[llvm] [VPlan] Use VPInstructionWithType for uniform casts. (PR #140623)

[David Sherwood via llvm-commits]

================
@@ -702,20 +702,20 @@ define i32 @narrowed_reduction(ptr %a, i1 %cmp) #0 {
 ; CHECK-NEXT:    [[TMP1:%.*]] = and <4 x i32> [[VEC_PHI1]], splat (i32 1)
 ; CHECK-NEXT:    [[TMP2:%.*]] = or <4 x i32> [[TMP0]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = or <4 x i32> [[TMP1]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP4:%.*]] = trunc <4 x i32> [[TMP2]] to <4 x i1>
----------------
david-arm wrote:

Looks like rename changes only?

https://github.com/llvm/llvm-project/pull/140623