[X86][SSE] Add AVX512 merge consecutive load tests

Add AVX512F/AVX512BW 512-bit tests.

Add AVX512F tests to existing 128/256-bit tests.

llvm-svn: 259410

1 files changed, 150 insertions, 33 deletions

diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
index 2621385c4ca..4f3427ab513 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
@@ -1,6 +1,7 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; 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
 
 define <4 x double> @merge_4f64_2f64_23(<2 x double>* %ptr) nounwind uwtable noinline ssp {
 ; AVX-LABEL: merge_4f64_2f64_23:
@@ -49,10 +50,20 @@ define <4 x double> @merge_4f64_f64_2345(double* %ptr) nounwind uwtable noinline
 }
 
 define <4 x double> @merge_4f64_f64_3zuu(double* %ptr) nounwind uwtable noinline ssp {
-; AVX-LABEL: merge_4f64_f64_3zuu:
-; AVX:       # BB#0:
-; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
-; AVX-NEXT:    retq
+; AVX1-LABEL: merge_4f64_f64_3zuu:
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: merge_4f64_f64_3zuu:
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_4f64_f64_3zuu:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovsd 24(%rdi), %xmm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 3
   %val0 = load double, double* %ptr0
   %res0 = insertelement <4 x double> undef, double %val0, i32 0
@@ -106,6 +117,14 @@ define <4 x double> @merge_4f64_f64_34z6(double* %ptr) nounwind uwtable noinline
 ; AVX2-NEXT:    vpslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_4f64_f64_34z6:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovdqu 24(%rdi), %xmm0
+; AVX512F-NEXT:    vmovsd 48(%rdi), %xmm1
+; 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
   %ptr0 = getelementptr inbounds double, double* %ptr, i64 3
   %ptr1 = getelementptr inbounds double, double* %ptr, i64 4
   %ptr3 = getelementptr inbounds double, double* %ptr, i64 6
@@ -153,10 +172,20 @@ define <4 x i64> @merge_4i64_i64_1234(i64* %ptr) nounwind uwtable noinline ssp {
 }
 
 define <4 x i64> @merge_4i64_i64_1zzu(i64* %ptr) nounwind uwtable noinline ssp {
-; AVX-LABEL: merge_4i64_i64_1zzu:
-; AVX:       # BB#0:
-; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
-; AVX-NEXT:    retq
+; AVX1-LABEL: merge_4i64_i64_1zzu:
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: merge_4i64_i64_1zzu:
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_4i64_i64_1zzu:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovq 8(%rdi), %xmm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %val0 = load i64, i64* %ptr0
   %res0 = insertelement <4 x i64> undef, i64 %val0, i32 0
@@ -182,13 +211,29 @@ define <4 x i64> @merge_4i64_i64_23zz(i64* %ptr) nounwind uwtable noinline ssp {
 }
 
 define <8 x float> @merge_8f32_2f32_23z5(<2 x float>* %ptr) nounwind uwtable noinline ssp {
-; AVX-LABEL: merge_8f32_2f32_23z5:
-; AVX:       # BB#0:
-; AVX-NEXT:    vmovupd 16(%rdi), %xmm0
-; AVX-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
-; AVX-NEXT:    vmovhpd 40(%rdi), %xmm1, %xmm1
-; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX-NEXT:    retq
+; AVX1-LABEL: merge_8f32_2f32_23z5:
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vmovupd 16(%rdi), %xmm0
+; AVX1-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
+; AVX1-NEXT:    vmovhpd 40(%rdi), %xmm1, %xmm1
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: merge_8f32_2f32_23z5:
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vmovupd 16(%rdi), %xmm0
+; AVX2-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
+; AVX2-NEXT:    vmovhpd 40(%rdi), %xmm1, %xmm1
+; AVX2-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_8f32_2f32_23z5:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovups 16(%rdi), %xmm0
+; AVX512F-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; AVX512F-NEXT:    vmovhpd 40(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX512F-NEXT:    retq
   %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
@@ -227,6 +272,13 @@ define <8 x float> @merge_8f32_f32_12zzuuzz(float* %ptr) nounwind uwtable noinli
 ; AVX2-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_8f32_f32_12zzuuzz:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; AVX512F-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; AVX512F-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr1 = getelementptr inbounds float, float* %ptr, i64 2
   %val0 = load float, float* %ptr0
@@ -241,15 +293,35 @@ define <8 x float> @merge_8f32_f32_12zzuuzz(float* %ptr) nounwind uwtable noinli
 }
 
 define <8 x float> @merge_8f32_f32_1u3u5zu8(float* %ptr) nounwind uwtable noinline ssp {
-; AVX-LABEL: merge_8f32_f32_1u3u5zu8:
-; AVX:       # BB#0:
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; AVX-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm2[0,1],xmm0[1,0]
-; AVX-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
-; AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX-NEXT:    retq
+; AVX1-LABEL: merge_8f32_f32_1u3u5zu8:
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX1-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; AVX1-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm2[0,1],xmm0[1,0]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: merge_8f32_f32_1u3u5zu8:
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX2-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; AVX2-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; AVX2-NEXT:    vshufps {{.*#+}} xmm0 = xmm2[0,1],xmm0[1,0]
+; AVX2-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
+; AVX2-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_8f32_f32_1u3u5zu8:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovss 32(%rdi), %xmm0
+; AVX512F-NEXT:    vmovss 20(%rdi), %xmm1
+; AVX512F-NEXT:    vshufps {{.*#+}} xmm0 = xmm1[0,1],xmm0[1,0]
+; AVX512F-NEXT:    vmovss 4(%rdi), %xmm1
+; AVX512F-NEXT:    vinsertps $32, 12(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds float, float* %ptr, i64 1
   %ptr2 = getelementptr inbounds float, float* %ptr, i64 3
   %ptr4 = getelementptr inbounds float, float* %ptr, i64 5
@@ -292,6 +364,13 @@ define <8 x i32> @merge_8i32_i32_56zz9uzz(i32* %ptr) nounwind uwtable noinline s
 ; AVX2-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_8i32_i32_56zz9uzz:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; AVX512F-NEXT:    vmovd 36(%rdi), %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 5
   %ptr1 = getelementptr inbounds i32, i32* %ptr, i64 6
   %ptr4 = getelementptr inbounds i32, i32* %ptr, i64 9
@@ -330,6 +409,17 @@ define <8 x i32> @merge_8i32_i32_1u3u5zu8(i32* %ptr) nounwind uwtable noinline s
 ; AVX2-NEXT:    vpinsrd $2, 12(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_8i32_i32_1u3u5zu8:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovd 20(%rdi), %xmm0
+; AVX512F-NEXT:    vmovd 32(%rdi), %xmm1
+; AVX512F-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,1,1,0]
+; AVX512F-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; AVX512F-NEXT:    vmovd 4(%rdi), %xmm1
+; AVX512F-NEXT:    vpinsrd $2, 12(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds i32, i32* %ptr, i64 1
   %ptr2 = getelementptr inbounds i32, i32* %ptr, i64 3
   %ptr4 = getelementptr inbounds i32, i32* %ptr, i64 5
@@ -362,6 +452,14 @@ define <16 x i16> @merge_16i16_i16_89zzzuuuuuuuuuuuz(i16* %ptr) nounwind uwtable
 ; AVX2-NEXT:    vpinsrw $1, 18(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_16i16_i16_89zzzuuuuuuuuuuuz:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; AVX512F-NEXT:    vpinsrw $0, 16(%rdi), %xmm0, %xmm1
+; AVX512F-NEXT:    vpinsrw $1, 18(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 8
   %ptr1 = getelementptr inbounds i16, i16* %ptr, i64 9
   %val0 = load i16, i16* %ptr0
@@ -435,6 +533,16 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF(i16* %ptr) nounwind uwtable
 ; AVX2-NEXT:    vpinsrw $7, 30(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
+;
+; AVX512F-LABEL: merge_16i16_i16_0uu3zzuuuuuzCuEF:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; AVX512F-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; AVX512F-NEXT:    vpinsrw $4, 24(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vpinsrw $6, 28(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vpinsrw $7, 30(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX512F-NEXT:    retq
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 0
   %ptr3 = getelementptr inbounds i16, i16* %ptr, i64 3
   %ptrC = getelementptr inbounds i16, i16* %ptr, i64 12
@@ -479,23 +587,32 @@ define <32 x i8> @merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu(i8* %ptr) nounw
 ; AVX1-LABEL: merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vpxor %xmm0, %xmm0, %xmm0
-; AVX1-NEXT:    vpinsrb $0, 4(%rdi), %xmm0, %xmm1
-; AVX1-NEXT:    vpinsrb $1, 5(%rdi), %xmm1, %xmm1
-; AVX1-NEXT:    vpinsrb $3, 7(%rdi), %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrb $0, 2(%rdi), %xmm0, %xmm1
+; AVX1-NEXT:    vpinsrb $1, 3(%rdi), %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrb $3, 5(%rdi), %xmm1, %xmm1
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpxor %xmm0, %xmm0, %xmm0
-; AVX2-NEXT:    vpinsrb $0, 4(%rdi), %xmm0, %xmm1
-; AVX2-NEXT:    vpinsrb $1, 5(%rdi), %xmm1, %xmm1
-; AVX2-NEXT:    vpinsrb $3, 7(%rdi), %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrb $0, 2(%rdi), %xmm0, %xmm1
+; AVX2-NEXT:    vpinsrb $1, 3(%rdi), %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrb $3, 5(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
-  %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 4
-  %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 5
-  %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 7
+;
+; AVX512F-LABEL: merge_32i8_i8_23u5uuuuuuuuuuzzzzuuuuuuuuuuuuuu:
+; AVX512F:       # BB#0:
+; AVX512F-NEXT:    vpxor %xmm0, %xmm0, %xmm0
+; AVX512F-NEXT:    vpinsrb $0, 2(%rdi), %xmm0, %xmm1
+; AVX512F-NEXT:    vpinsrb $1, 3(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vpinsrb $3, 5(%rdi), %xmm1, %xmm1
+; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
+; AVX512F-NEXT:    retq
+  %ptr0 = getelementptr inbounds i8, i8* %ptr, i64 2
+  %ptr1 = getelementptr inbounds i8, i8* %ptr, i64 3
+  %ptr3 = getelementptr inbounds i8, i8* %ptr, i64 5
   %val0 = load i8, i8* %ptr0
   %val1 = load i8, i8* %ptr1
   %val3 = load i8, i8* %ptr3