cl::Hidden);
static cl::opt<bool> ExperimentalVectorShuffleLowering(
- "x86-experimental-vector-shuffle-lowering", cl::init(false),
+ "x86-experimental-vector-shuffle-lowering", cl::init(true),
cl::desc("Enable an experimental vector shuffle lowering code path."),
cl::Hidden);
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=i386 -mattr=+sse2 | grep pinsrw
-
-; Test to make sure we actually insert the bottom element of the vector
-define <8 x i16> @a(<8 x i16> %a) nounwind {
-entry:
- shufflevector <8 x i16> %a, <8 x i16> zeroinitializer, <8 x i32> < i32 0, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8 >
- %add = add <8 x i16> %0, %a
- ret <8 x i16> %add
-}
-
+++ /dev/null
-; RUN: llc -mtriple=i386-apple-darwin10.0 -relocation-model=pic -asm-verbose=false \
-; RUN: -mcpu=generic -disable-fp-elim -mattr=-sse4.1,-sse3,+sse2 -post-RA-scheduler=false -regalloc=basic < %s | \
-; RUN: FileCheck %s
-; rdar://6808032
-
-; CHECK: pextrw $14
-; CHECK-NEXT: shrl $8
-; CHECK-NEXT: pinsrw
-
-define void @update(i8** %args_list) nounwind {
-entry:
- %cmp.i = icmp eq i32 0, 0 ; <i1> [#uses=1]
- br i1 %cmp.i, label %if.then.i, label %test_cl.exit
-
-if.then.i: ; preds = %entry
- %val = load <16 x i8> addrspace(1)* null ; <<16 x i8>> [#uses=8]
- %tmp10.i = shufflevector <16 x i8> <i8 0, i8 0, i8 0, i8 undef, i8 0, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 0, i8 0, i8 0, i8 undef, i8 undef, i8 undef>, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 undef, i32 4, i32 undef, i32 6, i32 undef, i32 29, i32 undef, i32 10, i32 11, i32 12, i32 undef, i32 undef, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp17.i = shufflevector <16 x i8> %tmp10.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 18, i32 4, i32 undef, i32 6, i32 undef, i32 8, i32 undef, i32 10, i32 11, i32 12, i32 undef, i32 undef, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp24.i = shufflevector <16 x i8> %tmp17.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 24, i32 6, i32 undef, i32 8, i32 undef, i32 10, i32 11, i32 12, i32 undef, i32 undef, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp31.i = shufflevector <16 x i8> %tmp24.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 undef, i32 8, i32 undef, i32 10, i32 11, i32 12, i32 21, i32 undef, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp38.i = shufflevector <16 x i8> %tmp31.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 27, i32 8, i32 undef, i32 10, i32 11, i32 12, i32 13, i32 undef, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp45.i = shufflevector <16 x i8> %tmp38.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 undef, i32 10, i32 11, i32 12, i32 13, i32 29, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp52.i = shufflevector <16 x i8> %tmp45.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 21, i32 10, i32 11, i32 12, i32 13, i32 14, i32 undef> ; <<16 x i8>> [#uses=1]
- %tmp59.i = shufflevector <16 x i8> %tmp52.i, <16 x i8> %val, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 20> ; <<16 x i8>> [#uses=1]
- store <16 x i8> %tmp59.i, <16 x i8> addrspace(1)* null
- ret void
-
-test_cl.exit: ; preds = %entry
- ret void
-}
;CHECK: vpxor
;CHECK: vinserti128
;CHECK: vpshufd
-;CHECK: vpshufd
+;CHECK: vpbroadcastd
;CHECK: vmulps
;CHECK: vmulps
;CHECK: ret
ret <4 x i64> %shuffle
}
-;;;
-;;; Check that some 256-bit vectors are xformed into 128 ops
-; CHECK: _A
-; CHECK: vshufpd $1
-; CHECK-NEXT: vextractf128 $1
-; CHECK-NEXT: vshufpd $1
-; CHECK-NEXT: vinsertf128 $1
-define <4 x i64> @A(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: _B
-; CHECK: vshufpd $1, %ymm
-define <4 x i64> @B(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 undef, i32 undef, i32 6>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: vmovlhps
-; CHECK-NEXT: vextractf128 $1
-; CHECK-NEXT: vmovlhps
-; CHECK-NEXT: vinsertf128 $1
-define <4 x i64> @C(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 undef, i32 0, i32 undef, i32 6>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: vpshufd $-96
-; CHECK: vpshufd $-6
-; CHECK: vinsertf128 $1
-define <8 x i32> @D(<8 x i32> %a, <8 x i32> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 10, i32 10, i32 11, i32 11>
- ret <8 x i32> %shuffle
-}
-
;;; Don't crash on movd
; CHECK: _VMOVZQI2PQI
; CHECK: vmovd (%
define <4 x float> @test_x86_avx_vpermil_ps(<4 x float> %a0) {
- ; CHECK: vpshufd
+ ; CHECK: vpermilps
%res = call <4 x float> @llvm.x86.avx.vpermil.ps(<4 x float> %a0, i8 7) ; <<4 x float>> [#uses=1]
ret <4 x float> %res
}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s
-
-; CHECK: vmovsldup
-define <8 x float> @movdupA(<8 x float> %src) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <8 x float> %src, <8 x float> undef, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
- ret <8 x float> %shuffle.i
-}
-
-; CHECK: vmovshdup
-define <8 x float> @movdupB(<8 x float> %src) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <8 x float> %src, <8 x float> undef, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
- ret <8 x float> %shuffle.i
-}
-
-; CHECK: vmovsldup
-define <4 x i64> @movdupC(<4 x i64> %src) nounwind uwtable readnone ssp {
-entry:
- %0 = bitcast <4 x i64> %src to <8 x float>
- %shuffle.i = shufflevector <8 x float> %0, <8 x float> undef, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
- %1 = bitcast <8 x float> %shuffle.i to <4 x i64>
- ret <4 x i64> %1
-}
-
-; CHECK: vmovshdup
-define <4 x i64> @movdupD(<4 x i64> %src) nounwind uwtable readnone ssp {
-entry:
- %0 = bitcast <4 x i64> %src to <8 x float>
- %shuffle.i = shufflevector <8 x float> %0, <8 x float> undef, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
- %1 = bitcast <8 x float> %shuffle.i to <4 x i64>
- ret <4 x i64> %1
-}
-
-; Check that there is a pattern (v2f64 (X86VBroadcast f64:$src)).
-
-; CHECK-LABEL: _vbroadcast_v128_f64
-; CHECK: vmovsd LCPI{{[0-9]+}}_0(%rip), %xmm[[R0:[0-9]+]]
-; CHECK: vmovddup %xmm[[R0]], %xmm{{[0-9]+}}
-
-@E1 = external global [5 x double], align 16
-@.str3 = external unnamed_addr constant [44 x i8], align 1
-
-define void @vbroadcast_v128_f64() #0 {
-entry:
- store <2 x double> <double -1.000000e+00, double -1.000000e+00>, <2 x double>* bitcast (double* getelementptr inbounds ([5 x double]* @E1, i64 0, i64 2) to <2 x double>*), align 16
- tail call void @foo1(double -1.000000e+00)
- ret void
-}
-
-declare void @foo1(double)
-
-attributes #0 = { optsize }
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s
-
-; PR11102
-define <4 x float> @test1(<4 x float> %a) nounwind {
- %b = shufflevector <4 x float> zeroinitializer, <4 x float> %a, <4 x i32> <i32 2, i32 5, i32 undef, i32 undef>
- ret <4 x float> %b
-; CHECK-LABEL: test1:
-;; TODO: This test could be improved by removing the xor instruction and
-;; having vinsertps zero out the needed elements.
-; CHECK: vxorps
-; CHECK: vinsertps
-}
-
-; rdar://10538417
-define <3 x i64> @test2(<2 x i64> %v) nounwind readnone {
-; CHECK-LABEL: test2:
-; CHECK: vinsertf128
- %1 = shufflevector <2 x i64> %v, <2 x i64> %v, <3 x i32> <i32 0, i32 1, i32 undef>
- %2 = shufflevector <3 x i64> zeroinitializer, <3 x i64> %1, <3 x i32> <i32 3, i32 4, i32 2>
- ret <3 x i64> %2
-; CHECK: ret
-}
-
-define <4 x i64> @test3(<4 x i64> %a, <4 x i64> %b) nounwind {
- %c = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 undef>
- ret <4 x i64> %c
-; CHECK-LABEL: test3:
-; CHECK: vblendpd
-; CHECK: ret
-}
-
-define <8 x float> @test4(float %a) nounwind {
- %b = insertelement <8 x float> zeroinitializer, float %a, i32 0
- ret <8 x float> %b
-; CHECK-LABEL: test4:
-; CHECK: vinsertf128
-}
-
-; rdar://10594409
-define <8 x float> @test5(float* nocapture %f) nounwind uwtable readonly ssp {
-entry:
- %0 = bitcast float* %f to <4 x float>*
- %1 = load <4 x float>* %0, align 16
-; CHECK: test5
-; CHECK: vmovaps
-; CHECK-NOT: vxorps
-; CHECK-NOT: vinsertf128
- %shuffle.i = shufflevector <4 x float> %1, <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 4, i32 4, i32 4>
- ret <8 x float> %shuffle.i
-}
-
-define <4 x double> @test6(double* nocapture %d) nounwind uwtable readonly ssp {
-entry:
- %0 = bitcast double* %d to <2 x double>*
- %1 = load <2 x double>* %0, align 16
-; CHECK: test6
-; CHECK: vmovaps
-; CHECK-NOT: vxorps
-; CHECK-NOT: vinsertf128
- %shuffle.i = shufflevector <2 x double> %1, <2 x double> <double 0.000000e+00, double undef>, <4 x i32> <i32 0, i32 1, i32 2, i32 2>
- ret <4 x double> %shuffle.i
-}
-
-define <16 x i16> @test7(<4 x i16> %a) nounwind {
-; CHECK: test7
- %b = shufflevector <4 x i16> %a, <4 x i16> undef, <16 x i32> <i32 1, i32 1, 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>
-; CHECK: ret
- ret <16 x i16> %b
-}
-
-; CHECK: test8
-define void @test8() {
-entry:
- %0 = load <16 x i64> addrspace(1)* null, align 128
- %1 = shufflevector <16 x i64> <i64 undef, i64 undef, i64 0, i64 undef, i64 0, i64 0, i64 0, i64 0, i64 0, i64 0, i64 undef, i64 0, i64 undef, i64 undef, i64 undef, i64 undef>, <16 x i64> %0, <16 x i32> <i32 17, i32 18, i32 2, i32 undef, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 undef, i32 11, i32 undef, i32 undef, i32 undef, i32 26>
- %2 = shufflevector <16 x i64> %1, <16 x i64> %0, <16 x i32> <i32 0, i32 1, i32 2, i32 30, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 undef, i32 11, i32 undef, i32 22, i32 20, i32 15>
- store <16 x i64> %2, <16 x i64> addrspace(1)* undef, align 128
-; CHECK: ret
- ret void
-}
-
-; Extract a value from a shufflevector..
-define i32 @test9(<4 x i32> %a) nounwind {
-; CHECK: test9
-; CHECK: vpextrd
- %b = shufflevector <4 x i32> %a, <4 x i32> undef, <8 x i32> <i32 1, i32 1, i32 2, i32 2, i32 3, i32 3, i32 undef, i32 4>
- %r = extractelement <8 x i32> %b, i32 2
-; CHECK: ret
- ret i32 %r
-}
-
-; Extract a value which is the result of an undef mask.
-define i32 @test10(<4 x i32> %a) nounwind {
-; CHECK: @test10
-; CHECK-NOT: {{^[^#]*[a-z]}}
-; CHECK: ret
- %b = shufflevector <4 x i32> %a, <4 x i32> undef, <8 x i32> <i32 1, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- %r = extractelement <8 x i32> %b, i32 2
- ret i32 %r
-}
-
-define <4 x float> @test11(<4 x float> %a) nounwind {
-; CHECK: test11
-; CHECK: vpshufd $27
- %tmp1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
- ret <4 x float> %tmp1
-}
-
-define <4 x float> @test12(<4 x float>* %a) nounwind {
-; CHECK: test12
-; CHECK: vpshufd
- %tmp0 = load <4 x float>* %a
- %tmp1 = shufflevector <4 x float> %tmp0, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
- ret <4 x float> %tmp1
-}
-
-define <4 x i32> @test13(<4 x i32> %a) nounwind {
-; CHECK: test13
-; CHECK: vpshufd $27
- %tmp1 = shufflevector <4 x i32> %a, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
- ret <4 x i32> %tmp1
-}
-
-define <4 x i32> @test14(<4 x i32>* %a) nounwind {
-; CHECK: test14
-; CHECK: vpshufd $27, (
- %tmp0 = load <4 x i32>* %a
- %tmp1 = shufflevector <4 x i32> %tmp0, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
- ret <4 x i32> %tmp1
-}
-
-; CHECK: test15
-; CHECK: vpshufd $8
-; CHECK: ret
-define <4 x i32> @test15(<2 x i32>%x) nounwind readnone {
- %x1 = shufflevector <2 x i32> %x, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
- ret <4 x i32>%x1
-}
-
-; rdar://10974078
-define <8 x float> @test16(float* nocapture %f) nounwind uwtable readonly ssp {
-entry:
- %0 = bitcast float* %f to <4 x float>*
- %1 = load <4 x float>* %0, align 8
-; CHECK: test16
-; CHECK: vmovups
-; CHECK-NOT: vxorps
-; CHECK-NOT: vinsertf128
- %shuffle.i = shufflevector <4 x float> %1, <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 4, i32 4, i32 4>
- ret <8 x float> %shuffle.i
-}
-
-; PR12413
-; CHECK: shuf1
-; CHECK: vpshufb
-; CHECK: vpshufb
-; CHECK: vpshufb
-; CHECK: vpshufb
-define <32 x i8> @shuf1(<32 x i8> %inval1, <32 x i8> %inval2) {
-entry:
- %0 = shufflevector <32 x i8> %inval1, <32 x i8> %inval2, <32 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30, i32 32, i32 34, i32 36, i32 38, i32 40, i32 42, i32 44, i32 46, i32 48, i32 50, i32 52, i32 54, i32 56, i32 58, i32 60, i32 62>
- ret <32 x i8> %0
-}
-
-; handle the case where only half of the 256-bits is splittable
-; CHECK: shuf2
-; CHECK: vpshufb
-; CHECK: vpshufb
-; CHECK: vpextrb
-; CHECK: vpextrb
-define <32 x i8> @shuf2(<32 x i8> %inval1, <32 x i8> %inval2) {
-entry:
- %0 = shufflevector <32 x i8> %inval1, <32 x i8> %inval2, <32 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30, i32 31, i32 34, i32 36, i32 38, i32 40, i32 42, i32 44, i32 46, i32 48, i32 50, i32 52, i32 54, i32 56, i32 58, i32 60, i32 62>
- ret <32 x i8> %0
-}
-
-; CHECK: blend1
-; CHECK: vblendps
-; CHECK: ret
-define <4 x i32> @blend1(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline {
- %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 7>
- ret <4 x i32> %t
-}
-
-; CHECK: blend2
-; CHECK: vblendps
-; CHECK: ret
-define <4 x i32> @blend2(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline {
- %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x i32> %t
-}
-
-; CHECK: blend2a
-; CHECK: vblendps
-; CHECK: ret
-define <4 x float> @blend2a(<4 x float> %a, <4 x float> %b) nounwind alwaysinline {
- %t = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x float> %t
-}
-
-; CHECK: blend3
-; CHECK-NOT: vblendps
-; CHECK: ret
-define <4 x i32> @blend3(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline {
- %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 1, i32 5, i32 2, i32 7>
- ret <4 x i32> %t
-}
-
-; CHECK: blend4
-; CHECK: vblendpd
-; CHECK: ret
-define <4 x i64> @blend4(<4 x i64> %a, <4 x i64> %b) nounwind alwaysinline {
- %t = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 7>
- ret <4 x i64> %t
-}
-
-; CHECK: narrow
-; CHECK: vpermilps
-; CHECK: ret
-define <16 x i16> @narrow(<16 x i16> %a) nounwind alwaysinline {
- %t = shufflevector <16 x i16> %a, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 1, i32 6, i32 7, i32 4, i32 5, i32 10, i32 11, i32 8, i32 undef, i32 14, i32 15, i32 undef, i32 undef>
- ret <16 x i16> %t
-}
-
-;CHECK-LABEL: test17:
-;CHECK-NOT: vinsertf128
-;CHECK: ret
-define <8 x float> @test17(<4 x float> %y) {
- %x = shufflevector <4 x float> %y, <4 x float> undef, <8 x i32> <i32 undef, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x float> %x
-}
-
-; CHECK: test18
-; CHECK: vmovshdup
-; CHECK: vblendps
-; CHECK: ret
-define <8 x float> @test18(<8 x float> %A, <8 x float>%B) nounwind {
- %S = shufflevector <8 x float> %A, <8 x float> %B, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
- ret <8 x float>%S
-}
-
-; CHECK: test19
-; CHECK: vmovsldup
-; CHECK: vblendps
-; CHECK: ret
-define <8 x float> @test19(<8 x float> %A, <8 x float>%B) nounwind {
- %S = shufflevector <8 x float> %A, <8 x float> %B, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
- ret <8 x float>%S
-}
-
-; rdar://12684358
-; Make sure loads happen before stores.
-; CHECK: swap8doubles
-; CHECK: vmovups {{[0-9]*}}(%rdi), %xmm{{[0-9]+}}
-; CHECK: vmovups {{[0-9]*}}(%rdi), %xmm{{[0-9]+}}
-; CHECK: vinsertf128 $1, {{[0-9]*}}(%rdi), %ymm{{[0-9]+}}
-; CHECK: vinsertf128 $1, {{[0-9]*}}(%rdi), %ymm{{[0-9]+}}
-; CHECK: vmovaps {{[0-9]*}}(%rsi), %ymm{{[0-9]+}}
-; CHECK: vmovaps {{[0-9]*}}(%rsi), %ymm{{[0-9]+}}
-; CHECK: vmovaps %xmm{{[0-9]+}}, {{[0-9]*}}(%rdi)
-; CHECK: vextractf128
-; CHECK: vmovaps %xmm{{[0-9]+}}, {{[0-9]*}}(%rdi)
-; CHECK: vextractf128
-; CHECK: vmovaps %ymm{{[0-9]+}}, {{[0-9]*}}(%rsi)
-; CHECK: vmovaps %ymm{{[0-9]+}}, {{[0-9]*}}(%rsi)
-define void @swap8doubles(double* nocapture %A, double* nocapture %C) nounwind uwtable ssp {
-entry:
- %add.ptr = getelementptr inbounds double* %A, i64 2
- %v.i = bitcast double* %A to <2 x double>*
- %0 = load <2 x double>* %v.i, align 1
- %shuffle.i.i = shufflevector <2 x double> %0, <2 x double> <double 0.000000e+00, double undef>, <4 x i32> <i32 0, i32 1, i32 2, i32 2>
- %v1.i = bitcast double* %add.ptr to <2 x double>*
- %1 = load <2 x double>* %v1.i, align 1
- %2 = tail call <4 x double> @llvm.x86.avx.vinsertf128.pd.256(<4 x double> %shuffle.i.i, <2 x double> %1, i8 1) nounwind
- %add.ptr1 = getelementptr inbounds double* %A, i64 6
- %add.ptr2 = getelementptr inbounds double* %A, i64 4
- %v.i27 = bitcast double* %add.ptr2 to <2 x double>*
- %3 = load <2 x double>* %v.i27, align 1
- %shuffle.i.i28 = shufflevector <2 x double> %3, <2 x double> <double 0.000000e+00, double undef>, <4 x i32> <i32 0, i32 1, i32 2, i32 2>
- %v1.i29 = bitcast double* %add.ptr1 to <2 x double>*
- %4 = load <2 x double>* %v1.i29, align 1
- %5 = tail call <4 x double> @llvm.x86.avx.vinsertf128.pd.256(<4 x double> %shuffle.i.i28, <2 x double> %4, i8 1) nounwind
- %6 = bitcast double* %C to <4 x double>*
- %7 = load <4 x double>* %6, align 32
- %add.ptr5 = getelementptr inbounds double* %C, i64 4
- %8 = bitcast double* %add.ptr5 to <4 x double>*
- %9 = load <4 x double>* %8, align 32
- %shuffle.i26 = shufflevector <4 x double> %7, <4 x double> undef, <2 x i32> <i32 0, i32 1>
- %10 = tail call <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double> %7, i8 1)
- %shuffle.i = shufflevector <4 x double> %9, <4 x double> undef, <2 x i32> <i32 0, i32 1>
- %11 = tail call <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double> %9, i8 1)
- store <2 x double> %shuffle.i26, <2 x double>* %v.i, align 16
- store <2 x double> %10, <2 x double>* %v1.i, align 16
- store <2 x double> %shuffle.i, <2 x double>* %v.i27, align 16
- store <2 x double> %11, <2 x double>* %v1.i29, align 16
- store <4 x double> %2, <4 x double>* %6, align 32
- store <4 x double> %5, <4 x double>* %8, align 32
- ret void
-}
-declare <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double>, i8) nounwind readnone
-declare <4 x double> @llvm.x86.avx.vinsertf128.pd.256(<4 x double>, <2 x double>, i8) nounwind readnone
-
-; this test case just should not fail
-define void @test20() {
- %a0 = insertelement <3 x double> <double 0.000000e+00, double 0.000000e+00, double undef>, double 0.000000e+00, i32 2
- store <3 x double> %a0, <3 x double>* undef, align 1
- %a1 = insertelement <3 x double> <double 0.000000e+00, double 0.000000e+00, double undef>, double undef, i32 2
- store <3 x double> %a1, <3 x double>* undef, align 1
- ret void
-}
-
-define <2 x i64> @test_insert_64_zext(<2 x i64> %i) {
-; CHECK-LABEL: test_insert_64_zext
-; CHECK-NOT: xor
-; CHECK: vmovq
- %1 = shufflevector <2 x i64> %i, <2 x i64> <i64 0, i64 undef>, <2 x i32> <i32 0, i32 2>
- ret <2 x i64> %1
-}
-
-;; Ensure we don't use insertps from non v4x32 vectors.
-;; On SSE4.1 it works because bigger vectors use more than 1 register.
-;; On AVX they get passed in a single register.
-;; FIXME: We could probably optimize this case, if we're only using the
-;; first 4 indices.
-define <4 x i32> @insert_from_diff_size(<8 x i32> %x) {
-; CHECK-LABEL: insert_from_diff_size:
-; CHECK-NOT: insertps
-; CHECK: ret
- %vecext = extractelement <8 x i32> %x, i32 0
- %vecinit = insertelement <4 x i32> undef, i32 %vecext, i32 0
- %vecinit1 = insertelement <4 x i32> %vecinit, i32 0, i32 1
- %vecinit2 = insertelement <4 x i32> %vecinit1, i32 0, i32 2
- %a.0 = extractelement <8 x i32> %x, i32 0
- %vecinit3 = insertelement <4 x i32> %vecinit2, i32 %a.0, i32 3
- ret <4 x i32> %vecinit3
-}
ret <32 x i8> %shuffle
}
-; CHECK: vpunpckhwd %xmm
-; CHECK-NEXT: vpshufd $85
+; CHECK: vpshufb {{.*}} ## xmm0 = xmm0[10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11]
; CHECK-NEXT: vinsertf128 $1
define <16 x i16> @funcB(<16 x i16> %a) nounwind uwtable readnone ssp {
entry:
}
; CHECK: vmovq
-; CHECK-NEXT: vmovlhps %xmm
+; CHECK-NEXT: vunpcklpd %xmm
; CHECK-NEXT: vinsertf128 $1
define <4 x i64> @funcC(i64 %q) nounwind uwtable readnone ssp {
entry:
ret <8 x float> %load_broadcast12281250
}
-; CHECK: vpshufd $0
+; CHECK: vpermilps $4
; CHECK-NEXT: vinsertf128 $1
define <8 x float> @funcF(i32 %val) nounwind {
%ret6 = insertelement <8 x i32> undef, i32 %val, i32 6
ret <8 x float> %tmp
}
-; CHECK: vpshufd $0
+; CHECK: vpermilps $0
; CHECK-NEXT: vinsertf128 $1
define <8 x float> @funcG(<8 x float> %a) nounwind uwtable readnone ssp {
entry:
}
; CHECK: vextractf128 $1
-; CHECK-NEXT: vpshufd
+; CHECK-NEXT: vpermilps $85
; CHECK-NEXT: vinsertf128 $1
define <8 x float> @funcH(<8 x float> %a) nounwind uwtable readnone ssp {
entry:
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s
-
-; CHECK: vmovddup %ymm
-define <4 x i64> @A(<4 x i64> %a) {
- %c = shufflevector <4 x i64> %a, <4 x i64> undef, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
- ret <4 x i64> %c
-}
-
-; CHECK: vmovddup (%
-define <4 x i64> @B(<4 x i64>* %ptr) {
- %a = load <4 x i64>* %ptr
- %c = shufflevector <4 x i64> %a, <4 x i64> undef, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
- ret <4 x i64> %c
-}
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX2
define <8 x float> @A(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: A:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
-; ALL-NEXT: retq
+; AVX1-LABEL: A:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: A:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
+; AVX2-NEXT: retq
entry:
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
ret <8 x float> %shuffle
define <8 x float> @B(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
; ALL-LABEL: B:
; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
+; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
; ALL-NEXT: retq
entry:
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 12, i32 13, i32 14, i32 15>
}
define <8 x float> @C(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: C:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; ALL-NEXT: retq
+; AVX1-LABEL: C:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: C:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
entry:
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
ret <8 x float> %shuffle
}
define <8 x float> @D(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: D:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
-; ALL-NEXT: retq
+; AVX1-LABEL: D:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: D:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX2-NEXT: retq
entry:
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 4, i32 5, i32 6, i32 7>
ret <8 x float> %shuffle
}
define <32 x i8> @E(<32 x i8> %a, <32 x i8> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: E:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
-; ALL-NEXT: retq
+; AVX1-LABEL: E:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: E:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX2-NEXT: retq
entry:
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
ret <32 x i8> %shuffle
}
define <4 x i64> @E2(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: E2:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1]
-; ALL-NEXT: retq
+; AVX1-LABEL: E2:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: E2:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[2,3,2,3]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
+; AVX2-NEXT: retq
entry:
%shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
ret <4 x i64> %shuffle
define <32 x i8> @Ei(<32 x i8> %a, <32 x i8> %b) nounwind uwtable readnone ssp {
; AVX1-LABEL: Ei:
; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-; AVX1-NEXT: vpaddb %xmm2, %xmm1, %xmm1
-; AVX1-NEXT: vpaddb %xmm2, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT: vpaddb {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: Ei:
; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vpaddb {{.*}}, %ymm0, %ymm0
-; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX2-NEXT: vpaddb {{.*}}(%rip), %ymm0, %ymm0
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,3,2,3]
; AVX2-NEXT: retq
entry:
; add forces execution domain
define <4 x i64> @E2i(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
; AVX1-LABEL: E2i:
; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,1]
-; AVX1-NEXT: vpaddq %xmm3, %xmm2, %xmm2
-; AVX1-NEXT: vpaddq %xmm3, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1]
+; AVX1-NEXT: vpaddq {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
; AVX1-NEXT: retq
;
; AVX2-LABEL: E2i:
; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vpbroadcastq {{.*}}, %ymm2
+; AVX2-NEXT: vpbroadcastq {{.*}}(%rip), %ymm2
; AVX2-NEXT: vpaddq %ymm2, %ymm0, %ymm0
-; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[2,3,2,3]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT: retq
entry:
; add forces execution domain
define <8 x i32> @E3i(<8 x i32> %a, <8 x i32> %b) nounwind uwtable readnone ssp {
; AVX1-LABEL: E3i:
; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,1,1,1]
-; AVX1-NEXT: vpaddd %xmm3, %xmm2, %xmm2
-; AVX1-NEXT: vpaddd %xmm3, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT: vpaddd {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
; AVX1-NEXT: retq
;
; AVX2-LABEL: E3i:
; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vpbroadcastd {{.*}}, %ymm2
+; AVX2-NEXT: vpbroadcastd {{.*}}(%rip), %ymm2
; AVX2-NEXT: vpaddd %ymm2, %ymm0, %ymm0
-; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,3,2,3]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; AVX2-NEXT: retq
entry:
; add forces execution domain
define <16 x i16> @E4i(<16 x i16> %a, <16 x i16> %b) nounwind uwtable readnone ssp {
; AVX1-LABEL: E4i:
; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,1,1,1,1,1,1,1]
-; AVX1-NEXT: vpaddw %xmm3, %xmm2, %xmm2
-; AVX1-NEXT: vpaddw %xmm3, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[0,1],ymm0[0,1]
+; AVX1-NEXT: vpaddw {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
; AVX1-NEXT: retq
;
; AVX2-LABEL: E4i:
; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vpaddw {{.*}}, %ymm0, %ymm0
-; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[0,1],ymm0[0,1]
+; AVX2-NEXT: vpaddw {{.*}}(%rip), %ymm0, %ymm0
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT: retq
entry:
; add forces execution domain
define <16 x i16> @E5i(<16 x i16>* %a, <16 x i16>* %b) nounwind uwtable readnone ssp {
; AVX1-LABEL: E5i:
; AVX1: ## BB#0: ## %entry
-; AVX1-NEXT: vmovaps (%rdi), %ymm0
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [1,1,1,1,1,1,1,1]
-; AVX1-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; AVX1-NEXT: vpaddw %xmm2, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = mem[0,1],ymm0[0,1]
+; AVX1-NEXT: vmovdqa (%rdi), %ymm0
+; AVX1-NEXT: vpaddw {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT: vmovapd (%rsi), %ymm1
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3]
; AVX1-NEXT: retq
;
; AVX2-LABEL: E5i:
; AVX2: ## BB#0: ## %entry
; AVX2-NEXT: vmovdqa (%rdi), %ymm0
-; AVX2-NEXT: vpaddw {{.*}}, %ymm0, %ymm0
-; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = mem[0,1],ymm0[0,1]
+; AVX2-NEXT: vmovdqa (%rsi), %ymm1
+; AVX2-NEXT: vpaddw {{.*}}(%rip), %ymm0, %ymm0
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT: retq
entry:
%c = load <16 x i16>* %a
;;;; Cases with undef indicies mixed in the mask
define <8 x float> @F(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: F:
-; ALL: ## BB#0: ## %entry
-; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
-; ALL-NEXT: retq
+; AVX1-LABEL: F:
+; AVX1: ## BB#0: ## %entry
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: F:
+; AVX2: ## BB#0: ## %entry
+; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,3]
+; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,1,0,1]
+; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
+; AVX2-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 9, i32 undef, i32 11>
ret <8 x float> %shuffle
; AVX1-LABEL: G:
; AVX1: ## BB#0: ## %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
-; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,3]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vpermilps {{.*#+}} ymm1 = ymm1[0,0,2,3,4,4,6,7]
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
; AVX1-NEXT: retq
;
; AVX2-LABEL: G:
; AVX2: ## BB#0: ## %entry
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm1
-; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,3]
-; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,3]
+; 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
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>
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s
-
-; CHECK: vpermilps
-define <8 x float> @funcA(<8 x float> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> undef, <8 x i32> <i32 1, i32 2, i32 3, i32 1, i32 5, i32 6, i32 7, i32 5>
- ret <8 x float> %shuffle
-}
-
-; CHECK: vpermilpd
-define <4 x double> @funcB(<4 x double> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x double> %a, <4 x double> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 3>
- ret <4 x double> %shuffle
-}
-
-; CHECK: vpermilps
-define <8 x i32> @funcC(<8 x i32> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x i32> %a, <8 x i32> undef, <8 x i32> <i32 1, i32 2, i32 3, i32 1, i32 5, i32 6, i32 7, i32 5>
- ret <8 x i32> %shuffle
-}
-
-; CHECK: vpermilpd
-define <4 x i64> @funcD(<4 x i64> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i64> %a, <4 x i64> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 3>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: vpermilpd
-define <4 x i64> @funcQ(<4 x i64>* %a) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <4 x i64>* %a
- %shuffle = shufflevector <4 x i64> %a2, <4 x i64> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 3>
- ret <4 x i64> %shuffle
-}
-
-; vpermil should match masks like this: <u,3,1,2,4,u,5,6>. Check that the
-; target specific mask was correctly generated.
-; CHECK: vpermilps $-100
-define <8 x float> @funcE(<8 x float> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> undef, <8 x i32> <i32 8, i32 3, i32 1, i32 2, i32 4, i32 8, i32 5, i32 6>
- ret <8 x float> %shuffle
-}
-
-; CHECK: palignr $8
-; CHECK: palignr $8
-define <8 x float> @funcF(<8 x float> %a) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> zeroinitializer, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
- ret <8 x float> %shuffle
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s
-
-; CHECK: vshufps $-53, %ymm
-define <8 x float> @A(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 3, i32 2, i32 8, i32 11, i32 7, i32 6, i32 12, i32 15>
- ret <8 x float> %shuffle
-}
-
-; CHECK: vshufps $-53, (%{{.*}}), %ymm
-define <8 x float> @A2(<8 x float>* %a, <8 x float>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <8 x float>* %a
- %b2 = load <8 x float>* %b
- %shuffle = shufflevector <8 x float> %a2, <8 x float> %b2, <8 x i32> <i32 3, i32 2, i32 8, i32 11, i32 7, i32 6, i32 12, i32 15>
- ret <8 x float> %shuffle
-}
-
-; CHECK: vshufps $-53, %ymm
-define <8 x i32> @A3(<8 x i32> %a, <8 x i32> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 3, i32 2, i32 8, i32 11, i32 7, i32 6, i32 12, i32 15>
- ret <8 x i32> %shuffle
-}
-
-; CHECK: vshufps $-53, (%{{.*}}), %ymm
-define <8 x i32> @A4(<8 x i32>* %a, <8 x i32>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <8 x i32>* %a
- %b2 = load <8 x i32>* %b
- %shuffle = shufflevector <8 x i32> %a2, <8 x i32> %b2, <8 x i32> <i32 3, i32 2, i32 8, i32 11, i32 7, i32 6, i32 12, i32 15>
- ret <8 x i32> %shuffle
-}
-
-; CHECK: vblendpd $10, %ymm
-define <4 x double> @B(<4 x double> %a, <4 x double> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x double> %shuffle
-}
-
-; CHECK: vblendpd $10, (%{{.*}}), %ymm
-define <4 x double> @B2(<4 x double>* %a, <4 x double>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <4 x double>* %a
- %b2 = load <4 x double>* %b
- %shuffle = shufflevector <4 x double> %a2, <4 x double> %b2, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x double> %shuffle
-}
-
-; CHECK: vblendpd $10, %ymm
-define <4 x i64> @B3(<4 x i64> %a, <4 x i64> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: vblendpd $10, (%{{.*}}), %ymm
-define <4 x i64> @B4(<4 x i64>* %a, <4 x i64>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <4 x i64>* %a
- %b2 = load <4 x i64>* %b
- %shuffle = shufflevector <4 x i64> %a2, <4 x i64> %b2, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x i64> %shuffle
-}
-
-; CHECK: vshufps $-53, %ymm
-define <8 x float> @C(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 3, i32 undef, i32 undef, i32 11, i32 undef, i32 6, i32 12, i32 undef>
- ret <8 x float> %shuffle
-}
-
-; CHECK: vblendpd $2, %ymm
-define <4 x double> @D(<4 x double> %a, <4 x double> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 undef>
- ret <4 x double> %shuffle
-}
-
-; CHECK: vshufps $-55, %ymm
-define <8 x float> @E(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 9, i32 10, i32 0, i32 3, i32 13, i32 14, i32 4, i32 7>
- ret <8 x float> %shuffle
-}
-
-; CHECK: vshufpd $8, %ymm
-define <4 x double> @F(<4 x double> %a, <4 x double> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 7>
- ret <4 x double> %shuffle
-}
-
-; CHECK: vshufps $-53, %xmm
-define <4 x float> @A128(<4 x float> %a, <4 x float> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 3, i32 2, i32 4, i32 7>
- ret <4 x float> %shuffle
-}
-
-; CHECK: vshufps $-53, (%{{.*}}), %xmm
-define <4 x float> @A2128(<4 x float>* %a, <4 x float>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <4 x float>* %a
- %b2 = load <4 x float>* %b
- %shuffle = shufflevector <4 x float> %a2, <4 x float> %b2, <4 x i32> <i32 3, i32 2, i32 4, i32 7>
- ret <4 x float> %shuffle
-}
-
-; CHECK: vshufps $-53, %xmm
-define <4 x i32> @A3128(<4 x i32> %a, <4 x i32> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 3, i32 2, i32 4, i32 7>
- ret <4 x i32> %shuffle
-}
-
-; CHECK: vshufps $-53, (%{{.*}}), %xmm
-define <4 x i32> @A4128(<4 x i32>* %a, <4 x i32>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <4 x i32>* %a
- %b2 = load <4 x i32>* %b
- %shuffle = shufflevector <4 x i32> %a2, <4 x i32> %b2, <4 x i32> <i32 3, i32 2, i32 4, i32 7>
- ret <4 x i32> %shuffle
-}
-
-; CHECK: vshufpd $1, %xmm
-define <2 x double> @B128(<2 x double> %a, <2 x double> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 1, i32 2>
- ret <2 x double> %shuffle
-}
-
-; CHECK: vshufpd $1, (%{{.*}}), %xmm
-define <2 x double> @B2128(<2 x double>* %a, <2 x double>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <2 x double>* %a
- %b2 = load <2 x double>* %b
- %shuffle = shufflevector <2 x double> %a2, <2 x double> %b2, <2 x i32> <i32 1, i32 2>
- ret <2 x double> %shuffle
-}
-
-; CHECK: vshufpd $1, %xmm
-define <2 x i64> @B3128(<2 x i64> %a, <2 x i64> %b) nounwind uwtable readnone ssp {
-entry:
- %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 1, i32 2>
- ret <2 x i64> %shuffle
-}
-
-; CHECK: vshufpd $1, (%{{.*}}), %xmm
-define <2 x i64> @B4128(<2 x i64>* %a, <2 x i64>* %b) nounwind uwtable readnone ssp {
-entry:
- %a2 = load <2 x i64>* %a
- %b2 = load <2 x i64>* %b
- %shuffle = shufflevector <2 x i64> %a2, <2 x i64> %b2, <2 x i32> <i32 1, i32 2>
- ret <2 x i64> %shuffle
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s
-
-define <8 x i32> @test1(<8 x i32> %A, <8 x i32> %B) nounwind {
-; CHECK-LABEL: test1:
-; CHECK: vpalignr $4
- %C = shufflevector <8 x i32> %A, <8 x i32> %B, <8 x i32> <i32 1, i32 2, i32 3, i32 8, i32 5, i32 6, i32 7, i32 12>
- ret <8 x i32> %C
-}
-
-define <8 x i32> @test2(<8 x i32> %A, <8 x i32> %B) nounwind {
-; CHECK-LABEL: test2:
-; CHECK: vpalignr $4
- %C = shufflevector <8 x i32> %A, <8 x i32> %B, <8 x i32> <i32 1, i32 2, i32 3, i32 8, i32 5, i32 6, i32 undef, i32 12>
- ret <8 x i32> %C
-}
-
-define <8 x i32> @test3(<8 x i32> %A, <8 x i32> %B) nounwind {
-; CHECK-LABEL: test3:
-; CHECK: vpalignr $4
- %C = shufflevector <8 x i32> %A, <8 x i32> %B, <8 x i32> <i32 1, i32 undef, i32 3, i32 8, i32 5, i32 6, i32 7, i32 12>
- ret <8 x i32> %C
-}
-;
-define <8 x i32> @test4(<8 x i32> %A, <8 x i32> %B) nounwind {
-; CHECK-LABEL: test4:
-; CHECK: vpalignr $8
- %C = shufflevector <8 x i32> %A, <8 x i32> %B, <8 x i32> <i32 10, i32 11, i32 undef, i32 1, i32 14, i32 15, i32 4, i32 5>
- ret <8 x i32> %C
-}
-
-define <16 x i16> @test5(<16 x i16> %A, <16 x i16> %B) nounwind {
-; CHECK-LABEL: test5:
-; CHECK: vpalignr $6
- %C = shufflevector <16 x i16> %A, <16 x i16> %B, <16 x i32> <i32 3, i32 4, i32 undef, i32 6, i32 7, i32 16, i32 17, i32 18, i32 11, i32 12, i32 13, i32 undef, i32 15, i32 24, i32 25, i32 26>
- ret <16 x i16> %C
-}
-
-define <16 x i16> @test6(<16 x i16> %A, <16 x i16> %B) nounwind {
-; CHECK-LABEL: test6:
-; CHECK: vpalignr $6
- %C = shufflevector <16 x i16> %A, <16 x i16> %B, <16 x i32> <i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 11, i32 12, i32 13, i32 undef, i32 15, i32 24, i32 25, i32 26>
- ret <16 x i16> %C
-}
-
-define <16 x i16> @test7(<16 x i16> %A, <16 x i16> %B) nounwind {
-; CHECK-LABEL: test7:
-; CHECK: vpalignr $6
- %C = shufflevector <16 x i16> %A, <16 x i16> %B, <16 x i32> <i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <16 x i16> %C
-}
-
-define <32 x i8> @test8(<32 x i8> %A, <32 x i8> %B) nounwind {
-; CHECK-LABEL: test8:
-; CHECK: vpalignr $5
- %C = shufflevector <32 x i8> %A, <32 x i8> %B, <32 x i32> <i32 5, i32 6, i32 7, i32 undef, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 32, i32 33, i32 34, i32 35, i32 36, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 48, i32 49, i32 50, i32 51, i32 52>
- ret <32 x i8> %C
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s
-
-; Make sure that we don't match this shuffle using the vpblendw YMM instruction.
-; The mask for the vpblendw instruction needs to be identical for both halves
-; of the YMM. Need to use two vpblendw instructions.
-
-; CHECK: vpblendw_test1
-; mask = 10010110,b = 150,d
-; CHECK: vpblendw $150, %ymm
-; CHECK: ret
-define <16 x i16> @vpblendw_test1(<16 x i16> %a, <16 x i16> %b) nounwind alwaysinline {
- %t = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 17, i32 18, i32 3, i32 20, i32 5, i32 6, i32 23,
- i32 8, i32 25, i32 26, i32 11, i32 28, i32 13, i32 14, i32 31>
- ret <16 x i16> %t
-}
-
-; CHECK: vpblendw_test2
-; mask1 = 00010110 = 22
-; mask2 = 10000000 = 128
-; CHECK: vpblendw $128, %xmm
-; CHECK: vpblendw $22, %xmm
-; CHECK: vinserti128
-; CHECK: ret
-define <16 x i16> @vpblendw_test2(<16 x i16> %a, <16 x i16> %b) nounwind alwaysinline {
- %t = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 17, i32 18, i32 3, i32 20, i32 5, i32 6, i32 7,
- i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 31>
- ret <16 x i16> %t
-}
-
-; CHECK: blend_test1
-; CHECK: vpblendd
-; CHECK: ret
-define <8 x i32> @blend_test1(<8 x i32> %a, <8 x i32> %b) nounwind alwaysinline {
- %t = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 9, i32 10, i32 3, i32 12, i32 5, i32 6, i32 7>
- ret <8 x i32> %t
-}
-
-; CHECK: blend_test2
-; CHECK: vpblendd
-; CHECK: ret
-define <8 x i32> @blend_test2(<8 x i32> %a, <8 x i32> %b) nounwind alwaysinline {
- %t = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 9, i32 10, i32 3, i32 12, i32 5, i32 6, i32 7>
- ret <8 x i32> %t
-}
-
-
-; CHECK: blend_test3
-; CHECK: vblendps
-; CHECK: ret
-define <8 x float> @blend_test3(<8 x float> %a, <8 x float> %b) nounwind alwaysinline {
- %t = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 9, i32 10, i32 3, i32 12, i32 5, i32 6, i32 7>
- ret <8 x float> %t
-}
-
-; CHECK: blend_test4
-; CHECK: vblendpd
-; CHECK: ret
-define <4 x i64> @blend_test4(<4 x i64> %a, <4 x i64> %b) nounwind alwaysinline {
- %t = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 5, i32 6, i32 3>
- ret <4 x i64> %t
-}
-
-;; 2 tests for shufflevectors that optimize to blend + immediate
-; CHECK-LABEL: @blend_test5
-; CHECK: vpblendd $10, %xmm1, %xmm0, %xmm0
-; CHECK: ret
-define <4 x i32> @blend_test5(<4 x i32> %a, <4 x i32> %b) {
- %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
- ret <4 x i32> %1
-}
-
-; CHECK-LABEL: @blend_test6
-; CHECK: vpblendw $134, %ymm1, %ymm0, %ymm0
-; CHECK: ret
-define <16 x i16> @blend_test6(<16 x i16> %a, <16 x i16> %b) {
- %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 17, i32 18, i32 3, i32 4, i32 5, i32 6, i32 23,
- i32 8, i32 25, i32 26, i32 11, i32 12, i32 13, i32 14, i32 31>
- ret <16 x i16> %1
-}
-
-; CHECK: vpshufhw $27, %ymm
-define <16 x i16> @vpshufhw(<16 x i16> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 5, i32 4, i32 8, i32 9, i32 10, i32 11, i32 15, i32 14, i32 13, i32 12>
- ret <16 x i16> %shuffle.i
-}
-
-; CHECK: vpshuflw $27, %ymm
-define <16 x i16> @vpshuflw(<16 x i16> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src1, <16 x i32> <i32 3, i32 undef, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7, i32 11, i32 10, i32 9, i32 8, i32 12, i32 13, i32 14, i32 15>
- ret <16 x i16> %shuffle.i
-}
-
-; CHECK: vpshufb_test
-; CHECK: vpshufb {{.*\(%r.*}}, %ymm
-; CHECK: ret
-define <32 x i8> @vpshufb_test(<32 x i8> %a) nounwind {
- %S = shufflevector <32 x i8> %a, <32 x i8> undef, <32 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 18, i32 19, i32 30, i32 16, i32 25, i32 23, i32 17, i32 25,
- i32 20, i32 19, i32 31, i32 17, i32 23, i32 undef, i32 29, i32 18>
- ret <32 x i8>%S
-}
-
-; CHECK: vpshufb1_test
-; CHECK: vpshufb {{.*\(%r.*}}, %ymm
-; CHECK: ret
-define <32 x i8> @vpshufb1_test(<32 x i8> %a) nounwind {
- %S = shufflevector <32 x i8> %a, <32 x i8> zeroinitializer, <32 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 1, i32 9, i32 36, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 18, i32 49, i32 30, i32 16, i32 25, i32 23, i32 17, i32 25,
- i32 20, i32 19, i32 31, i32 17, i32 23, i32 undef, i32 29, i32 18>
- ret <32 x i8>%S
-}
-
-
-; CHECK: vpshufb2_test
-; CHECK: vpshufb {{.*\(%r.*}}, %ymm
-; CHECK: ret
-define <32 x i8> @vpshufb2_test(<32 x i8> %a) nounwind {
- %S = shufflevector <32 x i8> zeroinitializer, <32 x i8> %a, <32 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 1, i32 9, i32 36, i32 11, i32 5, i32 13, i32 7, i32 15,
- i32 18, i32 49, i32 30, i32 16, i32 25, i32 23, i32 17, i32 25,
- i32 20, i32 19, i32 31, i32 17, i32 23, i32 undef, i32 29, i32 18>
- ret <32 x i8>%S
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s
-
-; CHECK: vpunpckhdq
-define <8 x i32> @unpackhidq1(<8 x i32> %src1, <8 x i32> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <8 x i32> %src1, <8 x i32> %src2, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
- ret <8 x i32> %shuffle.i
-}
-
-; CHECK: vpunpckhqdq
-define <4 x i64> @unpackhiqdq1(<4 x i64> %src1, <4 x i64> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <4 x i64> %src1, <4 x i64> %src2, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
- ret <4 x i64> %shuffle.i
-}
-
-; CHECK: vpunpckldq
-define <8 x i32> @unpacklodq1(<8 x i32> %src1, <8 x i32> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <8 x i32> %src1, <8 x i32> %src2, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
- ret <8 x i32> %shuffle.i
-}
-
-; CHECK: vpunpcklqdq
-define <4 x i64> @unpacklqdq1(<4 x i64> %src1, <4 x i64> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <4 x i64> %src1, <4 x i64> %src2, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
- ret <4 x i64> %shuffle.i
-}
-
-; CHECK: vpunpckhwd
-define <16 x i16> @unpackhwd(<16 x i16> %src1, <16 x i16> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src2, <16 x i32> <i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
- ret <16 x i16> %shuffle.i
-}
-
-; CHECK: vpunpcklwd
-define <16 x i16> @unpacklwd(<16 x i16> %src1, <16 x i16> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src2, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27>
- ret <16 x i16> %shuffle.i
-}
-
-; CHECK: vpunpckhbw
-define <32 x i8> @unpackhbw(<32 x i8> %src1, <32 x i8> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <32 x i8> %src1, <32 x i8> %src2, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
- ret <32 x i8> %shuffle.i
-}
-
-; CHECK: vpunpcklbw
-define <32 x i8> @unpacklbw(<32 x i8> %src1, <32 x i8> %src2) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <32 x i8> %src1, <32 x i8> %src2, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
- ret <32 x i8> %shuffle.i
-}
-
-; CHECK: vpunpckhdq
-define <8 x i32> @unpackhidq1_undef(<8 x i32> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <8 x i32> %src1, <8 x i32> %src1, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
- ret <8 x i32> %shuffle.i
-}
-
-; CHECK: vpunpckhqdq
-define <4 x i64> @unpackhiqdq1_undef(<4 x i64> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <4 x i64> %src1, <4 x i64> %src1, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
- ret <4 x i64> %shuffle.i
-}
-
-; CHECK: vpunpckhwd
-define <16 x i16> @unpackhwd_undef(<16 x i16> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src1, <16 x i32> <i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
- ret <16 x i16> %shuffle.i
-}
-
-; CHECK: vpunpcklwd
-define <16 x i16> @unpacklwd_undef(<16 x i16> %src1) nounwind uwtable readnone ssp {
-entry:
- %shuffle.i = shufflevector <16 x i16> %src1, <16 x i16> %src1, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27>
- ret <16 x i16> %shuffle.i
-}
-
}
;CHECK-LABEL: _inreg2xdouble:
-;CHECK: vpbroadcastq
+;CHECK: vunpcklpd
;CHECK: ret
define <2 x double> @_inreg2xdouble(<2 x double> %a) {
%b = shufflevector <2 x double> %a, <2 x double> undef, <2 x i32> zeroinitializer
define <8 x i64> @andqbrst(<8 x i64> %p1, i64* %ap) {
; CHECK-LABEL: andqbrst:
; CHECK: ## BB#0: ## %entry
-; CHECK-NEXT: vpandq (%rdi){1to8}, %zmm0, %zmm0
+; CHECK-NEXT: vmovq (%rdi), %xmm1
+; CHECK-NEXT: vpbroadcastq %xmm1, %ymm1
+; CHECK-NEXT: vinserti64x4 $1, %ymm1, %zmm1, %zmm1
+; CHECK-NEXT: vpandq %zmm1, %zmm0, %zmm0
; CHECK-NEXT: retq
entry:
%a = load i64* %ap, align 8
; CHECK-LABEL: test1:
; CHECK: ## BB#0:
; CHECK-NEXT: vmovd (%rdi), %xmm0
-; CHECK-NEXT: vmovdqa32 {{.*}}(%rip), %zmm1
-; CHECK-NEXT: vpxord %zmm2, %zmm2, %zmm2
-; CHECK-NEXT: vpermt2d %zmm0, %zmm1, %zmm2
-; CHECK-NEXT: vmovdqa32 {{.*}}(%rip), %zmm0
-; CHECK-NEXT: vpermd %zmm2, %zmm0, %zmm0
+; CHECK-NEXT: vpbroadcastq %xmm0, %ymm0
+; CHECK-NEXT: vpxor %ymm1, %ymm1, %ymm1
+; CHECK-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4],ymm1[5,6,7]
+; CHECK-NEXT: vinserti64x4 $1, %ymm1, %zmm0, %zmm0
; CHECK-NEXT: retq
%y = load i32* %x, align 4
%res = insertelement <16 x i32>zeroinitializer, i32 %y, i32 4
define <16 x float> @test3(<4 x float> %a) {
; CHECK-LABEL: test3:
; CHECK: ## BB#0:
-; CHECK-NEXT: vmovhlps %xmm0, %xmm0, %xmm1
+; CHECK-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
; CHECK-NEXT: vxorps %xmm2, %xmm2, %xmm2
; CHECK-NEXT: vmovss %xmm0, %xmm2, %xmm0
; CHECK-NEXT: vmovss %xmm1, %xmm2, %xmm1
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
-; CHECK: LCP
-; CHECK: .long 2
-; CHECK: .long 5
-; CHECK: .long 0
-; CHECK: .long 0
-; CHECK: .long 7
-; CHECK: .long 0
-; CHECK: .long 10
-; CHECK: .long 1
-; CHECK: .long 0
-; CHECK: .long 5
-; CHECK: .long 0
-; CHECK: .long 4
-; CHECK: .long 7
-; CHECK: .long 0
-; CHECK: .long 10
-; CHECK: .long 1
-; CHECK-LABEL: test1:
-; CHECK: vpermps
-; CHECK: ret
-define <16 x float> @test1(<16 x float> %a) nounwind {
- %c = shufflevector <16 x float> %a, <16 x float> undef, <16 x i32> <i32 2, i32 5, i32 undef, i32 undef, i32 7, i32 undef, i32 10, i32 1, i32 0, i32 5, i32 undef, i32 4, i32 7, i32 undef, i32 10, i32 1>
- ret <16 x float> %c
-}
-
-; CHECK-LABEL: test2:
-; CHECK: vpermd
-; CHECK: ret
-define <16 x i32> @test2(<16 x i32> %a) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> undef, <16 x i32> <i32 2, i32 5, i32 undef, i32 undef, i32 7, i32 undef, i32 10, i32 1, i32 0, i32 5, i32 undef, i32 4, i32 7, i32 undef, i32 10, i32 1>
- ret <16 x i32> %c
-}
-
-; CHECK-LABEL: test3:
-; CHECK: vpermq
-; CHECK: ret
-define <8 x i64> @test3(<8 x i64> %a) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> undef, <8 x i32> <i32 2, i32 5, i32 1, i32 undef, i32 7, i32 undef, i32 3, i32 1>
- ret <8 x i64> %c
-}
-
-; CHECK-LABEL: test4:
-; CHECK: vpermpd
-; CHECK: ret
-define <8 x double> @test4(<8 x double> %a) nounwind {
- %c = shufflevector <8 x double> %a, <8 x double> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x double> %c
-}
-
-; CHECK-LABEL: test5:
-; CHECK: vpermt2pd
-; CHECK: ret
-define <8 x double> @test5(<8 x double> %a, <8 x double> %b) nounwind {
- %c = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 2, i32 8, i32 0, i32 1, i32 6, i32 10, i32 4, i32 5>
- ret <8 x double> %c
-}
-
-; The reg variant of vpermt2 with a writemask
-; CHECK-LABEL: test5m:
-; CHECK: vpermt2pd {{.* {%k[1-7]} {z}}}
-define <8 x double> @test5m(<8 x double> %a, <8 x double> %b, i8 %mask) nounwind {
- %c = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 2, i32 8, i32 0, i32 1, i32 6, i32 10, i32 4, i32 5>
- %m = bitcast i8 %mask to <8 x i1>
- %res = select <8 x i1> %m, <8 x double> %c, <8 x double> zeroinitializer
- ret <8 x double> %res
-}
-
-; CHECK-LABEL: test6:
-; CHECK: vpermq $30
-; CHECK: ret
-define <8 x i64> @test6(<8 x i64> %a) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> undef, <8 x i32> <i32 2, i32 3, i32 1, i32 0, i32 6, i32 7, i32 5, i32 4>
- ret <8 x i64> %c
-}
-
-; CHECK-LABEL: test7:
-; CHECK: vpermt2q
-; CHECK: ret
-define <8 x i64> @test7(<8 x i64> %a, <8 x i64> %b) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 2, i32 8, i32 0, i32 1, i32 6, i32 10, i32 4, i32 5>
- ret <8 x i64> %c
-}
-
-; The reg variant of vpermt2 with a writemask
-; CHECK-LABEL: test7m:
-; CHECK: vpermt2q {{.* {%k[1-7]} {z}}}
-define <8 x i64> @test7m(<8 x i64> %a, <8 x i64> %b, i8 %mask) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 2, i32 8, i32 0, i32 1, i32 6, i32 10, i32 4, i32 5>
- %m = bitcast i8 %mask to <8 x i1>
- %res = select <8 x i1> %m, <8 x i64> %c, <8 x i64> zeroinitializer
- ret <8 x i64> %res
-}
-
-; The mem variant of vpermt2 with a writemask
-; CHECK-LABEL: test7mm:
-; CHECK: vpermt2q {{\(.*\).* {%k[1-7]} {z}}}
-define <8 x i64> @test7mm(<8 x i64> %a, <8 x i64> *%pb, i8 %mask) nounwind {
- %b = load <8 x i64>* %pb
- %c = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 2, i32 8, i32 0, i32 1, i32 6, i32 10, i32 4, i32 5>
- %m = bitcast i8 %mask to <8 x i1>
- %res = select <8 x i1> %m, <8 x i64> %c, <8 x i64> zeroinitializer
- ret <8 x i64> %res
-}
-
-; CHECK-LABEL: test8:
-; CHECK: vpermt2d
-; CHECK: ret
-define <16 x i32> @test8(<16 x i32> %a, <16 x i32> %b) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- ret <16 x i32> %c
-}
-
-; The reg variant of vpermt2 with a writemask
-; CHECK-LABEL: test8m:
-; CHECK: vpermt2d {{.* {%k[1-7]} {z}}}
-define <16 x i32> @test8m(<16 x i32> %a, <16 x i32> %b, i16 %mask) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- %m = bitcast i16 %mask to <16 x i1>
- %res = select <16 x i1> %m, <16 x i32> %c, <16 x i32> zeroinitializer
- ret <16 x i32> %res
-}
-
-; The mem variant of vpermt2 with a writemask
-; CHECK-LABEL: test8mm:
-; CHECK: vpermt2d {{\(.*\).* {%k[1-7]} {z}}}
-define <16 x i32> @test8mm(<16 x i32> %a, <16 x i32> *%pb, i16 %mask) nounwind {
- %b = load <16 x i32> * %pb
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- %m = bitcast i16 %mask to <16 x i1>
- %res = select <16 x i1> %m, <16 x i32> %c, <16 x i32> zeroinitializer
- ret <16 x i32> %res
-}
-
-; CHECK-LABEL: test9:
-; CHECK: vpermt2ps
-; CHECK: ret
-define <16 x float> @test9(<16 x float> %a, <16 x float> %b) nounwind {
- %c = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- ret <16 x float> %c
-}
-
-; The reg variant of vpermt2 with a writemask
-; CHECK-LABEL: test9m:
-; CHECK: vpermt2ps {{.*}} {%k{{.}}} {z}
-define <16 x float> @test9m(<16 x float> %a, <16 x float> %b, i16 %mask) nounwind {
- %c = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- %m = bitcast i16 %mask to <16 x i1>
- %res = select <16 x i1> %m, <16 x float> %c, <16 x float> zeroinitializer
- ret <16 x float> %res
-}
-
-; CHECK-LABEL: test10:
-; CHECK: vpermt2ps (
-; CHECK: ret
-define <16 x float> @test10(<16 x float> %a, <16 x float>* %b) nounwind {
- %c = load <16 x float>* %b
- %d = shufflevector <16 x float> %a, <16 x float> %c, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- ret <16 x float> %d
-}
-
-; CHECK-LABEL: test11:
-; CHECK: vpermt2d
-; CHECK: ret
-define <16 x i32> @test11(<16 x i32> %a, <16 x i32>* %b) nounwind {
- %c = load <16 x i32>* %b
- %d = shufflevector <16 x i32> %a, <16 x i32> %c, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24>
- ret <16 x i32> %d
-}
-
-; CHECK-LABEL: test12
-; CHECK: vmovlhps {{.*}}
-; CHECK: ret
-define <4 x i32> @test12(<4 x i32> %a, <4 x i32> %b) nounwind {
- %c = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
- ret <4 x i32> %c
-}
-
-; CHECK-LABEL: test13
-; CHECK: vpermilps $-79, %zmm
-; CHECK: ret
-define <16 x float> @test13(<16 x float> %a) {
- %b = shufflevector <16 x float> %a, <16 x float> undef, <16 x i32><i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
- ret <16 x float> %b
-}
-
-; CHECK-LABEL: test14
-; CHECK: vpermilpd $-53, %zmm
-; CHECK: ret
-define <8 x double> @test14(<8 x double> %a) {
- %b = shufflevector <8 x double> %a, <8 x double> undef, <8 x i32><i32 1, i32 1, i32 2, i32 3, i32 4, i32 4, i32 7, i32 7>
- ret <8 x double> %b
-}
-
-; CHECK-LABEL: test15
-; CHECK: vpshufd $-79, %zmm
-; CHECK: ret
-define <16 x i32> @test15(<16 x i32> %a) {
- %b = shufflevector <16 x i32> %a, <16 x i32> undef, <16 x i32><i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
- ret <16 x i32> %b
-}
-
-; CHECK-LABEL: valign_test_v16f32
-; CHECK: valignd $2, %zmm0, %zmm0
-; CHECK: ret
-define <16 x float> @valign_test_v16f32(<16 x float> %a, <16 x float> %b) nounwind {
- %c = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 2, i32 3, i32 undef, i32 undef, i32 6, i32 7, i32 undef, i32 undef, i32 10, i32 11, i32 undef, i32 undef, i32 14, i32 15, i32 undef, i32 undef>
- ret <16 x float> %c
-}
-
-; CHECK-LABEL: valign_test_v16i32
-; CHECK: valignd $2, %zmm0, %zmm0
-; CHECK: ret
-define <16 x i32> @valign_test_v16i32(<16 x i32> %a, <16 x i32> %b) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 2, i32 3, i32 undef, i32 undef, i32 6, i32 7, i32 undef, i32 undef, i32 10, i32 11, i32 undef, i32 undef, i32 14, i32 15, i32 undef, i32 undef>
- ret <16 x i32> %c
-}
-
-
-; CHECK-LABEL: test16
-; CHECK: valignq $2, %zmm0, %zmm1
-; CHECK: ret
-define <8 x double> @test16(<8 x double> %a, <8 x double> %b) nounwind {
- %c = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
- ret <8 x double> %c
-}
-
-; CHECK-LABEL: test16k
-; CHECK: valignq $2, %zmm0, %zmm1, %zmm2 {%k1}
-define <8 x i64> @test16k(<8 x i64> %a, <8 x i64> %b, <8 x i64> %src, i8 %mask) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
- %m = bitcast i8 %mask to <8 x i1>
- %res = select <8 x i1> %m, <8 x i64> %c, <8 x i64> %src
- ret <8 x i64> %res
-}
-
-; CHECK-LABEL: test16kz
-; CHECK: valignq $2, %zmm0, %zmm1, %zmm0 {%k1} {z}
-define <8 x i64> @test16kz(<8 x i64> %a, <8 x i64> %b, i8 %mask) nounwind {
- %c = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
- %m = bitcast i8 %mask to <8 x i1>
- %res = select <8 x i1> %m, <8 x i64> %c, <8 x i64> zeroinitializer
- ret <8 x i64> %res
-}
-
-; CHECK-LABEL: test17
-; CHECK: vshufpd $19, %zmm1, %zmm0
-; CHECK: ret
-define <8 x double> @test17(<8 x double> %a, <8 x double> %b) nounwind {
- %c = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 1, i32 9, i32 2, i32 10, i32 5, i32 undef, i32 undef, i32 undef>
- ret <8 x double> %c
-}
-
-; CHECK-LABEL: test18
-; CHECK: vpunpckhdq %zmm
-; CHECK: ret
-define <16 x i32> @test18(<16 x i32> %a, <16 x i32> %c) {
- %b = shufflevector <16 x i32> %a, <16 x i32> %c, <16 x i32><i32 2, i32 18, i32 3, i32 19,
- i32 6, i32 22, i32 7, i32 23,
- i32 10, i32 26, i32 11, i32 27,
- i32 14, i32 30, i32 15, i32 31>
- ret <16 x i32> %b
-}
-
-; CHECK-LABEL: test19
-; CHECK: vpunpckldq %zmm
-; CHECK: ret
-define <16 x i32> @test19(<16 x i32> %a, <16 x i32> %c) {
- %b = shufflevector <16 x i32> %a, <16 x i32> %c, <16 x i32><i32 0, i32 16, i32 1, i32 17,
- i32 4, i32 20, i32 5, i32 21,
- i32 8, i32 24, i32 9, i32 25,
- i32 12, i32 28, i32 13, i32 29>
- ret <16 x i32> %b
-}
-
-; CHECK-LABEL: test20
-; CHECK: vpunpckhqdq %zmm
-; CHECK: ret
-define <8 x i64> @test20(<8 x i64> %a, <8 x i64> %c) {
- %b = shufflevector <8 x i64> %a, <8 x i64> %c, <8 x i32><i32 1, i32 9,
- i32 3, i32 11,
- i32 5, i32 13,
- i32 7, i32 15>
- ret <8 x i64> %b
-}
-
-; CHECK-LABEL: test21
-; CHECK: vunpcklps %zmm
-; CHECK: ret
-define <16 x float> @test21(<16 x float> %a, <16 x float> %c) {
- %b = shufflevector <16 x float> %a, <16 x float> %c, <16 x i32><i32 0, i32 16, i32 1, i32 17,
- i32 4, i32 20, i32 5, i32 21,
- i32 8, i32 24, i32 9, i32 25,
- i32 12, i32 28, i32 13, i32 29>
- ret <16 x float> %b
-}
-
-; CHECK-LABEL: test22
-; CHECK: vmovhlps {{.*}}
-; CHECK: ret
-define <4 x i32> @test22(<4 x i32> %a, <4 x i32> %b) nounwind {
- %c = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 2, i32 3, i32 6, i32 7>
- ret <4 x i32> %c
-}
-
-; CHECK-LABEL: @test23
-; CHECK: vshufps $-112, %zmm
-; CHECK: ret
-define <16 x float> @test23(<16 x float> %a, <16 x float> %c) {
- %b = shufflevector <16 x float> %a, <16 x float> %c, <16 x i32><i32 0, i32 0, i32 17, i32 18, i32 4, i32 4, i32 21, i32 22, i32 8, i32 8, i32 25, i32 26, i32 12, i32 12, i32 29, i32 30>
- ret <16 x float> %b
-}
-
-; CHECK-LABEL: @test24
-; CHECK: vpermt2d
-; CHECK: ret
-define <16 x i32> @test24(<16 x i32> %a, <16 x i32> %b) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 19, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <16 x i32> %c
-}
-
-; CHECK-LABEL: @test25
-; CHECK: vshufps $52
-; CHECK: ret
-define <16 x i32> @test25(<16 x i32> %a, <16 x i32> %b) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 19, i32 undef, i32 4, i32 5, i32 23, i32 undef, i32 8, i32 9, i32 27, i32 undef, i32 12, i32 13, i32 undef, i32 undef>
- ret <16 x i32> %c
-}
-
-; CHECK-LABEL: @test26
-; CHECK: vmovshdup
-; CHECK: ret
-define <16 x i32> @test26(<16 x i32> %a) nounwind {
- %c = shufflevector <16 x i32> %a, <16 x i32> undef, <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 undef, i32 9, i32 9, i32 undef, i32 11, i32 13, i32 undef, i32 undef, i32 undef>
- ret <16 x i32> %c
-}
-
-; CHECK-LABEL: @test27
-; CHECK: ret
-define <16 x i32> @test27(<4 x i32>%a) {
- %res = shufflevector <4 x i32> %a, <4 x i32> undef, <16 x i32> <i32 0, i32 1, 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>
- ret <16 x i32> %res
-}
-
-; CHECK-LABEL: @test28
-; CHECK: vinserti64x4 $1
-; CHECK: ret
-define <16 x i32> @test28(<16 x i32>%x, <16 x i32>%y) {
- %res = shufflevector <16 x i32>%x, <16 x i32>%y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
- i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
- ret <16 x i32> %res
-}
-
-; CHECK-LABEL: @test29
-; CHECK: vinserti64x4 $0
-; CHECK: ret
-define <16 x i32> @test29(<16 x i32>%x, <16 x i32>%y) {
- %res = shufflevector <16 x i32>%x, <16 x i32>%y, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
- i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
- ret <16 x i32> %res
-}
-
define <16 x i32> @_inreg16xi32(i32 %a) {
; CHECK-LABEL: _inreg16xi32:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpbroadcastd %edi, %zmm0
+; CHECK-NEXT: vmovd %edi, %xmm0
+; CHECK-NEXT: vpbroadcastd %xmm0, %ymm0
+; CHECK-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = insertelement <16 x i32> undef, i32 %a, i32 0
%c = shufflevector <16 x i32> %b, <16 x i32> undef, <16 x i32> zeroinitializer
define <8 x i64> @_inreg8xi64(i64 %a) {
; CHECK-LABEL: _inreg8xi64:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpbroadcastq %rdi, %zmm0
+; CHECK-NEXT: vmovq %rdi, %xmm0
+; CHECK-NEXT: vpbroadcastq %xmm0, %ymm0
+; CHECK-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = insertelement <8 x i64> undef, i64 %a, i32 0
%c = shufflevector <8 x i64> %b, <8 x i64> undef, <8 x i32> zeroinitializer
define <16 x float> @_inreg16xfloat(float %a) {
; CHECK-LABEL: _inreg16xfloat:
; CHECK: ## BB#0:
-; CHECK-NEXT: vbroadcastss %xmm0, %zmm0
+; CHECK-NEXT: ## kill: XMM0<def> XMM0<kill> ZMM0<def>
+; CHECK-NEXT: vbroadcastss %xmm0, %ymm0
+; CHECK-NEXT: vinsertf64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = insertelement <16 x float> undef, float %a, i32 0
%c = shufflevector <16 x float> %b, <16 x float> undef, <16 x i32> zeroinitializer
define <8 x double> @_inreg8xdouble(double %a) {
; CHECK-LABEL: _inreg8xdouble:
; CHECK: ## BB#0:
-; CHECK-NEXT: vbroadcastsd %xmm0, %zmm0
+; CHECK-NEXT: ## kill: XMM0<def> XMM0<kill> ZMM0<def>
+; CHECK-NEXT: vbroadcastsd %xmm0, %ymm0
+; CHECK-NEXT: vinsertf64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = insertelement <8 x double> undef, double %a, i32 0
%c = shufflevector <8 x double> %b, <8 x double> undef, <8 x i32> zeroinitializer
define <16 x i32> @_xmm16xi32(<16 x i32> %a) {
; CHECK-LABEL: _xmm16xi32:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpbroadcastd %xmm0, %zmm0
+; CHECK-NEXT: vpbroadcastd %xmm0, %ymm0
+; CHECK-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = shufflevector <16 x i32> %a, <16 x i32> undef, <16 x i32> zeroinitializer
ret <16 x i32> %b
define <16 x float> @_xmm16xfloat(<16 x float> %a) {
; CHECK-LABEL: _xmm16xfloat:
; CHECK: ## BB#0:
-; CHECK-NEXT: vbroadcastss %xmm0, %zmm0
+; CHECK-NEXT: vbroadcastss %xmm0, %ymm0
+; CHECK-NEXT: vinsertf64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = shufflevector <16 x float> %a, <16 x float> undef, <16 x i32> zeroinitializer
ret <16 x float> %b
define <8 x i64> @test24(<8 x i64> %x, <8 x i64> %x1, i64* %yb.ptr) nounwind {
; CHECK-LABEL: test24:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpcmpeqq (%rdi){1to8}, %zmm0, %k1
+; CHECK-NEXT: vmovq (%rdi), %xmm2
+; CHECK-NEXT: vpbroadcastq %xmm2, %ymm2
+; CHECK-NEXT: vinserti64x4 $1, %ymm2, %zmm2, %zmm2
+; CHECK-NEXT: vpcmpeqq %zmm2, %zmm0, %k1
; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1}
; CHECK-NEXT: vmovaps %zmm1, %zmm0
; CHECK-NEXT: retq
define <16 x i32> @test25(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1) nounwind {
; CHECK-LABEL: test25:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpcmpled (%rdi){1to16}, %zmm0, %k1
+; CHECK-NEXT: vmovd (%rdi), %xmm2
+; CHECK-NEXT: vpbroadcastd %xmm2, %ymm2
+; CHECK-NEXT: vinserti64x4 $1, %ymm2, %zmm2, %zmm2
+; CHECK-NEXT: vpcmpled %zmm2, %zmm0, %k1
; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1}
; CHECK-NEXT: vmovaps %zmm1, %zmm0
; CHECK-NEXT: retq
define <16 x i32> @test26(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1, <16 x i32> %y1) nounwind {
; CHECK-LABEL: test26:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpcmpled %zmm1, %zmm2, %k1
-; CHECK-NEXT: vpcmpgtd (%rdi){1to16}, %zmm0, %k1 {%k1}
+; CHECK-NEXT: vmovd (%rdi), %xmm3
+; CHECK-NEXT: vpbroadcastd %xmm3, %ymm3
+; CHECK-NEXT: vinserti64x4 $1, %ymm3, %zmm3, %zmm3
+; CHECK-NEXT: vpcmpgtd %zmm3, %zmm0, %k1
+; CHECK-NEXT: vpcmpled %zmm1, %zmm2, %k1 {%k1}
; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1}
; CHECK-NEXT: vmovaps %zmm1, %zmm0
; CHECK-NEXT: retq
define <8 x i64> @test27(<8 x i64> %x, i64* %yb.ptr, <8 x i64> %x1, <8 x i64> %y1) nounwind {
; CHECK-LABEL: test27:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpcmpleq %zmm1, %zmm2, %k1
-; CHECK-NEXT: vpcmpleq (%rdi){1to8}, %zmm0, %k1 {%k1}
+; CHECK-NEXT: vmovq (%rdi), %xmm3
+; CHECK-NEXT: vpbroadcastq %xmm3, %ymm3
+; CHECK-NEXT: vinserti64x4 $1, %ymm3, %zmm3, %zmm3
+; CHECK-NEXT: vpcmpleq %zmm3, %zmm0, %k1
+; CHECK-NEXT: vpcmpleq %zmm1, %zmm2, %k1 {%k1}
; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1}
; CHECK-NEXT: vmovaps %zmm1, %zmm0
; CHECK-NEXT: retq
define <2 x i64> @test1(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test1:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: retq
%shuf1 = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <2 x i32><i32 0, i32 2>
%shuf2 = shufflevector <2 x i64> %b, <2 x i64> zeroinitializer, <2 x i32><i32 2, i32 1>
define <4 x i32> @test2(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test2:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm0[4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 4, i32 2, i32 3>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 1, i32 4, i32 4>
define <2 x i64> @test3(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test3:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm0[4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <2 x i32><i32 2, i32 1>
%shuf2 = shufflevector <2 x i64> %b, <2 x i64> zeroinitializer, <2 x i32><i32 0, i32 2>
define <4 x i32> @test4(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test4:
; CHECK: # BB#0:
-; CHECK-NEXT: movss %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 4, i32 4, i32 4>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 1, i32 2, i32 3>
define <4 x i32> @test5(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test5:
; CHECK: # BB#0:
-; CHECK-NEXT: movss %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7]
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 1, i32 2, i32 3>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 4, i32 4, i32 4>
define <4 x i32> @test6(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test6:
; CHECK: # BB#0:
-; CHECK-NEXT: blendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 1, i32 4, i32 4>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 4, i32 2, i32 3>
define <4 x i32> @test7(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test7:
; CHECK: # BB#0:
-; CHECK-NEXT: blendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: retq
%and1 = and <4 x i32> %a, <i32 -1, i32 -1, i32 0, i32 0>
%and2 = and <4 x i32> %b, <i32 0, i32 0, i32 -1, i32 -1>
define <2 x i64> @test8(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test8:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: retq
%and1 = and <2 x i64> %a, <i64 -1, i64 0>
%and2 = and <2 x i64> %b, <i64 0, i64 -1>
define <4 x i32> @test9(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test9:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm0[4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%and1 = and <4 x i32> %a, <i32 0, i32 0, i32 -1, i32 -1>
%and2 = and <4 x i32> %b, <i32 -1, i32 -1, i32 0, i32 0>
define <2 x i64> @test10(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test10:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm0[4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%and1 = and <2 x i64> %a, <i64 0, i64 -1>
%and2 = and <2 x i64> %b, <i64 -1, i64 0>
define <4 x i32> @test11(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test11:
; CHECK: # BB#0:
-; CHECK-NEXT: movss %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%and1 = and <4 x i32> %a, <i32 -1, i32 0, i32 0, i32 0>
%and2 = and <4 x i32> %b, <i32 0, i32 -1, i32 -1, i32 -1>
define <4 x i32> @test12(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test12:
; CHECK: # BB#0:
-; CHECK-NEXT: movss %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7]
; CHECK-NEXT: retq
%and1 = and <4 x i32> %a, <i32 0, i32 -1, i32 -1, i32 -1>
%and2 = and <4 x i32> %b, <i32 -1, i32 0, i32 0, i32 0>
; CHECK-LABEL: test17:
; CHECK: # BB#0:
; CHECK-NEXT: xorps %xmm2, %xmm2
-; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,0]
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
; CHECK-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm2[0,0]
-; CHECK-NEXT: por %xmm1, %xmm0
+; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,0],xmm0[0,2]
+; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; CHECK-NEXT: orps %xmm1, %xmm2
+; CHECK-NEXT: movaps %xmm2, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 0, i32 4, i32 2>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 1, i32 4, i32 4>
; CHECK: # BB#0:
; CHECK-NEXT: xorps %xmm2, %xmm2
; CHECK-NEXT: xorps %xmm3, %xmm3
-; CHECK-NEXT: blendps $1, %xmm0, %xmm3
-; CHECK-NEXT: shufps {{.*#+}} xmm3 = xmm3[2,0],xmm2[0,0]
-; CHECK-NEXT: blendps $1, %xmm1, %xmm2
-; CHECK-NEXT: orps %xmm3, %xmm2
+; CHECK-NEXT: blendps {{.*#+}} xmm3 = xmm0[0],xmm3[1,2,3]
+; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm3[1,0,1,1]
+; CHECK-NEXT: blendps {{.*#+}} xmm2 = xmm1[0],xmm2[1,2,3]
+; CHECK-NEXT: orps %xmm0, %xmm2
; CHECK-NEXT: movaps %xmm2, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 0, i32 4, i32 4>
; CHECK-LABEL: test19:
; CHECK: # BB#0:
; CHECK-NEXT: xorps %xmm2, %xmm2
-; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3],xmm2[0,0]
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; CHECK-NEXT: movdqa %xmm1, %xmm2
-; CHECK-NEXT: pslldq $8, %xmm2
+; CHECK-NEXT: xorps %xmm3, %xmm3
+; CHECK-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,0],xmm0[0,3]
+; CHECK-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[0,0]
; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm1[2,2]
-; CHECK-NEXT: por %xmm2, %xmm0
+; CHECK-NEXT: orps %xmm3, %xmm2
+; CHECK-NEXT: movaps %xmm2, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32><i32 4, i32 0, i32 4, i32 3>
%shuf2 = shufflevector <4 x i32> %b, <4 x i32> zeroinitializer, <4 x i32><i32 0, i32 4, i32 2, i32 2>
define <2 x i64> @test21(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test21:
; CHECK: # BB#0:
-; CHECK-NEXT: por %xmm1, %xmm0
-; CHECK-NEXT: pslldq $8, %xmm0
+; CHECK-NEXT: orps %xmm1, %xmm0
+; CHECK-NEXT: movq %xmm0, %xmm0
+; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; CHECK-NEXT: retq
%shuf1 = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <2 x i32><i32 2, i32 0>
%shuf2 = shufflevector <2 x i64> %b, <2 x i64> zeroinitializer, <2 x i32><i32 2, i32 0>
define <4 x i8> @test_crash(<4 x i8> %a, <4 x i8> %b) {
; CHECK-LABEL: test_crash:
; CHECK: # BB#0:
-; CHECK-NEXT: movsd %xmm1, %xmm0
+; CHECK-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm0[4,5,6,7]
+; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retq
%shuf1 = shufflevector <4 x i8> %a, <4 x i8> zeroinitializer, <4 x i32><i32 4, i32 4, i32 2, i32 3>
%shuf2 = shufflevector <4 x i8> %b, <4 x i8> zeroinitializer, <4 x i32><i32 0, i32 1, i32 4, i32 4>
; CHECK-LABEL: ExeDepsFix_broadcastsd_inreg
; ExeDepsFix works top down, thus it coalesces vpunpcklqdq domain with
; vpand and there is nothing more you can do to match vmaxpd.
-; CHECK: vmovlhps
-; CHECK: vandps
+; CHECK: vmovq
+; CHECK: vpbroadcastq
+; CHECK: vpand
; CHECK: vmaxpd
; CHECK: ret
define <2 x double> @ExeDepsFix_broadcastsd_inreg(<2 x double> %arg, <2 x double> %arg2, i64 %broadcastvalue) {
;
; This movs the entire vector, shuffling the high double down. If we fixed the
; FIXME above it would just move the high double directly.
-; CHECK: movups
-; CHECK: movhlps
-; CHECK: movlps
+; CHECK: movupd
+; CHECK: shufpd
+; CHECK: movlpd
bb:
%tmp13 = load <2 x double>* undef, align 1
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: cvtpd2ps 16(%eax), %xmm1
; CHECK-NEXT: cvtpd2ps (%eax), %xmm0
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: retl
;
; AVX-LABEL: test3:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: cvtpd2ps 16(%eax), %xmm1
; CHECK-NEXT: cvtpd2ps (%eax), %xmm0
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: cvtpd2ps 48(%eax), %xmm2
; CHECK-NEXT: cvtpd2ps 32(%eax), %xmm1
-; CHECK-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0],xmm2[0]
; CHECK-NEXT: retl
;
; AVX-LABEL: test4:
; CHECK: # BB#0:
; CHECK-NEXT: cvtpd2ps %xmm1, %xmm1
; CHECK-NEXT: cvtpd2ps %xmm0, %xmm0
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: retl
;
; AVX-LABEL: test3:
; CHECK: # BB#0:
; CHECK-NEXT: cvtpd2ps %xmm1, %xmm1
; CHECK-NEXT: cvtpd2ps %xmm0, %xmm0
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: cvtpd2ps %xmm3, %xmm3
; CHECK-NEXT: cvtpd2ps %xmm2, %xmm1
-; CHECK-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm3[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0],xmm3[0]
; CHECK-NEXT: retl
;
; AVX-LABEL: test4:
;
; CHECK-YONAH-LABEL: test3:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,2],xmm1[0,0]
+; CHECK-YONAH-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,2],xmm1[2,0]
; CHECK-YONAH-NEXT: retl
%C = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> < i32 1, i32 2, i32 undef, i32 4 >
ret <4 x i32> %C
;
; CHECK-YONAH-LABEL: test4:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,3],xmm0[0,1]
-; CHECK-YONAH-NEXT: movaps %xmm1, %xmm0
+; CHECK-YONAH-NEXT: shufpd {{.*#+}} xmm1 = xmm1[1],xmm0[0]
+; CHECK-YONAH-NEXT: movapd %xmm1, %xmm0
; CHECK-YONAH-NEXT: retl
%C = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> < i32 6, i32 7, i32 undef, i32 1 >
ret <4 x i32> %C
define <4 x float> @test5(<4 x float> %A, <4 x float> %B) nounwind {
; CHECK-LABEL: test5:
; CHECK: # BB#0:
-; CHECK-NEXT: palignr {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7]
+; CHECK-NEXT: shufpd {{.*#+}} xmm1 = xmm1[1],xmm0[0]
+; CHECK-NEXT: movapd %xmm1, %xmm0
; CHECK-NEXT: retl
;
; CHECK-YONAH-LABEL: test5:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,3],xmm0[0,1]
-; CHECK-YONAH-NEXT: movaps %xmm1, %xmm0
+; CHECK-YONAH-NEXT: shufpd {{.*#+}} xmm1 = xmm1[1],xmm0[0]
+; CHECK-YONAH-NEXT: movapd %xmm1, %xmm0
; CHECK-YONAH-NEXT: retl
%C = shufflevector <4 x float> %A, <4 x float> %B, <4 x i32> < i32 6, i32 7, i32 undef, i32 1 >
ret <4 x float> %C
;
; CHECK-YONAH-LABEL: test6:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: movapd %xmm0, %xmm2
-; CHECK-YONAH-NEXT: shufpd {{.*#+}} xmm2 = xmm2[1],xmm1[0]
-; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm1 = xmm2[0,0,0,2,4,5,6,7]
-; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,5,6]
-; CHECK-YONAH-NEXT: pextrw $3, %xmm0, %eax
-; CHECK-YONAH-NEXT: pinsrw $0, %eax, %xmm1
-; CHECK-YONAH-NEXT: pextrw $7, %xmm0, %eax
-; CHECK-YONAH-NEXT: pinsrw $4, %eax, %xmm1
-; CHECK-YONAH-NEXT: movdqa %xmm1, %xmm0
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,0,1]
+; CHECK-YONAH-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,7,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,0,1,2,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,0,2,1,4,5,6,7]
+; CHECK-YONAH-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-YONAH-NEXT: retl
%C = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 3, i32 4, i32 undef, i32 6, i32 7, i32 8, i32 9, i32 10 >
ret <8 x i16> %C
;
; CHECK-YONAH-LABEL: test7:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0]
-; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,0,0,4,5,6,7]
-; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
-; CHECK-YONAH-NEXT: movd %xmm1, %eax
-; CHECK-YONAH-NEXT: pinsrw $3, %eax, %xmm0
-; CHECK-YONAH-NEXT: pextrw $4, %xmm1, %eax
-; CHECK-YONAH-NEXT: pinsrw $7, %eax, %xmm0
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
+; CHECK-YONAH-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,1,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[3,1,2,0]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,7,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,1,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[1,2,3,0,4,5,6,7]
+; CHECK-YONAH-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-YONAH-NEXT: retl
%C = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 undef, i32 6, i32 undef, i32 8, i32 9, i32 10, i32 11, i32 12 >
ret <8 x i16> %C
;
; CHECK-YONAH-LABEL: test8:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: pushl %esi
+; CHECK-YONAH-NEXT: pxor %xmm3, %xmm3
; CHECK-YONAH-NEXT: movdqa %xmm0, %xmm2
-; CHECK-YONAH-NEXT: pextrw $4, %xmm2, %eax
-; CHECK-YONAH-NEXT: pextrw $5, %xmm2, %ecx
-; CHECK-YONAH-NEXT: shrdw $8, %cx, %ax
-; CHECK-YONAH-NEXT: pextrw $2, %xmm2, %edx
-; CHECK-YONAH-NEXT: pextrw $3, %xmm2, %esi
-; CHECK-YONAH-NEXT: shrdw $8, %si, %dx
-; CHECK-YONAH-NEXT: # kill: XMM0<def> XMM2<kill>
-; CHECK-YONAH-NEXT: pinsrw $0, %edx, %xmm0
-; CHECK-YONAH-NEXT: shrl $8, %esi
-; CHECK-YONAH-NEXT: pinsrw $1, %esi, %xmm0
-; CHECK-YONAH-NEXT: pinsrw $2, %eax, %xmm0
-; CHECK-YONAH-NEXT: pextrw $6, %xmm2, %eax
-; CHECK-YONAH-NEXT: shrdw $8, %ax, %cx
-; CHECK-YONAH-NEXT: pinsrw $3, %ecx, %xmm0
-; CHECK-YONAH-NEXT: pextrw $7, %xmm2, %ecx
-; CHECK-YONAH-NEXT: shrdw $8, %cx, %ax
-; CHECK-YONAH-NEXT: pinsrw $4, %eax, %xmm0
-; CHECK-YONAH-NEXT: pextrw $8, %xmm1, %eax
-; CHECK-YONAH-NEXT: shrdw $8, %ax, %cx
-; CHECK-YONAH-NEXT: pinsrw $5, %ecx, %xmm0
-; CHECK-YONAH-NEXT: pextrw $9, %xmm1, %ecx
-; CHECK-YONAH-NEXT: shrdw $8, %cx, %ax
-; CHECK-YONAH-NEXT: pinsrw $6, %eax, %xmm0
-; CHECK-YONAH-NEXT: pextrw $10, %xmm1, %eax
-; CHECK-YONAH-NEXT: shldw $8, %cx, %ax
-; CHECK-YONAH-NEXT: pinsrw $7, %eax, %xmm0
-; CHECK-YONAH-NEXT: popl %esi
+; CHECK-YONAH-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3],xmm2[4],xmm3[4],xmm2[5],xmm3[5],xmm2[6],xmm3[6],xmm2[7],xmm3[7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[1,2,3,3,4,5,6,7]
+; CHECK-YONAH-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm3[8],xmm0[9],xmm3[9],xmm0[10],xmm3[10],xmm0[11],xmm3[11],xmm0[12],xmm3[12],xmm0[13],xmm3[13],xmm0[14],xmm3[14],xmm0[15],xmm3[15]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm4 = xmm0[3,1,2,0]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,7,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,1,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm4 = xmm4[1,2,3,0,4,5,6,7]
+; CHECK-YONAH-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm4[0]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,2,3,3,4,5,6,7]
+; CHECK-YONAH-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[3,1,2,0]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,7,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,3,2,1]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,1,2,2,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,5,6,7,4]
+; CHECK-YONAH-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
+; CHECK-YONAH-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-YONAH-NEXT: packuswb %xmm0, %xmm2
+; CHECK-YONAH-NEXT: movdqa %xmm2, %xmm0
; CHECK-YONAH-NEXT: retl
%C = shufflevector <16 x i8> %A, <16 x i8> %B, <16 x i32> < i32 5, i32 6, i32 7, i32 undef, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20 >
ret <16 x i8> %C
define <8 x i16> @test9(<8 x i16> %A, <8 x i16> %B) nounwind {
; CHECK-LABEL: test9:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm1 = zero,zero,xmm1[4,5,6,7,8,9,10,11,12,13,14,15,0,1]
+; CHECK-NEXT: palignr {{.*#+}} xmm1 = xmm1[2,3,4,5,6,7,8,9,10,11,12,13,14,15,0,1]
; CHECK-NEXT: movdqa %xmm1, %xmm0
; CHECK-NEXT: retl
;
; CHECK-YONAH-LABEL: test9:
; CHECK-YONAH: # BB#0:
-; CHECK-YONAH-NEXT: pextrw $4, %xmm1, %eax
-; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,2,3,0,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,2,1,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,3,4,5,6,7]
+; CHECK-YONAH-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
+; CHECK-YONAH-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,1,2,4,5,6,7]
; CHECK-YONAH-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
-; CHECK-YONAH-NEXT: pinsrw $3, %eax, %xmm0
-; CHECK-YONAH-NEXT: movd %xmm1, %eax
-; CHECK-YONAH-NEXT: pinsrw $7, %eax, %xmm0
; CHECK-YONAH-NEXT: retl
%C = shufflevector <8 x i16> %B, <8 x i16> %A, <8 x i32> < i32 undef, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0 >
ret <8 x i16> %C
define <4 x i32> @a(<4 x i32> %i) nounwind {
; SSE2-LABEL: a:
; SSE2: movdqa {{.*}}, %[[X1:xmm[0-9]+]]
-; SSE2-NEXT: pshufd {{.*}} # [[X2:xmm[0-9]+]] = xmm0[1,0,3,0]
+; SSE2-NEXT: pshufd {{.*}} # [[X2:xmm[0-9]+]] = xmm0[1,1,3,3]
; SSE2-NEXT: pmuludq %[[X1]], %xmm0
; SSE2-NEXT: pmuludq %[[X1]], %[[X2]]
; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2],[[X2]][0,2]
-; SSE2-NEXT: pshufd {{.*}} # xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSE41-LABEL: a:
define <4 x i32> @c(<4 x i32> %i, <4 x i32> %j) nounwind {
; SSE2-LABEL: c:
-; SSE2: pshufd {{.*}} # [[X2:xmm[0-9]+]] = xmm0[1,0,3,0]
+; SSE2: pshufd {{.*}} # [[X2:xmm[0-9]+]] = xmm0[1,1,3,3]
; SSE2-NEXT: pmuludq %xmm1, %xmm0
-; SSE2-NEXT: pshufd {{.*}} # xmm1 = xmm1[1,0,3,0]
+; SSE2-NEXT: pshufd {{.*}} # xmm1 = xmm1[1,1,3,3]
; SSE2-NEXT: pmuludq %[[X2]], %xmm1
; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*}} # xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSE41-LABEL: c:
define <4 x i32> @e(<4 x i32> %i, <4 x i32> %j) nounwind {
; SSE2-LABEL: e:
-; SSE2: movdqa {{[0-9]*}}(%rsp), %[[X1:xmm[0-9]+]]
-; SSE2-NEXT: pshufd {{.*}} # xmm0 = [[X2]][1,0,3,0]
+; SSE2: movdqa {{[0-9]*}}(%rsp), %xmm0
+; SSE2-NEXT: pshufd {{.*}} # [[X1:xmm[0-9]+]] = xmm0[1,1,3,3]
; SSE2-NEXT: movdqa {{[0-9]*}}(%rsp), %[[X2:xmm[0-9]+]]
-; SSE2-NEXT: pmuludq %[[X2]], %[[X1]]
-; SSE2-NEXT: pshufd {{.*}} # [[X2]] = [[X2]][1,0,3,0]
-; SSE2-NEXT: pmuludq %xmm0, %[[X2]]
-; SSE2-NEXT: shufps {{.*}} # [[X1]] = [[X1]][0,2],[[X2]][0,2]
-; SSE2-NEXT: pshufd {{.*}} # xmm0 = [[X1]][0,2,1,3]
+; SSE2-NEXT: pmuludq %[[X2]], %xmm0
+; SSE2-NEXT: pshufd {{.*}} # [[X2]] = [[X2]][1,1,3,3]
+; SSE2-NEXT: pmuludq %[[X1]], %[[X2]]
+; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2],[[X2]][0,2]
+; SSE2-NEXT: shufps {{.*}} # xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: addq ${{[0-9]+}}, %rsp
; SSE2-NEXT: retq
;
entry:
; CHECK: v3f2d_ext_vec
; CHECK: cvtps2pd
-; CHECK: movhlps
+; CHECK: shufpd
; CHECK: cvtps2pd
; AVX: v3f2d_ext_vec
; AVX: vcvtps2pd
entry:
; CHECK: v4f2d_ext_vec
; CHECK: cvtps2pd
-; CHECK: movhlps
+; CHECK: shufpd
; CHECK: cvtps2pd
; AVX: v4f2d_ext_vec
; AVX: vcvtps2pd
; CHECK: v8f2d_ext_vec
; CHECK: cvtps2pd
; CHECK: cvtps2pd
-; CHECK: movhlps
+; CHECK: shufpd
; CHECK: cvtps2pd
-; CHECK: movhlps
+; CHECK: shufpd
; CHECK: cvtps2pd
; AVX: v8f2d_ext_vec
; AVX: vcvtps2pd
+++ /dev/null
-; RUN: llc -asm-verbose -mtriple=x86_64-unknown-unknown -mcpu=corei7 < %s | FileCheck %s
-define <16 x i8> @shuf(<16 x i8> %inval1) {
-entry:
- %0 = shufflevector <16 x i8> %inval1, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 4, i32 3, i32 2, i32 16, i32 16, i32 3, i32 4, i32 0, i32 4, i32 3, i32 2, i32 16, i32 16, i32 3, i32 4>
- ret <16 x i8> %0
-; CHECK: shuf
-; CHECK: # BB#0: # %entry
-; CHECK-NEXT: pshufb
-; CHECK-NEXT: ret
-}
; OSX_SINCOS-LABEL: test1:
; OSX_SINCOS: callq ___sincosf_stret
-; OSX_SINCOS: pshufd $1, %xmm0, %xmm1
+; OSX_SINCOS: pshufd {{.*}} ## xmm1 = xmm0[1,1,2,3]
; OSX_SINCOS: addss %xmm0, %xmm1
; OSX_NOOPT: test1
+++ /dev/null
-; RUN: llc < %s -mtriple=i386-apple-darwin -mattr=+sse2 -mcpu=nehalem | FileCheck %s
-; rdar://7434544
-
-define <2 x i64> @t2() nounwind {
-entry:
-; CHECK-LABEL: t2:
-; CHECK: pshufd $85, (%esp), %xmm0
- %array = alloca [8 x float], align 4
- %arrayidx = getelementptr inbounds [8 x float]* %array, i32 0, i32 1
- %tmp2 = load float* %arrayidx
- %vecinit = insertelement <4 x float> undef, float %tmp2, i32 0
- %vecinit5 = insertelement <4 x float> %vecinit, float %tmp2, i32 1
- %vecinit7 = insertelement <4 x float> %vecinit5, float %tmp2, i32 2
- %vecinit9 = insertelement <4 x float> %vecinit7, float %tmp2, i32 3
- %0 = bitcast <4 x float> %vecinit9 to <2 x i64>
- ret <2 x i64> %0
-}
define <4 x float> @a(<4 x float>* %y) nounwind {
; CHECK-LABEL: a:
; CHECK: # BB#0:
-; CHECK-NEXT: movdqu (%rdi), %xmm0
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-NEXT: movups (%rdi), %xmm0
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,2,1,0]
; CHECK-NEXT: retq
%x = load <4 x float>* %y, align 4
%a = extractelement <4 x float> %x, i32 0
; RUN: llc -mcpu=x86-64 -mattr=+sse2 < %s | FileCheck --check-prefix=SSE --check-prefix=SSE2 %s
-; RUN: llc -mcpu=x86-64 -mattr=+sse2 < %s -x86-experimental-vector-shuffle-lowering | FileCheck --check-prefix=SSE --check-prefix=SSE2 %s
; RUN: llc -mcpu=x86-64 -mattr=+sse4.1 < %s | FileCheck --check-prefix=SSE --check-prefix=SSE41 %s
-; RUN: llc -mcpu=x86-64 -mattr=+sse4.1 < %s -x86-experimental-vector-shuffle-lowering | FileCheck --check-prefix=SSE --check-prefix=SSE41 %s
; RUN: llc -mcpu=x86-64 -mattr=+avx < %s | FileCheck --check-prefix=AVX %s
-; RUN: llc -mcpu=x86-64 -mattr=+avx < %s -x86-experimental-vector-shuffle-lowering | FileCheck --check-prefix=AVX %s
target triple = "x86_64-unknown-unknown"
; CHECK-LABEL: test4:
; CHECK: # BB#0: # %entry
; CHECK-NEXT: movaps %xmm0, %xmm2
-; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,0,0,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,1,2,3]
; CHECK-NEXT: addss %xmm1, %xmm0
-; CHECK-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0,0,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1,2,3]
; CHECK-NEXT: subss %xmm1, %xmm2
; CHECK-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; CHECK-NEXT: ret
+++ /dev/null
-; RUN: llc < %s -march=x86-64 -mcpu=core2 | FileCheck %s
-
-define <4 x i32> @test1(<4 x i32> %x, <4 x i32> %y) {
- %m = mul <4 x i32> %x, %y
- ret <4 x i32> %m
-; CHECK-LABEL: test1:
-; CHECK: pshufd $49
-; CHECK: pmuludq
-; CHECK: pshufd $49
-; CHECK: pmuludq
-; CHECK: shufps $-120
-; CHECK: pshufd $-40
-; CHECK: ret
-}
; CHECK-LABEL: test4:
; CHECK: ## BB#0:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
-; CHECK-NEXT: movdqa %xmm0, (%eax)
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,1,3,3]
+; CHECK-NEXT: movaps %xmm0, (%eax)
; CHECK-NEXT: retl
%tmp5 = shufflevector <4 x float> %X, <4 x float> undef, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=1]
store <4 x float> %tmp5, <4 x float>* %res
define void @test12() nounwind {
; CHECK-LABEL: test12:
; CHECK: ## BB#0:
-; CHECK-NEXT: movaps 0, %xmm0
-; CHECK-NEXT: xorps %xmm1, %xmm1
-; CHECK-NEXT: movhlps {{.*#+}} xmm1 = xmm0[1],xmm1[1]
-; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],mem[2,3]
+; CHECK-NEXT: movapd 0, %xmm0
+; CHECK-NEXT: movaps {{.*#+}} xmm1 = [1.000000e+00,1.000000e+00,1.000000e+00,1.000000e+00]
+; CHECK-NEXT: movsd %xmm0, %xmm1
+; CHECK-NEXT: xorpd %xmm2, %xmm2
+; CHECK-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm2[1]
; CHECK-NEXT: addps %xmm1, %xmm0
; CHECK-NEXT: movaps %xmm0, 0
; CHECK-NEXT: retl
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %edx
; CHECK-NEXT: movaps (%edx), %xmm0
; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1],mem[0,1]
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
-; CHECK-NEXT: movdqa %xmm0, (%eax)
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; CHECK-NEXT: movaps %xmm0, (%eax)
; CHECK-NEXT: retl
%tmp3 = load <4 x float>* %B ; <<4 x float>> [#uses=1]
%tmp5 = load <4 x float>* %C ; <<4 x float>> [#uses=1]
; CHECK-NEXT: movaps %xmm2, %xmm0
; CHECK-NEXT: addps %xmm1, %xmm0
; CHECK-NEXT: subps %xmm1, %xmm2
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm2[0]
; CHECK-NEXT: retl
%tmp = load <4 x float>* %y ; <<4 x float>> [#uses=2]
%tmp5 = load <4 x float>* %x ; <<4 x float>> [#uses=2]
; CHECK: ## BB#0: ## %entry
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
-; CHECK-NEXT: movaps (%ecx), %xmm0
-; CHECK-NEXT: movaps (%eax), %xmm1
-; CHECK-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; CHECK-NEXT: movapd (%ecx), %xmm0
+; CHECK-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],mem[1]
; CHECK-NEXT: retl
entry:
%tmp = load <4 x float>* %y ; <<4 x float>> [#uses=1]
; CHECK: ## BB#0: ## %entry
; CHECK-NEXT: cvtpd2ps %xmm1, %xmm1
; CHECK-NEXT: cvtpd2ps %xmm0, %xmm0
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: retl
entry:
%double2float.i = fptrunc <4 x double> %0 to <4 x float>
define <4 x i32> @PR19721(<4 x i32> %i) {
; CHECK-LABEL: PR19721:
; CHECK: ## BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,0,0,0]
+; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
; CHECK-NEXT: movd %xmm1, %eax
-; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,0,0,0]
-; CHECK-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
-; CHECK-NEXT: movd %xmm0, %ecx
-; CHECK-NEXT: movd %xmm1, %edx
-; CHECK-NEXT: movd %edx, %xmm0
+; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3]
+; CHECK-NEXT: movd %xmm1, %ecx
+; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,2,3]
+; CHECK-NEXT: pxor %xmm0, %xmm0
+; CHECK-NEXT: movss %xmm1, %xmm0
; CHECK-NEXT: movd %ecx, %xmm1
-; CHECK-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; CHECK-NEXT: movd %eax, %xmm0
-; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,1]
+; CHECK-NEXT: movd %eax, %xmm2
+; CHECK-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0],xmm2[0,1]
; CHECK-NEXT: retl
%bc = bitcast <4 x i32> %i to i128
%insert = and i128 %bc, -4294967296
define <4 x i32> @test_mul(<4 x i32> %x, <4 x i32> %y) {
; CHECK-LABEL: test_mul:
; CHECK: ## BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,0,3,0]
+; CHECK-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
; CHECK-NEXT: pmuludq %xmm1, %xmm0
-; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
+; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-NEXT: pmuludq %xmm2, %xmm1
; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; CHECK-NEXT: retl
%m = mul <4 x i32> %x, %y
ret <4 x i32> %m
define void @t0(<8 x i16>* %dest, <8 x i16>* %old) nounwind {
; X64-LABEL: t0:
; X64: ## BB#0: ## %entry
-; X64-NEXT: movdqa (%rsi), %xmm0
-; X64-NEXT: pslldq $2, %xmm0
+; X64-NEXT: pxor %xmm0, %xmm0
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3]
; X64-NEXT: movdqa %xmm0, (%rdi)
; X64-NEXT: retq
entry:
; X64-LABEL: t1:
; X64: ## BB#0:
; X64-NEXT: movdqa (%rdi), %xmm0
-; X64-NEXT: pinsrw $0, (%rsi), %xmm0
+; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,1,2,3,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,0,2,3,4,5,6,7]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X64-NEXT: retq
%tmp1 = load <8 x i16>* %A
%tmp2 = load <8 x i16>* %B
define <8 x i16> @t2(<8 x i16> %A, <8 x i16> %B) nounwind {
; X64-LABEL: t2:
; X64: ## BB#0:
-; X64-NEXT: pextrw $1, %xmm1, %eax
-; X64-NEXT: pinsrw $0, %eax, %xmm0
-; X64-NEXT: pinsrw $3, %eax, %xmm0
+; X64-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,0,1]
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,2,3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,0,3,4,5,6,7]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
; X64-NEXT: retq
%tmp = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 9, i32 1, i32 2, i32 9, i32 4, i32 5, i32 6, i32 7 >
ret <8 x i16> %tmp
define <8 x i16> @t3(<8 x i16> %A, <8 x i16> %B) nounwind {
; X64-LABEL: t3:
; X64: ## BB#0:
-; X64-NEXT: pextrw $5, %xmm0, %eax
-; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,0,4,5,6,7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,2,0]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,6,5]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,2,0]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,1,4,5,6,7]
; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4]
-; X64-NEXT: pinsrw $3, %eax, %xmm0
; X64-NEXT: retq
%tmp = shufflevector <8 x i16> %A, <8 x i16> %A, <8 x i32> < i32 8, i32 3, i32 2, i32 13, i32 7, i32 6, i32 5, i32 4 >
ret <8 x i16> %tmp
define <8 x i16> @t4(<8 x i16> %A, <8 x i16> %B) nounwind {
; X64-LABEL: t4:
; X64: ## BB#0:
-; X64-NEXT: pextrw $7, %xmm0, %eax
-; X64-NEXT: pshufhw {{.*#+}} xmm1 = xmm0[0,1,2,3,4,5,6,5]
-; X64-NEXT: pinsrw $1, %eax, %xmm1
-; X64-NEXT: pextrw $1, %xmm0, %eax
-; X64-NEXT: pinsrw $4, %eax, %xmm1
-; X64-NEXT: movdqa %xmm1, %xmm0
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,5,4,7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,2,0]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,7,4,7]
; X64-NEXT: retq
%tmp = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 0, i32 7, i32 2, i32 3, i32 1, i32 5, i32 6, i32 5 >
ret <8 x i16> %tmp
define <8 x i16> @t5(<8 x i16> %A, <8 x i16> %B) nounwind {
; X64-LABEL: t5:
; X64: ## BB#0:
-; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
+; X64-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; X64-NEXT: movdqa %xmm1, %xmm0
; X64-NEXT: retq
%tmp = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 8, i32 9, i32 0, i32 1, i32 10, i32 11, i32 2, i32 3 >
ret <8 x i16> %tmp
define void @t9(<4 x float>* %r, <2 x i32>* %A) nounwind {
; X64-LABEL: t9:
; X64: ## BB#0:
-; X64-NEXT: movaps (%rdi), %xmm0
-; X64-NEXT: movhps (%rsi), %xmm0
-; X64-NEXT: movaps %xmm0, (%rdi)
+; X64-NEXT: movapd (%rdi), %xmm0
+; X64-NEXT: movhpd (%rsi), %xmm0
+; X64-NEXT: movapd %xmm0, (%rdi)
; X64-NEXT: retq
%tmp = load <4 x float>* %r
%tmp.upgrd.3 = bitcast <2 x i32>* %A to double*
; X64-LABEL: t10:
; X64: ## BB#0:
; X64-NEXT: movq _g1@{{.*}}(%rip), %rax
-; X64-NEXT: movdqa (%rax), %xmm0
-; X64-NEXT: pextrw $4, %xmm0, %eax
-; X64-NEXT: pextrw $6, %xmm0, %ecx
-; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
-; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,0,0,4,5,6,7]
-; X64-NEXT: pinsrw $2, %eax, %xmm0
-; X64-NEXT: pinsrw $3, %ecx, %xmm0
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = mem[0,2,2,3,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; X64-NEXT: movq _g2@{{.*}}(%rip), %rax
; X64-NEXT: movq %xmm0, (%rax)
; X64-NEXT: retq
define <8 x i16> @t11(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
; X64-LABEL: t11:
; X64: ## BB#0: ## %entry
-; X64-NEXT: movd %xmm1, %eax
-; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
-; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,0,0,0,4,5,6,7]
-; X64-NEXT: pinsrw $1, %eax, %xmm0
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,1,2,3,4,5,6,7]
; X64-NEXT: retq
entry:
%tmp7 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 1, i32 8, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef , i32 undef >
define <8 x i16> @t12(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
; X64-LABEL: t12:
; X64: ## BB#0: ## %entry
-; X64-NEXT: pextrw $3, %xmm1, %eax
-; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
-; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,4,4,4]
-; X64-NEXT: pinsrw $5, %eax, %xmm0
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,7,6,7]
; X64-NEXT: retq
entry:
%tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 0, i32 1, i32 undef, i32 undef, i32 3, i32 11, i32 undef , i32 undef >
define <8 x i16> @t13(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
; X64-LABEL: t13:
; X64: ## BB#0: ## %entry
-; X64-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm0[0]
-; X64-NEXT: pextrw $3, %xmm1, %eax
-; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,7,4,4]
-; X64-NEXT: pinsrw $4, %eax, %xmm0
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,2,2,3,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,7,6,7]
; X64-NEXT: retq
entry:
%tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 8, i32 9, i32 undef, i32 undef, i32 11, i32 3, i32 undef , i32 undef >
define <8 x i16> @t14(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
; X64-LABEL: t14:
; X64: ## BB#0: ## %entry
-; X64-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm0[0]
-; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,6,4,4]
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,2,2,3,4,5,6,7]
; X64-NEXT: retq
entry:
%tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 8, i32 9, i32 undef, i32 undef, i32 undef, i32 2, i32 undef , i32 undef >
define <8 x i16> @t15(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
; X64-LABEL: t15:
; X64: ## BB#0: ## %entry
-; X64-NEXT: pextrw $7, %xmm0, %eax
-; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,2,4,5,6,7]
-; X64-NEXT: pinsrw $2, %eax, %xmm0
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,2,3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,2,3,4,5,6,7]
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,0,3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,6,7]
; X64-NEXT: retq
entry:
%tmp8 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 undef, i32 undef, i32 7, i32 2, i32 8, i32 undef, i32 undef , i32 undef >
define <16 x i8> @t16(<16 x i8> %T0) nounwind readnone {
; X64-LABEL: t16:
; X64: ## BB#0: ## %entry
-; X64-NEXT: pextrw $8, %xmm0, %eax
-; X64-NEXT: pslldq $2, %xmm0
-; X64-NEXT: andl $65280, %eax ## imm = 0xFF00
-; X64-NEXT: pextrw $1, %xmm0, %ecx
-; X64-NEXT: movzbl %cl, %ecx
-; X64-NEXT: orl %eax, %ecx
-; X64-NEXT: pinsrw $1, %ecx, %xmm0
+; X64-NEXT: movdqa {{.*#+}} xmm1 = [0,0,0,0,1,1,1,1,0,0,0,0,0,0,0,0]
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-NEXT: pxor %xmm2, %xmm2
+; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[0,2,2,3,4,5,6,7]
+; X64-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
+; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X64-NEXT: packuswb %xmm0, %xmm0
; X64-NEXT: retq
entry:
%tmp8 = shufflevector <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 1, i8 1, i8 1, i8 1, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, <16 x i8> %T0, <16 x i32> < i32 0, i32 1, i32 16, 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 >
; X32-LABEL: ext_1:
; X32: ## BB#0:
; X32-NEXT: pushl %eax
-; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,0,0]
+; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,1,2,3]
; X32-NEXT: addss LCPI7_0, %xmm0
; X32-NEXT: movss %xmm0, (%esp)
; X32-NEXT: flds (%esp)
;
; X64-LABEL: ext_1:
; X64: ## BB#0:
-; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,0,0]
+; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,1,2,3]
; X64-NEXT: addss {{.*}}(%rip), %xmm0
; X64-NEXT: retq
%s = extractelement <4 x float> %v, i32 3
; X32-LABEL: ext_2:
; X32: ## BB#0:
; X32-NEXT: pushl %eax
-; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,0,0]
+; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,1,2,3]
; X32-NEXT: movss %xmm0, (%esp)
; X32-NEXT: flds (%esp)
; X32-NEXT: popl %eax
;
; X64-LABEL: ext_2:
; X64: ## BB#0:
-; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,0,0]
+; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,1,2,3]
; X64-NEXT: retq
%s = extractelement <4 x float> %v, i32 3
ret float %s
define <2 x float> @buildvector(<2 x float> %A, <2 x float> %B) nounwind {
; X32-LABEL: buildvector:
; X32: ## BB#0: ## %entry
-; X32-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,0,0,0]
-; X32-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,0,0]
+; X32-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,2,3]
; X32-NEXT: addss %xmm1, %xmm0
-; X32-NEXT: addss %xmm2, %xmm3
-; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[2,3]
+; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1,2,3]
+; X32-NEXT: addss %xmm2, %xmm1
+; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
; X32-NEXT: retl
;
; X64-LABEL: buildvector:
; X64: ## BB#0: ## %entry
-; X64-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,0,0,0]
-; X64-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,0,0]
+; X64-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,2,3]
; X64-NEXT: addss %xmm1, %xmm0
-; X64-NEXT: addss %xmm2, %xmm3
-; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[2,3]
+; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1,2,3]
+; X64-NEXT: addss %xmm2, %xmm1
+; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
; X64-NEXT: retq
entry:
%tmp7 = extractelement <2 x float> %A, i32 0
; X32-LABEL: pinsrd_from_shufflevector_i32:
; X32: ## BB#0: ## %entry
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: pinsrd $3, (%eax), %xmm0
+; X32-NEXT: insertps $48, (%eax), %xmm0
; X32-NEXT: retl
;
; X64-LABEL: pinsrd_from_shufflevector_i32:
; X64: ## BB#0: ## %entry
-; X64-NEXT: pinsrd $3, (%rdi), %xmm0
+; X64-NEXT: insertps $48, (%rdi), %xmm0
; X64-NEXT: retq
entry:
%0 = load <4 x i32>* %pb, align 16
; X32-LABEL: shuf_X00A:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm2, %xmm2
-; X32-NEXT: blendps $1, %xmm0, %xmm2
-; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[0]
+; X32-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0],zero,zero,xmm1[0]
; X32-NEXT: movaps %xmm2, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: shuf_X00A:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm2, %xmm2
-; X64-NEXT: blendps $1, %xmm0, %xmm2
-; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[0]
+; X64-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0],zero,zero,xmm1[0]
; X64-NEXT: movaps %xmm2, %xmm0
; X64-NEXT: retq
%vecext = extractelement <4 x float> %x, i32 0
; X32-LABEL: shuf_X00X:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm1, %xmm1
-; X32-NEXT: blendps $1, %xmm0, %xmm1
-; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0]
+; X32-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
+; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],zero,zero,xmm0[0]
; X32-NEXT: movaps %xmm1, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: shuf_X00X:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm1, %xmm1
-; X64-NEXT: blendps $1, %xmm0, %xmm1
-; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0]
+; X64-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
+; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],zero,zero,xmm0[0]
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: retq
%vecext = extractelement <4 x float> %x, i32 0
; X32-LABEL: shuf_X0YC:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm2, %xmm2
-; X32-NEXT: blendps $1, %xmm0, %xmm2
-; X32-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1,0]
+; X32-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0],zero,xmm0[1],zero
; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2]
; X32-NEXT: movaps %xmm2, %xmm0
; X32-NEXT: retl
; X64-LABEL: shuf_X0YC:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm2, %xmm2
-; X64-NEXT: blendps $1, %xmm0, %xmm2
-; X64-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1,0]
+; X64-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0],zero,xmm0[1],zero
; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2]
; X64-NEXT: movaps %xmm2, %xmm0
; X64-NEXT: retq
; X32-LABEL: i32_shuf_X00A:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm2, %xmm2
-; X32-NEXT: blendps $1, %xmm0, %xmm2
+; X32-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[0]
; X32-NEXT: movaps %xmm2, %xmm0
; X32-NEXT: retl
; X64-LABEL: i32_shuf_X00A:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm2, %xmm2
-; X64-NEXT: blendps $1, %xmm0, %xmm2
+; X64-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[0]
; X64-NEXT: movaps %xmm2, %xmm0
; X64-NEXT: retq
; X32-LABEL: i32_shuf_X00X:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm1, %xmm1
-; X32-NEXT: blendps $1, %xmm0, %xmm1
+; X32-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0]
; X32-NEXT: movaps %xmm1, %xmm0
; X32-NEXT: retl
; X64-LABEL: i32_shuf_X00X:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm1, %xmm1
-; X64-NEXT: blendps $1, %xmm0, %xmm1
+; X64-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0]
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: retq
; X32-LABEL: i32_shuf_X0YC:
; X32: ## BB#0:
; X32-NEXT: xorps %xmm2, %xmm2
-; X32-NEXT: blendps $1, %xmm0, %xmm2
-; X32-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1,0]
+; X32-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1],zero
; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2]
; X32-NEXT: movaps %xmm2, %xmm0
; X32-NEXT: retl
; X64-LABEL: i32_shuf_X0YC:
; X64: ## BB#0:
; X64-NEXT: xorps %xmm2, %xmm2
-; X64-NEXT: blendps $1, %xmm0, %xmm2
-; X64-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1,0]
+; X64-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
+; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1],zero
; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2]
; X64-NEXT: movaps %xmm2, %xmm0
; X64-NEXT: retq
; X32: ## BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
-; X32-NEXT: insertps $48, (%ecx,%eax,4), %xmm0
+; X32-NEXT: movss (%ecx,%eax,4), %xmm1
+; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: insertps_from_broadcast_loadf32:
; X64: ## BB#0:
-; X64-NEXT: insertps $48, (%rdi,%rsi,4), %xmm0
+; X64-NEXT: movss (%rdi,%rsi,4), %xmm1
+; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
; X64-NEXT: retq
%1 = getelementptr inbounds float* %fb, i64 %index
%2 = load float* %1, align 4
; X32-LABEL: insertps_from_broadcast_loadv4f32:
; X32: ## BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: insertps $48, (%eax), %xmm0
+; X32-NEXT: movups (%eax), %xmm1
+; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: insertps_from_broadcast_loadv4f32:
; X64: ## BB#0:
-; X64-NEXT: insertps $48, (%rdi), %xmm0
+; X64-NEXT: movups (%rdi), %xmm1
+; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
; X64-NEXT: retq
%1 = load <4 x float>* %b, align 4
%2 = extractelement <4 x float> %1, i32 0
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: movss (%ecx,%eax,4), %xmm4
-; X32-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,0,0,0]
+; X32-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,0,0,0]
; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm4[0]
; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm4[0]
; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm4[0]
; X64-LABEL: insertps_from_broadcast_multiple_use:
; X64: ## BB#0:
; X64-NEXT: movss (%rdi,%rsi,4), %xmm4
-; X64-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,0,0,0]
+; X64-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,0,0,0]
; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm4[0]
; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm4[0]
; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm4[0]
; X32: ## BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movss (%eax), %xmm1
-; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1],xmm0[0],xmm1[3]
+; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],zero,xmm0[0],xmm1[3]
; X32-NEXT: movaps %xmm1, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: insertps_with_undefs:
; X64: ## BB#0:
; X64-NEXT: movss (%rdi), %xmm1
-; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1],xmm0[0],xmm1[3]
+; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],zero,xmm0[0],xmm1[3]
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: retq
%1 = load float* %b, align 4
; X32-LABEL: pr20087:
; X32: ## BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: insertps $48, 8(%eax), %xmm0
+; X32-NEXT: insertps $-78, 8(%eax), %xmm0
; X32-NEXT: retl
;
; X64-LABEL: pr20087:
; X64: ## BB#0:
-; X64-NEXT: insertps $48, 8(%rdi), %xmm0
+; X64-NEXT: insertps $-78, 8(%rdi), %xmm0
; X64-NEXT: retq
%load = load <4 x float> *%ptr
%ret = shufflevector <4 x float> %load, <4 x float> %a, <4 x i32> <i32 4, i32 undef, i32 6, i32 2>
; X32-LABEL: insertps_pr20411:
; X32: ## BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: movaps {{.*#+}} xmm0 = [4,5,6,7]
-; X32-NEXT: pshufd {{.*#+}} xmm1 = mem[3,0,0,0]
-; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],xmm0[3],xmm1[2,3]
-; X32-NEXT: movups %xmm1, (%eax)
+; X32-NEXT: pshufd {{.*#+}} xmm0 = mem[3,1,2,3]
+; X32-NEXT: insertps $-36, LCPI49_1+12, %xmm0
+; X32-NEXT: movups %xmm0, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: insertps_pr20411:
; X64: ## BB#0:
-; X64-NEXT: movaps {{.*#+}} xmm0 = [4,5,6,7]
-; X64-NEXT: pshufd {{.*#+}} xmm1 = mem[3,0,0,0]
-; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0],xmm0[3],xmm1[2,3]
-; X64-NEXT: movups %xmm1, (%rdi)
+; X64-NEXT: pshufd {{.*#+}} xmm0 = mem[3,1,2,3]
+; X64-NEXT: insertps $-36, LCPI49_1+{{.*}}(%rip), %xmm0
+; X64-NEXT: movups %xmm0, (%rdi)
; X64-NEXT: retq
%gather_load = shufflevector <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>, <8 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%shuffle109 = shufflevector <4 x i32> <i32 4, i32 5, i32 6, i32 7>, <4 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; 4 5 6 7
define <4 x float> @swizzle_15(<4 x float> %v) {
; CHECK-LABEL: swizzle_15:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,3,2]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0,3,2]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 0, i32 1>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 0, i32 1>
define <4 x float> @swizzle_16(<4 x float> %v) {
; CHECK-LABEL: swizzle_16:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,3,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,1,3,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 3, i32 1, i32 0, i32 2>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 3, i32 1, i32 0, i32 2>
define <4 x float> @swizzle_17(<4 x float> %v) {
; CHECK-LABEL: swizzle_17:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,3,2]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0,3,2]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 1, i32 0>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 1, i32 0>
define <4 x float> @swizzle_18(<4 x float> %v) {
; CHECK-LABEL: swizzle_18:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,0,2]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,1,0,2]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 2, i32 1, i32 3, i32 0>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 2, i32 1, i32 3, i32 0>
define <4 x float> @swizzle_19(<4 x float> %v) {
; CHECK-LABEL: swizzle_19:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
+; CHECK-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
define <4 x float> @swizzle_20(<4 x float> %v) {
; CHECK-LABEL: swizzle_20:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,1,3]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0,1,3]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 1, i32 2, i32 0, i32 3>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 1, i32 2, i32 0, i32 3>
define <4 x float> @swizzle_21(<4 x float> %v) {
; CHECK-LABEL: swizzle_21:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,3,1]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,3,1]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 0, i32 3, i32 1, i32 2>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 0, i32 3, i32 1, i32 2>
define <4 x float> @swizzle_22(<4 x float> %v) {
; CHECK-LABEL: swizzle_22:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3,2,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 3, i32 0, i32 2, i32 1>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 3, i32 0, i32 2, i32 1>
define <4 x float> @swizzle_23(<4 x float> %v) {
; CHECK-LABEL: swizzle_23:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
+; CHECK-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
define <4 x float> @swizzle_24(<4 x float> %v) {
; CHECK-LABEL: swizzle_24:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,2,0,3]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,2,0,3]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 2, i32 0, i32 1, i32 3>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 2, i32 0, i32 1, i32 3>
define <4 x float> @swizzle_25(<4 x float> %v) {
; CHECK-LABEL: swizzle_25:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,2,1,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 2, i32 0, i32 3, i32 1>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 2, i32 0, i32 3, i32 1>
define <4 x float> @swizzle_26(<4 x float> %v) {
; CHECK-LABEL: swizzle_26:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,3,1,2]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3,1,2]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 0, i32 2, i32 3, i32 1>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 0, i32 2, i32 3, i32 1>
define <4 x float> @swizzle_27(<4 x float> %v) {
; CHECK-LABEL: swizzle_27:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,2,1,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 0, i32 2>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 0, i32 2>
define <4 x float> @swizzle_28(<4 x float> %v) {
; CHECK-LABEL: swizzle_28:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,2,1]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,0,2,1]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 2, i32 0>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 2, i32 0>
define <4 x float> @swizzle_29(<4 x float> %v) {
; CHECK-LABEL: swizzle_29:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3,2,0]
; CHECK-NEXT: retq
%1 = shufflevector <4 x float> %v, <4 x float> undef, <4 x i32> <i32 3, i32 1, i32 2, i32 0>
%2 = shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> <i32 1, i32 0, i32 2, i32 3>
define <8 x i16> @swizzle_33(<8 x i16> %v) {
; CHECK-LABEL: swizzle_33:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,5,2,3,6,7,0,1,10,11,14,15,12,13,8,9]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,1,3,0,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,7,6,4]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 4, i32 6, i32 5, i32 7, i32 2, i32 3, i32 1, i32 0>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 4, i32 6, i32 5, i32 7, i32 2, i32 3, i32 1, i32 0>
define <8 x i16> @swizzle_34(<8 x i16> %v) {
; CHECK-LABEL: swizzle_34:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,3,6,7,0,1,4,5,14,15,12,13,8,9,10,11]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,3,0,2,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,4,5]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 4, i32 7, i32 6, i32 5, i32 1, i32 2, i32 0, i32 3>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 4, i32 7, i32 6, i32 5, i32 1, i32 2, i32 0, i32 3>
define <8 x i16> @swizzle_35(<8 x i16> %v) {
; CHECK-LABEL: swizzle_35:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,5,2,3,0,1,6,7,8,9,10,11,14,15,12,13]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,1,0,3,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,7,6]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 7, i32 4, i32 6, i32 5, i32 1, i32 3, i32 0, i32 2>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 7, i32 4, i32 6, i32 5, i32 1, i32 3, i32 0, i32 2>
define <8 x i16> @swizzle_36(<8 x i16> %v) {
; CHECK-LABEL: swizzle_36:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1,6,7,4,5,2,3,8,9,12,13,10,11,14,15]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,1,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,5,7]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 4, i32 6, i32 7, i32 5, i32 0, i32 1, i32 3, i32 2>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 4, i32 6, i32 7, i32 5, i32 0, i32 1, i32 3, i32 2>
define <8 x i16> @swizzle_38(<8 x i16> %v) {
; CHECK-LABEL: swizzle_38:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,5,2,3,0,1,6,7,10,11,8,9,12,13,14,15]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,1,0,3,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,4,6,7]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 5, i32 6, i32 4, i32 7, i32 0, i32 2, i32 1, i32 3>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 5, i32 6, i32 4, i32 7, i32 0, i32 2, i32 1, i32 3>
define <8 x i16> @swizzle_39(<8 x i16> %v) {
; CHECK-LABEL: swizzle_39:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,5,6,7,2,3,0,1,14,15,12,13,8,9,10,11]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,3,1,0,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,4,5]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 5, i32 4, i32 6, i32 7, i32 3, i32 2, i32 1, i32 0>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 5, i32 4, i32 6, i32 7, i32 3, i32 2, i32 1, i32 0>
define <8 x i16> @swizzle_40(<8 x i16> %v) {
; CHECK-LABEL: swizzle_40:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[6,7,2,3,4,5,0,1,8,9,12,13,10,11,14,15]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,1,2,0,4,5,6,7]
+; CHECK-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,5,7]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 6, i32 4, i32 7, i32 5, i32 1, i32 0, i32 3, i32 2>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 6, i32 4, i32 7, i32 5, i32 1, i32 0, i32 3, i32 2>
define <8 x i16> @swizzle_41(<8 x i16> %v) {
; CHECK-LABEL: swizzle_41:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[6,7,4,5,2,3,0,1,12,13,14,15,8,9,10,11]
+; CHECK-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
+; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,3,2]
; CHECK-NEXT: retq
%1 = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 6, i32 7, i32 5, i32 4, i32 0, i32 1, i32 3, i32 2>
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 6, i32 7, i32 5, i32 4, i32 0, i32 1, i32 3, i32 2>
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | grep movlps
-; RUN: llc < %s -march=x86 -mattr=+sse2 | grep movsd
-; RUN: llc < %s -march=x86 -mattr=+sse2 | not grep movups
-; rdar://6523650
-
- %struct.vector4_t = type { <4 x float> }
-
-define void @swizzle(i8* nocapture %a, %struct.vector4_t* nocapture %b, %struct.vector4_t* nocapture %c) nounwind {
-entry:
- %0 = getelementptr %struct.vector4_t* %b, i32 0, i32 0 ; <<4 x float>*> [#uses=2]
- %1 = load <4 x float>* %0, align 4 ; <<4 x float>> [#uses=1]
- %tmp.i = bitcast i8* %a to double* ; <double*> [#uses=1]
- %tmp1.i = load double* %tmp.i ; <double> [#uses=1]
- %2 = insertelement <2 x double> undef, double %tmp1.i, i32 0 ; <<2 x double>> [#uses=1]
- %tmp2.i = bitcast <2 x double> %2 to <4 x float> ; <<4 x float>> [#uses=1]
- %3 = shufflevector <4 x float> %1, <4 x float> %tmp2.i, <4 x i32> < i32 4, i32 5, i32 2, i32 3 > ; <<4 x float>> [#uses=1]
- store <4 x float> %3, <4 x float>* %0, align 4
- ret void
-}
; Read 32-bits
;CHECK: pmovzxwq
;CHECK: paddq
-;CHECK: pshufb
+;CHECK: pshufd
;CHECK: movd
;CHECK: ret
define void @load_2_i16(<2 x i16>* %A) {
; CHECK-LABEL: test1:
; CHECK: # BB#0: # %entry
; CHECK-NEXT: pushl %eax
-; CHECK-NEXT: movsd {{.*}}, %xmm0
+; CHECK-NEXT: movsd .LCPI0_0, %xmm0
; CHECK-NEXT: movd {{[0-9]+}}(%esp), %xmm1
; CHECK-NEXT: orps %xmm0, %xmm1
; CHECK-NEXT: subsd %xmm0, %xmm1
; CHECK-NEXT: pushl %eax
; CHECK-NEXT: xorps %xmm1, %xmm1
; CHECK-NEXT: movss %xmm0, %xmm1
-; CHECK-NEXT: movsd {{.*}}, %xmm0
+; CHECK-NEXT: movsd .LCPI1_0, %xmm0
; CHECK-NEXT: orps %xmm0, %xmm1
; CHECK-NEXT: subsd %xmm0, %xmm1
; CHECK-NEXT: xorps %xmm0, %xmm0
define void @test1(<2 x float> %Q, float *%P2) nounwind {
; X64-LABEL: test1:
; X64: # BB#0:
-; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,0,0,0]
+; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,2,3]
; X64-NEXT: addss %xmm0, %xmm1
; X64-NEXT: movss %xmm1, (%rdi)
; X64-NEXT: retq
; X32-LABEL: test1:
; X32: # BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,0,0,0]
+; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,2,3]
; X32-NEXT: addss %xmm0, %xmm1
; X32-NEXT: movss %xmm1, (%eax)
; X32-NEXT: retl
; CHECK-WIDE-NEXT: vpmovzxbd %xmm0, %xmm1
; CHECK-WIDE-NEXT: vpslld $24, %xmm1, %xmm1
; CHECK-WIDE-NEXT: vpsrad $24, %xmm1, %xmm1
-; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
+; CHECK-WIDE-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; CHECK-WIDE-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
; CHECK-WIDE-NEXT: vpslld $24, %xmm0, %xmm0
; CHECK-WIDE-NEXT: vpsrad $24, %xmm0, %xmm0
; CHECK-WIDE-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
define <8 x float> @foo2_8(<8 x i8> %src) {
; CHECK-LABEL: foo2_8:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm0[4,4,5,5,6,6,7,7]
-; CHECK-NEXT: vpmovzxwd %xmm0, %xmm0
-; CHECK-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; CHECK-NEXT: vpmovzxwd %xmm0, %xmm1
+; CHECK-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
+; CHECK-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; CHECK-NEXT: vandps LCPI2_0, %ymm0, %ymm0
; CHECK-NEXT: vcvtdq2ps %ymm0, %ymm0
; CHECK-NEXT: retl
;
; CHECK-WIDE-LABEL: foo2_8:
; CHECK-WIDE: ## BB#0:
-; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm1 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
-; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
-; CHECK-WIDE-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; CHECK-WIDE-NEXT: vxorps %xmm1, %xmm1, %xmm1
+; CHECK-WIDE-NEXT: vextractf128 $1, %ymm1, %xmm2
+; CHECK-WIDE-NEXT: vmovdqa {{.*#+}} xmm3 = <1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u>
+; CHECK-WIDE-NEXT: vpshufb %xmm3, %xmm2, %xmm4
+; CHECK-WIDE-NEXT: vmovdqa {{.*#+}} xmm5 = <2,6,10,14,u,u,u,u,u,u,u,u,u,u,u,u>
+; CHECK-WIDE-NEXT: vpshufb %xmm5, %xmm2, %xmm2
+; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm6 = xmm0[4,5,6,7,u,u,u,u,u,u,u,u,u,u,u,u]
+; CHECK-WIDE-NEXT: vpunpcklbw {{.*#+}} xmm2 = xmm6[0],xmm2[0],xmm6[1],xmm2[1],xmm6[2],xmm2[2],xmm6[3],xmm2[3],xmm6[4],xmm2[4],xmm6[5],xmm2[5],xmm6[6],xmm2[6],xmm6[7],xmm2[7]
+; CHECK-WIDE-NEXT: vpunpcklbw {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1],xmm2[2],xmm4[2],xmm2[3],xmm4[3],xmm2[4],xmm4[4],xmm2[5],xmm4[5],xmm2[6],xmm4[6],xmm2[7],xmm4[7]
+; CHECK-WIDE-NEXT: vpshufb %xmm3, %xmm1, %xmm3
+; CHECK-WIDE-NEXT: vpshufb %xmm5, %xmm1, %xmm1
+; CHECK-WIDE-NEXT: vpmovzxbd %xmm0, %xmm0
+; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
+; CHECK-WIDE-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; CHECK-WIDE-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
+; CHECK-WIDE-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
; CHECK-WIDE-NEXT: vcvtdq2ps %ymm0, %ymm0
; CHECK-WIDE-NEXT: retl
%res = uitofp <8 x i8> %src to <8 x float>
;
; CHECK-WIDE-LABEL: foo2_4:
; CHECK-WIDE: ## BB#0:
-; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
-; CHECK-WIDE-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; CHECK-WIDE-NEXT: vpshufb {{.*#+}} xmm1 = zero,xmm1[1,2,3],zero,xmm1[5,6,7],zero,xmm1[9,10,11],zero,xmm1[13,14,15]
-; CHECK-WIDE-NEXT: vpor %xmm0, %xmm1, %xmm0
+; CHECK-WIDE-NEXT: vpmovzxbd %xmm0, %xmm0
; CHECK-WIDE-NEXT: vcvtdq2ps %xmm0, %xmm0
; CHECK-WIDE-NEXT: retl
%res = uitofp <4 x i8> %src to <4 x float>
; CHECK: ## BB#0:
; CHECK-NEXT: vcvttps2dq %ymm0, %ymm0
; CHECK-NEXT: vextractf128 $1, %ymm0, %xmm1
-; CHECK-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,128,128,128,128,128,128,128,128]
+; CHECK-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; CHECK-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; CHECK-NEXT: vpshufb %xmm2, %xmm0, %xmm0
-; CHECK-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retl
;
; CHECK-WIDE-LABEL: foo3_8:
; CHECK-WIDE: ## BB#0:
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[3,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[3,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
-; CHECK-WIDE-NEXT: vmovhlps {{.*#+}} xmm1 = xmm0[1,1]
+; CHECK-WIDE-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %ecx
; CHECK-WIDE-NEXT: movzbl %cl, %ecx
; CHECK-WIDE-NEXT: orl %eax, %ecx
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[1,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %edx
; CHECK-WIDE-NEXT: vpinsrw $0, %edx, %xmm0, %xmm1
; CHECK-WIDE-NEXT: vpinsrw $1, %ecx, %xmm1, %xmm1
; CHECK-WIDE-NEXT: vextractf128 $1, %ymm0, %xmm0
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm2 = xmm0[1,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[1,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm2, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %ecx
; CHECK-WIDE-NEXT: movzbl %cl, %ecx
; CHECK-WIDE-NEXT: orl %eax, %ecx
; CHECK-WIDE-NEXT: vpinsrw $2, %ecx, %xmm1, %xmm1
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm2 = xmm0[3,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[3,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm2, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
-; CHECK-WIDE-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; CHECK-WIDE-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %ecx
; CHECK-WIDE-NEXT: movzbl %cl, %ecx
; CHECK-WIDE-NEXT: orl %eax, %ecx
;
; CHECK-WIDE-LABEL: foo3_4:
; CHECK-WIDE: ## BB#0:
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[3,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[3,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
-; CHECK-WIDE-NEXT: vmovhlps {{.*#+}} xmm1 = xmm0[1,1]
+; CHECK-WIDE-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %ecx
; CHECK-WIDE-NEXT: movzbl %cl, %ecx
; CHECK-WIDE-NEXT: orl %eax, %ecx
-; CHECK-WIDE-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,0,0,0]
+; CHECK-WIDE-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[1,1,2,3]
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %edx
; RUN: llc < %s -mcpu=corei7 -march=x86 -mattr=+sse4.1 | FileCheck %s
define void @t1(float* %R, <4 x float>* %P1) nounwind {
-; CHECK-LABEL: @t1
-; CHECK: movl 4(%esp), %[[R0:e[abcd]x]]
-; CHECK-NEXT: movl 8(%esp), %[[R1:e[abcd]x]]
-; CHECK-NEXT: movss 12(%[[R1]]), %[[R2:xmm.*]]
-; CHECK-NEXT: movss %[[R2]], (%[[R0]])
+; CHECK-LABEL: t1:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; CHECK-NEXT: movss 12(%ecx), %xmm0
+; CHECK-NEXT: movss %xmm0, (%eax)
; CHECK-NEXT: retl
%X = load <4 x float>* %P1
}
define float @t2(<4 x float>* %P1) nounwind {
-; CHECK-LABEL: @t2
-; CHECK: movl 4(%esp), %[[R0:e[abcd]x]]
-; CHECK-NEXT: flds 8(%[[R0]])
+; CHECK-LABEL: t2:
+; CHECK: # BB#0:
+; CHECK-NEXT: pushl %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movapd (%eax), %xmm0
+; CHECK-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0]
+; CHECK-NEXT: movss %xmm0, (%esp)
+; CHECK-NEXT: flds (%esp)
+; CHECK-NEXT: popl %eax
; CHECK-NEXT: retl
%X = load <4 x float>* %P1
}
define void @t3(i32* %R, <4 x i32>* %P1) nounwind {
-; CHECK-LABEL: @t3
-; CHECK: movl 4(%esp), %[[R0:e[abcd]x]]
-; CHECK-NEXT: movl 8(%esp), %[[R1:e[abcd]x]]
-; CHECK-NEXT: movl 12(%[[R1]]), %[[R2:e[abcd]x]]
-; CHECK-NEXT: movl %[[R2]], (%[[R0]])
+; CHECK-LABEL: t3:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; CHECK-NEXT: movl 12(%ecx), %ecx
+; CHECK-NEXT: movl %ecx, (%eax)
; CHECK-NEXT: retl
%X = load <4 x i32>* %P1
}
define i32 @t4(<4 x i32>* %P1) nounwind {
-; CHECK-LABEL: @t4
-; CHECK: movl 4(%esp), %[[R0:e[abcd]x]]
-; CHECK-NEXT: movl 12(%[[R0]]), %eax
+; CHECK-LABEL: t4:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl 12(%eax), %eax
; CHECK-NEXT: retl
%X = load <4 x i32>* %P1
define void @test1(<4 x float>* %F, float* %f) nounwind {
; CHECK-LABEL: test1:
-; CHECK: addps %[[X:xmm[0-9]+]], %[[X]]
-; CHECK-NEXT: movss %[[X]], {{.*}}(%{{.*}})
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; CHECK-NEXT: movaps (%ecx), %xmm0
+; CHECK-NEXT: addps %xmm0, %xmm0
+; CHECK-NEXT: movss %xmm0, (%eax)
; CHECK-NEXT: retl
entry:
%tmp = load <4 x float>* %F ; <<4 x float>> [#uses=2]
define float @test2(<4 x float>* %F, float* %f) nounwind {
; CHECK-LABEL: test2:
-; CHECK: addps %[[X:xmm[0-9]+]], %[[X]]
-; CHECK-NEXT: movhlps %[[X]], %[[X2:xmm[0-9]+]]
-; CHECK-NEXT: movss %[[X2]], [[mem:.*\(%.*\)]]
-; CHECK-NEXT: flds [[mem]]
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: pushl %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movaps (%eax), %xmm0
+; CHECK-NEXT: addps %xmm0, %xmm0
+; CHECK-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0]
+; CHECK-NEXT: movss %xmm0, (%esp)
+; CHECK-NEXT: flds (%esp)
+; CHECK-NEXT: popl %eax
+; CHECK-NEXT: retl
entry:
%tmp = load <4 x float>* %F ; <<4 x float>> [#uses=2]
%tmp7 = fadd <4 x float> %tmp, %tmp ; <<4 x float>> [#uses=1]
define void @test3(float* %R, <4 x float>* %P1) nounwind {
; CHECK-LABEL: test3:
-; CHECK: movss {{.*}}(%{{.*}}), %[[X:xmm[0-9]+]]
-; CHECK-NEXT: movss %[[X]], {{.*}}(%{{.*}})
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; CHECK-NEXT: movss 12(%ecx), %xmm0
+; CHECK-NEXT: movss %xmm0, (%eax)
; CHECK-NEXT: retl
entry:
%X = load <4 x float>* %P1 ; <<4 x float>> [#uses=1]
define double @test4(double %A) nounwind {
; CHECK-LABEL: test4:
-; CHECK: calll {{.*}}foo
-; CHECK-NEXT: movhlps %[[X:xmm[0-9]+]], %[[X]]
-; CHECK-NEXT: addsd {{.*}}(%{{.*}}), %[[X2]]
-; CHECK-NEXT: movsd %[[X2]], [[mem:.*\(%.*\)]]
-; CHECK-NEXT: fldl [[mem]]
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: subl $12, %esp
+; CHECK-NEXT: calll foo
+; CHECK-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0]
+; CHECK-NEXT: addsd {{[0-9]+}}(%esp), %xmm0
+; CHECK-NEXT: movsd %xmm0, (%esp)
+; CHECK-NEXT: fldl (%esp)
+; CHECK-NEXT: addl $12, %esp
+; CHECK-NEXT: retl
entry:
%tmp1 = call <2 x double> @foo( ) ; <<2 x double>> [#uses=1]
%tmp2 = extractelement <2 x double> %tmp1, i32 1 ; <double> [#uses=1]
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
; CHECK-NEXT: shll $12, %ecx
; CHECK-NEXT: movd %ecx, %xmm0
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,1,1]
+; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,0,1]
; CHECK-NEXT: movlpd %xmm0, (%eax)
; CHECK-NEXT: retl
%tmp12 = shl i32 %a, 12
; CHECK-LABEL: t2:
; CHECK: # BB#0:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT: movdqa (%eax), %xmm0
-; CHECK-NEXT: pslldq $12, %xmm0
+; CHECK-NEXT: movaps (%eax), %xmm1
+; CHECK-NEXT: xorps %xmm0, %xmm0
+; CHECK-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[0,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[2,0]
; CHECK-NEXT: retl
%tmp1 = load <4 x float>* %P
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> zeroinitializer, <4 x i32> < i32 4, i32 4, i32 4, i32 0 >
; CHECK-LABEL: t3:
; CHECK: # BB#0:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT: movdqa (%eax), %xmm0
-; CHECK-NEXT: psrldq $8, %xmm0
+; CHECK-NEXT: movaps (%eax), %xmm0
+; CHECK-NEXT: xorps %xmm1, %xmm1
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,3],xmm1[0,0]
; CHECK-NEXT: retl
%tmp1 = load <4 x float>* %P
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> zeroinitializer, <4 x i32> < i32 2, i32 3, i32 4, i32 4 >
; CHECK-LABEL: t4:
; CHECK: # BB#0:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT: movdqa (%eax), %xmm0
-; CHECK-NEXT: psrldq $12, %xmm0
+; CHECK-NEXT: movaps (%eax), %xmm0
+; CHECK-NEXT: xorps %xmm1, %xmm1
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,0],xmm1[0,0]
+; CHECK-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,0]
; CHECK-NEXT: retl
%tmp1 = load <4 x float>* %P
%tmp2 = shufflevector <4 x float> zeroinitializer, <4 x float> %tmp1, <4 x i32> < i32 7, i32 0, i32 0, i32 0 >
define <16 x i8> @t5(<16 x i8> %x) nounwind {
; CHECK-LABEL: t5:
; CHECK: # BB#0:
-; CHECK-NEXT: psrldq $1, %xmm0
+; CHECK-NEXT: pxor %xmm1, %xmm1
+; CHECK-NEXT: pshufb {{.*#+}} xmm1 = zero,xmm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,1]
+; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero
+; CHECK-NEXT: por %xmm1, %xmm0
; CHECK-NEXT: retl
%s = shufflevector <16 x i8> %x, <16 x i8> zeroinitializer, <16 x i32> <i32 1, 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, i32 17>
ret <16 x i8> %s
define <16 x i8> @t7(<16 x i8> %x) nounwind {
; CHECK-LABEL: t7:
; CHECK: # BB#0:
-; CHECK-NEXT: pslldq $13, %xmm0
+; CHECK-NEXT: palignr {{.*#+}} xmm0 = xmm0[3,4,5,6,7,8,9,10,11,12,13,14,15,0,1,2]
; CHECK-NEXT: retl
%s = shufflevector <16 x i8> %x, <16 x i8> undef, <16 x i32> <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, i32 1, i32 2>
ret <16 x i8> %s
; RUN: llc < %s -march=x86 -mattr=+sse2 -mcpu=penryn | FileCheck %s
-; RUN: llc < %s -march=x86 -mattr=+sse2 -mcpu=penryn -x86-experimental-vector-shuffle-lowering | FileCheck %s --check-prefix=CHECK-EXP
define <4 x float> @test(float %a) {
; CHECK-LABEL: test:
-; CHECK: movss {{.*}}, %xmm0
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,1,1]
+; CHECK: insertps $29, {{.*}}, %xmm0
; CHECK-NEXT: retl
-;
-; CHECK-EXP-LABEL: test:
-; CHECK-EXP: insertps $29, {{.*}}, %xmm0
-; CHECK-EXP-NEXT: retl
entry:
%tmp = insertelement <4 x float> zeroinitializer, float %a, i32 1
; CHECK: movd {{.*}}, %xmm0
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,0,1]
; CHECK-NEXT: retl
-;
-; CHECK-EXP-LABEL: test2:
-; CHECK-EXP: movd {{.*}}, %xmm0
-; CHECK-EXP-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,0,1]
-; CHECK-EXP-NEXT: retl
entry:
%tmp7 = insertelement <4 x i32> zeroinitializer, i32 %a, i32 2
define <4 x float> @test3(<4 x float> %A) {
; CHECK-LABEL: test3:
-; CHECK: xorps %[[X1:xmm[0-9]+]], %[[X1]]
-; CHECK-NEXT: blendps $1, %xmm0, %[[X1]]
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = [[X1]][1,0,1,1]
+; CHECK: insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
; CHECK-NEXT: retl
-;
-; CHECK-EXP-LABEL: test3:
-; CHECK-EXP: insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
-; CHECK-EXP-NEXT: retl
%tmp0 = extractelement <4 x float> %A, i32 0
%tmp1 = insertelement <4 x float> <float 0.000000e+00, float undef, float undef, float undef >, float %tmp0, i32 1
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 -o %t
-; RUN: grep movlhps %t | count 1
-; RUN: grep movq %t | count 2
-
-define <4 x float> @test1(float %a, float %b) nounwind {
- %tmp = insertelement <4 x float> zeroinitializer, float %a, i32 0 ; <<4 x float>> [#uses=1]
- %tmp6 = insertelement <4 x float> %tmp, float 0.000000e+00, i32 1 ; <<4 x float>> [#uses=1]
- %tmp8 = insertelement <4 x float> %tmp6, float %b, i32 2 ; <<4 x float>> [#uses=1]
- %tmp9 = insertelement <4 x float> %tmp8, float 0.000000e+00, i32 3 ; <<4 x float>> [#uses=1]
- ret <4 x float> %tmp9
-}
-
-define <4 x float> @test2(float %a, float %b) nounwind {
- %tmp = insertelement <4 x float> zeroinitializer, float %a, i32 0 ; <<4 x float>> [#uses=1]
- %tmp7 = insertelement <4 x float> %tmp, float %b, i32 1 ; <<4 x float>> [#uses=1]
- %tmp8 = insertelement <4 x float> %tmp7, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
- %tmp9 = insertelement <4 x float> %tmp8, float 0.000000e+00, i32 3 ; <<4 x float>> [#uses=1]
- ret <4 x float> %tmp9
-}
-
-define <2 x i64> @test3(i32 %a, i32 %b) nounwind {
- %tmp = insertelement <4 x i32> zeroinitializer, i32 %a, i32 0 ; <<4 x i32>> [#uses=1]
- %tmp6 = insertelement <4 x i32> %tmp, i32 %b, i32 1 ; <<4 x i32>> [#uses=1]
- %tmp8 = insertelement <4 x i32> %tmp6, i32 0, i32 2 ; <<4 x i32>> [#uses=1]
- %tmp10 = insertelement <4 x i32> %tmp8, i32 0, i32 3 ; <<4 x i32>> [#uses=1]
- %tmp11 = bitcast <4 x i32> %tmp10 to <2 x i64> ; <<2 x i64>> [#uses=1]
- ret <2 x i64> %tmp11
-}
+++ /dev/null
-; RUN: llc < %s -march=x86-64 -mattr=-avx,-pad-short-functions | FileCheck %s
-
-; CHECK: test3
-; CHECK: movd
-; CHECK-NOT: movd
-; CHECK: {{movlhps.*%xmm0, %xmm0}}
-; CHECK-NEXT: ret
-
-define <2 x i64> @test3(i64 %A) nounwind {
-entry:
- %B = insertelement <2 x i64> undef, i64 %A, i32 1
- ret <2 x i64> %B
-}
-
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | grep movq
-
-define <4 x float> @t(float %X) nounwind {
- %tmp11 = insertelement <4 x float> undef, float %X, i32 0
- %tmp12 = insertelement <4 x float> %tmp11, float %X, i32 1
- %tmp27 = insertelement <4 x float> %tmp12, float 0.000000e+00, i32 2
- %tmp28 = insertelement <4 x float> %tmp27, float 0.000000e+00, i32 3
- ret <4 x float> %tmp28
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s
-
-; CHECK-NOT: xorp
-; CHECK: movd
-; CHECK-NOT: xorp
-
-define void @t1() nounwind {
- %tmp298.i.i = load <4 x float>* null, align 16
- %tmp304.i.i = bitcast <4 x float> %tmp298.i.i to <4 x i32>
- %tmp305.i.i = and <4 x i32> %tmp304.i.i, < i32 -1, i32 0, i32 0, i32 0 >
- store <4 x i32> %tmp305.i.i, <4 x i32>* null, align 16
- unreachable
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | grep movss
-; PR2472
-
-define <4 x i32> @a(<4 x i32> %a) nounwind {
-entry:
- %vecext = extractelement <4 x i32> %a, i32 0
- insertelement <4 x i32> zeroinitializer, i32 %vecext, i32 0
- %add = add <4 x i32> %a, %0
- ret <4 x i32> %add
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2
-; RUN: llc < %s -march=x86 -mattr=+sse2 -mtriple=i386-apple-darwin | not grep mov
-
-define <4 x i32> @test() nounwind {
- %tmp131 = call <2 x i64> @llvm.x86.sse2.psrl.dq( <2 x i64> < i64 -1, i64 -1 >, i32 96 ) ; <<2 x i64>> [#uses=1]
- %tmp137 = bitcast <2 x i64> %tmp131 to <4 x i32> ; <<4 x i32>> [#uses=1]
- %tmp138 = and <4 x i32> %tmp137, bitcast (<2 x i64> < i64 -1, i64 -1 > to <4 x i32>) ; <<4 x i32>> [#uses=1]
- ret <4 x i32> %tmp138
-}
-
-declare <2 x i64> @llvm.x86.sse2.psrl.dq(<2 x i64>, i32)
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2,-avx | FileCheck %s -check-prefix=X86-32
-; RUN: llc < %s -march=x86-64 -mattr=+sse2,-avx | FileCheck %s -check-prefix=X86-64
-
-define <4 x i32> @t1(i32 %a) nounwind {
-entry:
- %tmp = insertelement <4 x i32> undef, i32 %a, i32 0
- %tmp6 = shufflevector <4 x i32> zeroinitializer, <4 x i32> %tmp, <4 x i32> < i32 4, i32 1, i32 2, i32 3 > ; <<4 x i32>> [#uses=1]
- ret <4 x i32> %tmp6
-
-; X86-32-LABEL: t1:
-; X86-32: movd 4(%esp), %xmm0
-
-; X86-64-LABEL: t1:
-; X86-64: movd %e{{..}}, %xmm0
-}
-
-define <2 x i64> @t2(i64 %a) nounwind {
-entry:
- %tmp = insertelement <2 x i64> undef, i64 %a, i32 0
- %tmp6 = shufflevector <2 x i64> zeroinitializer, <2 x i64> %tmp, <2 x i32> < i32 2, i32 1 > ; <<4 x i32>> [#uses=1]
- ret <2 x i64> %tmp6
-
-; X86-32-LABEL: t2:
-; X86-32: movq 4(%esp), %xmm0
-
-; X86-64-LABEL: t2:
-; X86-64: movd %r{{..}}, %xmm0
-}
-
-define <2 x i64> @t3(<2 x i64>* %a) nounwind {
-entry:
- %tmp4 = load <2 x i64>* %a, align 16 ; <<2 x i64>> [#uses=1]
- %tmp6 = bitcast <2 x i64> %tmp4 to <4 x i32> ; <<4 x i32>> [#uses=1]
- %tmp7 = shufflevector <4 x i32> zeroinitializer, <4 x i32> %tmp6, <4 x i32> < i32 4, i32 5, i32 2, i32 3 > ; <<4 x i32>> [#uses=1]
- %tmp8 = bitcast <4 x i32> %tmp7 to <2 x i64> ; <<2 x i64>> [#uses=1]
- ret <2 x i64> %tmp8
-
-; X86-32-LABEL: t3:
-; X86-32: movl 4(%esp)
-; X86-32: movq
-
-; X86-64-LABEL: t3:
-; X86-64: movq ({{.*}}), %xmm0
-}
-
-define <2 x i64> @t4(<2 x i64> %a) nounwind {
-entry:
- %tmp5 = bitcast <2 x i64> %a to <4 x i32> ; <<4 x i32>> [#uses=1]
- %tmp6 = shufflevector <4 x i32> zeroinitializer, <4 x i32> %tmp5, <4 x i32> < i32 4, i32 5, i32 2, i32 3 > ; <<4 x i32>> [#uses=1]
- %tmp7 = bitcast <4 x i32> %tmp6 to <2 x i64> ; <<2 x i64>> [#uses=1]
- ret <2 x i64> %tmp7
-
-; X86-32-LABEL: t4:
-; X86-32: movq %xmm0, %xmm0
-
-; X86-64-LABEL: t4:
-; X86-64: movq {{.*}}, %xmm0
-}
-
-define <2 x i64> @t5(<2 x i64> %a) nounwind {
-entry:
- %tmp6 = shufflevector <2 x i64> zeroinitializer, <2 x i64> %a, <2 x i32> < i32 2, i32 1 > ; <<4 x i32>> [#uses=1]
- ret <2 x i64> %tmp6
-
-; X86-32-LABEL: t5:
-; X86-32: movq %xmm0, %xmm0
-
-; X86-64-LABEL: t5:
-; X86-64: movq {{.*}}, %xmm0
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2
-
-define <2 x i64> @t00(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 0, i32 0 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t01(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 0, i32 1 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t02(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 0, i32 2 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t03(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 0, i32 3 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t10(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 1, i32 0 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t11(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 1, i32 1 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t12(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 1, i32 2 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t13(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 1, i32 3 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t20(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 2, i32 0 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t21(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 2, i32 1 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t22(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 2, i32 2 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t23(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 2, i32 3 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t30(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 3, i32 0 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t31(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 3, i32 1 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t32(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 3, i32 2 >
- ret <2 x i64> %tmp
-}
-
-define <2 x i64> @t33(<2 x i64> %a, <2 x i64> %b) nounwind {
- %tmp = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> < i32 3, i32 3 >
- ret <2 x i64> %tmp
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=penryn -mattr=+sse,-sse2 -mtriple=i386-apple-darwin | FileCheck %s -check-prefix=sse
-; RUN: llc < %s -march=x86 -mcpu=penryn -mattr=+sse2 -mtriple=i386-apple-darwin | FileCheck %s -check-prefix=sse2
-
-; sse-LABEL: t1:
-; sse2-LABEL: t1:
-define <4 x float> @t1(<4 x float> %a, <4 x float> %b) nounwind {
-; sse: shufps
-; sse2: pshufd
-; sse2-NEXT: ret
- %tmp1 = shufflevector <4 x float> %b, <4 x float> undef, <4 x i32> zeroinitializer
- ret <4 x float> %tmp1
-}
-
-; sse-LABEL: t2:
-; sse2-LABEL: t2:
-define <4 x float> @t2(<4 x float> %A, <4 x float> %B) nounwind {
-; sse: shufps
-; sse2: pshufd
-; sse2-NEXT: ret
- %tmp = shufflevector <4 x float> %A, <4 x float> %B, <4 x i32> < i32 3, i32 3, i32 3, i32 3 >
- ret <4 x float> %tmp
-}
-
-; sse-LABEL: t3:
-; sse2-LABEL: t3:
-define <4 x float> @t3(<4 x float> %A, <4 x float> %B) nounwind {
-; sse: shufps
-; sse2: pshufd
-; sse2-NEXT: ret
- %tmp = shufflevector <4 x float> %A, <4 x float> %B, <4 x i32> < i32 4, i32 4, i32 4, i32 4 >
- ret <4 x float> %tmp
-}
-
-; sse-LABEL: t4:
-; sse2-LABEL: t4:
-define <4 x float> @t4(<4 x float> %A, <4 x float> %B) nounwind {
-
-; sse: shufps
-; sse2: pshufd
-; sse2-NEXT: ret
- %tmp = shufflevector <4 x float> %A, <4 x float> %B, <4 x i32> < i32 1, i32 3, i32 2, i32 0 >
- ret <4 x float> %tmp
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-linux -mattr=-avx | FileCheck %s
-; RUN: llc < %s -mtriple=x86_64-win32 -mattr=-avx | FileCheck %s
-; CHECK-NOT: xor
-; CHECK: movd {{%rdi|%rcx}}, %xmm0
-; CHECK-NOT: xor
-; PR2108
-
-define <2 x i64> @doload64(i64 %x) nounwind {
-entry:
- %tmp717 = bitcast i64 %x to double ; <double> [#uses=1]
- %tmp8 = insertelement <2 x double> undef, double %tmp717, i32 0 ; <<2 x double>> [#uses=1]
- %tmp9 = insertelement <2 x double> %tmp8, double 0.000000e+00, i32 1 ; <<2 x double>> [#uses=1]
- %tmp11 = bitcast <2 x double> %tmp9 to <2 x i64> ; <<2 x i64>> [#uses=1]
- ret <2 x i64> %tmp11
-}
-
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 -mtriple=i686-apple-darwin8.8.0 | grep mov | count 7
-
- %struct.vector4_t = type { <4 x float> }
-
-define void @swizzle(i8* %a, %struct.vector4_t* %b, %struct.vector4_t* %c) nounwind {
-entry:
- %tmp9 = getelementptr %struct.vector4_t* %b, i32 0, i32 0 ; <<4 x float>*> [#uses=2]
- %tmp10 = load <4 x float>* %tmp9, align 16 ; <<4 x float>> [#uses=1]
- %tmp14 = bitcast i8* %a to double* ; <double*> [#uses=1]
- %tmp15 = load double* %tmp14 ; <double> [#uses=1]
- %tmp16 = insertelement <2 x double> undef, double %tmp15, i32 0 ; <<2 x double>> [#uses=1]
- %tmp18 = bitcast <2 x double> %tmp16 to <4 x float> ; <<4 x float>> [#uses=1]
- %tmp19 = shufflevector <4 x float> %tmp10, <4 x float> %tmp18, <4 x i32> < i32 4, i32 5, i32 2, i32 3 > ; <<4 x float>> [#uses=1]
- store <4 x float> %tmp19, <4 x float>* %tmp9, align 16
- %tmp28 = getelementptr %struct.vector4_t* %c, i32 0, i32 0 ; <<4 x float>*> [#uses=2]
- %tmp29 = load <4 x float>* %tmp28, align 16 ; <<4 x float>> [#uses=1]
- %tmp26 = getelementptr i8* %a, i32 8 ; <i8*> [#uses=1]
- %tmp33 = bitcast i8* %tmp26 to double* ; <double*> [#uses=1]
- %tmp34 = load double* %tmp33 ; <double> [#uses=1]
- %tmp35 = insertelement <2 x double> undef, double %tmp34, i32 0 ; <<2 x double>> [#uses=1]
- %tmp37 = bitcast <2 x double> %tmp35 to <4 x float> ; <<4 x float>> [#uses=1]
- %tmp38 = shufflevector <4 x float> %tmp29, <4 x float> %tmp37, <4 x i32> < i32 4, i32 5, i32 2, i32 3 > ; <<4 x float>> [#uses=1]
- store <4 x float> %tmp38, <4 x float>* %tmp28, align 16
- ret void
-}
+++ /dev/null
-; REQUIRES: asserts
-; RUN: llc < %s -o /dev/null -march=x86 -mcpu=penryn -mattr=+sse2 -mtriple=i686-apple-darwin9 -stats -info-output-file - | grep asm-printer | grep 4
-; PR2485
-
-define <4 x i32> @t(<4 x i32> %a, <4 x i32> %b) nounwind {
-entry:
- %shuffle = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> < i32 4, i32 0, i32 0, i32 0 > ; <<4 x i32>> [#uses=1]
- ret <4 x i32> %shuffle
-}
+++ /dev/null
-; REQUIRES: asserts
-; RUN: llc < %s -o /dev/null -march=x86 -mcpu=corei7 -mtriple=i686-apple-darwin9 -stats -info-output-file - | grep asm-printer | grep 2
-
-define <4 x float> @func(<4 x float> %fp0, <4 x float> %fp1) nounwind {
-entry:
- shufflevector <4 x float> %fp0, <4 x float> %fp1, <4 x i32> < i32 0, i32 1, i32 2, i32 7 > ; <<4 x float>>:0 [#uses=1]
- ret <4 x float> %0
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=pentium-m | FileCheck %s
-
-define <4 x float> @t1(<4 x float> %a) nounwind {
-; CHECK: movlhps
- %tmp1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> < i32 0, i32 1, i32 0, i32 1 > ; <<4 x float>> [#uses=1]
- ret <4 x float> %tmp1
-}
-
-define <4 x i32> @t2(<4 x i32>* %a) nounwind {
-; CHECK: pshufd
-; CHECK: ret
- %tmp1 = load <4 x i32>* %a
- %tmp2 = shufflevector <4 x i32> %tmp1, <4 x i32> undef, <4 x i32> < i32 0, i32 1, i32 0, i32 1 > ; <<4 x i32>> [#uses=1]
- ret <4 x i32> %tmp2
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | not grep punpck
-; RUN: llc < %s -march=x86 -mattr=+sse2 | grep pshufd
-
-define i32 @t() nounwind {
-entry:
- %a = alloca <4 x i32> ; <<4 x i32>*> [#uses=2]
- %b = alloca <4 x i32> ; <<4 x i32>*> [#uses=5]
- store volatile <4 x i32> < i32 0, i32 1, i32 2, i32 3 >, <4 x i32>* %a
- %tmp = load <4 x i32>* %a ; <<4 x i32>> [#uses=1]
- store <4 x i32> %tmp, <4 x i32>* %b
- %tmp1 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %tmp2 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %punpckldq = shufflevector <4 x i32> %tmp1, <4 x i32> %tmp2, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x i32>> [#uses=1]
- store <4 x i32> %punpckldq, <4 x i32>* %b
- %tmp3 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %result = extractelement <4 x i32> %tmp3, i32 0 ; <i32> [#uses=1]
- ret i32 %result
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s
-
-define i32 @t() nounwind optsize {
-entry:
-; CHECK: punpckldq
- %a = alloca <4 x i32> ; <<4 x i32>*> [#uses=2]
- %b = alloca <4 x i32> ; <<4 x i32>*> [#uses=5]
- store volatile <4 x i32> < i32 0, i32 1, i32 2, i32 3 >, <4 x i32>* %a
- %tmp = load <4 x i32>* %a ; <<4 x i32>> [#uses=1]
- store <4 x i32> %tmp, <4 x i32>* %b
- %tmp1 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %tmp2 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %punpckldq = shufflevector <4 x i32> %tmp1, <4 x i32> %tmp2, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x i32>> [#uses=1]
- store <4 x i32> %punpckldq, <4 x i32>* %b
- %tmp3 = load <4 x i32>* %b ; <<4 x i32>> [#uses=1]
- %result = extractelement <4 x i32> %tmp3, i32 0 ; <i32> [#uses=1]
- ret i32 %result
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=sse4.1 -o %t
-; RUN: grep unpcklps %t | count 3
-; RUN: grep unpckhps %t | count 1
-
-; Transpose example using the more generic vector shuffle. We return
-; float8 instead of float16 since x86 can return that in register.
-; ModuleID = 'transpose2_opt.bc'
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
-target triple = "i386-apple-cl.1.0"
-@r0 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r1 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r2 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r3 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-
-define <8 x float> @__transpose2(<4 x float> %p0, <4 x float> %p1, <4 x float> %p2, <4 x float> %p3) nounwind {
-entry:
- %unpcklps = shufflevector <4 x float> %p0, <4 x float> %p2, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=2]
- %unpckhps = shufflevector <4 x float> %p0, <4 x float> %p2, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=2]
- %unpcklps8 = shufflevector <4 x float> %p1, <4 x float> %p3, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=2]
- %unpckhps11 = shufflevector <4 x float> %p1, <4 x float> %p3, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=2]
- %unpcklps14 = shufflevector <4 x float> %unpcklps, <4 x float> %unpcklps8, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=1]
- %unpcklps14a = shufflevector <4 x float> %unpcklps14, <4 x float> undef, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- %unpckhps17 = shufflevector <4 x float> %unpcklps, <4 x float> %unpcklps8, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=1]
- %unpckhps17a = shufflevector <4 x float> %unpckhps17, <4 x float> undef, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- %r1 = shufflevector <16 x float> %unpcklps14a, <16 x float> %unpckhps17a, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
- %unpcklps20 = shufflevector <4 x float> %unpckhps, <4 x float> %unpckhps11, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=1]
- %unpcklps20a = shufflevector <4 x float> %unpcklps20, <4 x float> undef, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- %r2 = shufflevector <16 x float> %r1, <16 x float> %unpcklps20a, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
- %unpckhps23 = shufflevector <4 x float> %unpckhps, <4 x float> %unpckhps11, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=1]
- %unpckhps23a = shufflevector <4 x float> %unpckhps23, <4 x float> undef, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- %r3 = shufflevector <16 x float> %r2, <16 x float> %unpckhps23a, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
- %r4 = shufflevector <16 x float> %r3, <16 x float> undef, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
- ret <8 x float> %r4
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=generic -mattr=sse4.1 | FileCheck %s
-; RUN: llc < %s -march=x86 -mcpu=atom | FileCheck -check-prefix=ATOM %s
-
-; Transpose example using the more generic vector shuffle. Return float8
-; instead of float16
-; ModuleID = 'transpose2_opt.bc'
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
-target triple = "i386-apple-cl.1.0"
-@r0 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r1 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r2 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-@r3 = common global <4 x float> zeroinitializer, align 16 ; <<4 x float>*> [#uses=1]
-
-define <8 x float> @__transpose2(<4 x float> %p0, <4 x float> %p1, <4 x float> %p2, <4 x float> %p3) nounwind {
-entry:
-; CHECK: transpose2
-; CHECK: unpckhps
-; CHECK: unpckhps
-; CHECK: unpcklps
-; CHECK: unpckhps
-; Different instruction order for Atom.
-; ATOM: transpose2
-; ATOM: unpckhps
-; ATOM: unpckhps
-; ATOM: unpckhps
-; ATOM: unpcklps
- %unpcklps = shufflevector <4 x float> %p0, <4 x float> %p2, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=2]
- %unpckhps = shufflevector <4 x float> %p0, <4 x float> %p2, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=2]
- %unpcklps8 = shufflevector <4 x float> %p1, <4 x float> %p3, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=2]
- %unpckhps11 = shufflevector <4 x float> %p1, <4 x float> %p3, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=2]
- %unpcklps14 = shufflevector <4 x float> %unpcklps, <4 x float> %unpcklps8, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=1]
- %unpckhps17 = shufflevector <4 x float> %unpcklps, <4 x float> %unpcklps8, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=1]
- %r1 = shufflevector <4 x float> %unpcklps14, <4 x float> %unpckhps17, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 >
- %unpcklps20 = shufflevector <4 x float> %unpckhps, <4 x float> %unpckhps11, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>> [#uses=1]
- %unpckhps23 = shufflevector <4 x float> %unpckhps, <4 x float> %unpckhps11, <4 x i32> < i32 2, i32 6, i32 3, i32 7 > ; <<4 x float>> [#uses=1]
- %r2 = shufflevector <4 x float> %unpcklps20, <4 x float> %unpckhps23, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 >
-; %r3 = shufflevector <8 x float> %r1, <8 x float> %r2, <16 x i32> < i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15 >;
- ret <8 x float> %r2
-}
-
-define <2 x i64> @lo_hi_shift(float* nocapture %x, float* nocapture %y) nounwind {
-entry:
-; movhps should happen before extractps to assure it gets the correct value.
-; CHECK: lo_hi_shift
-; CHECK: movhps ([[BASEREG:%[a-z]+]]),
-; CHECK: extractps ${{[0-9]+}}, %xmm{{[0-9]+}}, {{[0-9]*}}([[BASEREG]])
-; CHECK: extractps ${{[0-9]+}}, %xmm{{[0-9]+}}, {{[0-9]*}}([[BASEREG]])
-; ATOM: lo_hi_shift
-; ATOM: movhps ([[BASEREG:%[a-z]+]]),
-; ATOM: movd %xmm{{[0-9]+}}, {{[0-9]*}}([[BASEREG]])
-; ATOM: movd %xmm{{[0-9]+}}, {{[0-9]*}}([[BASEREG]])
- %v.i = bitcast float* %y to <4 x float>*
- %0 = load <4 x float>* %v.i, align 1
- %1 = bitcast float* %x to <1 x i64>*
- %.val = load <1 x i64>* %1, align 1
- %2 = bitcast <1 x i64> %.val to <2 x float>
- %shuffle.i = shufflevector <2 x float> %2, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
- %shuffle1.i = shufflevector <4 x float> %0, <4 x float> %shuffle.i, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
- %cast.i = bitcast <4 x float> %0 to <2 x i64>
- %extract.i = extractelement <2 x i64> %cast.i, i32 1
- %3 = bitcast float* %x to i64*
- store i64 %extract.i, i64* %3, align 4
- %4 = bitcast <4 x float> %0 to <16 x i8>
- %5 = bitcast <4 x float> %shuffle1.i to <16 x i8>
- %palignr = shufflevector <16 x i8> %5, <16 x i8> %4, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
- %6 = bitcast <16 x i8> %palignr to <2 x i64>
- ret <2 x i64> %6
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=penryn -mattr=sse4.1 | FileCheck %s
-
-; ModuleID = 'vec_shuffle-27.bc'
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
-target triple = "i686-apple-cl.1.0"
-
-define <8 x float> @my2filter4_1d(<4 x float> %a, <8 x float> %T0, <8 x float> %T1) nounwind readnone {
-entry:
-; CHECK: subps
-; CHECK: subps
-; CHECK: mulps
-; CHECK: mulps
-; CHECK: addps
-; CHECK: addps
- %tmp7 = shufflevector <4 x float> %a, <4 x float> undef, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3 > ; <<8 x float>> [#uses=1]
- %sub = fsub <8 x float> %T1, %T0 ; <<8 x float>> [#uses=1]
- %mul = fmul <8 x float> %sub, %tmp7 ; <<8 x float>> [#uses=1]
- %add = fadd <8 x float> %mul, %T0 ; <<8 x float>> [#uses=1]
- ret <8 x float> %add
-}
-
-; Test case for r122206
-define void @test2(<4 x i64>* %ap, <4 x i64>* %bp) nounwind {
-entry:
-; CHECK: movdqa
- %a = load <4 x i64> * %ap
- %b = load <4 x i64> * %bp
- %mulaa = mul <4 x i64> %a, %a
- %mulbb = mul <4 x i64> %b, %b
- %mulab = mul <4 x i64> %a, %b
- %vect1271 = shufflevector <4 x i64> %mulaa, <4 x i64> %mulbb, <4 x i32> <i32 0, i32 4, i32 undef, i32 undef>
- %vect1272 = shufflevector <4 x i64> %mulaa, <4 x i64> %mulbb, <4 x i32> <i32 1, i32 5, i32 undef, i32 undef>
- %vect1487 = shufflevector <4 x i64> %vect1271, <4 x i64> %mulab, <4 x i32> <i32 0, i32 1, i32 2, i32 4>
- %vect1488 = shufflevector <4 x i64> %vect1272, <4 x i64> %mulab, <4 x i32> <i32 0, i32 1, i32 2, i32 5>
- store <4 x i64> %vect1487, <4 x i64>* %ap
- store <4 x i64> %vect1488, <4 x i64>* %bp
- ret void;
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=core2 | FileCheck %s
-
-; CHECK: pshufb
-; CHECK-NOT: pshufb
-
-; FIXME: this test has a superfluous punpcklqdq pre-pshufb currently.
-; Don't XFAIL it because it's still better than the previous code.
-
-; Pack various elements via shuffles.
-define <8 x i16> @shuf1(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
-entry:
- %tmp7 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 1, i32 8, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef , i32 undef >
- ret <8 x i16> %tmp7
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mattr=+avx | FileCheck %s
-
-; CHECK: test
-; Test case when creating pshufhw, we incorrectly set the higher order bit
-; for an undef,
-define void @test(<8 x i16>* %dest, <8 x i16> %in) nounwind {
-entry:
-; CHECK-NOT: vmovaps
-; CHECK: vmovlpd
-; CHECK: vpshufhw $-95
- %0 = load <8 x i16>* %dest
- %1 = shufflevector <8 x i16> %0, <8 x i16> %in, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 13, i32 undef, i32 14, i32 14>
- store <8 x i16> %1, <8 x i16>* %dest
- ret void
-}
-
-; CHECK: test2
-; A test case where we shouldn't generate a punpckldq but a pshufd and a pslldq
-define void @test2(<4 x i32>* %dest, <4 x i32> %in) nounwind {
-entry:
-; CHECK-NOT: pslldq
-; CHECK: shufps
- %0 = shufflevector <4 x i32> %in, <4 x i32> <i32 0, i32 0, i32 0, i32 0>, <4 x i32> < i32 undef, i32 5, i32 undef, i32 2>
- store <4 x i32> %0, <4 x i32>* %dest
- ret void
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=core2 -o %t
-; RUN: grep pshufb %t | count 1
-
-define <8 x i16> @shuf3(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
-entry:
- %tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 0, i32 1, i32 undef, i32 undef, i32 3, i32 11, i32 undef , i32 undef >
- ret <8 x i16> %tmp9
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=core2 | grep pshufb | count 2
-
-define <8 x i16> @shuf2(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
-entry:
- %tmp8 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 undef, i32 undef, i32 7, i32 2, i32 8, i32 undef, i32 undef , i32 undef >
- ret <8 x i16> %tmp8
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=yonah -stack-alignment=16 -o %t
-; RUN: grep pextrw %t | count 12
-; RUN: grep pinsrw %t | count 13
-; RUN: grep rolw %t | count 13
-; RUN: not grep esp %t
-; RUN: not grep ebp %t
-; RUN: llc < %s -march=x86 -mcpu=core2 -stack-alignment=16 -o %t
-; RUN: grep pshufb %t | count 3
-
-define <16 x i8> @shuf1(<16 x i8> %T0) nounwind readnone {
-entry:
- %tmp8 = shufflevector <16 x i8> %T0, <16 x i8> undef, <16 x i32> < i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 12, i32 13, i32 15 , i32 14 >
- ret <16 x i8> %tmp8
-}
-
-define <16 x i8> @shuf2(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
-entry:
- %tmp8 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> < i32 undef, i32 undef, i32 3, i32 2, i32 17, i32 16, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 12, i32 13, i32 15 , i32 14 >
- ret <16 x i8> %tmp8
-}
+++ /dev/null
-; RUN: llc < %s -march=x86-64 -mcpu=penryn -mattr=sse4.1 | FileCheck %s
-
-define <8 x i16> @shuf6(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
-; CHECK: pshufb
-; CHECK-NOT: pshufb
-; CHECK: ret
-entry:
- %tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 3, i32 2, i32 0, i32 2, i32 1, i32 5, i32 6 , i32 undef >
- ret <8 x i16> %tmp9
-}
-
-define <8 x i16> @shuf7(<8 x i16> %t0) {
-; CHECK: pshufd
- %tmp10 = shufflevector <8 x i16> %t0, <8 x i16> undef, <8 x i32> < i32 undef, i32 2, i32 2, i32 2, i32 2, i32 2, i32 undef, i32 undef >
- ret <8 x i16> %tmp10
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-linux -mcpu=core2 | FileCheck %s
-; RUN: llc < %s -mtriple=x86_64-win32 -mcpu=core2 | FileCheck %s
-; RUN: llc -O0 < %s -march=x86 -mcpu=core2 | FileCheck %s --check-prefix=CHECK_O0
-
-define <4 x i32> @t00(<4 x i32>* %a0) nounwind ssp {
-entry:
-; CHECK: movaps ({{%rdi|%rcx}}), %[[XMM0:xmm[0-9]+]]
-; CHECK: movaps %[[XMM0]], %[[XMM1:xmm[0-9]+]]
-; CHECK-NEXT: movss %xmm{{[0-9]+}}, %[[XMM1]]
-; CHECK-NEXT: shufps $36, %[[XMM1]], %[[XMM0]]
- %0 = load <4 x i32>* undef, align 16
- %1 = load <4 x i32>* %a0, align 16
- %2 = shufflevector <4 x i32> %1, <4 x i32> %0, <4 x i32> <i32 0, i32 1, i32 2, i32 4>
- ret <4 x i32> %2
-}
-
-define void @t01(double* %a0) nounwind ssp {
-entry:
-; CHECK_O0: movsd (%eax), %xmm0
-; CHECK_O0: unpcklpd %xmm0, %xmm0
- %tmp93 = load double* %a0, align 8
- %vecinit94 = insertelement <2 x double> undef, double %tmp93, i32 1
- store <2 x double> %vecinit94, <2 x double>* undef
- ret void
-}
-
-define void @t02(<8 x i32>* %source, <2 x i32>* %dest) nounwind noinline {
-entry:
-; CHECK: t02
-; CHECK: movaps
-; CHECK: shufps
-; CHECK: pshufd
-; CHECK: movq
-; CHECK: ret
- %0 = bitcast <8 x i32>* %source to <4 x i32>*
- %arrayidx = getelementptr inbounds <4 x i32>* %0, i64 3
- %tmp2 = load <4 x i32>* %arrayidx, align 16
- %tmp3 = extractelement <4 x i32> %tmp2, i32 0
- %tmp5 = insertelement <2 x i32> <i32 undef, i32 0>, i32 %tmp3, i32 0
- %arrayidx7 = getelementptr inbounds <8 x i32>* %source, i64 1
- %1 = bitcast <8 x i32>* %arrayidx7 to <4 x i32>*
- %tmp8 = load <4 x i32>* %1, align 16
- %tmp9 = extractelement <4 x i32> %tmp8, i32 1
- %tmp11 = insertelement <2 x i32> %tmp5, i32 %tmp9, i32 1
- store <2 x i32> %tmp11, <2 x i32>* %dest, align 8
- ret void
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=corei7 | FileCheck %s
-
-define <2 x double> @ld(<2 x double> %p) nounwind optsize ssp {
-; CHECK: unpcklpd
- %shuffle = shufflevector <2 x double> %p, <2 x double> undef, <2 x i32> zeroinitializer
- ret <2 x double> %shuffle
-}
-
-define <2 x double> @hd(<2 x double> %p) nounwind optsize ssp {
-; CHECK: movhlps
- %shuffle = shufflevector <2 x double> %p, <2 x double> undef, <2 x i32> <i32 1, i32 1>
- ret <2 x double> %shuffle
-}
-
-define <2 x i64> @ldi(<2 x i64> %p) nounwind optsize ssp {
-; CHECK: punpcklqdq
- %shuffle = shufflevector <2 x i64> %p, <2 x i64> undef, <2 x i32> zeroinitializer
- ret <2 x i64> %shuffle
-}
-
-define <2 x i64> @hdi(<2 x i64> %p) nounwind optsize ssp {
-; CHECK: punpckhqdq
- %shuffle = shufflevector <2 x i64> %p, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
- ret <2 x i64> %shuffle
-}
-
-; rdar://10050549
-%struct.Float2 = type { float, float }
-
-define <4 x float> @loadhpi(%struct.Float2* %vPtr, <4 x float> %vecin1) nounwind readonly ssp {
-entry:
-; CHECK: loadhpi
-; CHECK-NOT: movq
-; CHECK: movhps (
- %tmp1 = bitcast %struct.Float2* %vPtr to <1 x i64>*
- %addptr7 = getelementptr inbounds <1 x i64>* %tmp1, i64 0
- %tmp2 = bitcast <1 x i64>* %addptr7 to float*
- %tmp3 = load float* %tmp2, align 4
- %vec = insertelement <4 x float> undef, float %tmp3, i32 0
- %addptr.i12 = getelementptr inbounds float* %tmp2, i64 1
- %tmp4 = load float* %addptr.i12, align 4
- %vecin2 = insertelement <4 x float> %vec, float %tmp4, i32 1
- %shuffle = shufflevector <4 x float> %vecin1, <4 x float> %vecin2, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
- ret <4 x float> %shuffle
-}
-
-; rdar://10119696
-; CHECK: f
-define <4 x float> @f(<4 x float> %x, double* nocapture %y) nounwind readonly ssp {
-entry:
- ; CHECK: movlps (%{{rdi|rdx}}), %xmm0
- %u110.i = load double* %y, align 1
- %tmp8.i = insertelement <2 x double> undef, double %u110.i, i32 0
- %tmp9.i = bitcast <2 x double> %tmp8.i to <4 x float>
- %shuffle.i = shufflevector <4 x float> %x, <4 x float> %tmp9.i, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
- ret <4 x float> %shuffle.i
-}
-
-define <4 x float> @loadhpi2(%struct.Float2* nocapture %vHiCoefPtr_0, %struct.Float2* nocapture %vLoCoefPtr_0, i32 %s) nounwind readonly ssp {
-entry:
-; CHECK: loadhpi2
-; CHECK: movhps (
-; CHECK-NOT: movlhps
- %0 = bitcast %struct.Float2* %vHiCoefPtr_0 to <1 x i64>*
- %idx.ext = sext i32 %s to i64
- %add.ptr = getelementptr inbounds <1 x i64>* %0, i64 %idx.ext
- %add.ptr.val = load <1 x i64>* %add.ptr, align 1
- %1 = bitcast <1 x i64> %add.ptr.val to <2 x float>
- %shuffle.i = shufflevector <2 x float> %1, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
- %2 = bitcast %struct.Float2* %vLoCoefPtr_0 to <1 x i64>*
- %add.ptr2 = getelementptr inbounds <1 x i64>* %2, i64 %idx.ext
- %add.ptr2.val = load <1 x i64>* %add.ptr2, align 1
- %3 = bitcast <1 x i64> %add.ptr2.val to <2 x float>
- %shuffle.i4 = shufflevector <2 x float> %3, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
- %shuffle1.i5 = shufflevector <4 x float> %shuffle.i, <4 x float> %shuffle.i4, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
- ret <4 x float> %shuffle1.i5
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-linux -mcpu=penryn | FileCheck %s
-; rdar://10050222, rdar://10134392
-
-define <4 x float> @t1(<4 x float> %a, <1 x i64>* nocapture %p) nounwind {
-entry:
-; CHECK-LABEL: t1:
-; CHECK: movlps (%rdi), %xmm0
-; CHECK: ret
- %p.val = load <1 x i64>* %p, align 1
- %0 = bitcast <1 x i64> %p.val to <2 x float>
- %shuffle.i = shufflevector <2 x float> %0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
- %shuffle1.i = shufflevector <4 x float> %a, <4 x float> %shuffle.i, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
- ret <4 x float> %shuffle1.i
-}
-
-define <4 x float> @t1a(<4 x float> %a, <1 x i64>* nocapture %p) nounwind {
-entry:
-; CHECK-LABEL: t1a:
-; CHECK: movlps (%rdi), %xmm0
-; CHECK: ret
- %0 = bitcast <1 x i64>* %p to double*
- %1 = load double* %0
- %2 = insertelement <2 x double> undef, double %1, i32 0
- %3 = bitcast <2 x double> %2 to <4 x float>
- %4 = shufflevector <4 x float> %a, <4 x float> %3, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
- ret <4 x float> %4
-}
-
-define void @t2(<1 x i64>* nocapture %p, <4 x float> %a) nounwind {
-entry:
-; CHECK-LABEL: t2:
-; CHECK: movlps %xmm0, (%rdi)
-; CHECK: ret
- %cast.i = bitcast <4 x float> %a to <2 x i64>
- %extract.i = extractelement <2 x i64> %cast.i, i32 0
- %0 = getelementptr inbounds <1 x i64>* %p, i64 0, i64 0
- store i64 %extract.i, i64* %0, align 8
- ret void
-}
-
-define void @t2a(<1 x i64>* nocapture %p, <4 x float> %a) nounwind {
-entry:
-; CHECK-LABEL: t2a:
-; CHECK: movlps %xmm0, (%rdi)
-; CHECK: ret
- %0 = bitcast <1 x i64>* %p to double*
- %1 = bitcast <4 x float> %a to <2 x double>
- %2 = extractelement <2 x double> %1, i32 0
- store double %2, double* %0
- ret void
-}
-
-; rdar://10436044
-define <2 x double> @t3() nounwind readonly {
-bb:
-; CHECK-LABEL: t3:
-; CHECK: movq (%rax), %xmm1
-; CHECK: punpcklqdq %xmm2, %xmm0
-; CHECK: movsd %xmm1, %xmm0
- %tmp0 = load i128* null, align 1
- %tmp1 = load <2 x i32>* undef, align 8
- %tmp2 = bitcast i128 %tmp0 to <16 x i8>
- %tmp3 = bitcast <2 x i32> %tmp1 to i64
- %tmp4 = insertelement <2 x i64> undef, i64 %tmp3, i32 0
- %tmp5 = bitcast <16 x i8> %tmp2 to <2 x double>
- %tmp6 = bitcast <2 x i64> %tmp4 to <2 x double>
- %tmp7 = shufflevector <2 x double> %tmp5, <2 x double> %tmp6, <2 x i32> <i32 2, i32 1>
- ret <2 x double> %tmp7
-}
-
-; rdar://10450317
-define <2 x i64> @t4() nounwind readonly {
-bb:
-; CHECK-LABEL: t4:
-; CHECK: movq (%rax), %xmm0
-; CHECK: punpcklqdq %{{xmm.}}, %[[XMM:xmm[0-9]]]
-; CHECK: movsd %[[XMM]], %xmm0
- %tmp0 = load i128* null, align 1
- %tmp1 = load <2 x i32>* undef, align 8
- %tmp2 = bitcast i128 %tmp0 to <16 x i8>
- %tmp3 = bitcast <2 x i32> %tmp1 to i64
- %tmp4 = insertelement <2 x i64> undef, i64 %tmp3, i32 0
- %tmp5 = bitcast <16 x i8> %tmp2 to <2 x i64>
- %tmp6 = shufflevector <2 x i64> %tmp4, <2 x i64> %tmp5, <2 x i32> <i32 2, i32 1>
- ret <2 x i64> %tmp6
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 | FileCheck %s
-
-define void @shuffle_v16i16(<16 x i16>* %a) {
-; CHECK-LABEL: shuffle_v16i16:
-; CHECK: vpshufb {{.*}}%ymm
-; CHECK-NOT: vpshufb {{.*}}%xmm
-entry:
- %0 = load <16 x i16>* %a, align 32
- %shuffle = shufflevector <16 x i16> %0, <16 x i16> undef, <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
- store <16 x i16> %shuffle, <16 x i16>* %a, align 32
- ret void
-}
-
-define void @shuffle_v16i16_lanecrossing(<16 x i16>* %a) {
-; CHECK-LABEL: shuffle_v16i16_lanecrossing:
-; CHECK-NOT: vpshufb {{.*}}%ymm
-entry:
- %0 = load <16 x i16>* %a, align 32
- %shuffle = shufflevector <16 x i16> %0, <16 x i16> undef, <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 13, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
- store <16 x i16> %shuffle, <16 x i16>* %a, align 32
- ret void
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx | FileCheck %s
-
-; Use buildFromShuffleMostly which allows this to be generated as two 128-bit
-; shuffles and an insert.
-
-; This is the (somewhat questionable) LLVM IR that is generated for:
-; x8.s0123456 = x8.s1234567; // x8 is a <8 x float> type
-; x8.s7 = f; // f is float
-
-
-define <8 x float> @test1(<8 x float> %a, float %b) {
-; CHECK-LABEL: test1:
-; CHECK: vinsertps
-; CHECK-NOT: vinsertps
-entry:
- %shift = shufflevector <8 x float> %a, <8 x float> undef, <7 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
- %extend = shufflevector <7 x float> %shift, <7 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 undef>
- %insert = insertelement <8 x float> %extend, float %b, i32 7
-
- ret <8 x float> %insert
-}
+++ /dev/null
-; RUN: llc < %s -mtriple=i686-linux -mcpu=core2 | FileCheck %s
-
-; CHECK: test_v4sf
-; CHECK: movq 8(%esp)
-; CHECK: pshufd $80
-define void @test_v4sf(<4 x float>* %P, float %X, float %Y) nounwind {
- %tmp = insertelement <4 x float> zeroinitializer, float %X, i32 0 ; <<4 x float>> [#uses=1]
- %tmp2 = insertelement <4 x float> %tmp, float %X, i32 1 ; <<4 x float>> [#uses=1]
- %tmp4 = insertelement <4 x float> %tmp2, float %Y, i32 2 ; <<4 x float>> [#uses=1]
- %tmp6 = insertelement <4 x float> %tmp4, float %Y, i32 3 ; <<4 x float>> [#uses=1]
- store <4 x float> %tmp6, <4 x float>* %P
- ret void
-}
-
-; CHECK: test_v2sd
-; CHECK: movups 8(%esp)
-; CHECK: movaps
-define void @test_v2sd(<2 x double>* %P, double %X, double %Y) nounwind {
- %tmp = insertelement <2 x double> zeroinitializer, double %X, i32 0 ; <<2 x double>> [#uses=1]
- %tmp2 = insertelement <2 x double> %tmp, double %Y, i32 1 ; <<2 x double>> [#uses=1]
- store <2 x double> %tmp2, <2 x double>* %P
- ret void
-}
-
-; CHECK: test_v8i16
-; CHECK: pshufhw $-58
-; CHECK: movdqa
-define void @test_v8i16(<2 x i64>* %res, <2 x i64>* %A) nounwind {
- %tmp = load <2 x i64>* %A ; <<2 x i64>> [#uses=1]
- %tmp.upgrd.1 = bitcast <2 x i64> %tmp to <8 x i16> ; <<8 x i16>> [#uses=8]
- %tmp.upgrd.2 = extractelement <8 x i16> %tmp.upgrd.1, i32 0 ; <i16> [#uses=1]
- %tmp1 = extractelement <8 x i16> %tmp.upgrd.1, i32 1 ; <i16> [#uses=1]
- %tmp2 = extractelement <8 x i16> %tmp.upgrd.1, i32 2 ; <i16> [#uses=1]
- %tmp3 = extractelement <8 x i16> %tmp.upgrd.1, i32 3 ; <i16> [#uses=1]
- %tmp4 = extractelement <8 x i16> %tmp.upgrd.1, i32 6 ; <i16> [#uses=1]
- %tmp5 = extractelement <8 x i16> %tmp.upgrd.1, i32 5 ; <i16> [#uses=1]
- %tmp6 = extractelement <8 x i16> %tmp.upgrd.1, i32 4 ; <i16> [#uses=1]
- %tmp7 = extractelement <8 x i16> %tmp.upgrd.1, i32 7 ; <i16> [#uses=1]
- %tmp8 = insertelement <8 x i16> undef, i16 %tmp.upgrd.2, i32 0 ; <<8 x i16>> [#uses=1]
- %tmp9 = insertelement <8 x i16> %tmp8, i16 %tmp1, i32 1 ; <<8 x i16>> [#uses=1]
- %tmp10 = insertelement <8 x i16> %tmp9, i16 %tmp2, i32 2 ; <<8 x i16>> [#uses=1]
- %tmp11 = insertelement <8 x i16> %tmp10, i16 %tmp3, i32 3 ; <<8 x i16>> [#uses=1]
- %tmp12 = insertelement <8 x i16> %tmp11, i16 %tmp4, i32 4 ; <<8 x i16>> [#uses=1]
- %tmp13 = insertelement <8 x i16> %tmp12, i16 %tmp5, i32 5 ; <<8 x i16>> [#uses=1]
- %tmp14 = insertelement <8 x i16> %tmp13, i16 %tmp6, i32 6 ; <<8 x i16>> [#uses=1]
- %tmp15 = insertelement <8 x i16> %tmp14, i16 %tmp7, i32 7 ; <<8 x i16>> [#uses=1]
- %tmp15.upgrd.3 = bitcast <8 x i16> %tmp15 to <2 x i64> ; <<2 x i64>> [#uses=1]
- store <2 x i64> %tmp15.upgrd.3, <2 x i64>* %res
- ret void
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=pentium4 -mattr=+sse2 | FileCheck %s
-
-define void @test(<2 x i64>* %P, i8 %x) nounwind {
- %tmp = insertelement <16 x i8> zeroinitializer, i8 %x, i32 0 ; <<16 x i8>> [#uses=1]
- %tmp36 = insertelement <16 x i8> %tmp, i8 %x, i32 1 ; <<16 x i8>> [#uses=1]
- %tmp38 = insertelement <16 x i8> %tmp36, i8 %x, i32 2 ; <<16 x i8>> [#uses=1]
- %tmp40 = insertelement <16 x i8> %tmp38, i8 %x, i32 3 ; <<16 x i8>> [#uses=1]
- %tmp42 = insertelement <16 x i8> %tmp40, i8 %x, i32 4 ; <<16 x i8>> [#uses=1]
- %tmp44 = insertelement <16 x i8> %tmp42, i8 %x, i32 5 ; <<16 x i8>> [#uses=1]
- %tmp46 = insertelement <16 x i8> %tmp44, i8 %x, i32 6 ; <<16 x i8>> [#uses=1]
- %tmp48 = insertelement <16 x i8> %tmp46, i8 %x, i32 7 ; <<16 x i8>> [#uses=1]
- %tmp50 = insertelement <16 x i8> %tmp48, i8 %x, i32 8 ; <<16 x i8>> [#uses=1]
- %tmp52 = insertelement <16 x i8> %tmp50, i8 %x, i32 9 ; <<16 x i8>> [#uses=1]
- %tmp54 = insertelement <16 x i8> %tmp52, i8 %x, i32 10 ; <<16 x i8>> [#uses=1]
- %tmp56 = insertelement <16 x i8> %tmp54, i8 %x, i32 11 ; <<16 x i8>> [#uses=1]
- %tmp58 = insertelement <16 x i8> %tmp56, i8 %x, i32 12 ; <<16 x i8>> [#uses=1]
- %tmp60 = insertelement <16 x i8> %tmp58, i8 %x, i32 13 ; <<16 x i8>> [#uses=1]
- %tmp62 = insertelement <16 x i8> %tmp60, i8 %x, i32 14 ; <<16 x i8>> [#uses=1]
- %tmp64 = insertelement <16 x i8> %tmp62, i8 %x, i32 15 ; <<16 x i8>> [#uses=1]
- %tmp68 = load <2 x i64>* %P ; <<2 x i64>> [#uses=1]
- %tmp71 = bitcast <2 x i64> %tmp68 to <16 x i8> ; <<16 x i8>> [#uses=1]
- %tmp73 = add <16 x i8> %tmp71, %tmp64 ; <<16 x i8>> [#uses=1]
- %tmp73.upgrd.1 = bitcast <16 x i8> %tmp73 to <2 x i64> ; <<2 x i64>> [#uses=1]
- store <2 x i64> %tmp73.upgrd.1, <2 x i64>* %P
- ret void
-
-; CHECK-LABEL: test:
-; CHECK-NOT: pshufd
-; CHECK: punpcklbw
-; CHECK: punpcklbw
-; CHECK: pshufd $0
-; CHECK-NOT: pshufd
-}
+++ /dev/null
-; RUN: llc <%s -mtriple=i686-unknown-unknown -mcpu=penryn -mattr=sse4.1 | FileCheck %s
-
-; Splat test for v8i16
-define <8 x i16> @shuf_8i16_0(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 0, i32 undef, i32 undef, i32 0, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_0:
-; CHECK: pshuflw $0
-}
-
-define <8 x i16> @shuf_8i16_1(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 1, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_1:
-; CHECK: pshuflw $5
-}
-
-define <8 x i16> @shuf_8i16_2(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 2, i32 undef, i32 undef, i32 2, i32 undef, i32 2, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_2:
-; CHECK: punpcklwd
-; CHECK-NEXT: pshufd $-86
-}
-
-define <8 x i16> @shuf_8i16_3(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 3, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_3:
-; CHECK: pshuflw $15
-}
-
-define <8 x i16> @shuf_8i16_4(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 4, i32 undef, i32 undef, i32 undef, i32 4, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_4:
-; CHECK: movhlps
-}
-
-define <8 x i16> @shuf_8i16_5(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 5, i32 undef, i32 undef, i32 5, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_5:
-; CHECK: punpckhwd
-; CHECK-NEXT: pshufd $85
-}
-
-define <8 x i16> @shuf_8i16_6(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 6, i32 6, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_6:
-; CHECK: punpckhwd
-; CHECK-NEXT: pshufd $-86
-}
-
-define <8 x i16> @shuf_8i16_7(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
- %tmp6 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> <i32 7, i32 undef, i32 undef, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
- ret <8 x i16> %tmp6
-
-; CHECK-LABEL: shuf_8i16_7:
-; CHECK: punpckhwd
-; CHECK-NEXT: pshufd $-1
-}
-
-; Splat test for v16i8
-define <16 x i8> @shuf_16i8_8(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 0, i32 undef, i32 undef, i32 0, i32 undef, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_8:
-; CHECK: pxor %[[X:xmm[0-9]+]], %[[X]]
-; CHECK-NEXT: pshufb %[[X]], %xmm0
-}
-
-define <16 x i8> @shuf_16i8_9(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 1, i32 1, 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 >
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_9:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-}
-
-define <16 x i8> @shuf_16i8_10(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 2, i32 undef, i32 undef, i32 2, i32 undef, 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> %tmp6
-
-; CHECK-LABEL: shuf_16i8_10:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
-}
-
-define <16 x i8> @shuf_16i8_11(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 3, i32 undef, i32 undef, i32 3, i32 undef, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_11:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3]
-}
-
-
-define <16 x i8> @shuf_16i8_12(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 4, i32 undef, i32 undef, i32 undef, i32 4, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef >
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_12:
-; CHECK: pshufd $5
-}
-
-define <16 x i8> @shuf_16i8_13(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 5, i32 undef, i32 undef, i32 5, i32 undef, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_13:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5]
-}
-
-define <16 x i8> @shuf_16i8_14(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 6, i32 undef, i32 undef, i32 6, i32 undef, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_14:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6]
-}
-
-define <16 x i8> @shuf_16i8_15(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 7, i32 undef, i32 undef, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef >
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_15:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
-}
-
-define <16 x i8> @shuf_16i8_16(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 8, i32 undef, i32 undef, i32 8, i32 undef, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_16:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8]
-}
-
-define <16 x i8> @shuf_16i8_17(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 9, i32 undef, i32 undef, i32 9, i32 undef, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9, i32 9>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_17:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9]
-}
-
-define <16 x i8> @shuf_16i8_18(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 10, i32 undef, i32 undef, i32 10, i32 undef, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_18:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[10,10,10,10,10,10,10,10,10,10,10,10,10,10,10,10]
-}
-
-define <16 x i8> @shuf_16i8_19(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 11, i32 undef, i32 undef, i32 11, i32 undef, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11, i32 11>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_19:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[11,11,11,11,11,11,11,11,11,11,11,11,11,11,11,11]
-}
-
-define <16 x i8> @shuf_16i8_20(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 12, i32 undef, i32 undef, i32 12, i32 undef, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12, i32 12>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_20:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[12,12,12,12,12,12,12,12,12,12,12,12,12,12,12,12]
-}
-
-define <16 x i8> @shuf_16i8_21(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 13, i32 undef, i32 undef, i32 13, i32 undef, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_21:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13]
-}
-
-define <16 x i8> @shuf_16i8_22(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 14, i32 undef, i32 undef, i32 14, i32 undef, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14, i32 14>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_22:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[14,14,14,14,14,14,14,14,14,14,14,14,14,14,14,14]
-}
-
-define <16 x i8> @shuf_16i8_23(<16 x i8> %T0, <16 x i8> %T1) nounwind readnone {
- %tmp6 = shufflevector <16 x i8> %T0, <16 x i8> %T1, <16 x i32> <i32 15, i32 undef, i32 undef, i32 15, i32 undef, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
- ret <16 x i8> %tmp6
-
-; CHECK-LABEL: shuf_16i8_23:
-; CHECK: pshufb {{.*}} # xmm0 = xmm0[15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
-}
+++ /dev/null
-; RUN: llc < %s -march=x86 -mcpu=pentium4 -mattr=+sse2 | FileCheck %s -check-prefix=SSE2
-; RUN: llc < %s -march=x86 -mcpu=pentium4 -mattr=+sse3 | FileCheck %s -check-prefix=SSE3
-; RUN: llc < %s -march=x86-64 -mattr=+avx | FileCheck %s -check-prefix=AVX
-
-define void @test_v4sf(<4 x float>* %P, <4 x float>* %Q, float %X) nounwind {
- %tmp = insertelement <4 x float> zeroinitializer, float %X, i32 0 ; <<4 x float>> [#uses=1]
- %tmp2 = insertelement <4 x float> %tmp, float %X, i32 1 ; <<4 x float>> [#uses=1]
- %tmp4 = insertelement <4 x float> %tmp2, float %X, i32 2 ; <<4 x float>> [#uses=1]
- %tmp6 = insertelement <4 x float> %tmp4, float %X, i32 3 ; <<4 x float>> [#uses=1]
- %tmp8 = load <4 x float>* %Q ; <<4 x float>> [#uses=1]
- %tmp10 = fmul <4 x float> %tmp8, %tmp6 ; <<4 x float>> [#uses=1]
- store <4 x float> %tmp10, <4 x float>* %P
- ret void
-
-; SSE2-LABEL: test_v4sf:
-; SSE2: pshufd $0
-
-; SSE3-LABEL: test_v4sf:
-; SSE3: pshufd $0
-}
-
-define void @test_v2sd(<2 x double>* %P, <2 x double>* %Q, double %X) nounwind {
- %tmp = insertelement <2 x double> zeroinitializer, double %X, i32 0 ; <<2 x double>> [#uses=1]
- %tmp2 = insertelement <2 x double> %tmp, double %X, i32 1 ; <<2 x double>> [#uses=1]
- %tmp4 = load <2 x double>* %Q ; <<2 x double>> [#uses=1]
- %tmp6 = fmul <2 x double> %tmp4, %tmp2 ; <<2 x double>> [#uses=1]
- store <2 x double> %tmp6, <2 x double>* %P
- ret void
-
-; SSE2-LABEL: test_v2sd:
-; SSE2: movlhps
-
-; SSE3-LABEL: test_v2sd:
-; SSE3: movddup
-}
-
-; Fold extract of a load into the load's address computation. This avoids spilling to the stack.
-define <4 x float> @load_extract_splat(<4 x float>* nocapture readonly %ptr, i64 %i, i64 %j) nounwind {
- %1 = getelementptr inbounds <4 x float>* %ptr, i64 %i
- %2 = load <4 x float>* %1, align 16
- %3 = trunc i64 %j to i32
- %4 = extractelement <4 x float> %2, i32 %3
- %5 = insertelement <4 x float> undef, float %4, i32 0
- %6 = insertelement <4 x float> %5, float %4, i32 1
- %7 = insertelement <4 x float> %6, float %4, i32 2
- %8 = insertelement <4 x float> %7, float %4, i32 3
- ret <4 x float> %8
-
-; AVX-LABEL: load_extract_splat
-; AVX-NOT: rsp
-; AVX: vbroadcastss
-}
-
-; Fold extract of a load into the load's address computation. This avoids spilling to the stack.
-define <4 x float> @load_extract_splat1(<4 x float>* nocapture readonly %ptr, i64 %i, i64 %j) nounwind {
- %1 = getelementptr inbounds <4 x float>* %ptr, i64 %i
- %2 = load <4 x float>* %1, align 16
- %3 = extractelement <4 x float> %2, i64 %j
- %4 = insertelement <4 x float> undef, float %3, i32 0
- %5 = insertelement <4 x float> %4, float %3, i32 1
- %6 = insertelement <4 x float> %5, float %3, i32 2
- %7 = insertelement <4 x float> %6, float %3, i32 3
- ret <4 x float> %7
-
-; AVX-LABEL: load_extract_splat1
-; AVX-NOT: movs
-; AVX: vbroadcastss
-}
define <4 x float> @vsel_float(<4 x float> %v1, <4 x float> %v2) {
; SSE2-LABEL: vsel_float:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_float:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
define <4 x i8> @vsel_4xi8(<4 x i8> %v1, <4 x i8> %v2) {
; SSE2-LABEL: vsel_4xi8:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_4xi8:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: vsel_4xi8:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2],xmm0[3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5],xmm0[6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: vsel_4xi8:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2],xmm0[3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5],xmm0[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: vsel_4xi8:
define <4 x i16> @vsel_4xi16(<4 x i16> %v1, <4 x i16> %v2) {
; SSE2-LABEL: vsel_4xi16:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_4xi16:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: vsel_4xi16:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: vsel_4xi16:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: vsel_4xi16:
define <4 x i32> @vsel_i32(<4 x i32> %v1, <4 x i32> %v2) {
; SSE2-LABEL: vsel_i32:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_i32:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: vsel_i32:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: vsel_i32:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: vsel_i32:
define <8 x i16> @vsel_8xi16(<8 x i16> %v1, <8 x i16> %v2) {
; SSE2-LABEL: vsel_8xi16:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_8xi16:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: vsel_8xi16:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3],xmm0[4],xmm1[5,6,7]
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3],xmm0[4],xmm1[5,6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: vsel_8xi16:
define <16 x i8> @vsel_i8(<16 x i8> %v1, <16 x i8> %v2) {
; SSE2-LABEL: vsel_i8:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: andps {{.*}}, %xmm1
-; SSE2-NEXT: andps {{.*}}, %xmm0
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm1
+; SSE2-NEXT: andps {{.*}}(%rip), %xmm0
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: vsel_i8:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: andps {{.*}}, %xmm1
-; SSSE3-NEXT: andps {{.*}}, %xmm0
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm1
+; SSSE3-NEXT: andps {{.*}}(%rip), %xmm0
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-NEXT: blendpd {{.*#+}} xmm3 = xmm7[0,1]
; SSE41-NEXT: retq
;
-; AVX-LABEL: vsel_i648:
-; AVX: # BB#0: # %entry
-; AVX-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3]
-; AVX-NEXT: vblendpd {{.*#+}} ymm1 = ymm1[0],ymm3[1,2,3]
-; AVX-NEXT: retq
+; AVX1-LABEL: vsel_i648:
+; AVX1: # BB#0: # %entry
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3]
+; AVX1-NEXT: vblendpd {{.*#+}} ymm1 = ymm1[0],ymm3[1,2,3]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: vsel_i648:
+; AVX2: # BB#0: # %entry
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm2[2,3,4,5,6,7]
+; AVX2-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0,1],ymm3[2,3,4,5,6,7]
+; AVX2-NEXT: retq
entry:
%vsel = select <8 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <8 x i64> %v1, <8 x i64> %v2
ret <8 x i64> %vsel
;
; SSE41-LABEL: constant_blendvps_avx:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm2[0,1,2],xmm0[3]
-; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm3[0,1,2],xmm1[3]
+; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0,1,2],xmm0[3]
+; SSE41-NEXT: blendps {{.*#+}} xmm3 = xmm3[0,1,2],xmm1[3]
+; SSE41-NEXT: movaps %xmm2, %xmm0
+; SSE41-NEXT: movaps %xmm3, %xmm1
; SSE41-NEXT: retq
;
; AVX-LABEL: constant_blendvps_avx:
;
; AVX1-LABEL: constant_pblendvb_avx2:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vandps {{.*}}, %ymm1, %ymm1
-; AVX1-NEXT: vandps {{.*}}, %ymm0, %ymm0
+; AVX1-NEXT: vandps {{.*}}(%rip), %ymm1, %ymm1
+; AVX1-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm0
; AVX1-NEXT: vorps %ymm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; SSE2-LABEL: blend_shufflevector_4xfloat:
; SSE2: # BB#0: # %entry
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: blend_shufflevector_4xfloat:
; SSSE3: # BB#0: # %entry
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: blend_shufflevector_4xfloat:
;
; SSE41-LABEL: blend_shufflevector_8xfloat:
; SSE41: # BB#0: # %entry
+; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3]
; SSE41-NEXT: blendps {{.*#+}} xmm3 = xmm3[0,1],xmm1[2],xmm3[3]
-; SSE41-NEXT: movss %xmm0, %xmm2
; SSE41-NEXT: movaps %xmm2, %xmm0
; SSE41-NEXT: movaps %xmm3, %xmm1
; SSE41-NEXT: retq
}
define <4 x double> @blend_shufflevector_4xdouble(<4 x double> %a, <4 x double> %b) {
-; SSE-LABEL: blend_shufflevector_4xdouble:
-; SSE: # BB#0: # %entry
-; SSE-NEXT: movsd %xmm0, %xmm2
-; SSE-NEXT: movaps %xmm2, %xmm0
-; SSE-NEXT: retq
+; SSE2-LABEL: blend_shufflevector_4xdouble:
+; SSE2: # BB#0: # %entry
+; SSE2-NEXT: movsd %xmm0, %xmm2
+; SSE2-NEXT: movaps %xmm2, %xmm0
+; SSE2-NEXT: retq
+;
+; SSSE3-LABEL: blend_shufflevector_4xdouble:
+; SSSE3: # BB#0: # %entry
+; SSSE3-NEXT: movsd %xmm0, %xmm2
+; SSSE3-NEXT: movaps %xmm2, %xmm0
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: blend_shufflevector_4xdouble:
+; SSE41: # BB#0: # %entry
+; SSE41-NEXT: blendpd {{.*#+}} xmm0 = xmm0[0],xmm2[1]
+; SSE41-NEXT: retq
;
; AVX-LABEL: blend_shufflevector_4xdouble:
; AVX: # BB#0: # %entry
}
define <4 x i64> @blend_shufflevector_4xi64(<4 x i64> %a, <4 x i64> %b) {
-; SSE-LABEL: blend_shufflevector_4xi64:
-; SSE: # BB#0: # %entry
-; SSE-NEXT: movsd %xmm2, %xmm0
-; SSE-NEXT: movaps %xmm3, %xmm1
-; SSE-NEXT: retq
+; SSE2-LABEL: blend_shufflevector_4xi64:
+; SSE2: # BB#0: # %entry
+; SSE2-NEXT: movsd %xmm2, %xmm0
+; SSE2-NEXT: movaps %xmm3, %xmm1
+; SSE2-NEXT: retq
;
-; AVX-LABEL: blend_shufflevector_4xi64:
-; AVX: # BB#0: # %entry
-; AVX-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
-; AVX-NEXT: retq
+; SSSE3-LABEL: blend_shufflevector_4xi64:
+; SSSE3: # BB#0: # %entry
+; SSSE3-NEXT: movsd %xmm2, %xmm0
+; SSSE3-NEXT: movaps %xmm3, %xmm1
+; SSSE3-NEXT: retq
+;
+; SSE41-LABEL: blend_shufflevector_4xi64:
+; SSE41: # BB#0: # %entry
+; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm0[4,5,6,7]
+; SSE41-NEXT: movdqa %xmm2, %xmm0
+; SSE41-NEXT: movaps %xmm3, %xmm1
+; SSE41-NEXT: retq
+;
+; AVX1-LABEL: blend_shufflevector_4xi64:
+; AVX1: # BB#0: # %entry
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2,3]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: blend_shufflevector_4xi64:
+; AVX2: # BB#0: # %entry
+; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3],ymm1[4,5,6,7]
+; AVX2-NEXT: retq
entry:
%select = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 7>
ret <4 x i64> %select
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [613566757,613566757,613566757,613566757]
; SSE41-NEXT: movdqa %xmm0, %xmm2
; SSE41-NEXT: pmuludq %xmm1, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
; SSE41-NEXT: pmuludq %xmm1, %xmm3
; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm2[0,2,1,3]
-; SSE41-NEXT: psubd %xmm1, %xmm0
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE41-NEXT: psubd %xmm2, %xmm0
; SSE41-NEXT: psrld $1, %xmm0
-; SSE41-NEXT: paddd %xmm1, %xmm0
+; SSE41-NEXT: paddd %xmm2, %xmm0
; SSE41-NEXT: psrld $2, %xmm0
; SSE41-NEXT: retq
;
; SSE-NEXT: movdqa {{.*#+}} xmm1 = [613566757,613566757,613566757,613566757]
; SSE-NEXT: movdqa %xmm0, %xmm2
; SSE-NEXT: pmuludq %xmm1, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
; SSE-NEXT: pmuludq %xmm1, %xmm3
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm2[0,2,1,3]
-; SSE-NEXT: psubd %xmm1, %xmm0
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE-NEXT: psubd %xmm2, %xmm0
; SSE-NEXT: psrld $1, %xmm0
-; SSE-NEXT: paddd %xmm1, %xmm0
+; SSE-NEXT: paddd %xmm2, %xmm0
; SSE-NEXT: psrld $2, %xmm0
; SSE-NEXT: retq
;
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %xmm1
; AVX-NEXT: vpmuludq %xmm1, %xmm0, %xmm2
-; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; AVX-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,0,3,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; AVX-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
; AVX-NEXT: vpmuludq %xmm1, %xmm3, %xmm1
; AVX-NEXT: vshufps {{.*#+}} xmm1 = xmm2[1,3],xmm1[1,3]
-; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; AVX-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
; AVX-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpsrld $1, %xmm0, %xmm0
; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [613566757,613566757,613566757,613566757]
; SSE41-NEXT: movdqa %xmm0, %xmm3
; SSE41-NEXT: pmuludq %xmm2, %xmm3
-; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
; SSE41-NEXT: pmuludq %xmm4, %xmm5
; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm5[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
; SSE41-NEXT: psubd %xmm3, %xmm0
; SSE41-NEXT: psrld $1, %xmm0
; SSE41-NEXT: paddd %xmm3, %xmm0
; SSE41-NEXT: psrld $2, %xmm0
; SSE41-NEXT: pmuludq %xmm1, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,1,3,3]
; SSE41-NEXT: pmuludq %xmm4, %xmm3
; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE41-NEXT: psubd %xmm2, %xmm1
; SSE41-NEXT: psrld $1, %xmm1
; SSE41-NEXT: paddd %xmm2, %xmm1
; SSE-NEXT: movdqa {{.*#+}} xmm2 = [613566757,613566757,613566757,613566757]
; SSE-NEXT: movdqa %xmm0, %xmm3
; SSE-NEXT: pmuludq %xmm2, %xmm3
-; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm5
; SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm5[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
; SSE-NEXT: psubd %xmm3, %xmm0
; SSE-NEXT: psrld $1, %xmm0
; SSE-NEXT: paddd %xmm3, %xmm0
; SSE-NEXT: psrld $2, %xmm0
; SSE-NEXT: pmuludq %xmm1, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm3
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE-NEXT: psubd %xmm2, %xmm1
; SSE-NEXT: psrld $1, %xmm1
; SSE-NEXT: paddd %xmm2, %xmm1
; AVX-LABEL: test2:
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm2 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
-; AVX-NEXT: vpalignr {{.*#+}} ymm3 = ymm0[4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3,20,21,22,23,24,25,26,27,28,29,30,31,16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm2 = ymm1[1,1,3,3,5,5,7,7]
+; AVX-NEXT: vpshufd {{.*#+}} ymm3 = ymm0[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpmuludq %ymm2, %ymm3, %ymm2
; AVX-NEXT: vpmuludq %ymm1, %ymm0, %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm1 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],ymm2[1],ymm1[2],ymm2[3],ymm1[4],ymm2[5],ymm1[6],ymm2[7]
; AVX-NEXT: vpsubd %ymm1, %ymm0, %ymm0
; AVX-NEXT: vpsrld $1, %ymm0, %ymm0
define <4 x i32> @test8(<4 x i32> %a) {
; SSE41-LABEL: test8:
; SSE41: # BB#0:
-; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [2454267027,2454267027,2454267027,2454267027]
-; SSE41-NEXT: movdqa %xmm0, %xmm2
-; SSE41-NEXT: pmuldq %xmm1, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,0,3,0]
-; SSE41-NEXT: pmuldq %xmm1, %xmm3
-; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm2[0,2,1,3]
+; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [2454267027,2454267027,2454267027,2454267027]
+; SSE41-NEXT: movdqa %xmm0, %xmm1
+; SSE41-NEXT: pmuldq %xmm2, %xmm1
+; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; SSE41-NEXT: pmuldq %xmm2, %xmm3
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,3],xmm3[1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
; SSE41-NEXT: paddd %xmm0, %xmm1
; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: psrld $31, %xmm0
;
; SSE-LABEL: test8:
; SSE: # BB#0:
-; SSE-NEXT: movdqa {{.*#+}} xmm1 = [2454267027,2454267027,2454267027,2454267027]
-; SSE-NEXT: movdqa %xmm1, %xmm2
-; SSE-NEXT: psrad $31, %xmm2
-; SSE-NEXT: pand %xmm0, %xmm2
+; SSE-NEXT: movdqa {{.*#+}} xmm2 = [2454267027,2454267027,2454267027,2454267027]
+; SSE-NEXT: movdqa %xmm2, %xmm1
+; SSE-NEXT: psrad $31, %xmm1
+; SSE-NEXT: pand %xmm0, %xmm1
; SSE-NEXT: movdqa %xmm0, %xmm3
; SSE-NEXT: psrad $31, %xmm3
-; SSE-NEXT: pand %xmm1, %xmm3
-; SSE-NEXT: paddd %xmm2, %xmm3
-; SSE-NEXT: movdqa %xmm0, %xmm2
-; SSE-NEXT: pmuludq %xmm1, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,0,3,0]
-; SSE-NEXT: pmuludq %xmm1, %xmm4
-; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm4[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm2[0,2,1,3]
+; SSE-NEXT: pand %xmm2, %xmm3
+; SSE-NEXT: paddd %xmm1, %xmm3
+; SSE-NEXT: movdqa %xmm0, %xmm1
+; SSE-NEXT: pmuludq %xmm2, %xmm1
+; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
+; SSE-NEXT: pmuludq %xmm2, %xmm4
+; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,3],xmm4[1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
; SSE-NEXT: psubd %xmm3, %xmm1
; SSE-NEXT: paddd %xmm0, %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %xmm1
; AVX-NEXT: vpmuldq %xmm1, %xmm0, %xmm2
-; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; AVX-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,0,3,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; AVX-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
; AVX-NEXT: vpmuldq %xmm1, %xmm3, %xmm1
; AVX-NEXT: vshufps {{.*#+}} xmm1 = xmm2[1,3],xmm1[1,3]
-; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; AVX-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
; AVX-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; AVX-NEXT: vpsrld $31, %xmm0, %xmm1
; AVX-NEXT: vpsrad $2, %xmm0, %xmm0
define <8 x i32> @test9(<8 x i32> %a) {
; SSE41-LABEL: test9:
; SSE41: # BB#0:
-; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [2454267027,2454267027,2454267027,2454267027]
-; SSE41-NEXT: movdqa %xmm0, %xmm2
-; SSE41-NEXT: pmuldq %xmm3, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm3[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE41-NEXT: movdqa %xmm1, %xmm2
+; SSE41-NEXT: movdqa %xmm0, %xmm3
+; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [2454267027,2454267027,2454267027,2454267027]
+; SSE41-NEXT: # kill: XMM0<def> XMM3<kill>
+; SSE41-NEXT: pmuldq %xmm1, %xmm0
+; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm1[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm3[1,1,3,3]
; SSE41-NEXT: pmuldq %xmm4, %xmm5
-; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm5[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
-; SSE41-NEXT: paddd %xmm0, %xmm2
-; SSE41-NEXT: movdqa %xmm2, %xmm0
-; SSE41-NEXT: psrld $31, %xmm0
-; SSE41-NEXT: psrad $2, %xmm2
-; SSE41-NEXT: paddd %xmm0, %xmm2
-; SSE41-NEXT: pmuldq %xmm1, %xmm3
-; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,0,3,0]
-; SSE41-NEXT: pmuldq %xmm4, %xmm0
-; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm0[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
-; SSE41-NEXT: paddd %xmm1, %xmm3
-; SSE41-NEXT: movdqa %xmm3, %xmm0
-; SSE41-NEXT: psrld $31, %xmm0
-; SSE41-NEXT: psrad $2, %xmm3
-; SSE41-NEXT: paddd %xmm0, %xmm3
-; SSE41-NEXT: movdqa %xmm2, %xmm0
-; SSE41-NEXT: movdqa %xmm3, %xmm1
+; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm5[1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE41-NEXT: paddd %xmm3, %xmm0
+; SSE41-NEXT: movdqa %xmm0, %xmm3
+; SSE41-NEXT: psrld $31, %xmm3
+; SSE41-NEXT: psrad $2, %xmm0
+; SSE41-NEXT: paddd %xmm3, %xmm0
+; SSE41-NEXT: pmuldq %xmm2, %xmm1
+; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm2[1,1,3,3]
+; SSE41-NEXT: pmuldq %xmm4, %xmm3
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,3],xmm3[1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; SSE41-NEXT: paddd %xmm2, %xmm1
+; SSE41-NEXT: movdqa %xmm1, %xmm2
+; SSE41-NEXT: psrld $31, %xmm2
+; SSE41-NEXT: psrad $2, %xmm1
+; SSE41-NEXT: paddd %xmm2, %xmm1
; SSE41-NEXT: retq
;
; SSE-LABEL: test9:
; SSE: # BB#0:
-; SSE-NEXT: movdqa %xmm0, %xmm2
-; SSE-NEXT: movdqa {{.*#+}} xmm3 = [2454267027,2454267027,2454267027,2454267027]
-; SSE-NEXT: movdqa %xmm3, %xmm4
+; SSE-NEXT: movdqa %xmm1, %xmm2
+; SSE-NEXT: movdqa %xmm0, %xmm3
+; SSE-NEXT: movdqa {{.*#+}} xmm1 = [2454267027,2454267027,2454267027,2454267027]
+; SSE-NEXT: movdqa %xmm1, %xmm4
; SSE-NEXT: psrad $31, %xmm4
; SSE-NEXT: movdqa %xmm4, %xmm0
-; SSE-NEXT: pand %xmm2, %xmm0
-; SSE-NEXT: movdqa %xmm2, %xmm5
+; SSE-NEXT: pand %xmm3, %xmm0
+; SSE-NEXT: movdqa %xmm3, %xmm5
; SSE-NEXT: psrad $31, %xmm5
-; SSE-NEXT: pand %xmm3, %xmm5
+; SSE-NEXT: pand %xmm1, %xmm5
; SSE-NEXT: paddd %xmm0, %xmm5
-; SSE-NEXT: movdqa %xmm2, %xmm0
-; SSE-NEXT: pmuludq %xmm3, %xmm0
-; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm3[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,0,3,0]
+; SSE-NEXT: movdqa %xmm3, %xmm0
+; SSE-NEXT: pmuludq %xmm1, %xmm0
+; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3]
; SSE-NEXT: pmuludq %xmm6, %xmm7
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm7[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE-NEXT: psubd %xmm5, %xmm0
-; SSE-NEXT: paddd %xmm2, %xmm0
-; SSE-NEXT: movdqa %xmm0, %xmm2
-; SSE-NEXT: psrld $31, %xmm2
+; SSE-NEXT: paddd %xmm3, %xmm0
+; SSE-NEXT: movdqa %xmm0, %xmm3
+; SSE-NEXT: psrld $31, %xmm3
; SSE-NEXT: psrad $2, %xmm0
-; SSE-NEXT: paddd %xmm2, %xmm0
-; SSE-NEXT: pand %xmm1, %xmm4
-; SSE-NEXT: movdqa %xmm1, %xmm5
-; SSE-NEXT: psrad $31, %xmm5
-; SSE-NEXT: pand %xmm3, %xmm5
-; SSE-NEXT: paddd %xmm4, %xmm5
-; SSE-NEXT: pmuludq %xmm1, %xmm3
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,0,3,0]
-; SSE-NEXT: pmuludq %xmm6, %xmm2
-; SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm2[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[0,2,1,3]
-; SSE-NEXT: psubd %xmm5, %xmm2
-; SSE-NEXT: paddd %xmm1, %xmm2
-; SSE-NEXT: movdqa %xmm2, %xmm1
-; SSE-NEXT: psrld $31, %xmm1
-; SSE-NEXT: psrad $2, %xmm2
-; SSE-NEXT: paddd %xmm1, %xmm2
-; SSE-NEXT: movdqa %xmm2, %xmm1
+; SSE-NEXT: paddd %xmm3, %xmm0
+; SSE-NEXT: pand %xmm2, %xmm4
+; SSE-NEXT: movdqa %xmm2, %xmm3
+; SSE-NEXT: psrad $31, %xmm3
+; SSE-NEXT: pand %xmm1, %xmm3
+; SSE-NEXT: paddd %xmm4, %xmm3
+; SSE-NEXT: pmuludq %xmm2, %xmm1
+; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE-NEXT: pmuludq %xmm6, %xmm4
+; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,3],xmm4[1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; SSE-NEXT: psubd %xmm3, %xmm1
+; SSE-NEXT: paddd %xmm2, %xmm1
+; SSE-NEXT: movdqa %xmm1, %xmm2
+; SSE-NEXT: psrld $31, %xmm2
+; SSE-NEXT: psrad $2, %xmm1
+; SSE-NEXT: paddd %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX-LABEL: test9:
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm2 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
-; AVX-NEXT: vpalignr {{.*#+}} ymm3 = ymm0[4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3,20,21,22,23,24,25,26,27,28,29,30,31,16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm2 = ymm1[1,1,3,3,5,5,7,7]
+; AVX-NEXT: vpshufd {{.*#+}} ymm3 = ymm0[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpmuldq %ymm2, %ymm3, %ymm2
; AVX-NEXT: vpmuldq %ymm1, %ymm0, %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm1 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],ymm2[1],ymm1[2],ymm2[3],ymm1[4],ymm2[5],ymm1[6],ymm2[7]
; AVX-NEXT: vpaddd %ymm0, %ymm1, %ymm0
; AVX-NEXT: vpsrld $31, %ymm0, %ymm1
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [613566757,613566757,613566757,613566757]
; SSE41-NEXT: movdqa %xmm0, %xmm3
; SSE41-NEXT: pmuludq %xmm2, %xmm3
-; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
; SSE41-NEXT: pmuludq %xmm4, %xmm5
; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm5[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
; SSE41-NEXT: movdqa %xmm0, %xmm5
; SSE41-NEXT: psubd %xmm3, %xmm5
; SSE41-NEXT: psrld $1, %xmm5
; SSE41-NEXT: pmulld %xmm3, %xmm5
; SSE41-NEXT: psubd %xmm5, %xmm0
; SSE41-NEXT: pmuludq %xmm1, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm1[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm1[1,1,3,3]
; SSE41-NEXT: pmuludq %xmm4, %xmm5
; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm5[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE41-NEXT: movdqa %xmm1, %xmm4
; SSE41-NEXT: psubd %xmm2, %xmm4
; SSE41-NEXT: psrld $1, %xmm4
; SSE-NEXT: movdqa {{.*#+}} xmm2 = [613566757,613566757,613566757,613566757]
; SSE-NEXT: movdqa %xmm0, %xmm3
; SSE-NEXT: pmuludq %xmm2, %xmm3
-; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm5
; SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm5[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
; SSE-NEXT: movdqa %xmm0, %xmm5
; SSE-NEXT: psubd %xmm3, %xmm5
; SSE-NEXT: psrld $1, %xmm5
; SSE-NEXT: paddd %xmm3, %xmm5
; SSE-NEXT: psrld $2, %xmm5
; SSE-NEXT: movdqa {{.*#+}} xmm3 = [7,7,7,7]
-; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3]
; SSE-NEXT: pmuludq %xmm3, %xmm5
; SSE-NEXT: pmuludq %xmm3, %xmm6
; SSE-NEXT: shufps {{.*#+}} xmm5 = xmm5[0,2],xmm6[0,2]
-; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm5[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm5 = xmm5[0,2,1,3]
; SSE-NEXT: psubd %xmm5, %xmm0
; SSE-NEXT: pmuludq %xmm1, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm1[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm1[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm5
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm5[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE-NEXT: movdqa %xmm1, %xmm4
; SSE-NEXT: psubd %xmm2, %xmm4
; SSE-NEXT: psrld $1, %xmm4
; SSE-NEXT: paddd %xmm2, %xmm4
; SSE-NEXT: psrld $2, %xmm4
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm4[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm4[1,1,3,3]
; SSE-NEXT: pmuludq %xmm3, %xmm4
; SSE-NEXT: pmuludq %xmm3, %xmm2
; SSE-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,2],xmm2[0,2]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm4[0,2,1,3]
-; SSE-NEXT: psubd %xmm2, %xmm1
+; SSE-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,2,1,3]
+; SSE-NEXT: psubd %xmm4, %xmm1
; SSE-NEXT: retq
;
; AVX-LABEL: test10:
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm2 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
-; AVX-NEXT: vpalignr {{.*#+}} ymm3 = ymm0[4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3,20,21,22,23,24,25,26,27,28,29,30,31,16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm2 = ymm1[1,1,3,3,5,5,7,7]
+; AVX-NEXT: vpshufd {{.*#+}} ymm3 = ymm0[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpmuludq %ymm2, %ymm3, %ymm2
; AVX-NEXT: vpmuludq %ymm1, %ymm0, %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm1 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],ymm2[1],ymm1[2],ymm2[3],ymm1[4],ymm2[5],ymm1[6],ymm2[7]
; AVX-NEXT: vpsubd %ymm1, %ymm0, %ymm2
; AVX-NEXT: vpsrld $1, %ymm2, %ymm2
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [2454267027,2454267027,2454267027,2454267027]
; SSE41-NEXT: movdqa %xmm0, %xmm3
; SSE41-NEXT: pmuldq %xmm2, %xmm3
-; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,0,3,0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSE41-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
; SSE41-NEXT: pmuldq %xmm4, %xmm5
; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[1,3],xmm5[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,2,1,3]
; SSE41-NEXT: paddd %xmm0, %xmm3
; SSE41-NEXT: movdqa %xmm3, %xmm5
; SSE41-NEXT: psrld $31, %xmm5
; SSE41-NEXT: pmulld %xmm5, %xmm3
; SSE41-NEXT: psubd %xmm3, %xmm0
; SSE41-NEXT: pmuldq %xmm1, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,1,3,3]
; SSE41-NEXT: pmuldq %xmm4, %xmm3
; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE41-NEXT: paddd %xmm1, %xmm2
; SSE41-NEXT: movdqa %xmm2, %xmm3
; SSE41-NEXT: psrld $31, %xmm3
; SSE-NEXT: psrad $31, %xmm6
; SSE-NEXT: pand %xmm2, %xmm6
; SSE-NEXT: paddd %xmm4, %xmm6
-; SSE-NEXT: movdqa %xmm0, %xmm4
-; SSE-NEXT: pmuludq %xmm2, %xmm4
-; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm2[1,0,3,0]
-; SSE-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,0,3,0]
-; SSE-NEXT: pmuludq %xmm5, %xmm7
-; SSE-NEXT: shufps {{.*#+}} xmm4 = xmm4[1,3],xmm7[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm7 = xmm4[0,2,1,3]
+; SSE-NEXT: movdqa %xmm0, %xmm7
+; SSE-NEXT: pmuludq %xmm2, %xmm7
+; SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm2[1,1,3,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
+; SSE-NEXT: pmuludq %xmm5, %xmm4
+; SSE-NEXT: shufps {{.*#+}} xmm7 = xmm7[1,3],xmm4[1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm7 = xmm7[0,2,1,3]
; SSE-NEXT: psubd %xmm6, %xmm7
; SSE-NEXT: paddd %xmm0, %xmm7
; SSE-NEXT: movdqa %xmm7, %xmm4
; SSE-NEXT: psrad $2, %xmm7
; SSE-NEXT: paddd %xmm4, %xmm7
; SSE-NEXT: movdqa {{.*#+}} xmm4 = [7,7,7,7]
-; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm7[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm7[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm7
; SSE-NEXT: pmuludq %xmm4, %xmm6
; SSE-NEXT: shufps {{.*#+}} xmm7 = xmm7[0,2],xmm6[0,2]
-; SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm7[0,2,1,3]
-; SSE-NEXT: psubd %xmm6, %xmm0
+; SSE-NEXT: shufps {{.*#+}} xmm7 = xmm7[0,2,1,3]
+; SSE-NEXT: psubd %xmm7, %xmm0
; SSE-NEXT: pand %xmm1, %xmm3
; SSE-NEXT: movdqa %xmm1, %xmm6
; SSE-NEXT: psrad $31, %xmm6
; SSE-NEXT: pand %xmm2, %xmm6
; SSE-NEXT: paddd %xmm3, %xmm6
; SSE-NEXT: pmuludq %xmm1, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,1,3,3]
; SSE-NEXT: pmuludq %xmm5, %xmm3
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE-NEXT: psubd %xmm6, %xmm2
; SSE-NEXT: paddd %xmm1, %xmm2
; SSE-NEXT: movdqa %xmm2, %xmm3
; SSE-NEXT: psrld $31, %xmm3
; SSE-NEXT: psrad $2, %xmm2
; SSE-NEXT: paddd %xmm3, %xmm2
-; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm2[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm2[1,1,3,3]
; SSE-NEXT: pmuludq %xmm4, %xmm2
; SSE-NEXT: pmuludq %xmm4, %xmm3
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
; SSE-NEXT: psubd %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX-LABEL: test11:
; AVX: # BB#0:
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm2 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
-; AVX-NEXT: vpalignr {{.*#+}} ymm3 = ymm0[4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3,20,21,22,23,24,25,26,27,28,29,30,31,16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm2 = ymm1[1,1,3,3,5,5,7,7]
+; AVX-NEXT: vpshufd {{.*#+}} ymm3 = ymm0[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpmuldq %ymm2, %ymm3, %ymm2
; AVX-NEXT: vpmuldq %ymm1, %ymm0, %ymm1
-; AVX-NEXT: vpalignr {{.*#+}} ymm1 = ymm1[4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0,1,2,3],ymm1[20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16,17,18,19]
+; AVX-NEXT: vpshufd {{.*#+}} ymm1 = ymm1[1,1,3,3,5,5,7,7]
; AVX-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],ymm2[1],ymm1[2],ymm2[3],ymm1[4],ymm2[5],ymm1[6],ymm2[7]
; AVX-NEXT: vpaddd %ymm0, %ymm1, %ymm1
; AVX-NEXT: vpsrld $31, %ymm1, %ymm2
; SSE41-LABEL: PR20355:
; SSE41: # BB#0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [1431655766,1431655766,1431655766,1431655766]
-; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
; SSE41-NEXT: pmuldq %xmm1, %xmm0
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; SSE41-NEXT: pmuldq %xmm2, %xmm1
; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,2,1,3]
+; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE41-NEXT: movaps %xmm0, %xmm1
+; SSE41-NEXT: psrld $31, %xmm1
+; SSE41-NEXT: paddd %xmm0, %xmm1
; SSE41-NEXT: movdqa %xmm1, %xmm0
-; SSE41-NEXT: psrld $31, %xmm0
-; SSE41-NEXT: paddd %xmm1, %xmm0
; SSE41-NEXT: retq
;
; SSE-LABEL: PR20355:
; SSE-NEXT: psrad $31, %xmm3
; SSE-NEXT: pand %xmm1, %xmm3
; SSE-NEXT: paddd %xmm2, %xmm3
-; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
; SSE-NEXT: pmuludq %xmm1, %xmm0
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; SSE-NEXT: pmuludq %xmm2, %xmm1
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,2,1,3]
-; SSE-NEXT: psubd %xmm3, %xmm1
+; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE-NEXT: psubd %xmm3, %xmm0
+; SSE-NEXT: movdqa %xmm0, %xmm1
+; SSE-NEXT: psrld $31, %xmm1
+; SSE-NEXT: paddd %xmm0, %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
-; SSE-NEXT: psrld $31, %xmm0
-; SSE-NEXT: paddd %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: PR20355:
; AVX: # BB#0: # %entry
; AVX-NEXT: vpbroadcastd {{.*}}(%rip), %xmm1
; AVX-NEXT: vpmuldq %xmm1, %xmm0, %xmm2
-; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,0,3,0]
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,0,3,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX-NEXT: vpmuldq %xmm1, %xmm0, %xmm0
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm2[1,3],xmm0[1,3]
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; AVX-NEXT: vpsrld $31, %xmm0, %xmm1
; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0
; AVX-NEXT: retq
; SSE2-LABEL: sext_8i16_to_8i32:
; SSE2: # BB#0: # %entry
; SSE2-NEXT: movdqa %xmm0, %xmm1
-; SSE2-NEXT: # kill: XMM0<def> XMM1<kill>
+; SSE2-NEXT: # kill: XMM0<def> XMM1<kill>
; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
; SSE2-NEXT: pslld $16, %xmm0
; SSE2-NEXT: psrad $16, %xmm0
-; SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
; SSE2-NEXT: pslld $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: retq
; SSSE3-LABEL: sext_8i16_to_8i32:
; SSSE3: # BB#0: # %entry
; SSSE3-NEXT: movdqa %xmm0, %xmm1
-; SSSE3-NEXT: # kill: XMM0<def> XMM1<kill>
+; SSSE3-NEXT: # kill: XMM0<def> XMM1<kill>
; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
; SSSE3-NEXT: pslld $16, %xmm0
; SSSE3-NEXT: psrad $16, %xmm0
-; SSSE3-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
; SSSE3-NEXT: pslld $16, %xmm1
; SSSE3-NEXT: psrad $16, %xmm1
; SSSE3-NEXT: retq
; SSE41: # BB#0: # %entry
; SSE41-NEXT: movdqa %xmm0, %xmm1
; SSE41-NEXT: pmovzxwd %xmm1, %xmm0
-; SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; SSE41-NEXT: pslld $16, %xmm1
-; SSE41-NEXT: psrad $16, %xmm1
; SSE41-NEXT: pslld $16, %xmm0
; SSE41-NEXT: psrad $16, %xmm0
+; SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
+; SSE41-NEXT: pslld $16, %xmm1
+; SSE41-NEXT: psrad $16, %xmm1
; SSE41-NEXT: retq
;
; AVX1-LABEL: sext_8i16_to_8i32:
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vpmovsxwd %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxwd %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41: # BB#0: # %entry
; X32-SSE41-NEXT: movdqa %xmm0, %xmm1
; X32-SSE41-NEXT: pmovzxwd %xmm1, %xmm0
-; X32-SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X32-SSE41-NEXT: pslld $16, %xmm1
-; X32-SSE41-NEXT: psrad $16, %xmm1
; X32-SSE41-NEXT: pslld $16, %xmm0
; X32-SSE41-NEXT: psrad $16, %xmm0
+; X32-SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
+; X32-SSE41-NEXT: pslld $16, %xmm1
+; X32-SSE41-NEXT: psrad $16, %xmm1
; X32-SSE41-NEXT: retl
entry:
%B = sext <8 x i16> %A to <8 x i32>
define <4 x i64> @sext_4i32_to_4i64(<4 x i32> %A) nounwind uwtable readnone ssp {
; SSE2-LABEL: sext_4i32_to_4i64:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm2
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm0
;
; SSSE3-LABEL: sext_4i32_to_4i64:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm2
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm0
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm2
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE41-NEXT: pextrq $1, %xmm0, %rax
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm3
; AVX1-LABEL: sext_4i32_to_4i64:
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: pinsrd $1, %eax, %xmm2
; X32-SSE41-NEXT: sarl $31, %ecx
; X32-SSE41-NEXT: pinsrd $3, %ecx, %xmm2
-; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; X32-SSE41-NEXT: movd %xmm1, %eax
; X32-SSE41-NEXT: sarl $31, %eax
; X32-SSE41-NEXT: pextrd $2, %xmm1, %ecx
define <4 x i32> @load_sext_test2(<4 x i8> *%ptr) {
; SSE2-LABEL: load_sext_test2:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movl %eax, %ecx
-; SSE2-NEXT: shll $8, %ecx
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pextrw $1, %xmm0, %edx
-; SSE2-NEXT: pinsrw $1, %ecx, %xmm0
-; SSE2-NEXT: pinsrw $3, %eax, %xmm0
-; SSE2-NEXT: movl %edx, %eax
-; SSE2-NEXT: shll $8, %eax
-; SSE2-NEXT: pinsrw $5, %eax, %xmm0
-; SSE2-NEXT: pinsrw $7, %edx, %xmm0
+; SSE2-NEXT: movd (%rdi), %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
; SSE2-NEXT: psrad $24, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: load_sext_test2:
; SSSE3: # BB#0: # %entry
; SSSE3-NEXT: movd (%rdi), %xmm0
-; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = zero,zero,zero,xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3]
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
; SSSE3-NEXT: psrad $24, %xmm0
; SSSE3-NEXT: retq
;
; SSE2: # BB#0:
; SSE2-NEXT: pslld $31, %xmm0
; SSE2-NEXT: psrad $31, %xmm0
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm2
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm0
; SSSE3: # BB#0:
; SSSE3-NEXT: pslld $31, %xmm0
; SSSE3-NEXT: psrad $31, %xmm0
-; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm2
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm0
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm2
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE41-NEXT: pextrq $1, %xmm0, %rax
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm3
; AVX1-NEXT: vpslld $31, %xmm0, %xmm0
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm0
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: pinsrd $1, %eax, %xmm2
; X32-SSE41-NEXT: sarl $31, %ecx
; X32-SSE41-NEXT: pinsrd $3, %ecx, %xmm2
-; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; X32-SSE41-NEXT: movd %xmm1, %eax
; X32-SSE41-NEXT: sarl $31, %eax
; X32-SSE41-NEXT: pextrd $2, %xmm1, %ecx
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT: psllw $8, %xmm0
; SSE2-NEXT: psraw $8, %xmm0
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: psllw $8, %xmm1
; SSE2-NEXT: psraw $8, %xmm1
; SSE2-NEXT: retq
; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSSE3-NEXT: psllw $8, %xmm0
; SSSE3-NEXT: psraw $8, %xmm0
-; SSSE3-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
+; SSSE3-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSSE3-NEXT: psllw $8, %xmm1
; SSSE3-NEXT: psraw $8, %xmm1
; SSSE3-NEXT: retq
; SSE41: # BB#0: # %entry
; SSE41-NEXT: movdqa (%rdi), %xmm1
; SSE41-NEXT: pmovzxbw %xmm1, %xmm0
-; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
-; SSE41-NEXT: psllw $8, %xmm1
-; SSE41-NEXT: psraw $8, %xmm1
; SSE41-NEXT: psllw $8, %xmm0
; SSE41-NEXT: psraw $8, %xmm0
+; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; SSE41-NEXT: psllw $8, %xmm1
+; SSE41-NEXT: psraw $8, %xmm1
; SSE41-NEXT: retq
;
; AVX1-LABEL: sext_16i8_to_16i16:
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vmovdqa (%rdi), %xmm0
; AVX1-NEXT: vpmovsxbw %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxbw %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-SSE41-NEXT: movdqa (%eax), %xmm1
; X32-SSE41-NEXT: pmovzxbw %xmm1, %xmm0
-; X32-SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
-; X32-SSE41-NEXT: psllw $8, %xmm1
-; X32-SSE41-NEXT: psraw $8, %xmm1
; X32-SSE41-NEXT: psllw $8, %xmm0
; X32-SSE41-NEXT: psraw $8, %xmm0
+; X32-SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; X32-SSE41-NEXT: psllw $8, %xmm1
+; X32-SSE41-NEXT: psraw $8, %xmm1
; X32-SSE41-NEXT: retl
entry:
%X = load <16 x i8>* %ptr
; SSE2: # BB#0:
; SSE2-NEXT: pslld $24, %xmm0
; SSE2-NEXT: psrad $24, %xmm0
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm2
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSE2-NEXT: movd %xmm1, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm1
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSE2-NEXT: movd %xmm0, %rax
; SSE2-NEXT: cltq
; SSE2-NEXT: movd %rax, %xmm0
; SSSE3: # BB#0:
; SSSE3-NEXT: pslld $24, %xmm0
; SSSE3-NEXT: psrad $24, %xmm0
-; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,1,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm2
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1]
; SSSE3-NEXT: movd %xmm1, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm1
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSSE3-NEXT: movd %xmm0, %rax
; SSSE3-NEXT: cltq
; SSSE3-NEXT: movd %rax, %xmm0
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm2
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; SSE41-NEXT: pextrq $1, %xmm0, %rax
; SSE41-NEXT: cltq
; SSE41-NEXT: movd %rax, %xmm3
; AVX1-NEXT: vpslld $24, %xmm0, %xmm0
; AVX1-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: pinsrd $1, %eax, %xmm2
; X32-SSE41-NEXT: sarl $31, %ecx
; X32-SSE41-NEXT: pinsrd $3, %ecx, %xmm2
-; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; X32-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; X32-SSE41-NEXT: movd %xmm1, %eax
; X32-SSE41-NEXT: sarl $31, %eax
; X32-SSE41-NEXT: pextrd $2, %xmm1, %ecx
define <4 x i64> @load_sext_4i8_to_4i64(<4 x i8> *%ptr) {
; SSE2-LABEL: load_sext_4i8_to_4i64:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movd %eax, %xmm1
-; SSE2-NEXT: pextrw $1, %xmm1, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm1
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm1
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm1
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm1
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,1,0]
+; SSE2-NEXT: movd (%rdi), %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,1,1,3]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movsbq %al, %rax
; SSE2-NEXT: movd %rax, %xmm0
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movsbq %al, %rax
; SSE2-NEXT: movd %rax, %xmm2
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movsbq %al, %rax
; SSE2-NEXT: movd %rax, %xmm1
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movsbq %al, %rax
; SSE2-NEXT: movd %rax, %xmm2
; SSSE3-LABEL: load_sext_4i8_to_4i64:
; SSSE3: # BB#0: # %entry
; SSSE3-NEXT: movd (%rdi), %xmm1
-; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero
-; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,1,0]
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,1,1,3]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movsbq %al, %rax
; SSSE3-NEXT: movd %rax, %xmm0
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movsbq %al, %rax
; SSSE3-NEXT: movd %rax, %xmm2
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movsbq %al, %rax
; SSSE3-NEXT: movd %rax, %xmm1
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movsbq %al, %rax
; SSSE3-NEXT: movd %rax, %xmm2
;
; SSE41-LABEL: load_sext_4i8_to_4i64:
; SSE41: # BB#0: # %entry
-; SSE41-NEXT: movd (%rdi), %xmm0
-; SSE41-NEXT: pmovzxbd %xmm0, %xmm1
-; SSE41-NEXT: pmovzxbq %xmm0, %xmm0
+; SSE41-NEXT: pmovzxbd (%rdi), %xmm1
+; SSE41-NEXT: pmovzxdq %xmm1, %xmm0
; SSE41-NEXT: pextrq $1, %xmm0, %rax
; SSE41-NEXT: movsbq %al, %rax
; SSE41-NEXT: movd %rax, %xmm2
; SSE41-NEXT: movsbq %al, %rax
; SSE41-NEXT: movd %rax, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,2,3,3]
; SSE41-NEXT: pextrq $1, %xmm1, %rax
; SSE41-NEXT: movsbq %al, %rax
; SSE41-NEXT: movd %rax, %xmm2
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vpmovsxbd (%rdi), %xmm0
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: pinsrd $2, %eax, %xmm0
; X32-SSE41-NEXT: sarl $31, %eax
; X32-SSE41-NEXT: pinsrd $3, %eax, %xmm0
-; X32-SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; X32-SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; X32-SSE41-NEXT: movd %xmm2, %eax
; X32-SSE41-NEXT: movsbl %al, %eax
; X32-SSE41-NEXT: movd %eax, %xmm1
; SSE2: # BB#0: # %entry
; SSE2-NEXT: movq (%rdi), %xmm1
; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,1,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,1,1,3]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movswq %ax, %rax
; SSE2-NEXT: movd %rax, %xmm0
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movswq %ax, %rax
; SSE2-NEXT: movd %rax, %xmm2
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movswq %ax, %rax
; SSE2-NEXT: movd %rax, %xmm1
-; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSE2-NEXT: movd %xmm2, %rax
; SSE2-NEXT: movswq %ax, %rax
; SSE2-NEXT: movd %rax, %xmm2
; SSSE3: # BB#0: # %entry
; SSSE3-NEXT: movq (%rdi), %xmm1
; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,1,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,1,1,3]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movswq %ax, %rax
; SSSE3-NEXT: movd %rax, %xmm0
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movswq %ax, %rax
; SSSE3-NEXT: movd %rax, %xmm2
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movswq %ax, %rax
; SSSE3-NEXT: movd %rax, %xmm1
-; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
; SSSE3-NEXT: movd %xmm2, %rax
; SSSE3-NEXT: movswq %ax, %rax
; SSSE3-NEXT: movd %rax, %xmm2
; SSE41-NEXT: movswq %ax, %rax
; SSE41-NEXT: movd %rax, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,2,3,3]
; SSE41-NEXT: pextrq $1, %xmm1, %rax
; SSE41-NEXT: movswq %ax, %rax
; SSE41-NEXT: movd %rax, %xmm2
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vpmovsxwd (%rdi), %xmm0
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm1
-; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX1-NEXT: vpmovsxdq %xmm0, %xmm0
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
; X32-SSE41-NEXT: pinsrd $2, %eax, %xmm0
; X32-SSE41-NEXT: sarl $31, %eax
; X32-SSE41-NEXT: pinsrd $3, %eax, %xmm0
-; X32-SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,0,3,0]
+; X32-SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,2,3,3]
; X32-SSE41-NEXT: movd %xmm2, %eax
; X32-SSE41-NEXT: cwtl
; X32-SSE41-NEXT: movd %eax, %xmm1
; SSE2: # BB#0:
; SSE2-NEXT: andps %xmm1, %xmm0
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_bitwise_ops_test1b:
; SSSE3: # BB#0:
; SSSE3-NEXT: andps %xmm1, %xmm0
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_bitwise_ops_test1b:
; SSE41: # BB#0:
-; SSE41-NEXT: andps %xmm1, %xmm0
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
+; SSE41-NEXT: pand %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test1b:
; AVX1: # BB#0:
-; AVX1-NEXT: vandps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
+; AVX1-NEXT: vpand %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test1b:
; SSE2: # BB#0:
; SSE2-NEXT: orps %xmm1, %xmm0
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_bitwise_ops_test2b:
; SSSE3: # BB#0:
; SSSE3-NEXT: orps %xmm1, %xmm0
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_bitwise_ops_test2b:
; SSE41: # BB#0:
-; SSE41-NEXT: orps %xmm1, %xmm0
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
+; SSE41-NEXT: por %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test2b:
; AVX1: # BB#0:
-; AVX1-NEXT: vorps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
+; AVX1-NEXT: vpor %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test2b:
; SSE2-NEXT: xorps %xmm1, %xmm0
; SSE2-NEXT: xorps %xmm1, %xmm1
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_bitwise_ops_test3b:
; SSSE3-NEXT: xorps %xmm1, %xmm0
; SSSE3-NEXT: xorps %xmm1, %xmm1
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_bitwise_ops_test3b:
; SSE41: # BB#0:
-; SSE41-NEXT: xorps %xmm1, %xmm0
-; SSE41-NEXT: xorps %xmm1, %xmm1
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; SSE41-NEXT: pxor %xmm1, %xmm0
+; SSE41-NEXT: pxor %xmm1, %xmm1
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test3b:
; AVX1: # BB#0:
-; AVX1-NEXT: vxorps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vxorps %xmm1, %xmm1, %xmm1
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; AVX1-NEXT: vpxor %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test3b:
;
; SSE41-LABEL: combine_bitwise_ops_test4b:
; SSE41: # BB#0:
-; SSE41-NEXT: andps %xmm1, %xmm0
-; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
-; SSE41-NEXT: movaps %xmm2, %xmm0
+; SSE41-NEXT: pand %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm0[2,3],xmm2[4,5],xmm0[6,7]
+; SSE41-NEXT: movdqa %xmm2, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test4b:
; AVX1: # BB#0:
-; AVX1-NEXT: vandps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
+; AVX1-NEXT: vpand %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm2[0,1],xmm0[2,3],xmm2[4,5],xmm0[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test4b:
;
; SSE41-LABEL: combine_bitwise_ops_test5b:
; SSE41: # BB#0:
-; SSE41-NEXT: orps %xmm1, %xmm0
-; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
-; SSE41-NEXT: movaps %xmm2, %xmm0
+; SSE41-NEXT: por %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm0[2,3],xmm2[4,5],xmm0[6,7]
+; SSE41-NEXT: movdqa %xmm2, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test5b:
; AVX1: # BB#0:
-; AVX1-NEXT: vorps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
+; AVX1-NEXT: vpor %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm2[0,1],xmm0[2,3],xmm2[4,5],xmm0[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test5b:
; SSE2-NEXT: xorps %xmm1, %xmm0
; SSE2-NEXT: xorps %xmm1, %xmm1
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm0[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_bitwise_ops_test6b:
; SSSE3-NEXT: xorps %xmm1, %xmm0
; SSSE3-NEXT: xorps %xmm1, %xmm1
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm0[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,1,3]
+; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_bitwise_ops_test6b:
; SSE41: # BB#0:
-; SSE41-NEXT: xorps %xmm1, %xmm0
-; SSE41-NEXT: xorps %xmm1, %xmm1
-; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
-; SSE41-NEXT: movaps %xmm1, %xmm0
+; SSE41-NEXT: pxor %xmm1, %xmm0
+; SSE41-NEXT: pxor %xmm1, %xmm1
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3],xmm1[4,5],xmm0[6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_bitwise_ops_test6b:
; AVX1: # BB#0:
-; AVX1-NEXT: vxorps %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vxorps %xmm1, %xmm1, %xmm1
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
+; AVX1-NEXT: vpxor %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3],xmm1[4,5],xmm0[6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_bitwise_ops_test6b:
define <4 x i32> @combine_nested_undef_test1(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test1:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,0,0,1]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,1,0,1]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test1:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[3,0,0,1]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[3,1,0,1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 4, i32 3, i32 1>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 4, i32 0, i32 3>
define <4 x i32> @combine_nested_undef_test2(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test2:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,0,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test2:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,0,0,3]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 5, i32 2, i32 3>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 4, i32 0, i32 3>
define <4 x i32> @combine_nested_undef_test3(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test3:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,0,3]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test3:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,0,0,3]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 6, i32 2, i32 3>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 4, i32 0, i32 3>
define <4 x i32> @combine_nested_undef_test4(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test4:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,1]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; SSE-NEXT: retq
;
-; AVX-LABEL: combine_nested_undef_test4:
-; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,1]
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_nested_undef_test4:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_nested_undef_test4:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastq %xmm0, %xmm0
+; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 4, i32 7, i32 1>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 4, i32 4, i32 0, i32 3>
ret <4 x i32> %2
define <4 x i32> @combine_nested_undef_test5(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test5:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test5:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 5, i32 5, i32 2, i32 3>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 4, i32 4, i32 3>
define <4 x i32> @combine_nested_undef_test6(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test6:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,0,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test6:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,0,0,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 6, i32 2, i32 4>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 4, i32 0, i32 4>
define <4 x i32> @combine_nested_undef_test8(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test8:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,3,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test8:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,0,3,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 1, i32 4, i32 3, i32 4>
define <4 x i32> @combine_nested_undef_test9(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test9:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,0,2]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,2]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test9:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,3,0,2]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,3,2,2]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 3, i32 2, i32 5>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 1, i32 4, i32 2>
define <4 x i32> @combine_nested_undef_test10(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test10:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,1,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,1,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test10:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,0,1,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,1,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 1, i32 5, i32 5>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 4, i32 1, i32 4>
define <4 x i32> @combine_nested_undef_test11(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test11:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,2,1]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,2,1]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test11:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,0,2,1]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 2, i32 5, i32 4>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 4, i32 1, i32 0>
define <4 x i32> @combine_nested_undef_test12(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test12:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; SSE-NEXT: retq
;
; AVX1-LABEL: combine_nested_undef_test12:
; AVX1: # BB#0:
-; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_nested_undef_test12:
; AVX2: # BB#0:
-; AVX2-NEXT: vbroadcastss %xmm0, %xmm0
+; AVX2-NEXT: vpbroadcastq %xmm0, %xmm0
; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 0, i32 0, i32 2, i32 4>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 1, i32 4, i32 0, i32 4>
define <4 x i32> @combine_nested_undef_test15(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test15:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[0,0]
; SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,1]
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,1,0,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test15:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; AVX-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[0,0]
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm1[2,0],xmm0[3,1]
; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; AVX-NEXT: retq
; SSE2-LABEL: combine_nested_undef_test16:
; SSE2: # BB#0:
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,2,0,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_nested_undef_test16:
; SSSE3: # BB#0:
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,2,0,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_nested_undef_test16:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_nested_undef_test16:
; AVX1: # BB#0:
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,1,0,3]
; AVX1-NEXT: retq
;
define <4 x i32> @combine_nested_undef_test19(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test19:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[0,0]
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,2]
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,0,0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test19:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[0,0]
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,2]
; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,0,0,0]
; AVX-NEXT: retq
define <4 x i32> @combine_nested_undef_test24(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test24:
; SSE: # BB#0:
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,3,2,0]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,3,2,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test24:
; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[0,3,2,0]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[0,3,2,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 4, i32 1, i32 6, i32 7>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 3, i32 2, i32 4>
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; SSE-NEXT: retq
;
-; AVX-LABEL: combine_nested_undef_test25:
-; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_nested_undef_test25:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_nested_undef_test25:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastq %xmm0, %xmm0
+; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %B, <4 x i32> %A, <4 x i32> <i32 1, i32 5, i32 2, i32 4>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 1, i32 3, i32 1>
ret <4 x i32> %2
define <4 x i32> @combine_nested_undef_test26(<4 x i32> %A, <4 x i32> %B) {
; SSE-LABEL: combine_nested_undef_test26:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_nested_undef_test26:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %B, <4 x i32> %A, <4 x i32> <i32 1, i32 2, i32 6, i32 7>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 2, i32 3>
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; SSE-NEXT: retq
;
-; AVX-LABEL: combine_nested_undef_test27:
-; AVX: # BB#0:
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_nested_undef_test27:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_nested_undef_test27:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpbroadcastq %xmm0, %xmm0
+; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %B, <4 x i32> %A, <4 x i32> <i32 2, i32 1, i32 5, i32 4>
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 3, i32 2>
ret <4 x i32> %2
define <4 x float> @combine_test1(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_test1:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm1[1,3]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test1:
; SSE2-LABEL: combine_test2:
; SSE2: # BB#0:
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSE2-NEXT: retq
; SSSE3-LABEL: combine_test2:
; SSSE3: # BB#0:
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test2:
; SSE41: # BB#0:
-; SSE41-NEXT: movss %xmm0, %xmm1
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test2:
; AVX: # BB#0:
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
+; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
define <4 x float> @combine_test3(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_test3:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test3:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 1, i32 7>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 2, i32 4, i32 1>
define <4 x float> @combine_test4(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_test4:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: unpckhpd {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movapd %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test4:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
define <4 x float> @combine_test5(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_test5:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm0[2,0]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm2[2,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,0]
+; SSE2-NEXT: movaps %xmm2, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test5:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm0[2,0]
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
+; SSSE3-NEXT: movaps %xmm1, %xmm2
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm2[2,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,0]
+; SSSE3-NEXT: movaps %xmm2, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test5:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test5:
define <4 x i32> @combine_test6(<4 x i32> %a, <4 x i32> %b) {
; SSE2-LABEL: combine_test6:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm1[1,3]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test6:
; SSE2-LABEL: combine_test7:
; SSE2: # BB#0:
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSE2-NEXT: retq
; SSSE3-LABEL: combine_test7:
; SSSE3: # BB#0:
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test7:
; SSE41: # BB#0:
-; SSE41-NEXT: movss %xmm0, %xmm1
-; SSE41-NEXT: movaps %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
-; AVX-LABEL: combine_test7:
-; AVX: # BB#0:
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_test7:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_test7:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
+; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
ret <4 x i32> %2
define <4 x i32> @combine_test8(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test8:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test8:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 1, i32 7>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 0, i32 2, i32 4, i32 1>
define <4 x i32> @combine_test9(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test9:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test9:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
define <4 x i32> @combine_test10(<4 x i32> %a, <4 x i32> %b) {
; SSE2-LABEL: combine_test10:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm0[2,0]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm2[2,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,0]
+; SSE2-NEXT: movaps %xmm2, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test10:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm0[2,0]
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
+; SSSE3-NEXT: movaps %xmm1, %xmm2
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm2[2,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,0]
+; SSSE3-NEXT: movaps %xmm2, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test10:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_test10:
; AVX1: # BB#0:
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_test10:
;
; SSE41-LABEL: combine_test12:
; SSE41: # BB#0:
-; SSE41-NEXT: movss %xmm0, %xmm1
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test12:
; AVX: # BB#0:
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
+; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
%2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
define <4 x float> @combine_test13(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_test13:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test13:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
%2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
define <4 x float> @combine_test14(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_test14:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm1 = xmm0[1],xmm1[1]
-; SSE-NEXT: movaps %xmm1, %xmm0
+; SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test14:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1],xmm1[1]
+; AVX-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 6, i32 7, i32 5, i32 5>
%2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
;
; SSE41-LABEL: combine_test15:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test15:
;
; SSE41-LABEL: combine_test17:
; SSE41: # BB#0:
-; SSE41-NEXT: movss %xmm0, %xmm1
-; SSE41-NEXT: movaps %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
-; AVX-LABEL: combine_test17:
-; AVX: # BB#0:
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_test17:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_test17:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
+; AVX2-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
%2 = shufflevector <4 x i32> %1, <4 x i32> %a, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
ret <4 x i32> %2
define <4 x i32> @combine_test18(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test18:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test18:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
%2 = shufflevector <4 x i32> %1, <4 x i32> %a, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
define <4 x i32> @combine_test19(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test19:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm1 = xmm0[1],xmm1[1]
-; SSE-NEXT: movaps %xmm1, %xmm0
+; SSE-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test19:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1],xmm1[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 6, i32 7, i32 5, i32 5>
%2 = shufflevector <4 x i32> %1, <4 x i32> %a, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
;
; SSE41-LABEL: combine_test20:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7]
+; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_test20:
; AVX1: # BB#0:
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_test20:
define <4 x float> @combine_test1b(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_test1b:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm0[0,0]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[2,0]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm2[0,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm2[2,0]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test1b:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm0[0,0]
-; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[2,0]
+; SSSE3-NEXT: movaps %xmm1, %xmm2
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm2[0,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm2[2,0]
; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test1b:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
-; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm0[0,0]
-; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[2,0]
+; SSE41-NEXT: movaps %xmm1, %xmm2
+; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm2[0,0]
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm2[2,0]
; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
define <4 x float> @combine_test2b(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_test2b:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[1,1]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[1,1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,1,3]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test2b:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[1,1]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: movaps %xmm1, %xmm2
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[1,1]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,1,3]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test2b:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
-; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[1,1]
-; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE41-NEXT: movaps %xmm1, %xmm2
+; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[1,1]
+; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,1,3]
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test2b:
; AVX: # BB#0:
; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[1,1]
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 0, i32 5>
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm0[3,0]
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm2[0,2]
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[3,3]
-; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test3b:
; AVX-NEXT: vshufps {{.*#+}} xmm2 = xmm1[2,0],xmm0[3,0]
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm2[0,2]
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[3,3]
-; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 0, i32 6, i32 3>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 7, i32 2, i32 7>
define <4 x float> @combine_test4b(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_test4b:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0],xmm1[3,0]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[0,2]
+; SSE2-NEXT: movaps %xmm1, %xmm2
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm1[3,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm2[0,2]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test4b:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0],xmm1[3,0]
-; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[0,2]
+; SSSE3-NEXT: movaps %xmm1, %xmm2
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm0[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm1[3,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm2[0,2]
; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test4b:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
-; SSE41-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0],xmm1[3,0]
-; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[0,2]
+; SSE41-NEXT: movaps %xmm1, %xmm2
+; SSE41-NEXT: blendps {{.*#+}} xmm2 = xmm2[0],xmm0[1],xmm2[2],xmm0[3]
+; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm1[3,0]
+; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm2[0,2]
; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
define <4 x i8> @combine_test1c(<4 x i8>* %a, <4 x i8>* %b) {
; SSE2-LABEL: combine_test1c:
; SSE2: # BB#0:
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pextrw $1, %xmm0, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm0
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm0
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm0
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm0
-; SSE2-NEXT: movl (%rsi), %eax
-; SSE2-NEXT: movd %eax, %xmm1
-; SSE2-NEXT: pextrw $1, %xmm1, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm1
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm1
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm1
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm1
+; SSE2-NEXT: movd (%rdi), %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSE2-NEXT: movd (%rsi), %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSE2-NEXT: retq
; SSSE3-LABEL: combine_test1c:
; SSSE3: # BB#0:
; SSSE3-NEXT: movd (%rdi), %xmm0
-; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,128,128,128,1,128,128,128,2,128,128,128,3,128,128,128]
-; SSSE3-NEXT: pshufb %xmm1, %xmm0
-; SSSE3-NEXT: movd (%rsi), %xmm2
-; SSSE3-NEXT: pshufb %xmm1, %xmm2
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[1,3]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
-; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm0[3,0]
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm2[0,2]
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSSE3-NEXT: movd (%rsi), %xmm1
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[3,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[0,2]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test1c:
; SSE41: # BB#0:
; SSE41-NEXT: pmovzxbd (%rdi), %xmm1
; SSE41-NEXT: pmovzxbd (%rsi), %xmm0
-; SSE41-NEXT: movss %xmm1, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7]
; SSE41-NEXT: retq
;
-; AVX-LABEL: combine_test1c:
-; AVX: # BB#0:
-; AVX-NEXT: vpmovzxbd (%rdi), %xmm0
-; AVX-NEXT: vpmovzxbd (%rsi), %xmm1
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
-; AVX-NEXT: retq
+; AVX1-LABEL: combine_test1c:
+; AVX1: # BB#0:
+; AVX1-NEXT: vpmovzxbd (%rdi), %xmm0
+; AVX1-NEXT: vpmovzxbd (%rsi), %xmm1
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
+; AVX1-NEXT: retq
+;
+; AVX2-LABEL: combine_test1c:
+; AVX2: # BB#0:
+; AVX2-NEXT: vpmovzxbd (%rdi), %xmm0
+; AVX2-NEXT: vpmovzxbd (%rsi), %xmm1
+; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
+; AVX2-NEXT: retq
%A = load <4 x i8>* %a
%B = load <4 x i8>* %b
%1 = shufflevector <4 x i8> %A, <4 x i8> %B, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
define <4 x i8> @combine_test2c(<4 x i8>* %a, <4 x i8>* %b) {
; SSE2-LABEL: combine_test2c:
; SSE2: # BB#0:
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pextrw $1, %xmm0, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm0
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm0
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm0
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm0
-; SSE2-NEXT: movl (%rsi), %eax
-; SSE2-NEXT: movd %eax, %xmm1
-; SSE2-NEXT: pextrw $1, %xmm1, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm1
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm1
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm1
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm1
-; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE2-NEXT: movd (%rdi), %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSE2-NEXT: movd (%rsi), %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test2c:
; SSSE3: # BB#0:
; SSSE3-NEXT: movd (%rdi), %xmm0
-; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,128,128,128,1,128,128,128,2,128,128,128,3,128,128,128]
-; SSSE3-NEXT: pshufb %xmm1, %xmm0
-; SSSE3-NEXT: movd (%rsi), %xmm2
-; SSSE3-NEXT: pshufb %xmm1, %xmm2
-; SSSE3-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSSE3-NEXT: movd (%rsi), %xmm1
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test2c:
; SSE41: # BB#0:
; SSE41-NEXT: pmovzxbd (%rdi), %xmm0
; SSE41-NEXT: pmovzxbd (%rsi), %xmm1
-; SSE41-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test2c:
; AVX: # BB#0:
; AVX-NEXT: vpmovzxbd (%rdi), %xmm0
; AVX-NEXT: vpmovzxbd (%rsi), %xmm1
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%A = load <4 x i8>* %a
%B = load <4 x i8>* %b
define <4 x i8> @combine_test3c(<4 x i8>* %a, <4 x i8>* %b) {
; SSE2-LABEL: combine_test3c:
; SSE2: # BB#0:
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pextrw $1, %xmm0, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm0
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm0
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm0
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm0
-; SSE2-NEXT: movl (%rsi), %eax
-; SSE2-NEXT: movd %eax, %xmm1
-; SSE2-NEXT: pextrw $1, %xmm1, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm1
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm1
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm1
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm1
-; SSE2-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE2-NEXT: movd (%rdi), %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSE2-NEXT: movd (%rsi), %xmm0
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSE2-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test3c:
; SSSE3: # BB#0:
-; SSSE3-NEXT: movd (%rdi), %xmm0
-; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,128,128,128,1,128,128,128,2,128,128,128,3,128,128,128]
-; SSSE3-NEXT: pshufb %xmm1, %xmm0
-; SSSE3-NEXT: movd (%rsi), %xmm2
-; SSSE3-NEXT: pshufb %xmm1, %xmm2
-; SSSE3-NEXT: movhlps {{.*#+}} xmm0 = xmm2[1],xmm0[1]
+; SSSE3-NEXT: movd (%rdi), %xmm1
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSSE3-NEXT: movd (%rsi), %xmm0
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
+; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test3c:
; SSE41: # BB#0:
-; SSE41-NEXT: pmovzxbd (%rdi), %xmm0
-; SSE41-NEXT: pmovzxbd (%rsi), %xmm1
-; SSE41-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE41-NEXT: pmovzxbd (%rdi), %xmm1
+; SSE41-NEXT: pmovzxbd (%rsi), %xmm0
+; SSE41-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_test3c:
; AVX: # BB#0:
; AVX-NEXT: vpmovzxbd (%rdi), %xmm0
; AVX-NEXT: vpmovzxbd (%rsi), %xmm1
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%A = load <4 x i8>* %a
%B = load <4 x i8>* %b
define <4 x i8> @combine_test4c(<4 x i8>* %a, <4 x i8>* %b) {
; SSE2-LABEL: combine_test4c:
; SSE2: # BB#0:
-; SSE2-NEXT: movl (%rdi), %eax
-; SSE2-NEXT: movd %eax, %xmm0
-; SSE2-NEXT: pextrw $1, %xmm0, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm0
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm0
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm0
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm0
-; SSE2-NEXT: movl (%rsi), %eax
-; SSE2-NEXT: movd %eax, %xmm1
-; SSE2-NEXT: pextrw $1, %xmm1, %ecx
-; SSE2-NEXT: pinsrw $0, %eax, %xmm1
-; SSE2-NEXT: movzbl %ah, %eax
-; SSE2-NEXT: pinsrw $2, %eax, %xmm1
-; SSE2-NEXT: pinsrw $4, %ecx, %xmm1
-; SSE2-NEXT: shrl $8, %ecx
-; SSE2-NEXT: pinsrw $6, %ecx, %xmm1
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[0,2]
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
-; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[3,0],xmm0[2,0]
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
+; SSE2-NEXT: movd (%rdi), %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; SSE2-NEXT: movd (%rsi), %xmm2
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
+; SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
+; SSE2-NEXT: movdqa %xmm2, %xmm0
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[3,0],xmm0[2,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_test4c:
; SSSE3: # BB#0:
-; SSSE3-NEXT: movd (%rdi), %xmm0
-; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,128,128,128,1,128,128,128,2,128,128,128,3,128,128,128]
-; SSSE3-NEXT: pshufb %xmm1, %xmm0
+; SSSE3-NEXT: movd (%rdi), %xmm1
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
; SSSE3-NEXT: movd (%rsi), %xmm2
-; SSSE3-NEXT: pshufb %xmm1, %xmm2
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[0,2]
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,0,3,1]
+; SSSE3-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
+; SSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
+; SSSE3-NEXT: movdqa %xmm2, %xmm0
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[1,3]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[3,0],xmm0[2,0]
; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_test4c:
; SSE41: # BB#0:
-; SSE41-NEXT: pmovzxbd (%rdi), %xmm0
-; SSE41-NEXT: pmovzxbd (%rsi), %xmm1
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; SSE41-NEXT: pmovzxbd (%rdi), %xmm1
+; SSE41-NEXT: pmovzxbd (%rsi), %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: combine_test4c:
; AVX1: # BB#0:
; AVX1-NEXT: vpmovzxbd (%rdi), %xmm0
; AVX1-NEXT: vpmovzxbd (%rsi), %xmm1
-; AVX1-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7]
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_test4c:
define <4 x float> @combine_blend_01(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_blend_01:
; SSE2: # BB#0:
-; SSE2-NEXT: movss %xmm1, %xmm0
+; SSE2-NEXT: movsd %xmm1, %xmm0
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm0[0,0]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[2,3]
; SSE2-NEXT: movaps %xmm1, %xmm0
;
; SSSE3-LABEL: combine_blend_01:
; SSSE3: # BB#0:
-; SSSE3-NEXT: movss %xmm1, %xmm0
+; SSSE3-NEXT: movsd %xmm1, %xmm0
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,0],xmm0[0,0]
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[2,0],xmm0[2,3]
; SSSE3-NEXT: movaps %xmm1, %xmm0
;
; SSE41-LABEL: combine_blend_01:
; SSE41: # BB#0:
-; SSE41-NEXT: movsd %xmm1, %xmm0
+; SSE41-NEXT: blendpd {{.*#+}} xmm1 = xmm1[0],xmm0[1]
+; SSE41-NEXT: movapd %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_blend_01:
; AVX: # BB#0:
-; AVX-NEXT: vmovsd %xmm1, %xmm0, %xmm0
+; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
; AVX-NEXT: retq
%shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 undef, i32 2, i32 3>
%shuffle6 = shufflevector <4 x float> %shuffle, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 3>
;
; SSE41-LABEL: combine_blend_02:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm1[0],xmm0[1],xmm1[2],xmm0[3]
+; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_blend_02:
; SSE2-LABEL: combine_blend_123:
; SSE2: # BB#0:
; SSE2-NEXT: movaps %xmm1, %xmm2
-; SSE2-NEXT: movss %xmm0, %xmm2
-; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,3]
-; SSE2-NEXT: movaps %xmm2, %xmm0
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,0],xmm0[0,0]
+; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm0[2,3]
+; SSE2-NEXT: movsd %xmm2, %xmm1
+; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_blend_123:
; SSSE3: # BB#0:
; SSSE3-NEXT: movaps %xmm1, %xmm2
-; SSSE3-NEXT: movss %xmm0, %xmm2
-; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,3]
-; SSSE3-NEXT: movaps %xmm2, %xmm0
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,0],xmm0[0,0]
+; SSSE3-NEXT: shufps {{.*#+}} xmm2 = xmm2[2,0],xmm0[2,3]
+; SSSE3-NEXT: movsd %xmm2, %xmm1
+; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_blend_123:
; SSE41: # BB#0:
-; SSE41-NEXT: movss %xmm0, %xmm1
+; SSE41-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
; SSE41-NEXT: movaps %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_blend_123:
; AVX: # BB#0:
-; AVX-NEXT: vmovss %xmm0, %xmm1, %xmm0
+; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
; AVX-NEXT: retq
%shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 undef, i32 undef>
%shuffle6 = shufflevector <4 x float> %shuffle, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 undef>
define <4 x i32> @combine_test_movhl_1(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test_movhl_1:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test_movhl_1:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 2, i32 7, i32 5, i32 3>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 6, i32 1, i32 0, i32 3>
define <4 x i32> @combine_test_movhl_2(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test_movhl_2:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test_movhl_2:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 2, i32 0, i32 3, i32 6>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 3, i32 7, i32 0, i32 2>
define <4 x i32> @combine_test_movhl_3(<4 x i32> %a, <4 x i32> %b) {
; SSE-LABEL: combine_test_movhl_3:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_test_movhl_3:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 7, i32 6, i32 3, i32 2>
%2 = shufflevector <4 x i32> %1, <4 x i32> %b, <4 x i32> <i32 6, i32 0, i32 3, i32 2>
define <4 x float> @combine_undef_input_test1(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_undef_input_test1:
; SSE2: # BB#0:
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,3,1]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,3,1]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[1,2]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_undef_input_test1:
; SSSE3: # BB#0:
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,3,1]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,3,1]
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[1,2]
; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_undef_input_test1:
; SSE41: # BB#0:
-; SSE41-NEXT: movsd %xmm1, %xmm0
+; SSE41-NEXT: blendpd {{.*#+}} xmm1 = xmm1[0],xmm0[1]
+; SSE41-NEXT: movapd %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test1:
; AVX: # BB#0:
-; AVX-NEXT: vmovsd %xmm1, %xmm0, %xmm0
+; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 4, i32 2, i32 3, i32 1>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 1, i32 2>
define <4 x float> @combine_undef_input_test2(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test2:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test2:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 6, i32 0, i32 1, i32 7>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 1, i32 2, i32 4, i32 5>
define <4 x float> @combine_undef_input_test3(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test3:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test3:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 5, i32 1, i32 7>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 2, i32 4, i32 1>
define <4 x float> @combine_undef_input_test4(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test4:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: unpckhpd {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movapd %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test4:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
define <4 x float> @combine_undef_input_test5(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_undef_input_test5:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSE2-NEXT: movsd %xmm0, %xmm1
+; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_undef_input_test5:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSSE3-NEXT: movsd %xmm0, %xmm1
+; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_undef_input_test5:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSE41-NEXT: blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test5:
; AVX: # BB#0:
-; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 4, i32 1, i32 3>
%2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 2, i32 6, i32 7>
define <4 x float> @combine_undef_input_test11(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_undef_input_test11:
; SSE2: # BB#0:
-; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,3,1]
+; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,3,1]
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[1,2]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_undef_input_test11:
; SSSE3: # BB#0:
-; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,3,1]
+; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,3,1]
; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[1,2]
; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_undef_input_test11:
; SSE41: # BB#0:
-; SSE41-NEXT: movsd %xmm1, %xmm0
+; SSE41-NEXT: blendpd {{.*#+}} xmm1 = xmm1[0],xmm0[1]
+; SSE41-NEXT: movapd %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test11:
; AVX: # BB#0:
-; AVX-NEXT: vmovsd %xmm1, %xmm0, %xmm0
+; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 4, i32 2, i32 3, i32 1>
%2 = shufflevector <4 x float> %b, <4 x float> %1, <4 x i32> <i32 0, i32 1, i32 5, i32 6>
define <4 x float> @combine_undef_input_test12(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test12:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test12:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 6, i32 0, i32 1, i32 7>
%2 = shufflevector <4 x float> %b, <4 x float> %1, <4 x i32> <i32 5, i32 6, i32 0, i32 1>
define <4 x float> @combine_undef_input_test13(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test13:
; SSE: # BB#0:
-; SSE-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test13:
; AVX: # BB#0:
-; AVX-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 5, i32 1, i32 7>
%2 = shufflevector <4 x float> %b, <4 x float> %1, <4 x i32> <i32 4, i32 5, i32 0, i32 5>
define <4 x float> @combine_undef_input_test14(<4 x float> %a, <4 x float> %b) {
; SSE-LABEL: combine_undef_input_test14:
; SSE: # BB#0:
-; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; SSE-NEXT: unpckhpd {{.*#+}} xmm1 = xmm1[1],xmm0[1]
+; SSE-NEXT: movapd %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test14:
; AVX: # BB#0:
-; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm1[1],xmm0[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
%2 = shufflevector <4 x float> %b, <4 x float> %1, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
define <4 x float> @combine_undef_input_test15(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: combine_undef_input_test15:
; SSE2: # BB#0:
-; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSE2-NEXT: movsd %xmm0, %xmm1
+; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: combine_undef_input_test15:
; SSSE3: # BB#0:
-; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSSE3-NEXT: movsd %xmm0, %xmm1
+; SSSE3-NEXT: movaps %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: combine_undef_input_test15:
; SSE41: # BB#0:
-; SSE41-NEXT: blendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; SSE41-NEXT: blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
; SSE41-NEXT: retq
;
; AVX-LABEL: combine_undef_input_test15:
; AVX: # BB#0:
-; AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
; AVX-NEXT: retq
%1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 4, i32 1, i32 3>
%2 = shufflevector <4 x float> %b, <4 x float> %1, <4 x i32> <i32 4, i32 6, i32 2, i32 3>
;
; AVX1-LABEL: zext_8i16_to_8i32:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm2 = xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
-; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
+; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
+; AVX1-NEXT: vpmovzxwd %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: zext_8i16_to_8i32:
define <4 x i64> @zext_4i32_to_4i64(<4 x i32> %A) nounwind uwtable readnone ssp {
; SSE2-LABEL: zext_4i32_to_4i64:
; SSE2: # BB#0: # %entry
-; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,0,1,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
; SSE2-NEXT: pand %xmm3, %xmm2
-; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; SSE2-NEXT: pand %xmm3, %xmm1
; SSE2-NEXT: movdqa %xmm2, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: zext_4i32_to_4i64:
; SSSE3: # BB#0: # %entry
-; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,0,1,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
; SSSE3-NEXT: pand %xmm3, %xmm2
-; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; SSSE3-NEXT: pand %xmm3, %xmm1
; SSSE3-NEXT: movdqa %xmm2, %xmm0
; SSSE3-NEXT: retq
; SSE41-NEXT: pmovzxdq %xmm0, %xmm2
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
; SSE41-NEXT: pand %xmm3, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,0,3,0]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
; SSE41-NEXT: pand %xmm3, %xmm1
; SSE41-NEXT: movdqa %xmm2, %xmm0
; SSE41-NEXT: retq
; AVX1-LABEL: zext_4i32_to_4i64:
; AVX1: # BB#0: # %entry
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; AVX1-NEXT: vpunpckhdq {{.*#+}} xmm2 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; AVX1-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT: vpunpckhdq {{.*#+}} xmm1 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; AVX1-NEXT: vpmovzxdq %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: zext_4i32_to_4i64:
;
; AVX1-LABEL: zext_8i8_to_8i32:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm0[4,4,5,5,6,6,7,7]
-; AVX1-NEXT: vpmovzxwd %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: vandps .{{.*}}, %ymm0, %ymm0
+; AVX1-NEXT: vpmovzxwd %xmm0, %xmm1
+; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT: vandps {{.*}}, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: zext_8i8_to_8i32:
; AVX2: # BB#0: # %entry
; AVX2-NEXT: vpmovzxwd %xmm0, %ymm0
-; AVX2-NEXT: vpbroadcastd .{{.*}}, %ymm1
+; AVX2-NEXT: vpbroadcastd {{.*}}, %ymm1
; AVX2-NEXT: vpand %ymm1, %ymm0, %ymm0
; AVX2-NEXT: retq
entry:
;
; AVX1-LABEL: zext_16i8_to_16i16:
; AVX1: # BB#0: # %entry
-; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
-; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
+; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; AVX1-NEXT: vpmovzxbw %xmm0, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: zext_16i8_to_16i16:
define <4 x float> @select_of_shuffles_0(<2 x float> %a0, <2 x float> %b0, <2 x float> %a1, <2 x float> %b1) {
; CHECK-LABEL: select_of_shuffles_0:
; CHECK: # BB#0:
-; CHECK-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; CHECK-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm3[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; CHECK-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0],xmm3[0]
; CHECK-NEXT: subps %xmm1, %xmm0
; CHECK-NEXT: retq
%1 = shufflevector <2 x float> %a0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
; CHECK-LABEL: shuf3:
; CHECK: # BB#0: # %entry
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,0,0]
-; CHECK-NEXT: movdqa %xmm0, (%eax)
+; CHECK-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; CHECK-NEXT: movaps %xmm1, (%eax)
; CHECK-NEXT: retl
entry:
%shuffle.i.i.i12 = shufflevector <4 x float> %tmp10, <4 x float> %vecinit15, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
define <8 x i8> @shuf4(<4 x i8> %a, <4 x i8> %b) nounwind readnone {
; CHECK-LABEL: shuf4:
; CHECK: # BB#0:
-; CHECK-NEXT: pshufb {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0],zero,xmm1[4],zero,xmm1[8],zero,xmm1[12],zero
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0],zero,xmm0[4],zero,xmm0[8],zero,xmm0[12],zero,zero,zero,zero,zero,zero,zero,zero,zero
-; CHECK-NEXT: por %xmm1, %xmm0
+; CHECK-NEXT: movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; CHECK-NEXT: pshufb %xmm2, %xmm1
+; CHECK-NEXT: pshufb %xmm2, %xmm0
+; CHECK-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: retl
%vshuf = shufflevector <4 x i8> %a, <4 x i8> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i8> %vshuf
; CHECK: # BB#0:
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: movdqa {{.*#+}} xmm0 = <4,33,u,u,u,u,u,u>
-; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,0,0,8,10,12,14],zero,zero,zero,zero,zero,zero,zero,zero
+; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; CHECK-NEXT: movlpd %xmm0, (%eax)
; CHECK-NEXT: retl
%v = shufflevector <2 x i8> <i8 4, i8 33>, <2 x i8> undef, <8 x i32> <i32 1, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
projects / oota-llvm.git / commitdiff