diff --git a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
index 4cd61ee165aad5..442e54eaba8f64 100644
--- a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
+++ b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
@@ -8304,20 +8304,31 @@ class BoUpSLP::ShuffleCostEstimator : public BaseShuffleAnalysis {
         });
     // FIXME: this must be moved to TTI for better estimation.
     unsigned EltsPerVector = getPartNumElems(VL.size(), NumParts);
-    auto CheckPerRegistersShuffle =
-        [&](MutableArrayRef<int> Mask,
-            SmallVector<int> Indices) -> std::optional<TTI::ShuffleKind> {
+    auto CheckPerRegistersShuffle = [&](MutableArrayRef<int> Mask,
+                                        SmallVectorImpl<unsigned> &Indices)
+        -> std::optional<TTI::ShuffleKind> {
       if (NumElts <= EltsPerVector)
         return std::nullopt;
+      int OffsetReg0 =
+          alignDown(std::accumulate(Mask.begin(), Mask.end(), INT_MAX,
+                                    [](int S, int I) {
+                                      if (I == PoisonMaskElem)
+                                        return S;
+                                      return std::min(S, I);
+                                    }),
+                    EltsPerVector);
+      int OffsetReg1 = OffsetReg0;
       DenseSet<int> RegIndices;
       // Check that if trying to permute same single/2 input vectors.
       TTI::ShuffleKind ShuffleKind = TTI::SK_PermuteSingleSrc;
       int FirstRegId = -1;
-      Indices.assign(1, -1);
-      for (int &I : Mask) {
+      Indices.assign(1, OffsetReg0);
+      for (auto [Pos, I] : enumerate(Mask)) {
         if (I == PoisonMaskElem)
           continue;
-        int RegId = (I / NumElts) * NumParts + (I % NumElts) / EltsPerVector;
+        int Idx = I - OffsetReg0;
+        int RegId =
+            (Idx / NumElts) * NumParts + (Idx % NumElts) / EltsPerVector;
         if (FirstRegId < 0)
           FirstRegId = RegId;
         RegIndices.insert(RegId);
@@ -8325,14 +8336,25 @@ class BoUpSLP::ShuffleCostEstimator : public BaseShuffleAnalysis {
           return std::nullopt;
         if (RegIndices.size() == 2) {
           ShuffleKind = TTI::SK_PermuteTwoSrc;
-          if (Indices.size() == 1)
-            Indices.push_back(-1);
+          if (Indices.size() == 1) {
+            OffsetReg1 = alignDown(
+                std::accumulate(
+                    std::next(Mask.begin(), Pos), Mask.end(), INT_MAX,
+                    [&](int S, int I) {
+                      if (I == PoisonMaskElem)
+                        return S;
+                      int RegId = ((I - OffsetReg0) / NumElts) * NumParts +
+                                  ((I - OffsetReg0) % NumElts) / EltsPerVector;
+                      if (RegId == FirstRegId)
+                        return S;
+                      return std::min(S, I);
+                    }),
+                EltsPerVector);
+            Indices.push_back(OffsetReg1);
+          }
+          Idx = I - OffsetReg1;
         }
-        if (RegId == FirstRegId)
-          Indices.front() = I % NumElts;
-        else
-          Indices.back() = I % NumElts;
-        I = (I % NumElts) % EltsPerVector +
+        I = (Idx % NumElts) % EltsPerVector +
             (RegId == FirstRegId ? 0 : EltsPerVector);
       }
       return ShuffleKind;
@@ -8349,7 +8371,7 @@ class BoUpSLP::ShuffleCostEstimator : public BaseShuffleAnalysis {
           Part * EltsPerVector, getNumElems(Mask.size(), EltsPerVector, Part));
       SmallVector<int> SubMask(EltsPerVector, PoisonMaskElem);
       copy(MaskSlice, SubMask.begin());
-      SmallVector<int> Indices;
+      SmallVector<unsigned, 2> Indices;
       std::optional<TTI::ShuffleKind> RegShuffleKind =
           CheckPerRegistersShuffle(SubMask, Indices);
       if (!RegShuffleKind) {
@@ -8367,12 +8389,21 @@ class BoUpSLP::ShuffleCostEstimator : public BaseShuffleAnalysis {
                                  FixedVectorType::get(ScalarTy, EltsPerVector),
                                  SubMask);
       }
-      for (int Idx : Indices) {
+      for (unsigned Idx : Indices) {
         Cost += ::getShuffleCost(TTI, TTI::SK_ExtractSubvector,
                                  FixedVectorType::get(ScalarTy, NumElts),
                                  std::nullopt, CostKind, Idx,
                                  FixedVectorType::get(ScalarTy, EltsPerVector));
       }
+      // Second attempt to check, if just a permute is better estimated than
+      // subvector extract.
+      SubMask.assign(NumElts, PoisonMaskElem);
+      copy(MaskSlice, SubMask.begin());
+      InstructionCost OriginalCost =
+          ::getShuffleCost(TTI, *ShuffleKinds[Part],
+                           FixedVectorType::get(ScalarTy, NumElts), SubMask);
+      if (OriginalCost < Cost)
+        Cost = OriginalCost;
     }
     return Cost;
   }
diff --git a/llvm/test/Transforms/SLPVectorizer/RISCV/math-function.ll b/llvm/test/Transforms/SLPVectorizer/RISCV/math-function.ll
index 9608608a180982..059e4c38b519bd 100644
--- a/llvm/test/Transforms/SLPVectorizer/RISCV/math-function.ll
+++ b/llvm/test/Transforms/SLPVectorizer/RISCV/math-function.ll
@@ -155,11 +155,13 @@ define <4 x float> @exp_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @expf(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.exp.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @expf(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @expf(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @exp_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -171,11 +173,13 @@ define <4 x float> @exp_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @expf(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.exp.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @expf(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @expf(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
@@ -208,11 +212,13 @@ define <4 x float> @int_exp_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.exp.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @int_exp_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -224,11 +230,13 @@ define <4 x float> @int_exp_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.exp.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.exp.f32(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
@@ -261,11 +269,13 @@ define <4 x float> @log_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @logf(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.log.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @logf(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @logf(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @log_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -277,11 +287,13 @@ define <4 x float> @log_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @logf(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.log.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @logf(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @logf(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
@@ -314,11 +326,13 @@ define <4 x float> @int_log_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.log.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @int_log_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -330,11 +344,13 @@ define <4 x float> @int_log_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.log.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.log.f32(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
@@ -367,11 +383,13 @@ define <4 x float> @sin_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @sinf(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.sin.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @sinf(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @sinf(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @sin_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -383,11 +401,13 @@ define <4 x float> @sin_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @sinf(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.sin.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @sinf(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @sinf(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
@@ -420,11 +440,13 @@ define <4 x float> @int_sin_4x(ptr %a) {
 ; CHECK-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; CHECK-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_1]])
 ; CHECK-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.sin.v2f32(<2 x float> [[TMP3]])
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    ret <4 x float> [[VECINS_31]]
+; CHECK-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; CHECK-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_2]])
+; CHECK-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; CHECK-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; CHECK-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_3]])
+; CHECK-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; CHECK-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 ; DEFAULT-LABEL: define <4 x float> @int_sin_4x
 ; DEFAULT-SAME: (ptr [[A:%.*]]) #[[ATTR1]] {
@@ -436,11 +458,13 @@ define <4 x float> @int_sin_4x(ptr %a) {
 ; DEFAULT-NEXT:    [[VECEXT_1:%.*]] = extractelement <4 x float> [[TMP0]], i32 1
 ; DEFAULT-NEXT:    [[TMP2:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_1]])
 ; DEFAULT-NEXT:    [[VECINS_1:%.*]] = insertelement <4 x float> [[VECINS]], float [[TMP2]], i32 1
-; DEFAULT-NEXT:    [[TMP3:%.*]] = shufflevector <4 x float> [[TMP0]], <4 x float> poison, <2 x i32> <i32 2, i32 3>
-; DEFAULT-NEXT:    [[TMP4:%.*]] = call fast <2 x float> @llvm.sin.v2f32(<2 x float> [[TMP3]])
-; DEFAULT-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; DEFAULT-NEXT:    [[VECINS_31:%.*]] = shufflevector <4 x float> [[VECINS_1]], <4 x float> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; DEFAULT-NEXT:    ret <4 x float> [[VECINS_31]]
+; DEFAULT-NEXT:    [[VECEXT_2:%.*]] = extractelement <4 x float> [[TMP0]], i32 2
+; DEFAULT-NEXT:    [[TMP3:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_2]])
+; DEFAULT-NEXT:    [[VECINS_2:%.*]] = insertelement <4 x float> [[VECINS_1]], float [[TMP3]], i32 2
+; DEFAULT-NEXT:    [[VECEXT_3:%.*]] = extractelement <4 x float> [[TMP0]], i32 3
+; DEFAULT-NEXT:    [[TMP4:%.*]] = tail call fast float @llvm.sin.f32(float [[VECEXT_3]])
+; DEFAULT-NEXT:    [[VECINS_3:%.*]] = insertelement <4 x float> [[VECINS_2]], float [[TMP4]], i32 3
+; DEFAULT-NEXT:    ret <4 x float> [[VECINS_3]]
 ;
 entry:
   %0 = load <4 x float>, ptr %a, align 16
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls-inseltpoison.ll b/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls-inseltpoison.ll
index 45ce1eec2cbfcd..6c21cc1cfc5be8 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls-inseltpoison.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls-inseltpoison.ll
@@ -51,23 +51,25 @@ define <8 x float> @ceil_floor(<8 x float> %a) {
 ;
 ; AVX-LABEL: @ceil_floor(
 ; AVX-NEXT:    [[A0:%.*]] = extractelement <8 x float> [[A:%.*]], i64 0
+; AVX-NEXT:    [[A1:%.*]] = extractelement <8 x float> [[A]], i64 1
+; AVX-NEXT:    [[A2:%.*]] = extractelement <8 x float> [[A]], i64 2
 ; AVX-NEXT:    [[A3:%.*]] = extractelement <8 x float> [[A]], i64 3
 ; AVX-NEXT:    [[AB0:%.*]] = call float @llvm.ceil.f32(float [[A0]])
-; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 1, i32 2>
-; AVX-NEXT:    [[TMP2:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP1]])
+; AVX-NEXT:    [[AB1:%.*]] = call float @llvm.floor.f32(float [[A1]])
+; AVX-NEXT:    [[AB2:%.*]] = call float @llvm.floor.f32(float [[A2]])
 ; AVX-NEXT:    [[AB3:%.*]] = call float @llvm.ceil.f32(float [[A3]])
-; AVX-NEXT:    [[TMP3:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:    [[TMP4:%.*]] = call <2 x float> @llvm.ceil.v2f32(<2 x float> [[TMP3]])
-; AVX-NEXT:    [[TMP5:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 6, i32 7>
-; AVX-NEXT:    [[TMP6:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP5]])
+; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:    [[TMP2:%.*]] = call <2 x float> @llvm.ceil.v2f32(<2 x float> [[TMP1]])
+; AVX-NEXT:    [[TMP3:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:    [[TMP4:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP3]])
 ; AVX-NEXT:    [[R0:%.*]] = insertelement <8 x float> poison, float [[AB0]], i64 0
-; AVX-NEXT:    [[TMP7:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R23:%.*]] = shufflevector <8 x float> [[R0]], <8 x float> [[TMP7]], <8 x i32> <i32 0, i32 8, i32 9, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R3:%.*]] = insertelement <8 x float> [[R23]], float [[AB3]], i64 3
-; AVX-NEXT:    [[TMP8:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R52:%.*]] = shufflevector <8 x float> [[R3]], <8 x float> [[TMP8]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 poison, i32 poison>
-; AVX-NEXT:    [[TMP9:%.*]] = shufflevector <2 x float> [[TMP6]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R71:%.*]] = shufflevector <8 x float> [[R52]], <8 x float> [[TMP9]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
+; AVX-NEXT:    [[R1:%.*]] = insertelement <8 x float> [[R0]], float [[AB1]], i64 1
+; AVX-NEXT:    [[R2:%.*]] = insertelement <8 x float> [[R1]], float [[AB2]], i64 2
+; AVX-NEXT:    [[R3:%.*]] = insertelement <8 x float> [[R2]], float [[AB3]], i64 3
+; AVX-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; AVX-NEXT:    [[R52:%.*]] = shufflevector <8 x float> [[R3]], <8 x float> [[TMP5]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 poison, i32 poison>
+; AVX-NEXT:    [[TMP6:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; AVX-NEXT:    [[R71:%.*]] = shufflevector <8 x float> [[R52]], <8 x float> [[TMP6]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; AVX-NEXT:    ret <8 x float> [[R71]]
 ;
 ; AVX2-LABEL: @ceil_floor(
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls.ll b/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls.ll
index b8b284b9595a41..bc5bcee361168a 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/alternate-calls.ll
@@ -51,23 +51,25 @@ define <8 x float> @ceil_floor(<8 x float> %a) {
 ;
 ; AVX-LABEL: @ceil_floor(
 ; AVX-NEXT:    [[A0:%.*]] = extractelement <8 x float> [[A:%.*]], i64 0
+; AVX-NEXT:    [[A1:%.*]] = extractelement <8 x float> [[A]], i64 1
+; AVX-NEXT:    [[A2:%.*]] = extractelement <8 x float> [[A]], i64 2
 ; AVX-NEXT:    [[A3:%.*]] = extractelement <8 x float> [[A]], i64 3
 ; AVX-NEXT:    [[AB0:%.*]] = call float @llvm.ceil.f32(float [[A0]])
-; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 1, i32 2>
-; AVX-NEXT:    [[TMP2:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP1]])
+; AVX-NEXT:    [[AB1:%.*]] = call float @llvm.floor.f32(float [[A1]])
+; AVX-NEXT:    [[AB2:%.*]] = call float @llvm.floor.f32(float [[A2]])
 ; AVX-NEXT:    [[AB3:%.*]] = call float @llvm.ceil.f32(float [[A3]])
-; AVX-NEXT:    [[TMP3:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:    [[TMP4:%.*]] = call <2 x float> @llvm.ceil.v2f32(<2 x float> [[TMP3]])
-; AVX-NEXT:    [[TMP5:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 6, i32 7>
-; AVX-NEXT:    [[TMP6:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP5]])
+; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:    [[TMP2:%.*]] = call <2 x float> @llvm.ceil.v2f32(<2 x float> [[TMP1]])
+; AVX-NEXT:    [[TMP3:%.*]] = shufflevector <8 x float> [[A]], <8 x float> poison, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:    [[TMP4:%.*]] = call <2 x float> @llvm.floor.v2f32(<2 x float> [[TMP3]])
 ; AVX-NEXT:    [[R0:%.*]] = insertelement <8 x float> poison, float [[AB0]], i64 0
-; AVX-NEXT:    [[TMP7:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R23:%.*]] = shufflevector <8 x float> [[R0]], <8 x float> [[TMP7]], <8 x i32> <i32 0, i32 8, i32 9, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R3:%.*]] = insertelement <8 x float> [[R23]], float [[AB3]], i64 3
-; AVX-NEXT:    [[TMP8:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R52:%.*]] = shufflevector <8 x float> [[R3]], <8 x float> [[TMP8]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 poison, i32 poison>
-; AVX-NEXT:    [[TMP9:%.*]] = shufflevector <2 x float> [[TMP6]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT:    [[R71:%.*]] = shufflevector <8 x float> [[R52]], <8 x float> [[TMP9]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
+; AVX-NEXT:    [[R1:%.*]] = insertelement <8 x float> [[R0]], float [[AB1]], i64 1
+; AVX-NEXT:    [[R2:%.*]] = insertelement <8 x float> [[R1]], float [[AB2]], i64 2
+; AVX-NEXT:    [[R3:%.*]] = insertelement <8 x float> [[R2]], float [[AB3]], i64 3
+; AVX-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; AVX-NEXT:    [[R52:%.*]] = shufflevector <8 x float> [[R3]], <8 x float> [[TMP5]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 poison, i32 poison>
+; AVX-NEXT:    [[TMP6:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; AVX-NEXT:    [[R71:%.*]] = shufflevector <8 x float> [[R52]], <8 x float> [[TMP6]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; AVX-NEXT:    ret <8 x float> [[R71]]
 ;
 ; AVX2-LABEL: @ceil_floor(