[X86][SSE] Added i686 target tests to make sure we are correctly loading consecutive entries as 64-bit integers

llvm-svn: 259794

3 files changed, 504 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
index 459ec10e4e5..6f1daad843f 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
@@ -4,6 +4,9 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX --check-prefix=AVX512F
+;
+; Just one 32-bit run to make sure we do reasonable things.
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=X32-SSE
 
 define <2 x double> @merge_2f64_f64_23(double* %ptr) nounwind uwtable noinline ssp {
 ; SSE-LABEL: merge_2f64_f64_23:
@@ -15,6 +18,12 @@ define <2 x double> @merge_2f64_f64_23(double* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 16(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_2f64_f64_23:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups 16(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 2
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 3
   %val0 = load double, double* %ptr0
@@ -34,6 +43,12 @@ define <2 x i64> @merge_2i64_i64_12(i64* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 8(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_2i64_i64_12:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups 8(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %ptr1 = getelementptr inbounds i64, i64* %ptr, i64 2
   %val0 = load i64, i64* %ptr0
@@ -53,6 +68,12 @@ define <4 x float> @merge_4f32_f32_2345(float* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 8(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_2345:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups 8(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 2
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 3
   %ptr2 = getelementptr inbounds float, float* %ptr, i64 4
@@ -78,6 +99,12 @@ define <4 x float> @merge_4f32_f32_3zuu(float* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_3zuu:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 3
   %val0 = load float, float* %ptr0
   %res0 = insertelement <4 x float> undef, float %val0, i32 0
@@ -95,6 +122,12 @@ define <4 x float> @merge_4f32_f32_34uu(float* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_34uu:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 3
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 4
   %val0 = load float, float* %ptr0
@@ -118,6 +151,14 @@ define <4 x float> @merge_4f32_f32_34z6(float* %ptr) nounwind uwtable noinline s
 ; AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[1,0]
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_34z6:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[1,0]
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 3
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 4
   %ptr3 = getelementptr inbounds float, float* %ptr, i64 6
@@ -140,6 +181,12 @@ define <4 x float> @merge_4f32_f32_45zz(float* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_45zz:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 4
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 5
   %val0 = load float, float* %ptr0
@@ -169,6 +216,13 @@ define <4 x float> @merge_4f32_f32_012u(float* %ptr) nounwind uwtable noinline s
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_012u:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    insertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 0
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr2 = getelementptr inbounds float, float* %ptr, i64 2
@@ -203,6 +257,13 @@ define <4 x float> @merge_4f32_f32_019u(float* %ptr) nounwind uwtable noinline s
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4f32_f32_019u:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    insertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 0
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr2 = getelementptr inbounds float, float* %ptr, i64 9
@@ -226,6 +287,12 @@ define <4 x i32> @merge_4i32_i32_23u5(i32* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 8(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4i32_i32_23u5:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups 8(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 2
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptr3 = getelementptr inbounds i32, i32* %ptr, i64 5
@@ -248,6 +315,12 @@ define <4 x i32> @merge_4i32_i32_3zuu(i32* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4i32_i32_3zuu:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 3
   %val0 = load i32, i32* %ptr0
   %res0 = insertelement <4 x i32> undef, i32 %val0, i32 0
@@ -265,6 +338,12 @@ define <4 x i32> @merge_4i32_i32_34uu(i32* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4i32_i32_34uu:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 4
   %val0 = load i32, i32* %ptr0
@@ -284,6 +363,12 @@ define <4 x i32> @merge_4i32_i32_45zz(i32* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_4i32_i32_45zz:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 4
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 5
   %val0 = load i32, i32* %ptr0
@@ -303,6 +388,12 @@ define <8 x i16> @merge_8i16_i16_23u567u9(i16* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 4(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_8i16_i16_23u567u9:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups 4(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 2
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 3
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 5
@@ -336,6 +427,13 @@ define <8 x i16> @merge_8i16_i16_34uuuuuu(i16* %ptr) nounwind uwtable noinline s
 ; AVX-NEXT:    vpinsrw $0, 6(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    vpinsrw $1, 8(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_8i16_i16_34uuuuuu:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    pinsrw $0, 6(%eax), %xmm0
+; X32-SSE-NEXT:    pinsrw $1, 8(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 3
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 4
   %val0 = load i16, i16* %ptr0
@@ -355,6 +453,12 @@ define <8 x i16> @merge_8i16_i16_45u7zzzz(i16* %ptr) nounwind uwtable noinline s
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_8i16_i16_45u7zzzz:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 4
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 5
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 7
@@ -381,6 +485,12 @@ define <16 x i8> @merge_16i8_i8_01u3456789ABCDuF(i8* %ptr) nounwind uwtable noin
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups (%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_16i8_i8_01u3456789ABCDuF:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movups (%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 0
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 1
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 3
@@ -455,6 +565,15 @@ define <16 x i8> @merge_16i8_i8_01u3uuzzuuuuuzzz(i8* %ptr) nounwind uwtable noin
 ; AVX-NEXT:    vpinsrb $1, 1(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    vpinsrb $3, 3(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_16i8_i8_01u3uuzzuuuuuzzz:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    pxor %xmm0, %xmm0
+; X32-SSE-NEXT:    pinsrb $0, (%eax), %xmm0
+; X32-SSE-NEXT:    pinsrb $1, 1(%eax), %xmm0
+; X32-SSE-NEXT:    pinsrb $3, 3(%eax), %xmm0
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 0
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 1
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 3
@@ -482,6 +601,12 @@ define <16 x i8> @merge_16i8_i8_0123uu67uuuuuzzz(i8* %ptr) nounwind uwtable noin
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-SSE-LABEL: merge_16i8_i8_0123uu67uuuuuzzz:
+; X32-SSE:       # BB#0:
+; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X32-SSE-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 0
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 1
   %ptr2 = getelementptr inbounds i8, i8* %ptr, i64 2
diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
index 2f43adaf300..3106939c2bb 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
@@ -2,12 +2,21 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX --check-prefix=AVX512F
+;
+; Just one 32-bit run to make sure we do reasonable things.
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=X32-AVX
 
 define <4 x double> @merge_4f64_2f64_23(<2 x double>* %ptr) nounwind uwtable noinline ssp {
 ; AVX-LABEL: merge_4f64_2f64_23:
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 32(%rdi), %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_2f64_23:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 32(%eax), %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 2
   %ptr1 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 3
   %val0 = load <2 x double>, <2 x double>* %ptr0
@@ -23,6 +32,14 @@ define <4 x double> @merge_4f64_2f64_2z(<2 x double>* %ptr) nounwind uwtable noi
 ; AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_2f64_2z:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovaps 32(%eax), %xmm0
+; X32-AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 2
   %val0 = load <2 x double>, <2 x double>* %ptr0
   %res = shufflevector <2 x double> %val0, <2 x double> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -34,6 +51,12 @@ define <4 x double> @merge_4f64_f64_2345(double* %ptr) nounwind uwtable noinline
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 16(%rdi), %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_f64_2345:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 16(%eax), %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 2
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr2 = getelementptr inbounds double, double* %ptr, i64 4
@@ -54,6 +77,12 @@ define <4 x double> @merge_4f64_f64_3zuu(double* %ptr) nounwind uwtable noinline
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_f64_3zuu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 3
   %val0 = load double, double* %ptr0
   %res0 = insertelement <4 x double> undef, double %val0, i32 0
@@ -66,6 +95,12 @@ define <4 x double> @merge_4f64_f64_34uu(double* %ptr) nounwind uwtable noinline
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 24(%rdi), %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_f64_34uu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 24(%eax), %xmm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 4
   %val0 = load double, double* %ptr0
@@ -82,6 +117,14 @@ define <4 x double> @merge_4f64_f64_45zz(double* %ptr) nounwind uwtable noinline
 ; AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_f64_45zz:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 32(%eax), %xmm0
+; X32-AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 4
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 5
   %val0 = load double, double* %ptr0
@@ -115,6 +158,15 @@ define <4 x double> @merge_4f64_f64_34z6(double* %ptr) nounwind uwtable noinline
 ; AVX512F-NEXT:    vpslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
 ; AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4f64_f64_34z6:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 24(%eax), %xmm0
+; X32-AVX-NEXT:    vmovsd {{.*#+}} xmm1 = mem[0],zero
+; X32-AVX-NEXT:    vpslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 4
   %ptr3 = getelementptr inbounds double, double* %ptr, i64 6
@@ -135,6 +187,14 @@ define <4 x i64> @merge_4i64_2i64_3z(<2 x i64>* %ptr) nounwind uwtable noinline
 ; AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4i64_2i64_3z:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovaps 48(%eax), %xmm0
+; X32-AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x i64>, <2 x i64>* %ptr, i64 3
   %val0 = load <2 x i64>, <2 x i64>* %ptr0
   %res = shufflevector <2 x i64> %val0, <2 x i64> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -146,6 +206,12 @@ define <4 x i64> @merge_4i64_i64_1234(i64* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups 8(%rdi), %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4i64_i64_1234:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 8(%eax), %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %ptr1 = getelementptr inbounds i64, i64* %ptr, i64 2
   %ptr2 = getelementptr inbounds i64, i64* %ptr, i64 3
@@ -166,6 +232,12 @@ define <4 x i64> @merge_4i64_i64_1zzu(i64* %ptr) nounwind uwtable noinline ssp {
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4i64_i64_1zzu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %val0 = load i64, i64* %ptr0
   %res0 = insertelement <4 x i64> undef, i64 %val0, i32 0
@@ -181,6 +253,14 @@ define <4 x i64> @merge_4i64_i64_23zz(i64* %ptr) nounwind uwtable noinline ssp {
 ; AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_4i64_i64_23zz:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups 16(%eax), %xmm0
+; X32-AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 2
   %ptr1 = getelementptr inbounds i64, i64* %ptr, i64 3
   %val0 = load i64, i64* %ptr0
@@ -214,6 +294,16 @@ define <8 x float> @merge_8f32_2f32_23z5(<2 x float>* %ptr) nounwind uwtable noi
 ; AVX512F-NEXT:    vmovhpd 40(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8f32_2f32_23z5:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; X32-AVX-NEXT:    vmovupd 16(%eax), %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x float>, <2 x float>* %ptr, i64 2
   %ptr1 = getelementptr inbounds <2 x float>, <2 x float>* %ptr, i64 3
   %ptr3 = getelementptr inbounds <2 x float>, <2 x float>* %ptr, i64 5
@@ -232,6 +322,13 @@ define <8 x float> @merge_8f32_4f32_z2(<4 x float>* %ptr) nounwind uwtable noinl
 ; AVX-NEXT:    vxorps %xmm0, %xmm0, %xmm0
 ; AVX-NEXT:    vinsertf128 $1, 32(%rdi), %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8f32_4f32_z2:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vxorps %xmm0, %xmm0, %xmm0
+; X32-AVX-NEXT:    vinsertf128 $1, 32(%eax), %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr1 = getelementptr inbounds <4 x float>, <4 x float>* %ptr, i64 2
   %val1 = load <4 x float>, <4 x float>* %ptr1
   %res = shufflevector <4 x float> zeroinitializer, <4 x float> %val1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -243,6 +340,12 @@ define <8 x float> @merge_8f32_f32_12zzuuzz(float* %ptr) nounwind uwtable noinli
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8f32_f32_12zzuuzz:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 2
   %val0 = load float, float* %ptr0
@@ -286,6 +389,17 @@ define <8 x float> @merge_8f32_f32_1u3u5zu8(float* %ptr) nounwind uwtable noinli
 ; AVX512F-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
 ; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8f32_f32_1u3u5zu8:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm2[0,1],xmm0[1,0]
+; X32-AVX-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr2 = getelementptr inbounds float, float* %ptr, i64 3
   %ptr4 = getelementptr inbounds float, float* %ptr, i64 5
@@ -308,6 +422,13 @@ define <8 x i32> @merge_8i32_4i32_z3(<4 x i32>* %ptr) nounwind uwtable noinline
 ; AVX-NEXT:    vxorps %xmm0, %xmm0, %xmm0
 ; AVX-NEXT:    vinsertf128 $1, 48(%rdi), %ymm0, %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8i32_4i32_z3:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vxorps %xmm0, %xmm0, %xmm0
+; X32-AVX-NEXT:    vinsertf128 $1, 48(%eax), %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr1 = getelementptr inbounds <4 x i32>, <4 x i32>* %ptr, i64 3
   %val1 = load <4 x i32>, <4 x i32>* %ptr1
   %res = shufflevector <4 x i32> zeroinitializer, <4 x i32> %val1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -335,6 +456,14 @@ define <8 x i32> @merge_8i32_i32_56zz9uzz(i32* %ptr) nounwind uwtable noinline s
 ; AVX512F-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
 ; AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8i32_i32_56zz9uzz:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 5
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 6
   %ptr4 = getelementptr inbounds i32, i32* %ptr, i64 9
@@ -384,6 +513,18 @@ define <8 x i32> @merge_8i32_i32_1u3u5zu8(i32* %ptr) nounwind uwtable noinline s
 ; AVX512F-NEXT:    vpinsrd $2, 12(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_8i32_i32_1u3u5zu8:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,1,1,0]
+; X32-AVX-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
+; X32-AVX-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX-NEXT:    vpinsrd $2, 12(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 1
   %ptr2 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptr4 = getelementptr inbounds i32, i32* %ptr, i64 5
@@ -424,6 +565,15 @@ define <16 x i16> @merge_16i16_i16_89zzzuuuuuuuuuuuz(i16* %ptr) nounwind uwtable
 ; AVX512F-NEXT:    vpinsrw $1, 18(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_16i16_i16_89zzzuuuuuuuuuuuz:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; X32-AVX-NEXT:    vpinsrw $0, 16(%eax), %xmm0, %xmm1
+; X32-AVX-NEXT:    vpinsrw $1, 18(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 8
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 9
   %val0 = load i16, i16* %ptr0
@@ -442,6 +592,12 @@ define <16 x i16> @merge_16i16_i16_45u7uuuuuuuuuuuu(i16* %ptr) nounwind uwtable
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_16i16_i16_45u7uuuuuuuuuuuu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 4
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 5
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 7
@@ -459,6 +615,12 @@ define <16 x i16> @merge_16i16_i16_0uu3uuuuuuuuCuEF(i16* %ptr) nounwind uwtable
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovups (%rdi), %ymm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_16i16_i16_0uu3uuuuuuuuCuEF:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovups (%eax), %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 0
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 3
   %ptrC = getelementptr inbounds i16, i16* %ptr, i64 12
@@ -507,6 +669,17 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF(i16* %ptr) nounwind uwtable
 ; AVX512F-NEXT:    vpinsrw $7, 30(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_16i16_i16_0uu3zzuuuuuzCuEF:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; X32-AVX-NEXT:    vpinsrw $4, 24(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vpinsrw $6, 28(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vpinsrw $7, 30(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 0
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 3
   %ptrC = getelementptr inbounds i16, i16* %ptr, i64 12
@@ -535,6 +708,14 @@ define <32 x i8> @merge_32i8_i8_45u7uuuuuuuuuuuuuuuuuuuuuuuuuuuu(i8* %ptr) nounw
 ; AVX-NEXT:    vpinsrb $1, 5(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    vpinsrb $3, 7(%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_32i8_i8_45u7uuuuuuuuuuuuuuuuuuuuuuuuuuuu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vpinsrb $0, 4(%eax), %xmm0, %xmm0
+; X32-AVX-NEXT:    vpinsrb $1, 5(%eax), %xmm0, %xmm0
+; X32-AVX-NEXT:    vpinsrb $3, 7(%eax), %xmm0, %xmm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 4
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 5
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 7
@@ -574,6 +755,16 @@ define <32 x i8> @merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu(i8* %ptr) nounw
 ; AVX512F-NEXT:    vpinsrb $3, 5(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX512F-NEXT:    retq
+;
+; X32-AVX-LABEL: merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu:
+; X32-AVX:       # BB#0:
+; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; X32-AVX-NEXT:    vpinsrb $0, 2(%eax), %xmm0, %xmm1
+; X32-AVX-NEXT:    vpinsrb $1, 3(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vpinsrb $3, 5(%eax), %xmm1, %xmm1
+; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 2
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 3
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 5
diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
index 811d3da005f..9ce19ab1236 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
@@ -1,6 +1,9 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512f | FileCheck %s --check-prefix=ALL --check-prefix=AVX512F
 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512BW
+;
+; Just one 32-bit run to make sure we do reasonable things.
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=X32-AVX512F
 
 define <8 x double> @merge_8f64_2f64_12u4(<2 x double>* %ptr) nounwind uwtable noinline ssp {
 ; ALL-LABEL: merge_8f64_2f64_12u4:
@@ -9,6 +12,14 @@ define <8 x double> @merge_8f64_2f64_12u4(<2 x double>* %ptr) nounwind uwtable n
 ; ALL-NEXT:    vinsertf64x4 $0, 16(%rdi), %zmm0, %zmm1
 ; ALL-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_2f64_12u4:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vinsertf128 $1, 64(%eax), %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vinsertf64x4 $0, 16(%eax), %zmm0, %zmm1
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 1
   %ptr1 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 2
   %ptr3 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 4
@@ -29,6 +40,15 @@ define <8 x double> @merge_8f64_2f64_23z5(<2 x double>* %ptr) nounwind uwtable n
 ; ALL-NEXT:    vinsertf64x4 $0, 32(%rdi), %zmm0, %zmm1
 ; ALL-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_2f64_23z5:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vxorpd %xmm0, %xmm0, %xmm0
+; X32-AVX512F-NEXT:    vinsertf128 $1, 80(%eax), %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vinsertf64x4 $0, 32(%eax), %zmm0, %zmm1
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 2
   %ptr1 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 3
   %ptr3 = getelementptr inbounds <2 x double>, <2 x double>* %ptr, i64 5
@@ -47,6 +67,13 @@ define <8 x double> @merge_8f64_4f64_z2(<4 x double>* %ptr) nounwind uwtable noi
 ; ALL-NEXT:    vxorpd %ymm0, %ymm0, %ymm0
 ; ALL-NEXT:    vinsertf64x4 $1, 64(%rdi), %zmm0, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_4f64_z2:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vxorpd %ymm0, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, 64(%eax), %zmm0, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr1 = getelementptr inbounds <4 x double>, <4 x double>* %ptr, i64 2
   %val1 = load <4 x double>, <4 x double>* %ptr1
   %res = shufflevector <4 x double> zeroinitializer, <4 x double> %val1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -58,6 +85,12 @@ define <8 x double> @merge_8f64_f64_23uuuuu9(double* %ptr) nounwind uwtable noin
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovupd 16(%rdi), %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_f64_23uuuuu9:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovupd 16(%eax), %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 2
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr7 = getelementptr inbounds double, double* %ptr, i64 9
@@ -79,6 +112,16 @@ define <8 x double> @merge_8f64_f64_12zzuuzz(double* %ptr) nounwind uwtable noin
 ; ALL-NEXT:    vxorpd %ymm1, %ymm1, %ymm1
 ; ALL-NEXT:    vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_f64_12zzuuzz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovupd 8(%eax), %xmm0
+; X32-AVX512F-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
+; X32-AVX512F-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vxorpd %ymm1, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 1
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 2
   %val0 = load double, double* %ptr0
@@ -103,6 +146,18 @@ define <8 x double> @merge_8f64_f64_1u3u5zu8(double* %ptr) nounwind uwtable noin
 ; ALL-NEXT:    vinsertf128 $1, %xmm2, %ymm1, %ymm1
 ; ALL-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8f64_f64_1u3u5zu8:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    vmovddup {{.*#+}} xmm1 = mem[0,0]
+; X32-AVX512F-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vmovsd {{.*#+}} xmm1 = mem[0],zero
+; X32-AVX512F-NEXT:    vmovsd {{.*#+}} xmm2 = mem[0],zero
+; X32-AVX512F-NEXT:    vinsertf128 $1, %xmm2, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 1
   %ptr2 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr4 = getelementptr inbounds double, double* %ptr, i64 5
@@ -125,6 +180,13 @@ define <8 x i64> @merge_8i64_4i64_z3(<4 x i64>* %ptr) nounwind uwtable noinline
 ; ALL-NEXT:    vpxor %ymm0, %ymm0, %ymm0
 ; ALL-NEXT:    vinserti64x4 $1, 96(%rdi), %zmm0, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8i64_4i64_z3:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vpxor %ymm0, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vinserti64x4 $1, 96(%eax), %zmm0, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr1 = getelementptr inbounds <4 x i64>, <4 x i64>* %ptr, i64 3
   %val1 = load <4 x i64>, <4 x i64>* %ptr1
   %res = shufflevector <4 x i64> zeroinitializer, <4 x i64> %val1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -140,6 +202,16 @@ define <8 x i64> @merge_8i64_i64_56zz9uzz(i64* %ptr) nounwind uwtable noinline s
 ; ALL-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; ALL-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8i64_i64_56zz9uzz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovdqu 40(%eax), %xmm0
+; X32-AVX512F-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; X32-AVX512F-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 5
   %ptr1 = getelementptr inbounds i64, i64* %ptr, i64 6
   %ptr4 = getelementptr inbounds i64, i64* %ptr, i64 9
@@ -167,6 +239,19 @@ define <8 x i64> @merge_8i64_i64_1u3u5zu8(i64* %ptr) nounwind uwtable noinline s
 ; ALL-NEXT:    vinserti128 $1, %xmm2, %ymm1, %ymm1
 ; ALL-NEXT:    vinserti64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_8i64_i64_1u3u5zu8:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vpinsrd $2, 64(%eax), %xmm0, %xmm0
+; X32-AVX512F-NEXT:    vpinsrd $3, 68(%eax), %xmm0, %xmm0
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm2, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    vinserti64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %ptr2 = getelementptr inbounds i64, i64* %ptr, i64 3
   %ptr4 = getelementptr inbounds i64, i64* %ptr, i64 5
@@ -188,6 +273,12 @@ define <16 x float> @merge_16f32_f32_89zzzuuuuuuuuuuuz(float* %ptr) nounwind uwt
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16f32_f32_89zzzuuuuuuuuuuuz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 8
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 9
   %val0 = load float, float* %ptr0
@@ -206,6 +297,12 @@ define <16 x float> @merge_16f32_f32_45u7uuuuuuuuuuuu(float* %ptr) nounwind uwta
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovups 16(%rdi), %xmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16f32_f32_45u7uuuuuuuuuuuu:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovups 16(%eax), %xmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 4
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 5
   %ptr3 = getelementptr inbounds float, float* %ptr, i64 7
@@ -223,6 +320,12 @@ define <16 x float> @merge_16f32_f32_0uu3uuuuuuuuCuEF(float* %ptr) nounwind uwta
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovups (%rdi), %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16f32_f32_0uu3uuuuuuuuCuEF:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovups (%eax), %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 0
   %ptr3 = getelementptr inbounds float, float* %ptr, i64 3
   %ptrC = getelementptr inbounds float, float* %ptr, i64 12
@@ -253,6 +356,19 @@ define <16 x float> @merge_16f32_f32_0uu3zzuuuuuzCuEF(float* %ptr) nounwind uwta
 ; ALL-NEXT:    vinsertf128 $1, %xmm2, %ymm1, %ymm1
 ; ALL-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16f32_f32_0uu3zzuuuuuzCuEF:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX512F-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; X32-AVX512F-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vmovupd (%eax), %xmm1
+; X32-AVX512F-NEXT:    vxorpd %xmm2, %xmm2, %xmm2
+; X32-AVX512F-NEXT:    vinsertf128 $1, %xmm2, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 0
   %ptr3 = getelementptr inbounds float, float* %ptr, i64 3
   %ptrC = getelementptr inbounds float, float* %ptr, i64 12
@@ -279,6 +395,12 @@ define <16 x i32> @merge_16i32_i32_12zzzuuuuuuuuuuuz(i32* %ptr) nounwind uwtable
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16i32_i32_12zzzuuuuuuuuuuuz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 1
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 2
   %val0 = load i32, i32* %ptr0
@@ -297,6 +419,12 @@ define <16 x i32> @merge_16i32_i32_23u5uuuuuuuuuuuu(i32* %ptr) nounwind uwtable
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovups 8(%rdi), %xmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16i32_i32_23u5uuuuuuuuuuuu:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovups 8(%eax), %xmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 2
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptr3 = getelementptr inbounds i32, i32* %ptr, i64 5
@@ -314,6 +442,12 @@ define <16 x i32> @merge_16i32_i32_0uu3uuuuuuuuCuEF(i32* %ptr) nounwind uwtable
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovdqu32 (%rdi), %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16i32_i32_0uu3uuuuuuuuCuEF:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovdqu32 (%eax), %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 0
   %ptr3 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptrC = getelementptr inbounds i32, i32* %ptr, i64 12
@@ -344,6 +478,19 @@ define <16 x i32> @merge_16i32_i32_0uu3zzuuuuuzCuEF(i32* %ptr) nounwind uwtable
 ; ALL-NEXT:    vinserti128 $1, %xmm2, %ymm1, %ymm1
 ; ALL-NEXT:    vinserti64x4 $1, %ymm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_16i32_i32_0uu3zzuuuuuzCuEF:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-AVX512F-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; X32-AVX512F-NEXT:    vmovdqu (%eax), %xmm1
+; X32-AVX512F-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm2, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    vinserti64x4 $1, %ymm0, %zmm1, %zmm0
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 0
   %ptr3 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptrC = getelementptr inbounds i32, i32* %ptr, i64 12
@@ -376,6 +523,13 @@ define <32 x i16> @merge_32i16_i16_12u4uuuuuuuuuuuuuuuuuuuuuuuuuuzz(i16* %ptr) n
 ; AVX512BW:       # BB#0:
 ; AVX512BW-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX512BW-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_32i16_i16_12u4uuuuuuuuuuuuuuuuuuuuuuuuuuzz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    vxorps %ymm1, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 1
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 2
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 4
@@ -395,6 +549,12 @@ define <32 x i16> @merge_32i16_i16_45u7uuuuuuuuuuuuuuuuuuuuuuuuuuuu(i16* %ptr) n
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; ALL-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_32i16_i16_45u7uuuuuuuuuuuuuuuuuuuuuuuuuuuu:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 4
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 5
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 7
@@ -426,6 +586,16 @@ define <32 x i16> @merge_32i16_i16_23uzuuuuuuuuuuzzzzuuuuuuuuuuuuuu(i16* %ptr) n
 ; AVX512BW-NEXT:    vpxor %ymm1, %ymm1, %ymm1
 ; AVX512BW-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_32i16_i16_23uzuuuuuuuuuuzzzzuuuuuuuuuuuuuu:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; X32-AVX512F-NEXT:    vpinsrw $0, 4(%eax), %xmm0, %xmm1
+; X32-AVX512F-NEXT:    vpinsrw $1, 6(%eax), %xmm1, %xmm1
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX512F-NEXT:    vpxor %ymm1, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 2
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 3
   %val0 = load i16, i16* %ptr0
@@ -451,6 +621,13 @@ define <64 x i8> @merge_64i8_i8_12u4uuu8uuuuuuzzzzuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu
 ; AVX512BW:       # BB#0:
 ; AVX512BW-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX512BW-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_64i8_i8_12u4uuu8uuuuuuzzzzuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; X32-AVX512F-NEXT:    vxorps %ymm1, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 1
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 2
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 4
@@ -492,6 +669,17 @@ define <64 x i8> @merge_64i8_i8_12u4uuuuuuuuuuzzzzuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu
 ; AVX512BW-NEXT:    vpxor %ymm1, %ymm1, %ymm1
 ; AVX512BW-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    retq
+;
+; X32-AVX512F-LABEL: merge_64i8_i8_12u4uuuuuuuuuuzzzzuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuz:
+; X32-AVX512F:       # BB#0:
+; X32-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-AVX512F-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; X32-AVX512F-NEXT:    vpinsrb $0, 1(%eax), %xmm0, %xmm1
+; X32-AVX512F-NEXT:    vpinsrb $1, 2(%eax), %xmm1, %xmm1
+; X32-AVX512F-NEXT:    vpinsrb $3, 4(%eax), %xmm1, %xmm1
+; X32-AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; X32-AVX512F-NEXT:    vpxor %ymm1, %ymm1, %ymm1
+; X32-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 1
   %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 2
   %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 4