[llvm] AMDGPU: VALU data fast-forwarding needs no s_delay (PR #205481)

[Jay Foad via llvm-commits]

================
@@ -46,8 +46,8 @@ define amdgpu_ps void @s_fabs_f16_salu_use(half inreg %in, i32 inreg %val, ptr a
 ; GFX12:       ; %bb.0:
 ; GFX12-NEXT:    s_and_b32 s0, s0, 0x7fff
 ; GFX12-NEXT:    s_cmp_eq_u32 s1, 0
+; GFX12-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
----------------
jayfoad wrote:

New delay between two SALU instructions that implicitly write and read scc.

https://github.com/llvm/llvm-project/pull/205481