+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+ssse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSSE3
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41
ret <16 x i8> %shuffle
}
+define <16 x i8> @shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz(<16 x i8> %a) {
+; SSE-LABEL: shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz:
+; SSE: # BB#0:
+; SSE-NEXT: andps {{.*}}(%rip), %xmm0
+; SSE-NEXT: retq
+;
+; AVX-LABEL: shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz:
+; AVX: # BB#0:
+; AVX-NEXT: vandps {{.*}}(%rip), %xmm0, %xmm0
+; AVX-NEXT: retq
+ %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 19, i32 4, i32 5, i32 6, i32 23, i32 8, i32 9, i32 10, i32 27, i32 12, i32 13, i32 14, i32 31>
+ ret <16 x i8> %shuffle
+}
+
define <16 x i8> @shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31(<16 x i8> %a, <16 x i8> %b) {
; SSE2-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
; SSE2: # BB#0:
%bitcast8 = bitcast <8 x i16> %shuffle16 to <16 x i8>
ret <16 x i8> %bitcast8
}
+
+define <16 x i8> @insert_dup_mem_v16i8_i32(i32* %ptr) {
+; SSE2-LABEL: insert_dup_mem_v16i8_i32:
+; SSE2: # BB#0:
+; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,3]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_mem_v16i8_i32:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSSE3-NEXT: pxor %xmm1, %xmm1
+; SSSE3-NEXT: pshufb %xmm1, %xmm0
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_mem_v16i8_i32:
+; SSE41: # BB#0:
+; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE41-NEXT: pxor %xmm1, %xmm1
+; SSE41-NEXT: pshufb %xmm1, %xmm0
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_mem_v16i8_i32:
+; AVX1: # BB#0:
+; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_mem_v16i8_i32:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastb (%rdi), %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i32, i32* %ptr, align 4
+ %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
+ %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
+ %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> zeroinitializer
+ ret <16 x i8> %tmp3
+}
+
+define <16 x i8> @insert_dup_mem_v16i8_sext_i8(i8* %ptr) {
+; SSE2-LABEL: insert_dup_mem_v16i8_sext_i8:
+; SSE2: # BB#0:
+; SSE2-NEXT: movsbl (%rdi), %eax
+; SSE2-NEXT: movd %eax, %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,3]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_mem_v16i8_sext_i8:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movsbl (%rdi), %eax
+; SSSE3-NEXT: movd %eax, %xmm0
+; SSSE3-NEXT: pxor %xmm1, %xmm1
+; SSSE3-NEXT: pshufb %xmm1, %xmm0
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_mem_v16i8_sext_i8:
+; SSE41: # BB#0:
+; SSE41-NEXT: movsbl (%rdi), %eax
+; SSE41-NEXT: movd %eax, %xmm0
+; SSE41-NEXT: pxor %xmm1, %xmm1
+; SSE41-NEXT: pshufb %xmm1, %xmm0
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_mem_v16i8_sext_i8:
+; AVX1: # BB#0:
+; AVX1-NEXT: movsbl (%rdi), %eax
+; AVX1-NEXT: vmovd %eax, %xmm0
+; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_mem_v16i8_sext_i8:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastb (%rdi), %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i8, i8* %ptr, align 1
+ %tmp1 = sext i8 %tmp to i32
+ %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
+ %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
+ %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> zeroinitializer
+ ret <16 x i8> %tmp4
+}
+
+define <16 x i8> @insert_dup_elt1_mem_v16i8_i32(i32* %ptr) {
+; SSE2-LABEL: insert_dup_elt1_mem_v16i8_i32:
+; SSE2: # BB#0:
+; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,3]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,5]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_i32:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_elt1_mem_v16i8_i32:
+; SSE41: # BB#0:
+; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_elt1_mem_v16i8_i32:
+; AVX1: # BB#0:
+; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_elt1_mem_v16i8_i32:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastb 1(%rdi), %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i32, i32* %ptr, align 4
+ %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
+ %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
+ %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
+ ret <16 x i8> %tmp3
+}
+
+define <16 x i8> @insert_dup_elt2_mem_v16i8_i32(i32* %ptr) {
+; SSE2-LABEL: insert_dup_elt2_mem_v16i8_i32:
+; SSE2: # BB#0:
+; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,2,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,6,6,6]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_i32:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_elt2_mem_v16i8_i32:
+; SSE41: # BB#0:
+; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_elt2_mem_v16i8_i32:
+; AVX1: # BB#0:
+; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_elt2_mem_v16i8_i32:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastb 2(%rdi), %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i32, i32* %ptr, align 4
+ %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
+ %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
+ %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
+ ret <16 x i8> %tmp3
+}
+
+define <16 x i8> @insert_dup_elt1_mem_v16i8_sext_i8(i8* %ptr) {
+; SSE2-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
+; SSE2: # BB#0:
+; SSE2-NEXT: movsbl (%rdi), %eax
+; SSE2-NEXT: movd %eax, %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,3]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,5]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movsbl (%rdi), %eax
+; SSSE3-NEXT: movd %eax, %xmm0
+; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
+; SSE41: # BB#0:
+; SSE41-NEXT: movsbl (%rdi), %eax
+; SSE41-NEXT: movd %eax, %xmm0
+; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
+; AVX1: # BB#0:
+; AVX1-NEXT: movsbl (%rdi), %eax
+; AVX1-NEXT: vmovd %eax, %xmm0
+; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
+; AVX2: # BB#0:
+; AVX2-NEXT: movsbl (%rdi), %eax
+; AVX2-NEXT: shrl $8, %eax
+; AVX2-NEXT: vmovd %eax, %xmm0
+; AVX2-NEXT: vpbroadcastb %xmm0, %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i8, i8* %ptr, align 1
+ %tmp1 = sext i8 %tmp to i32
+ %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
+ %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
+ %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
+ ret <16 x i8> %tmp4
+}
+
+define <16 x i8> @insert_dup_elt2_mem_v16i8_sext_i8(i8* %ptr) {
+; SSE2-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
+; SSE2: # BB#0:
+; SSE2-NEXT: movsbl (%rdi), %eax
+; SSE2-NEXT: movd %eax, %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,2,1]
+; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
+; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,6,6,6]
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
+; SSSE3: # BB#0:
+; SSSE3-NEXT: movsbl (%rdi), %eax
+; SSSE3-NEXT: movd %eax, %xmm0
+; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
+; SSE41: # BB#0:
+; SSE41-NEXT: movsbl (%rdi), %eax
+; SSE41-NEXT: movd %eax, %xmm0
+; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
+; AVX1: # BB#0:
+; AVX1-NEXT: movsbl (%rdi), %eax
+; AVX1-NEXT: vmovd %eax, %xmm0
+; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
+; AVX2: # BB#0:
+; AVX2-NEXT: movsbl (%rdi), %eax
+; AVX2-NEXT: shrl $16, %eax
+; AVX2-NEXT: vmovd %eax, %xmm0
+; AVX2-NEXT: vpbroadcastb %xmm0, %xmm0
+; AVX2-NEXT: retq
+ %tmp = load i8, i8* %ptr, align 1
+ %tmp1 = sext i8 %tmp to i32
+ %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
+ %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
+ %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
+ ret <16 x i8> %tmp4
+}