[X86][AVX] Tests tidyup

Cleanup/regenerate some tests for some upcoming patches.

llvm-svn: 255432

2 files changed, 68 insertions, 69 deletions

diff --git a/llvm/test/CodeGen/X86/2012-01-12-extract-sv.ll b/llvm/test/CodeGen/X86/2012-01-12-extract-sv.ll
index 677c902668b..92ec107a007 100644
--- a/llvm/test/CodeGen/X86/2012-01-12-extract-sv.ll
+++ b/llvm/test/CodeGen/X86/2012-01-12-extract-sv.ll
@@ -2,20 +2,20 @@
 
 define void @endless_loop() {
 ; CHECK-LABEL: endless_loop:
-; CHECK-NEXT:     # BB#0:
-; CHECK-NEXT:	vmovaps	(%eax), %ymm0
-; CHECK-NEXT:	vextractf128	$1, %ymm0, %xmm0
-; CHECK-NEXT:	vmovsldup	%xmm0, %xmm0    # xmm0 = xmm0[0,0,2,2]
-; CHECK-NEXT:	vmovddup	%xmm0, %xmm1    # xmm1 = xmm0[0,0]
-; CHECK-NEXT:	vinsertf128	$1, %xmm1, %ymm0, %ymm1
-; CHECK-NEXT:	vxorps	%xmm2, %xmm2, %xmm2
-; CHECK-NEXT:	vblendps	$128, %ymm1, %ymm2, %ymm1 # ymm1 = ymm2[0,1,2,3,4,5,6],ymm1[7]
-; CHECK-NEXT:	vxorps	%ymm2, %ymm2, %ymm2
-; CHECK-NEXT:	vblendps	$1, %ymm0, %ymm2, %ymm0 # ymm0 = ymm0[0],ymm2[1,2,3,4,5,6,7]
-; CHECK-NEXT:	vmovaps	%ymm0, (%eax)
-; CHECK-NEXT:	vmovaps	%ymm1, (%eax)
-; CHECK-NEXT:	vzeroupper
-; CHECK-NEXT:	retl
+; CHECK-NEXT:  # BB#0:
+; CHECK-NEXT:    vmovaps (%eax), %ymm0
+; CHECK-NEXT:    vextractf128 $1, %ymm0, %xmm0
+; CHECK-NEXT:    vmovsldup {{.*#+}} xmm0 = xmm0[0,0,2,2]
+; CHECK-NEXT:    vmovddup {{.*#+}} xmm1 = xmm0[0,0]
+; CHECK-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm1
+; CHECK-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; CHECK-NEXT:    vblendps {{.*#+}} ymm1 = ymm2[0,1,2,3,4,5,6],ymm1[7]
+; CHECK-NEXT:    vxorps %ymm2, %ymm2, %ymm2
+; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3,4,5,6,7]
+; CHECK-NEXT:    vmovaps %ymm0, (%eax)
+; CHECK-NEXT:    vmovaps %ymm1, (%eax)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retl
 entry:
   %0 = load <8 x i32>, <8 x i32> addrspace(1)* undef, align 32
   %1 = shufflevector <8 x i32> %0, <8 x i32> undef, <16 x i32> <i32 4, i32 4, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
diff --git a/llvm/test/CodeGen/X86/vec_extract-avx.ll b/llvm/test/CodeGen/X86/vec_extract-avx.ll
index ea34fde4a0e..abb07233d35 100644
--- a/llvm/test/CodeGen/X86/vec_extract-avx.ll
+++ b/llvm/test/CodeGen/X86/vec_extract-avx.ll
@@ -1,6 +1,5 @@
-target triple = "x86_64-unknown-unknown"
-
-; RUN: llc < %s -march=x86-64 -mattr=+avx | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s
 
 ; When extracting multiple consecutive elements from a larger
 ; vector into a smaller one, do it efficiently. We should use
@@ -9,6 +8,11 @@ target triple = "x86_64-unknown-unknown"
 
 ; Extracting the low elements only requires using the right kind of store.
 define void @low_v8f32_to_v4f32(<8 x float> %v, <4 x float>* %ptr) {
+; CHECK-LABEL: low_v8f32_to_v4f32:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vmovaps %xmm0, (%rdi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ext0 = extractelement <8 x float> %v, i32 0
   %ext1 = extractelement <8 x float> %v, i32 1
   %ext2 = extractelement <8 x float> %v, i32 2
@@ -19,15 +23,15 @@ define void @low_v8f32_to_v4f32(<8 x float> %v, <4 x float>* %ptr) {
   %ins3 = insertelement <4 x float> %ins2, float %ext3, i32 3
   store <4 x float> %ins3, <4 x float>* %ptr, align 16
   ret void
-
-; CHECK-LABEL: low_v8f32_to_v4f32
-; CHECK: vmovaps
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 ; Extracting the high elements requires just one AVX instruction.
 define void @high_v8f32_to_v4f32(<8 x float> %v, <4 x float>* %ptr) {
+; CHECK-LABEL: high_v8f32_to_v4f32:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vextractf128 $1, %ymm0, (%rdi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ext0 = extractelement <8 x float> %v, i32 4
   %ext1 = extractelement <8 x float> %v, i32 5
   %ext2 = extractelement <8 x float> %v, i32 6
@@ -38,17 +42,17 @@ define void @high_v8f32_to_v4f32(<8 x float> %v, <4 x float>* %ptr) {
   %ins3 = insertelement <4 x float> %ins2, float %ext3, i32 3
   store <4 x float> %ins3, <4 x float>* %ptr, align 16
   ret void
-
-; CHECK-LABEL: high_v8f32_to_v4f32
-; CHECK: vextractf128
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 ; Make sure element type doesn't alter the codegen. Note that
 ; if we were actually using the vector in this function and
 ; have AVX2, we should generate vextracti128 (the int version).
 define void @high_v8i32_to_v4i32(<8 x i32> %v, <4 x i32>* %ptr) {
+; CHECK-LABEL: high_v8i32_to_v4i32:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vextractf128 $1, %ymm0, (%rdi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ext0 = extractelement <8 x i32> %v, i32 4
   %ext1 = extractelement <8 x i32> %v, i32 5
   %ext2 = extractelement <8 x i32> %v, i32 6
@@ -59,91 +63,86 @@ define void @high_v8i32_to_v4i32(<8 x i32> %v, <4 x i32>* %ptr) {
   %ins3 = insertelement <4 x i32> %ins2, i32 %ext3, i32 3
   store <4 x i32> %ins3, <4 x i32>* %ptr, align 16
   ret void
-
-; CHECK-LABEL: high_v8i32_to_v4i32
-; CHECK: vextractf128
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 ; Make sure that element size doesn't alter the codegen.
 define void @high_v4f64_to_v2f64(<4 x double> %v, <2 x double>* %ptr) {
+; CHECK-LABEL: high_v4f64_to_v2f64:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vextractf128 $1, %ymm0, (%rdi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ext0 = extractelement <4 x double> %v, i32 2
   %ext1 = extractelement <4 x double> %v, i32 3
   %ins0 = insertelement <2 x double> undef, double %ext0, i32 0
   %ins1 = insertelement <2 x double> %ins0, double %ext1, i32 1
   store <2 x double> %ins1, <2 x double>* %ptr, align 16
   ret void
-
-; CHECK-LABEL: high_v4f64_to_v2f64
-; CHECK: vextractf128
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 ; PR25320 Make sure that a widened (possibly legalized) vector correctly zero-extends upper elements.
 ; FIXME - Ideally these should just call VMOVD/VMOVQ/VMOVSS/VMOVSD
 
 define void @legal_vzmovl_2i32_8i32(<2 x i32>* %in, <8 x i32>* %out) {
+; CHECK-LABEL: legal_vzmovl_2i32_8i32:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vpmovzxdq {{.*#+}} xmm0 = mem[0],zero,mem[1],zero
+; CHECK-NEXT:    vxorps %ymm1, %ymm1, %ymm1
+; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7]
+; CHECK-NEXT:    vmovaps %ymm0, (%rsi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ld = load <2 x i32>, <2 x i32>* %in, align 8
   %ext = extractelement <2 x i32> %ld, i64 0
   %ins = insertelement <8 x i32> <i32 undef, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>, i32 %ext, i64 0
   store <8 x i32> %ins, <8 x i32>* %out, align 32
   ret void
-
-; CHECK-LABEL: legal_vzmovl_2i32_8i32
-; CHECK: vpmovzxdq {{.*#+}} xmm0 = mem[0],zero,mem[1],zero
-; CHECK-NEXT: vxorps %ymm1, %ymm1, %ymm1
-; CHECK-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7]
-; CHECK-NEXT: vmovaps %ymm0, (%rsi)
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 define void @legal_vzmovl_2i64_4i64(<2 x i64>* %in, <4 x i64>* %out) {
+; CHECK-LABEL: legal_vzmovl_2i64_4i64:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vmovupd (%rdi), %xmm0
+; CHECK-NEXT:    vxorpd %ymm1, %ymm1, %ymm1
+; CHECK-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
+; CHECK-NEXT:    vmovapd %ymm0, (%rsi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ld = load <2 x i64>, <2 x i64>* %in, align 8
   %ext = extractelement <2 x i64> %ld, i64 0
   %ins = insertelement <4 x i64> <i64 undef, i64 0, i64 0, i64 0>, i64 %ext, i64 0
   store <4 x i64> %ins, <4 x i64>* %out, align 32
   ret void
-
-; CHECK-LABEL: legal_vzmovl_2i64_4i64
-; CHECK: vmovupd (%rdi), %xmm0
-; CHECK-NEXT: vxorpd %ymm1, %ymm1, %ymm1
-; CHECK-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
-; CHECK-NEXT: vmovapd %ymm0, (%rsi)
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 define void @legal_vzmovl_2f32_8f32(<2 x float>* %in, <8 x float>* %out) {
+; CHECK-LABEL: legal_vzmovl_2f32_8f32:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
+; CHECK-NEXT:    vxorps %ymm1, %ymm1, %ymm1
+; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7]
+; CHECK-NEXT:    vmovaps %ymm0, (%rsi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ld = load <2 x float>, <2 x float>* %in, align 8
   %ext = extractelement <2 x float> %ld, i64 0
   %ins = insertelement <8 x float> <float undef, float 0.0, float 0.0, float 0.0, float 0.0, float 0.0, float 0.0, float 0.0>, float %ext, i64 0
   store <8 x float> %ins, <8 x float>* %out, align 32
   ret void
-
-; CHECK-LABEL: legal_vzmovl_2f32_8f32
-; CHECK: vmovq {{.*#+}} xmm0 = mem[0],zero
-; CHECK-NEXT: vxorps %ymm1, %ymm1, %ymm1
-; CHECK-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7]
-; CHECK-NEXT: vmovaps %ymm0, (%rsi)
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }
 
 define void @legal_vzmovl_2f64_4f64(<2 x double>* %in, <4 x double>* %out) {
+; CHECK-LABEL: legal_vzmovl_2f64_4f64:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vmovupd (%rdi), %xmm0
+; CHECK-NEXT:    vxorpd %ymm1, %ymm1, %ymm1
+; CHECK-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
+; CHECK-NEXT:    vmovapd %ymm0, (%rsi)
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
   %ld = load <2 x double>, <2 x double>* %in, align 8
   %ext = extractelement <2 x double> %ld, i64 0
   %ins = insertelement <4 x double> <double undef, double 0.0, double 0.0, double 0.0>, double %ext, i64 0
   store <4 x double> %ins, <4 x double>* %out, align 32
   ret void
-
-; CHECK-LABEL: legal_vzmovl_2f64_4f64
-; CHECK: vmovupd (%rdi), %xmm0
-; CHECK-NEXT: vxorpd %ymm1, %ymm1, %ymm1
-; CHECK-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
-; CHECK-NEXT: vmovapd %ymm0, (%rsi)
-; CHECK-NEXT: vzeroupper
-; CHECK-NEXT: retq
 }