[llvm] [AMDGPU] Vectorize i8 Shuffles (PR #105850)

[Matt Arsenault via llvm-commits]

================
@@ -363,11 +363,67 @@ bb:
   ret <4 x i16> %ins.3
 }
 
+define <4 x i8> @uadd_sat_v4i8(<4 x i8> %arg0, <4 x i8> %arg1, ptr addrspace(1) %dst) {
+; GCN-LABEL: @uadd_sat_v4i8(
+; GCN-NEXT:  bb:
+; GCN-NEXT:    [[TMP0:%.*]] = call <4 x i8> @llvm.uadd.sat.v4i8(<4 x i8> [[ARG0:%.*]], <4 x i8> [[ARG1:%.*]])
----------------
arsenm wrote:

This is a bad side effect: https://godbolt.org/z/aKxncz77e

I expect similar degradation for all of the other cases 

https://github.com/llvm/llvm-project/pull/105850