summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/X86
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/test/CodeGen/X86')
-rw-r--r--llvm/test/CodeGen/X86/avx-vperm2x128.ll19
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll23
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll20
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll86
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll92
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll32
-rw-r--r--llvm/test/CodeGen/X86/vector-shuffle-combining.ll11
7 files changed, 82 insertions, 201 deletions
diff --git a/llvm/test/CodeGen/X86/avx-vperm2x128.ll b/llvm/test/CodeGen/X86/avx-vperm2x128.ll
index a103405a85d..43303ca57c4 100644
--- a/llvm/test/CodeGen/X86/avx-vperm2x128.ll
+++ b/llvm/test/CodeGen/X86/avx-vperm2x128.ll
@@ -182,20 +182,11 @@ entry:
;;;; Cases we must not select vperm2f128
define <8 x float> @G(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; AVX1-LABEL: G:
-; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[0,0,2,3]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: G:
-; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2-NEXT: vpermilps {{.*#+}} ymm1 = ymm1[0,0,2,3,4,4,6,7]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: G:
+; ALL: ## BB#0: ## %entry
+; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[0,0,2,3,4,4,6,7]
+; ALL-NEXT: retq
entry:
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 undef, i32 undef, i32 6, i32 7, i32 undef, i32 12, i32 undef, i32 15>
ret <8 x float> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll b/llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll
index b3fba4f349b..2a6efda0dfa 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll
@@ -1278,11 +1278,8 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_04_04_04_04_16_16_16_16_20_20_20_2
;
; AVX2-LABEL: shuffle_v16i16_00_00_00_00_04_04_04_04_16_16_16_16_20_20_20_20:
; AVX2: # BB#0:
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,0,1,0,1,0,1,8,9,8,9,8,9,8,9,16,17,16,17,16,17,16,17,24,25,24,25,24,25,24,25]
; AVX2-NEXT: retq
%shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 16, i32 16, i32 16, i32 16, i32 20, i32 20, i32 20, i32 20>
ret <16 x i16> %shuffle
@@ -1301,13 +1298,8 @@ define <16 x i16> @shuffle_v16i16_08_08_08_08_12_12_12_12_16_16_16_16_20_20_20_2
;
; AVX2-LABEL: shuffle_v16i16_08_08_08_08_12_12_12_12_16_16_16_16_20_20_20_20:
; AVX2: # BB#0:
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
-; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm1
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,0,1,0,1,0,1,8,9,8,9,8,9,8,9,16,17,16,17,16,17,16,17,24,25,24,25,24,25,24,25]
; AVX2-NEXT: retq
%shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12, i32 16, i32 16, i32 16, i32 16, i32 20, i32 20, i32 20, i32 20>
ret <16 x i16> %shuffle
@@ -1327,13 +1319,8 @@ define <16 x i16> @shuffle_v16i16_08_08_08_08_12_12_12_12_24_24_24_24_28_28_28_2
;
; AVX2-LABEL: shuffle_v16i16_08_08_08_08_12_12_12_12_24_24_24_24_28_28_28_28:
; AVX2: # BB#0:
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
-; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm1
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
-; AVX2-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
-; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,0,1,0,1,0,1,8,9,8,9,8,9,8,9,16,17,16,17,16,17,16,17,24,25,24,25,24,25,24,25]
; AVX2-NEXT: retq
%shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12, i32 24, i32 24, i32 24, i32 24, i32 28, i32 28, i32 28, i32 28>
ret <16 x i16> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll b/llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll
index 1c09ba0861d..d8390fdfb35 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll
@@ -1572,10 +1572,8 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08_
;
; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08_32_32_32_32_32_32_32_32_40_40_40_40_40_40_40_40:
; AVX2: # BB#0:
-; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
-; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
-; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8,16,16,16,16,16,16,16,16,24,24,24,24,24,24,24,24]
; AVX2-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40>
ret <32 x i8> %shuffle
@@ -1593,12 +1591,8 @@ define <32 x i8> @shuffle_v32i8_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24_
;
; AVX2-LABEL: shuffle_v32i8_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24_32_32_32_32_32_32_32_32_40_40_40_40_40_40_40_40:
; AVX2: # BB#0:
-; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
-; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
-; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm1
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8,16,16,16,16,16,16,16,16,24,24,24,24,24,24,24,24]
; AVX2-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40, i32 40>
ret <32 x i8> %shuffle
@@ -1617,12 +1611,8 @@ define <32 x i8> @shuffle_v32i8_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24_
;
; AVX2-LABEL: shuffle_v32i8_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24_48_48_48_48_48_48_48_48_56_56_56_56_56_56_56_56:
; AVX2: # BB#0:
-; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm1
-; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
-; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
-; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8,16,16,16,16,16,16,16,16,24,24,24,24,24,24,24,24]
; AVX2-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 56, i32 56, i32 56, i32 56, i32 56, i32 56, i32 56, i32 56>
ret <32 x i8> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll b/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
index 18b9970f070..19f3c7743ca 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
@@ -301,58 +301,31 @@ define <4 x double> @shuffle_v4f64_0167(<4 x double> %a, <4 x double> %b) {
}
define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_1054:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v4f64_1054:
-; AVX2: # BB#0:
-; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,1,1,0]
-; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v4f64_1054:
+; ALL: # BB#0:
+; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT: retq
%shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
ret <4 x double> %shuffle
}
define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_3254:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v4f64_3254:
-; AVX2: # BB#0:
-; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,1,1,0]
-; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,2,3]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v4f64_3254:
+; ALL: # BB#0:
+; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT: retq
%shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
ret <4 x double> %shuffle
}
define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_3276:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v4f64_3276:
-; AVX2: # BB#0:
-; AVX2-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,3,2]
-; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,2,3]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v4f64_3276:
+; ALL: # BB#0:
+; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT: retq
%shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
ret <4 x double> %shuffle
}
@@ -687,16 +660,14 @@ define <4 x i64> @shuffle_v4i64_1251(<4 x i64> %a, <4 x i64> %b) {
define <4 x i64> @shuffle_v4i64_1054(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_1054:
; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v4i64_1054:
; AVX2: # BB#0:
-; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,0]
+; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; AVX2-NEXT: retq
%shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
ret <4 x i64> %shuffle
@@ -705,17 +676,14 @@ define <4 x i64> @shuffle_v4i64_1054(<4 x i64> %a, <4 x i64> %b) {
define <4 x i64> @shuffle_v4i64_3254(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_3254:
; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v4i64_3254:
; AVX2: # BB#0:
-; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,0]
-; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,2,3]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
; AVX2-NEXT: retq
%shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
ret <4 x i64> %shuffle
@@ -724,18 +692,14 @@ define <4 x i64> @shuffle_v4i64_3254(<4 x i64> %a, <4 x i64> %b) {
define <4 x i64> @shuffle_v4i64_3276(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_3276:
; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v4i64_3276:
; AVX2: # BB#0:
-; AVX2-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[2,3,0,1,6,7,4,5]
-; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,2,3]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
; AVX2-NEXT: retq
%shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
ret <4 x i64> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll b/llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll
index ded8232dad2..d04adf9f7e8 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll
@@ -736,20 +736,11 @@ define <8 x float> @shuffle_v8f32_76543210(<8 x float> %a, <8 x float> %b) {
}
define <8 x float> @shuffle_v8f32_3210ba98(<8 x float> %a, <8 x float> %b) {
-; AVX1-LABEL: shuffle_v8f32_3210ba98:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8f32_3210ba98:
-; AVX2: # BB#0:
-; AVX2-NEXT: vmovaps {{.*#+}} ymm2 = <u,u,u,u,3,2,1,0>
-; AVX2-NEXT: vpermps %ymm1, %ymm2, %ymm1
-; AVX2-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v8f32_3210ba98:
+; ALL: # BB#0:
+; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; ALL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 11, i32 10, i32 9, i32 8>
ret <8 x float> %shuffle
}
@@ -774,43 +765,21 @@ define <8 x float> @shuffle_v8f32_3210fedc(<8 x float> %a, <8 x float> %b) {
}
define <8 x float> @shuffle_v8f32_7654fedc(<8 x float> %a, <8 x float> %b) {
-; AVX1-LABEL: shuffle_v8f32_7654fedc:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8f32_7654fedc:
-; AVX2: # BB#0:
-; AVX2-NEXT: vmovaps {{.*#+}} ymm2 = <7,6,5,4,u,u,u,u>
-; AVX2-NEXT: vpermps %ymm0, %ymm2, %ymm0
-; AVX2-NEXT: vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v8f32_7654fedc:
+; ALL: # BB#0:
+; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; ALL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 15, i32 14, i32 13, i32 12>
ret <8 x float> %shuffle
}
define <8 x float> @shuffle_v8f32_fedc7654(<8 x float> %a, <8 x float> %b) {
-; AVX1-LABEL: shuffle_v8f32_fedc7654:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8f32_fedc7654:
-; AVX2: # BB#0:
-; AVX2-NEXT: vmovaps {{.*#+}} ymm2 = <7,6,5,4,u,u,u,u>
-; AVX2-NEXT: vpermps %ymm1, %ymm2, %ymm1
-; AVX2-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
-; AVX2-NEXT: retq
+; ALL-LABEL: shuffle_v8f32_fedc7654:
+; ALL: # BB#0:
+; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[2,3]
+; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; ALL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 15, i32 14, i32 13, i32 12, i32 7, i32 6, i32 5, i32 4>
ret <8 x float> %shuffle
}
@@ -1789,17 +1758,14 @@ define <8 x i32> @shuffle_v8i32_76543210(<8 x i32> %a, <8 x i32> %b) {
define <8 x i32> @shuffle_v8i32_3210ba98(<8 x i32> %a, <8 x i32> %b) {
; AVX1-LABEL: shuffle_v8i32_3210ba98:
; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v8i32_3210ba98:
; AVX2: # BB#0:
-; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = <u,u,u,u,3,2,1,0>
-; AVX2-NEXT: vpermd %ymm1, %ymm2, %ymm1
+; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; AVX2-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 11, i32 10, i32 9, i32 8>
ret <8 x i32> %shuffle
@@ -1827,19 +1793,14 @@ define <8 x i32> @shuffle_v8i32_3210fedc(<8 x i32> %a, <8 x i32> %b) {
define <8 x i32> @shuffle_v8i32_7654fedc(<8 x i32> %a, <8 x i32> %b) {
; AVX1-LABEL: shuffle_v8i32_7654fedc:
; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v8i32_7654fedc:
; AVX2: # BB#0:
-; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = <7,6,5,4,u,u,u,u>
-; AVX2-NEXT: vpermd %ymm0, %ymm2, %ymm0
-; AVX2-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
; AVX2-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 15, i32 14, i32 13, i32 12>
ret <8 x i32> %shuffle
@@ -1848,19 +1809,14 @@ define <8 x i32> @shuffle_v8i32_7654fedc(<8 x i32> %a, <8 x i32> %b) {
define <8 x i32> @shuffle_v8i32_fedc7654(<8 x i32> %a, <8 x i32> %b) {
; AVX1-LABEL: shuffle_v8i32_fedc7654:
; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[2,3]
+; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
; AVX1-NEXT: retq
;
; AVX2-LABEL: shuffle_v8i32_fedc7654:
; AVX2: # BB#0:
-; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = <7,6,5,4,u,u,u,u>
-; AVX2-NEXT: vpermd %ymm1, %ymm2, %ymm1
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[2,3]
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 15, i32 14, i32 13, i32 12, i32 7, i32 6, i32 5, i32 4>
ret <8 x i32> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll b/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
index 31cabbcaf5f..6671735248a 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
@@ -208,15 +208,13 @@ define <8 x double> @shuffle_v8f64_9832dc76(<8 x double> %a, <8 x double> %b) {
define <8 x double> @shuffle_v8f64_9810dc54(<8 x double> %a, <8 x double> %b) {
; ALL-LABEL: shuffle_v8f64_9810dc54:
; ALL: # BB#0:
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm2
-; ALL-NEXT: vpermpd {{.*#+}} ymm2 = ymm2[0,1,1,0]
-; ALL-NEXT: vextractf64x4 $1, %zmm1, %ymm3
-; ALL-NEXT: vpermilpd {{.*#+}} ymm3 = ymm3[1,0,2,2]
-; ALL-NEXT: vblendpd {{.*#+}} ymm2 = ymm3[0,1],ymm2[2,3]
-; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,0]
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[1,0,2,2]
-; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
-; ALL-NEXT: vinsertf64x4 $1, %ymm2, %zmm0, %zmm0
+; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm2
+; ALL-NEXT: vpermilpd {{.*#+}} ymm2 = ymm2[1,0,3,2]
+; ALL-NEXT: vextractf64x4 $1, %zmm1, %ymm1
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0
+; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm2, %zmm0
; ALL-NEXT: retq
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 9, i32 8, i32 1, i32 0, i32 13, i32 12, i32 5, i32 4>
ret <8 x double> %shuffle
@@ -930,15 +928,13 @@ define <8 x i64> @shuffle_v8i64_9832dc76(<8 x i64> %a, <8 x i64> %b) {
define <8 x i64> @shuffle_v8i64_9810dc54(<8 x i64> %a, <8 x i64> %b) {
; ALL-LABEL: shuffle_v8i64_9810dc54:
; ALL: # BB#0:
-; ALL-NEXT: vextracti64x4 $1, %zmm0, %ymm2
-; ALL-NEXT: vpermq {{.*#+}} ymm2 = ymm2[0,1,1,0]
-; ALL-NEXT: vextracti64x4 $1, %zmm1, %ymm3
-; ALL-NEXT: vpshufd {{.*#+}} ymm3 = ymm3[2,3,0,1,6,7,4,5]
-; ALL-NEXT: vpblendd {{.*#+}} ymm2 = ymm3[0,1,2,3],ymm2[4,5,6,7]
-; ALL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,0]
-; ALL-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[2,3,0,1,6,7,4,5]
-; ALL-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
-; ALL-NEXT: vinserti64x4 $1, %ymm2, %zmm0, %zmm0
+; ALL-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm2
+; ALL-NEXT: vpshufd {{.*#+}} ymm2 = ymm2[2,3,0,1,6,7,4,5]
+; ALL-NEXT: vextracti64x4 $1, %zmm1, %ymm1
+; ALL-NEXT: vextracti64x4 $1, %zmm0, %ymm0
+; ALL-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
+; ALL-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
+; ALL-NEXT: vinserti64x4 $1, %ymm0, %zmm2, %zmm0
; ALL-NEXT: retq
%shuffle = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 9, i32 8, i32 1, i32 0, i32 13, i32 12, i32 5, i32 4>
ret <8 x i64> %shuffle
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining.ll
index 22a67492110..5f6e5c60022 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-combining.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-combining.ll
@@ -2463,19 +2463,16 @@ define <8 x i32> @combine_unneeded_subvector2(<8 x i32> %a, <8 x i32> %b) {
; AVX1: # BB#0:
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
; AVX1-NEXT: vpaddd {{.*}}(%rip), %xmm0, %xmm0
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[3,2,1,0]
-; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[2,3]
+; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_unneeded_subvector2:
; AVX2: # BB#0:
; AVX2-NEXT: vpaddd {{.*}}(%rip), %ymm0, %ymm0
-; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = <7,6,5,4,u,u,u,u>
-; AVX2-NEXT: vpermd %ymm1, %ymm2, %ymm1
+; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[2,3]
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
-; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT: retq
%c = add <8 x i32> %a, <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
%d = shufflevector <8 x i32> %b, <8 x i32> %c, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 15, i32 14, i32 13, i32 12>
OpenPOWER on IntegriCloud