[llvm] [X86] Fold expand(splat,passthrough,mask) -> select(splat,passthrough,mask) (PR #180238)

[Phoebe Wang via llvm-commits]

================
@@ -1035,28 +1035,25 @@ define <8 x double> @concat_vpermilvar_v8f64_v4f64(<4 x double> %a0, <4 x double
   ret <8 x double> %res
 }
 
-define <16 x float> @combine_vexpandd_of_broadcast(float %x, <16 x float> %y, i16 %m) {
-; X86-LABEL: combine_vexpandd_of_broadcast:
+define <16 x float> @combine_vexpand_of_broadcast(float %x, <16 x float> %y, i16 %m) {
----------------
phoebewang wrote:

Is the d intended for 32-bit? I see below tests also have d there.

https://github.com/llvm/llvm-project/pull/180238