diff --git a/llvm/test/CodeGen/X86/x86-interleaved-access.ll b/llvm/test/CodeGen/X86/x86-interleaved-access.ll
index 4c81fae4ca927..fc81e86585384 100644
--- a/llvm/test/CodeGen/X86/x86-interleaved-access.ll
+++ b/llvm/test/CodeGen/X86/x86-interleaved-access.ll
@@ -1627,3 +1627,305 @@ define void @interleaved_store_vf64_i8_stride4(<64 x i8> %a, <64 x i8> %b, <64 x
 store <256 x i8> %interleaved, <256 x i8>* %p
 ret void
 }
+
+define void @splat2_v4f64_load_store(<4 x double>* %s, <8 x double>* %d) {
+; AVX1-LABEL: splat2_v4f64_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovupd (%rdi), %ymm0
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm1 = ymm0[0,1,0,1]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm1 = ymm1[0,0,3,3]
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[0,0,3,3]
+; AVX1-NEXT:    vmovupd %ymm0, 32(%rsi)
+; AVX1-NEXT:    vmovupd %ymm1, (%rsi)
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat2_v4f64_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm1 = ymm0[0,0,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,2,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat2_v4f64_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vmovups (%rdi), %ymm0
+; AVX512-NEXT:    vmovaps {{.*#+}} zmm1 = [0,0,1,1,2,2,3,3]
+; AVX512-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
+; AVX512-NEXT:    vmovups %zmm0, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <4 x double>, <4 x double>* %s, align 8
+  %x2 = shufflevector <4 x double> %x, <4 x double> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+  %r = shufflevector <8 x double> %x2, <8 x double> undef, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
+  store <8 x double> %r, <8 x double>* %d, align 8
+  ret void
+}
+
+define void @splat2_v4i64_load_store(<4 x i64>* %s, <8 x i64>* %d) {
+; AVX1-LABEL: splat2_v4i64_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovupd (%rdi), %ymm0
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm1 = ymm0[0,1,0,1]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm1 = ymm1[0,0,3,3]
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[0,0,3,3]
+; AVX1-NEXT:    vmovupd %ymm0, 32(%rsi)
+; AVX1-NEXT:    vmovupd %ymm1, (%rsi)
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat2_v4i64_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm1 = ymm0[0,0,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,2,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat2_v4i64_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vmovups (%rdi), %ymm0
+; AVX512-NEXT:    vmovaps {{.*#+}} zmm1 = [0,0,1,1,2,2,3,3]
+; AVX512-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
+; AVX512-NEXT:    vmovups %zmm0, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <4 x i64>, <4 x i64>* %s, align 8
+  %x2 = shufflevector <4 x i64> %x, <4 x i64> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+  %r = shufflevector <8 x i64> %x2, <8 x i64> undef, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
+  store <8 x i64> %r, <8 x i64>* %d, align 8
+  ret void
+}
+
+define void @splat4_v8f32_load_store(<8 x float>* %s, <32 x float>* %d) {
+; AVX1-LABEL: splat4_v8f32_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovups (%rdi), %xmm0
+; AVX1-NEXT:    vmovups 16(%rdi), %xmm1
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm2 = xmm1[0,0,0,0]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm3 = xmm1[1,1,1,1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm4 = xmm1[2,2,2,2]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm1[3,3,3,3]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm5 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm6 = xmm0[1,1,1,1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm7 = xmm0[2,2,2,2]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,3,3,3]
+; AVX1-NEXT:    vmovups %xmm0, 48(%rsi)
+; AVX1-NEXT:    vmovups %xmm7, 32(%rsi)
+; AVX1-NEXT:    vmovups %xmm6, 16(%rsi)
+; AVX1-NEXT:    vmovups %xmm5, (%rsi)
+; AVX1-NEXT:    vmovups %xmm1, 112(%rsi)
+; AVX1-NEXT:    vmovups %xmm4, 96(%rsi)
+; AVX1-NEXT:    vmovups %xmm3, 80(%rsi)
+; AVX1-NEXT:    vmovups %xmm2, 64(%rsi)
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat4_v8f32_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vmovups (%rdi), %xmm1
+; AVX2-NEXT:    vpermilps {{.*#+}} xmm2 = xmm1[0,0,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm2 = ymm2[0,0,1,1]
+; AVX2-NEXT:    vpermilps {{.*#+}} xmm1 = xmm1[2,2,3,3]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm1 = ymm1[0,0,1,1]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm3 = ymm0[0,0,1,1,4,4,5,5]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm3 = ymm3[2,2,3,3]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,2,3,3,6,6,7,7]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,2,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 96(%rsi)
+; AVX2-NEXT:    vmovups %ymm3, 64(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm2, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat4_v8f32_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vbroadcastf64x4 {{.*#+}} zmm0 = mem[0,1,2,3,0,1,2,3]
+; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,8,0,8,1,9,1,9,2,10,2,10,3,11,3,11]
+; AVX512-NEXT:    vpermd %zmm0, %zmm1, %zmm1
+; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [4,12,4,12,5,13,5,13,6,14,6,14,7,15,7,15]
+; AVX512-NEXT:    vpermd %zmm0, %zmm2, %zmm0
+; AVX512-NEXT:    vmovdqu64 %zmm0, 64(%rsi)
+; AVX512-NEXT:    vmovdqu64 %zmm1, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <8 x float>, <8 x float>* %s, align 4
+  %x2 = shufflevector <8 x float> %x, <8 x float> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  %x4 = shufflevector <16 x float> %x2, <16 x float> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  %r = shufflevector <32 x float> %x4, <32 x float> undef, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
+  store <32 x float> %r, <32 x float>* %d, align 4
+  ret void
+}
+
+define void @splat4_v8i32_load_store(<8 x i32>* %s, <32 x i32>* %d) {
+; AVX1-LABEL: splat4_v8i32_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovups (%rdi), %xmm0
+; AVX1-NEXT:    vmovups 16(%rdi), %xmm1
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm2 = xmm1[0,0,0,0]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm3 = xmm1[1,1,1,1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm4 = xmm1[2,2,2,2]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm1[3,3,3,3]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm5 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm6 = xmm0[1,1,1,1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm7 = xmm0[2,2,2,2]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,3,3,3]
+; AVX1-NEXT:    vmovups %xmm0, 48(%rsi)
+; AVX1-NEXT:    vmovups %xmm7, 32(%rsi)
+; AVX1-NEXT:    vmovups %xmm6, 16(%rsi)
+; AVX1-NEXT:    vmovups %xmm5, (%rsi)
+; AVX1-NEXT:    vmovups %xmm1, 112(%rsi)
+; AVX1-NEXT:    vmovups %xmm4, 96(%rsi)
+; AVX1-NEXT:    vmovups %xmm3, 80(%rsi)
+; AVX1-NEXT:    vmovups %xmm2, 64(%rsi)
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat4_v8i32_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vmovups (%rdi), %xmm1
+; AVX2-NEXT:    vpermilps {{.*#+}} xmm2 = xmm1[0,0,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm2 = ymm2[0,0,1,1]
+; AVX2-NEXT:    vpermilps {{.*#+}} xmm1 = xmm1[2,2,3,3]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm1 = ymm1[0,0,1,1]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm3 = ymm0[0,0,1,1,4,4,5,5]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm3 = ymm3[2,2,3,3]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,2,3,3,6,6,7,7]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,2,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 96(%rsi)
+; AVX2-NEXT:    vmovups %ymm3, 64(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm2, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat4_v8i32_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vbroadcasti64x4 {{.*#+}} zmm0 = mem[0,1,2,3,0,1,2,3]
+; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,8,0,8,1,9,1,9,2,10,2,10,3,11,3,11]
+; AVX512-NEXT:    vpermd %zmm0, %zmm1, %zmm1
+; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [4,12,4,12,5,13,5,13,6,14,6,14,7,15,7,15]
+; AVX512-NEXT:    vpermd %zmm0, %zmm2, %zmm0
+; AVX512-NEXT:    vmovdqu64 %zmm0, 64(%rsi)
+; AVX512-NEXT:    vmovdqu64 %zmm1, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <8 x i32>, <8 x i32>* %s, align 4
+  %x2 = shufflevector <8 x i32> %x, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  %x4 = shufflevector <16 x i32> %x2, <16 x i32> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  %r = shufflevector <32 x i32> %x4, <32 x i32> undef, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
+  store <32 x i32> %r, <32 x i32>* %d, align 4
+  ret void
+}
+
+define void @splat4_v4f64_load_store(<4 x double>* %s, <16 x double>* %d) {
+; AVX1-LABEL: splat4_v4f64_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovupd (%rdi), %ymm0
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm1 = ymm0[0,1,0,1]
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX1-NEXT:    vmovddup {{.*#+}} ymm2 = ymm1[0,0,2,2]
+; AVX1-NEXT:    vmovddup {{.*#+}} ymm3 = ymm0[0,0,2,2]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm1 = ymm1[1,1,3,3]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3]
+; AVX1-NEXT:    vmovupd %ymm0, 96(%rsi)
+; AVX1-NEXT:    vmovupd %ymm3, 64(%rsi)
+; AVX1-NEXT:    vmovupd %ymm1, 32(%rsi)
+; AVX1-NEXT:    vmovupd %ymm2, (%rsi)
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat4_v4f64_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vbroadcastsd (%rdi), %ymm1
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm2 = ymm0[2,2,2,2]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm3 = ymm0[1,1,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 96(%rsi)
+; AVX2-NEXT:    vmovups %ymm2, 64(%rsi)
+; AVX2-NEXT:    vmovups %ymm3, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat4_v4f64_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vmovups (%rdi), %ymm0
+; AVX512-NEXT:    vbroadcastsd (%rdi), %ymm1
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm2 = ymm0[2,2,2,2]
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm3 = ymm0[1,1,1,1]
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,3]
+; AVX512-NEXT:    vinsertf64x4 $1, %ymm3, %zmm1, %zmm1
+; AVX512-NEXT:    vinsertf64x4 $1, %ymm0, %zmm2, %zmm0
+; AVX512-NEXT:    vmovups %zmm0, 64(%rsi)
+; AVX512-NEXT:    vmovups %zmm1, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <4 x double>, <4 x double>* %s, align 8
+  %x2 = shufflevector <4 x double> %x, <4 x double> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+  %x4 = shufflevector <8 x double> %x2, <8 x double> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  %r = shufflevector <16 x double> %x4, <16 x double> undef, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
+  store <16 x double> %r, <16 x double>* %d, align 8
+  ret void
+}
+
+define void @splat4_v4i64_load_store(<4 x i64>* %s, <16 x i64>* %d) {
+; AVX1-LABEL: splat4_v4i64_load_store:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovupd (%rdi), %ymm0
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm1 = ymm0[0,1,0,1]
+; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX1-NEXT:    vmovddup {{.*#+}} ymm2 = ymm1[0,0,2,2]
+; AVX1-NEXT:    vmovddup {{.*#+}} ymm3 = ymm0[0,0,2,2]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm1 = ymm1[1,1,3,3]
+; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3]
+; AVX1-NEXT:    vmovupd %ymm0, 96(%rsi)
+; AVX1-NEXT:    vmovupd %ymm3, 64(%rsi)
+; AVX1-NEXT:    vmovupd %ymm1, 32(%rsi)
+; AVX1-NEXT:    vmovupd %ymm2, (%rsi)
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: splat4_v4i64_load_store:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vmovups (%rdi), %ymm0
+; AVX2-NEXT:    vbroadcastsd (%rdi), %ymm1
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm2 = ymm0[2,2,2,2]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm3 = ymm0[1,1,1,1]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,3]
+; AVX2-NEXT:    vmovups %ymm0, 96(%rsi)
+; AVX2-NEXT:    vmovups %ymm2, 64(%rsi)
+; AVX2-NEXT:    vmovups %ymm3, 32(%rsi)
+; AVX2-NEXT:    vmovups %ymm1, (%rsi)
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: splat4_v4i64_load_store:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vmovups (%rdi), %ymm0
+; AVX512-NEXT:    vbroadcastsd (%rdi), %ymm1
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm2 = ymm0[2,2,2,2]
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm3 = ymm0[1,1,1,1]
+; AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,3]
+; AVX512-NEXT:    vinsertf64x4 $1, %ymm3, %zmm1, %zmm1
+; AVX512-NEXT:    vinsertf64x4 $1, %ymm0, %zmm2, %zmm0
+; AVX512-NEXT:    vmovups %zmm0, 64(%rsi)
+; AVX512-NEXT:    vmovups %zmm1, (%rsi)
+; AVX512-NEXT:    vzeroupper
+; AVX512-NEXT:    retq
+  %x = load <4 x i64>, <4 x i64>* %s, align 8
+  %x2 = shufflevector <4 x i64> %x, <4 x i64> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+  %x4 = shufflevector <8 x i64> %x2, <8 x i64> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  %r = shufflevector <16 x i64> %x4, <16 x i64> undef, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
+  store <16 x i64> %r, <16 x i64>* %d, align 8
+  ret void
+}