,6 +66,7 @@ define amdgpu_kernel void @sqrt_v2bf16(ptr addrspace(1) %r, ptr addrspace(1) %a)
6666
; GFX12-TRUE16-NEXT: s_mov_b32 s5, s1
6767
; GFX12-TRUE16-NEXT: s_wait_loadcnt 0x0
6868
; GFX12-TRUE16-NEXT: v_sqrt_bf16_e32 v1.l, v0.l
69+
; GFX12-TRUE16-NEXT: v_nop
6970
; GFX12-TRUE16-NEXT: v_lshrrev_b32_e32 v0, 16, v0
7071
; GFX12-TRUE16-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(TRANS32_DEP_2)
7172
; GFX12-TRUE16-NEXT: v_sqrt_bf16_e32 v0.l, v0.l
@@ -90,6 +91,7 @@ define amdgpu_kernel void @sqrt_v2bf16(ptr addrspace(1) %r, ptr addrspace(1) %a)
9091
; GFX12-FAKE16-NEXT: s_mov_b32 s5, s1
9192
; GFX12-FAKE16-NEXT: s_wait_loadcnt 0x0
9293
; GFX12-FAKE16-NEXT: v_sqrt_bf16_e32 v1, v0
94+
; GFX12-FAKE16-NEXT: v_nop
9395
; GFX12-FAKE16-NEXT: v_lshrrev_b32_e32 v0, 16, v0
9496
; GFX12-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(TRANS32_DEP_2)
9597
; GFX12-FAKE16-NEXT: v_sqrt_bf16_e32 v0, v0

0 commit comments

Comments
 (0)