[llvm] [AMDGPU] Filter candidates of LiveRegOptimizer for profitable cases (PR #124624)

[Jeffrey Byrnes via llvm-commits]

================
@@ -63,19 +55,34 @@ define amdgpu_kernel void @v4i8_liveout(ptr addrspace(1) %src1, ptr addrspace(1)
 ; GFX906:       ; %bb.0: ; %entry
 ; GFX906-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
 ; GFX906-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
-; GFX906-NEXT:    v_lshlrev_b32_e32 v2, 2, v0
+; GFX906-NEXT:    v_lshlrev_b32_e32 v5, 2, v0
 ; GFX906-NEXT:    v_cmp_gt_u32_e32 vcc, 15, v0
 ; GFX906-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX906-NEXT:    global_load_dword v1, v2, s[0:1]
+; GFX906-NEXT:    global_load_dword v1, v5, s[0:1]
+; GFX906-NEXT:    s_waitcnt vmcnt(0)
+; GFX906-NEXT:    v_lshrrev_b32_e32 v2, 8, v1
+; GFX906-NEXT:    v_lshrrev_b32_e32 v3, 16, v1
+; GFX906-NEXT:    v_lshrrev_b32_e32 v4, 24, v1
----------------
jrbyrnes wrote:

LRO should coerce multiple tests in this file (analogous to below)

https://github.com/llvm/llvm-project/pull/124624