[llvm] [LV] Handle FSub Partial Reductions (PR #191186)

[Benjamin Maxwell via llvm-commits]

================
@@ -1308,35 +1308,35 @@ define float @fcmp_fadd_fsub(ptr nocapture %a, i32 %n) {
 ; CHECK:       [[VECTOR_PH]]:
 ; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 4
 ; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF]]
-; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
-; CHECK:       [[VECTOR_BODY]]:
-; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x float> [ zeroinitializer, %[[VECTOR_PH]] ], [ [[PREDPHI1:%.*]], %[[VECTOR_BODY]] ]
----------------
MacDue wrote:

Can you revert all the changes in this file. I don't think anything has actually changed other than the names in `%[[..]]`. 

https://github.com/llvm/llvm-project/pull/191186