[SLP] Improve reordering in presence of constant only nodes.

We can skip the analysis of the constant nodes, their order should not
affect the ordering of the trees/subtrees.

Differential Revision: https://reviews.llvm.org/D127775

Author: alexey-bataev

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

@@ -3717,10 +3717,7 @@ void BoUpSLP::reorderTopToBottom() {
                      EI.UserTE->isAltShuffle() && EI.UserTE->Idx != 0;
             }))
           return;
-        if (UserTE->UserTreeIndices.empty())
-          UserTE = nullptr;
-        else
-          UserTE = UserTE->UserTreeIndices.back().UserTE;
+        UserTE = UserTE->UserTreeIndices.back().UserTE;
         ++Cnt;
       }
       VFToOrderedEntries[TE->Scalars.size()].insert(TE.get());
@@ -3885,15 +3882,17 @@ bool BoUpSLP::canReorderOperands(
     }
     ArrayRef<Value *> VL = UserTE->getOperand(I);
     TreeEntry *Gather = nullptr;
-    if (count_if(ReorderableGathers, [VL, &Gather](TreeEntry *TE) {
-          assert(TE->State != TreeEntry::Vectorize &&
-                 "Only non-vectorized nodes are expected.");
-          if (TE->isSame(VL)) {
-            Gather = TE;
-            return true;
-          }
-          return false;
-        }) > 1)
+    if (count_if(ReorderableGathers,
+                 [VL, &Gather](TreeEntry *TE) {
+                   assert(TE->State != TreeEntry::Vectorize &&
+                          "Only non-vectorized nodes are expected.");
+                   if (TE->isSame(VL)) {
+                     Gather = TE;
+                     return true;
+                   }
+                   return false;
+                 }) > 1 &&
+        !all_of(VL, isConstant))
       return false;
     if (Gather)
       GatherOps.push_back(Gather);

llvm/test/Transforms/SLPVectorizer/X86/reorder_diamond_match.ll

@@ -8,14 +8,13 @@ define void @test() {
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP1]] to <4 x i8>*
 ; CHECK-NEXT:    [[TMP4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP3]], align 1
 ; CHECK-NEXT:    [[TMP5:%.*]] = zext <4 x i8> [[TMP4]] to <4 x i32>
-; CHECK-NEXT:    [[SHUFFLE:%.*]] = shufflevector <4 x i32> [[TMP5]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
-; CHECK-NEXT:    [[TMP6:%.*]] = sub nsw <4 x i32> zeroinitializer, [[SHUFFLE]]
+; CHECK-NEXT:    [[TMP6:%.*]] = sub nsw <4 x i32> zeroinitializer, [[TMP5]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = shl nsw <4 x i32> [[TMP6]], zeroinitializer
 ; CHECK-NEXT:    [[TMP8:%.*]] = add nsw <4 x i32> [[TMP7]], zeroinitializer
 ; CHECK-NEXT:    [[TMP9:%.*]] = shufflevector <4 x i32> [[TMP8]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
 ; CHECK-NEXT:    [[TMP10:%.*]] = add nsw <4 x i32> [[TMP8]], [[TMP9]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = sub nsw <4 x i32> [[TMP8]], [[TMP9]]
-; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <4 x i32> [[TMP10]], <4 x i32> [[TMP11]], <4 x i32> <i32 0, i32 5, i32 2, i32 7>
+; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <4 x i32> [[TMP10]], <4 x i32> [[TMP11]], <4 x i32> <i32 1, i32 4, i32 3, i32 6>
 ; CHECK-NEXT:    [[TMP13:%.*]] = add nsw <4 x i32> zeroinitializer, [[TMP12]]
 ; CHECK-NEXT:    [[TMP14:%.*]] = sub nsw <4 x i32> zeroinitializer, [[TMP12]]
 ; CHECK-NEXT:    [[TMP15:%.*]] = shufflevector <4 x i32> [[TMP13]], <4 x i32> [[TMP14]], <4 x i32> <i32 0, i32 1, i32 6, i32 7>

llvm/test/Transforms/SLPVectorizer/X86/scatter-vectorize-reorder.ll

@@ -17,14 +17,14 @@ define void @test() {
 ; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x float> [[TMP5]], float [[TMP2]], i32 1
 ; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x float> <float poison, float 0.000000e+00>, float [[TMP2]], i32 0
 ; CHECK-NEXT:    [[TMP8:%.*]] = call <2 x float> @llvm.fmuladd.v2f32(<2 x float> [[TMP4]], <2 x float> [[TMP6]], <2 x float> [[TMP7]])
-; CHECK-NEXT:    [[SHUFFLE:%.*]] = shufflevector <2 x float> [[TMP8]], <2 x float> poison, <2 x i32> <i32 1, i32 0>
 ; CHECK-NEXT:    br i1 false, label [[BB2:%.*]], label [[BB3:%.*]]
 ; CHECK:       bb2:
-; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x float> [[SHUFFLE]], zeroinitializer
+; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x float> [[TMP8]], zeroinitializer
 ; CHECK-NEXT:    br label [[BB3]]
 ; CHECK:       bb3:
 ; CHECK-NEXT:    [[TMP10:%.*]] = phi <2 x float> [ [[TMP9]], [[BB2]] ], [ zeroinitializer, [[BB1]] ]
-; CHECK-NEXT:    [[TMP11:%.*]] = fadd <2 x float> [[TMP1]], [[TMP10]]
+; CHECK-NEXT:    [[SHUFFLE:%.*]] = shufflevector <2 x float> [[TMP10]], <2 x float> poison, <2 x i32> <i32 1, i32 0>
+; CHECK-NEXT:    [[TMP11:%.*]] = fadd <2 x float> [[TMP1]], [[SHUFFLE]]
 ; CHECK-NEXT:    [[TMP12:%.*]] = fadd <2 x float> [[TMP11]], zeroinitializer
 ; CHECK-NEXT:    [[TMP13:%.*]] = fsub <2 x float> [[TMP12]], zeroinitializer
 ; CHECK-NEXT:    [[TMP14:%.*]] = fsub <2 x float> [[TMP13]], zeroinitializer