[llvm] [VPlan] Run CSE closer to VPlan::execute. (PR #160572)

[Ramkumar Ramachandra via llvm-commits]

================
@@ -2675,8 +2675,7 @@ define i32 @predicated_not_dominates_reduction_twoadd(ptr nocapture noundef read
 ; CHECK-NEXT:    [[TMP5:%.*]] = select <4 x i1> [[TMP6]], <4 x i32> [[TMP13]], <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP5]])
 ; CHECK-NEXT:    [[TMP8:%.*]] = add i32 [[VEC_PHI]], [[TMP7]]
-; CHECK-NEXT:    [[TMP18:%.*]] = zext <4 x i8> [[TMP4]] to <4 x i32>
----------------
artagnon wrote:

Can you unstage all the changes prior to this one, as they're not relevant?

https://github.com/llvm/llvm-project/pull/160572