From: Simon Pilgrim <llvm-dev@redking.me.uk>
Date: Tue, 27 Oct 2015 21:18:45 +0000 (+0000)
Subject: [X86][AVX512] Test UNPCK with non-sequential scalars
X-Git-Url: http://plrg.eecs.uci.edu/git/?p=oota-llvm.git;a=commitdiff_plain;h=d694139c3adfd04c3eddfee57e155e4e8dac1321

[X86][AVX512] Test UNPCK with non-sequential scalars

Missing tests for r251297

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251453 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/test/CodeGen/X86/vector-shuffle-512-v16.ll b/test/CodeGen/X86/vector-shuffle-512-v16.ll
index a621a369d3e..89ab7391263 100644
--- a/test/CodeGen/X86/vector-shuffle-512-v16.ll
+++ b/test/CodeGen/X86/vector-shuffle-512-v16.ll
@@ -12,6 +12,16 @@ define <16 x float> @shuffle_v16f32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d
   ret <16 x float> %shuffle
 }
 
+define <16 x float> @shuffle_v16f32_00_zz_01_zz_04_zz_05_zz_08_zz_09_zz_0c_zz_0d_zz(<16 x float> %a, <16 x float> %b) {
+; ALL-LABEL: shuffle_v16f32_00_zz_01_zz_04_zz_05_zz_08_zz_09_zz_0c_zz_0d_zz:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; ALL-NEXT:    vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x float> %a, <16 x float> zeroinitializer, <16 x i32><i32 0, i32 16, i32 1, i32 16, i32 4, i32 16, i32 5, i32 16, i32 8, i32 16, i32 9, i32 16, i32 12, i32 16, i32 13, i32 16>
+  ret <16 x float> %shuffle
+}
+
 define <16 x float> @shuffle_v16f32_vunpcklps_swap(<16 x float> %a, <16 x float> %b) {
 ; ALL-LABEL: shuffle_v16f32_vunpcklps_swap:
 ; ALL:       # BB#0:
@@ -30,6 +40,16 @@ define <16 x i32> @shuffle_v16i32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d_1
   ret <16 x i32> %shuffle
 }
 
+define <16 x i32> @shuffle_v16i32_zz_10_zz_11_zz_14_zz_15_zz_18_zz_19_zz_1c_zz_1d(<16 x i32> %a, <16 x i32> %b) {
+; ALL-LABEL: shuffle_v16i32_zz_10_zz_11_zz_14_zz_15_zz_18_zz_19_zz_1c_zz_1d:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; ALL-NEXT:    vpunpckldq {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x i32> zeroinitializer, <16 x i32> %b, <16 x i32><i32 15, i32 16, i32 13, i32 17, i32 11, i32 20, i32 9, i32 21, i32 7, i32 24, i32 5, i32 25, i32 3, i32 28, i32 1, i32 29>
+  ret <16 x i32> %shuffle
+}
+
 define <16 x float> @shuffle_v16f32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f(<16 x float> %a, <16 x float> %b) {
 ; ALL-LABEL: shuffle_v16f32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f:
 ; ALL:       # BB#0:
@@ -39,6 +59,16 @@ define <16 x float> @shuffle_v16f32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f
   ret <16 x float> %shuffle
 }
 
+define <16 x float> @shuffle_v16f32_zz_12_zz_13_zz_16_zz_17_zz_1a_zz_1b_zz_1e_zz_1f(<16 x float> %a, <16 x float> %b) {
+; ALL-LABEL: shuffle_v16f32_zz_12_zz_13_zz_16_zz_17_zz_1a_zz_1b_zz_1e_zz_1f:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; ALL-NEXT:    vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x float> zeroinitializer, <16 x float> %b, <16 x i32><i32 0, i32 18, i32 0, i32 19, i32 4, i32 22, i32 4, i32 23, i32 6, i32 26, i32 6, i32 27, i32 8, i32 30, i32 8, i32 31>
+  ret <16 x float> %shuffle
+}
+
 define <16 x i32> @shuffle_v16i32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f(<16 x i32> %a, <16 x i32> %b) {
 ; ALL-LABEL: shuffle_v16i32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f:
 ; ALL:       # BB#0:
@@ -48,6 +78,16 @@ define <16 x i32> @shuffle_v16i32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1
   ret <16 x i32> %shuffle
 }
 
+define <16 x i32> @shuffle_v16i32_02_zz_03_zz_06_zz_07_zz_0a_zz_0b_zz_0e_zz_0f_zz(<16 x i32> %a, <16 x i32> %b) {
+; ALL-LABEL: shuffle_v16i32_02_zz_03_zz_06_zz_07_zz_0a_zz_0b_zz_0e_zz_0f_zz:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; ALL-NEXT:    vpunpckhdq {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; ALL-NEXT:    retq
+  %shuffle = shufflevector <16 x i32> %a, <16 x i32> zeroinitializer, <16 x i32><i32 2, i32 30, i32 3, i32 28, i32 6, i32 26, i32 7, i32 24, i32 10, i32 22, i32 11, i32 20, i32 14, i32 18, i32 15, i32 16>
+  ret <16 x i32> %shuffle
+}
+
 define <16 x float> @shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01(<16 x float> %a)  {
 ; ALL-LABEL: shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01:
 ; ALL:       # BB#0:
diff --git a/test/CodeGen/X86/vector-shuffle-512-v8.ll b/test/CodeGen/X86/vector-shuffle-512-v8.ll
index 0192014d374..2a8fc618f8f 100644
--- a/test/CodeGen/X86/vector-shuffle-512-v8.ll
+++ b/test/CodeGen/X86/vector-shuffle-512-v8.ll
@@ -1962,6 +1962,23 @@ define <8 x double> @shuffle_v8f64_082a4c6e(<8 x double> %a, <8 x double> %b) {
   ret <8 x double> %shuffle
 }
 
+define <8 x double> @shuffle_v8f64_0z2z4z6z(<8 x double> %a, <8 x double> %b) {
+;
+; AVX512F-LABEL: shuffle_v8f64_0z2z4z6z:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; AVX512F-NEXT:    vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; AVX512F-NEXT:    retq
+;
+; AVX512F-32-LABEL: shuffle_v8f64_0z2z4z6z:
+; AVX512F-32:       # BB#0:
+; AVX512F-32-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; AVX512F-32-NEXT:    vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; AVX512F-32-NEXT:    retl
+  %shuffle = shufflevector <8 x double> %a, <8 x double> zeroinitializer, <8 x i32><i32 0, i32 8, i32 2, i32 8, i32 4, i32 8, i32 6, i32 8>
+  ret <8 x double> %shuffle
+}
+
 define <8 x i64> @shuffle_v8i64_082a4c6e(<8 x i64> %a, <8 x i64> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8i64_082a4c6e:
@@ -1977,6 +1994,23 @@ define <8 x i64> @shuffle_v8i64_082a4c6e(<8 x i64> %a, <8 x i64> %b) {
   ret <8 x i64> %shuffle
 }
 
+define <8 x i64> @shuffle_v8i64_z8zazcze(<8 x i64> %a, <8 x i64> %b) {
+;
+; AVX512F-LABEL: shuffle_v8i64_z8zazcze:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; AVX512F-NEXT:    vpunpcklqdq {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; AVX512F-NEXT:    retq
+;
+; AVX512F-32-LABEL: shuffle_v8i64_z8zazcze:
+; AVX512F-32:       # BB#0:
+; AVX512F-32-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; AVX512F-32-NEXT:    vpunpcklqdq {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; AVX512F-32-NEXT:    retl
+  %shuffle = shufflevector <8 x i64> zeroinitializer, <8 x i64> %b, <8 x i32><i32 7, i32 8, i32 5, i32 10, i32 3, i32 12, i32 1, i32 14>
+  ret <8 x i64> %shuffle
+}
+
 define <8 x double> @shuffle_v8f64_193b5d7f(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_193b5d7f:
@@ -1992,6 +2026,23 @@ define <8 x double> @shuffle_v8f64_193b5d7f(<8 x double> %a, <8 x double> %b) {
   ret <8 x double> %shuffle
 }
 
+define <8 x double> @shuffle_v8f64_z9zbzdzf(<8 x double> %a, <8 x double> %b) {
+;
+; AVX512F-LABEL: shuffle_v8f64_z9zbzdzf:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; AVX512F-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; AVX512F-NEXT:    retq
+;
+; AVX512F-32-LABEL: shuffle_v8f64_z9zbzdzf:
+; AVX512F-32:       # BB#0:
+; AVX512F-32-NEXT:    vpxord %zmm0, %zmm0, %zmm0
+; AVX512F-32-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; AVX512F-32-NEXT:    retl
+  %shuffle = shufflevector <8 x double> zeroinitializer, <8 x double> %b, <8 x i32><i32 0, i32 9, i32 0, i32 11, i32 0, i32 13, i32 0, i32 15>
+  ret <8 x double> %shuffle
+}
+
 define <8 x i64> @shuffle_v8i64_193b5d7f(<8 x i64> %a, <8 x i64> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8i64_193b5d7f:
@@ -2007,6 +2058,23 @@ define <8 x i64> @shuffle_v8i64_193b5d7f(<8 x i64> %a, <8 x i64> %b) {
   ret <8 x i64> %shuffle
 }
 
+define <8 x i64> @shuffle_v8i64_1z3z5z7z(<8 x i64> %a, <8 x i64> %b) {
+;
+; AVX512F-LABEL: shuffle_v8i64_1z3z5z7z:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; AVX512F-NEXT:    vpunpckhqdq {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; AVX512F-NEXT:    retq
+;
+; AVX512F-32-LABEL: shuffle_v8i64_1z3z5z7z:
+; AVX512F-32:       # BB#0:
+; AVX512F-32-NEXT:    vpxord %zmm1, %zmm1, %zmm1
+; AVX512F-32-NEXT:    vpunpckhqdq {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; AVX512F-32-NEXT:    retl
+  %shuffle = shufflevector <8 x i64> %a, <8 x i64> zeroinitializer, <8 x i32><i32 1, i32 8, i32 3, i32 15, i32 5, i32 8, i32 7, i32 15>
+  ret <8 x i64> %shuffle
+}
+
 define <8 x double> @test_vshuff64x2_512(<8 x double> %x, <8 x double> %x1) nounwind {
 ; AVX512F-LABEL: test_vshuff64x2_512:
 ; AVX512F:       # BB#0:
@@ -2033,7 +2101,7 @@ define <8 x double> @test_vshuff64x2_512_maskz(<8 x double> %x, <8 x double> %x1
 ; AVX512F-32-LABEL: test_vshuff64x2_512_maskz:
 ; AVX512F-32:       # BB#0:
 ; AVX512F-32-NEXT:    vpmovsxwq %xmm2, %zmm2
-; AVX512F-32-NEXT:    vpandq .LCPI118_0, %zmm2, %zmm2
+; AVX512F-32-NEXT:    vpandq .LCPI122_0, %zmm2, %zmm2
 ; AVX512F-32-NEXT:    vptestmq %zmm2, %zmm2, %k1
 ; AVX512F-32-NEXT:    vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,4,5],zmm1[2,3,0,1]
 ; AVX512F-32-NEXT:    retl
@@ -2054,7 +2122,7 @@ define <8 x i64> @test_vshufi64x2_512_mask(<8 x i64> %x, <8 x i64> %x1, <8 x i1>
 ; AVX512F-32-LABEL: test_vshufi64x2_512_mask:
 ; AVX512F-32:       # BB#0:
 ; AVX512F-32-NEXT:    vpmovsxwq %xmm2, %zmm2
-; AVX512F-32-NEXT:    vpandq .LCPI119_0, %zmm2, %zmm2
+; AVX512F-32-NEXT:    vpandq .LCPI123_0, %zmm2, %zmm2
 ; AVX512F-32-NEXT:    vptestmq %zmm2, %zmm2, %k1
 ; AVX512F-32-NEXT:    vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,4,5],zmm1[2,3,0,1]
 ; AVX512F-32-NEXT:    retl
@@ -2091,7 +2159,7 @@ define <8 x double> @test_vshuff64x2_512_mem_mask(<8 x double> %x, <8 x double>
 ; AVX512F-32-LABEL: test_vshuff64x2_512_mem_mask:
 ; AVX512F-32:       # BB#0:
 ; AVX512F-32-NEXT:    vpmovsxwq %xmm1, %zmm1
-; AVX512F-32-NEXT:    vpandq .LCPI121_0, %zmm1, %zmm1
+; AVX512F-32-NEXT:    vpandq .LCPI125_0, %zmm1, %zmm1
 ; AVX512F-32-NEXT:    vptestmq %zmm1, %zmm1, %k1
 ; AVX512F-32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; AVX512F-32-NEXT:    vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,4,5],mem[2,3,0,1]
@@ -2114,7 +2182,7 @@ define <8 x double> @test_vshuff64x2_512_mem_maskz(<8 x double> %x, <8 x double>
 ; AVX512F-32-LABEL: test_vshuff64x2_512_mem_maskz:
 ; AVX512F-32:       # BB#0:
 ; AVX512F-32-NEXT:    vpmovsxwq %xmm1, %zmm1
-; AVX512F-32-NEXT:    vpandq .LCPI122_0, %zmm1, %zmm1
+; AVX512F-32-NEXT:    vpandq .LCPI126_0, %zmm1, %zmm1
 ; AVX512F-32-NEXT:    vptestmq %zmm1, %zmm1, %k1
 ; AVX512F-32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; AVX512F-32-NEXT:    vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,4,5],mem[2,3,0,1]