[llvm] [AMDGPU][True16][CodeGen] true16 isel pattern for fma_mix_f16/bf16 (PR #159648)

[Stanislav Mekhanoshin via llvm-commits]

================
@@ -65,10 +65,9 @@ define <2 x half> @v_mad_mixhi_f16_f16lo_f16lo_f16lo_constlo(half %src0, half %s
 ; SDAG-GFX11-TRUE16-LABEL: v_mad_mixhi_f16_f16lo_f16lo_f16lo_constlo:
 ; SDAG-GFX11-TRUE16:       ; %bb.0:
 ; SDAG-GFX11-TRUE16-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SDAG-GFX11-TRUE16-NEXT:    v_mov_b16_e32 v3.l, 0x3c00
-; SDAG-GFX11-TRUE16-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
-; SDAG-GFX11-TRUE16-NEXT:    v_fma_mixhi_f16 v3, v0, v1, v2 op_sel_hi:[1,1,1]
-; SDAG-GFX11-TRUE16-NEXT:    v_mov_b32_e32 v0, v3
+; SDAG-GFX11-TRUE16-NEXT:    v_fma_mixlo_f16 v0, v0, v1, v2 op_sel_hi:[1,1,1]
----------------
rampitec wrote:

Is this right it changed from v_fma_mixhi_f16 to v_fma_mixlo_f16?

https://github.com/llvm/llvm-project/pull/159648