diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 6eb190fa10a..9c02303ed11 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -11765,6 +11765,10 @@ static SDValue lowerV16F32VectorShuffle(SDLoc DL, ArrayRef<int> Mask,
       return DAG.getNode(X86ISD::MOVSLDUP, DL, MVT::v16f32, V1);
     if (isShuffleEquivalent(V1, V2, RepeatedMask, {1, 1, 3, 3}))
       return DAG.getNode(X86ISD::MOVSHDUP, DL, MVT::v16f32, V1);
+
+    if (V2.isUndef())
+      return DAG.getNode(X86ISD::VPERMILPI, DL, MVT::v16f32, V1,
+                         getV4X86ShuffleImm8ForMask(RepeatedMask, DL, DAG));
   }
 
   if (SDValue Unpck =
diff --git a/test/CodeGen/X86/vector-shuffle-512-v16.ll b/test/CodeGen/X86/vector-shuffle-512-v16.ll
index 5f01909eb31..1f242fef117 100644
--- a/test/CodeGen/X86/vector-shuffle-512-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-512-v16.ll
@@ -107,6 +107,33 @@ define <16 x float> @shuffle_v16f32_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15
   ret <16 x float> %shuffle
 }
 
+define <16 x float> @shuffle_v16f32_00_01_00_01_06_07_06_07_08_09_10_11_12_13_12_13(<16 x float> %a, <16 x float> %b) {
+; ALL-LABEL: shuffle_v16f32_00_01_00_01_06_07_06_07_08_09_10_11_12_13_12_13:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpermilpd {{.*#+}} zmm0 = zmm0[0,0,3,3,4,5,6,6]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 6, i32 7, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 12, i32 13>
+  ret <16 x float> %shuffle
+}
+
+define <16 x float> @shuffle_v16f32_00_00_02_00_04_04_06_04_08_08_10_08_12_12_14_12(<16 x float> %a, <16 x float> %b) {
+; ALL-LABEL: shuffle_v16f32_00_00_02_00_04_04_06_04_08_08_10_08_12_12_14_12:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpermilps {{.*#+}} zmm0 = zmm0[0,0,2,0,4,4,6,4,8,8,10,8,12,12,14,12]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 0, i32 2, i32 0, i32 4, i32 4, i32 6, i32 4, i32 8, i32 8, i32 10, i32 8, i32 12, i32 12, i32 14, i32 12>
+  ret <16 x float> %shuffle
+}
+
+define <16 x float> @shuffle_v16f32_03_uu_uu_uu_uu_04_uu_uu_uu_uu_11_uu_uu_uu_uu_12(<16 x float> %a, <16 x float> %b) {
+; ALL-LABEL: shuffle_v16f32_03_uu_uu_uu_uu_04_uu_uu_uu_uu_11_uu_uu_uu_uu_12:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpermilps {{.*#+}} zmm0 = zmm0[3,0,3,0,7,4,7,4,11,8,11,8,15,12,15,12]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 4, i32 undef, i32 undef, i32 undef, i32 undef, i32 11, i32 undef, i32 undef, i32 undef, i32 undef, i32 12>
+  ret <16 x float> %shuffle
+}
+
 define <16 x i32> @shuffle_v16i32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<16 x i32> %a, <16 x i32> %b) {
 ; ALL-LABEL: shuffle_v16i32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
 ; ALL:       # BB#0: