break;
case X86ISD::MOVSLDUP:
DecodeMOVSLDUPMask(VT, Mask);
+ IsUnary = true;
break;
case X86ISD::MOVSHDUP:
DecodeMOVSHDUPMask(VT, Mask);
+ IsUnary = true;
break;
case X86ISD::MOVDDUP:
+ DecodeMOVDDUPMask(VT, Mask);
+ IsUnary = true;
+ break;
case X86ISD::MOVLHPD:
case X86ISD::MOVLPD:
case X86ISD::MOVLPS:
assert(Mask.size() == 2 && "Unexpected mask size for v2 shuffle!");
if (isSingleInputShuffleMask(Mask)) {
+ // Use low duplicate instructions for masks that match their pattern.
+ if (Subtarget->hasSSE3())
+ if (isShuffleEquivalent(Mask, 0, 0))
+ return DAG.getNode(X86ISD::MOVDDUP, DL, MVT::v2f64, V1);
+
// Straight shuffle of a single input vector. Simulate this by using the
// single input as both of the "inputs" to this instruction..
unsigned SHUFPDMask = (Mask[0] == 1) | ((Mask[1] == 1) << 1);
Mask, Subtarget, DAG))
return Broadcast;
+ // Use even/odd duplicate instructions for masks that match their pattern.
+ if (Subtarget->hasSSE3()) {
+ if (isShuffleEquivalent(Mask, 0, 0, 2, 2))
+ return DAG.getNode(X86ISD::MOVSLDUP, DL, MVT::v4f32, V1);
+ if (isShuffleEquivalent(Mask, 1, 1, 3, 3))
+ return DAG.getNode(X86ISD::MOVSHDUP, DL, MVT::v4f32, V1);
+ }
+
if (Subtarget->hasAVX()) {
// If we have AVX, we can use VPERMILPS which will allow folding a load
// into the shuffle.
Mask, Subtarget, DAG))
return Broadcast;
+ // Use low duplicate instructions for masks that match their pattern.
+ if (isShuffleEquivalent(Mask, 0, 0, 2, 2))
+ return DAG.getNode(X86ISD::MOVDDUP, DL, MVT::v4f64, V1);
+
if (!is128BitLaneCrossingShuffleMask(MVT::v4f64, Mask)) {
// Non-half-crossing single input shuffles can be lowerid with an
// interleaved permutation.
if (is128BitLaneRepeatedShuffleMask(MVT::v8f32, Mask, RepeatedMask)) {
assert(RepeatedMask.size() == 4 &&
"Repeated masks must be half the mask width!");
+
+ // Use even/odd duplicate instructions for masks that match their pattern.
+ if (isShuffleEquivalent(Mask, 0, 0, 2, 2, 4, 4, 6, 6))
+ return DAG.getNode(X86ISD::MOVSLDUP, DL, MVT::v8f32, V1);
+ if (isShuffleEquivalent(Mask, 1, 1, 3, 3, 5, 5, 7, 7))
+ return DAG.getNode(X86ISD::MOVSHDUP, DL, MVT::v8f32, V1);
+
if (isSingleInputShuffleMask(Mask))
return DAG.getNode(X86ISD::VPERMILPI, DL, MVT::v8f32, V1,
getV4X86ShuffleImm8ForMask(RepeatedMask, DAG));
: InVec.getOperand(1);
// If inputs to shuffle are the same for both ops, then allow 2 uses
- unsigned AllowedUses = InVec.getOperand(0) == InVec.getOperand(1) ? 2 : 1;
+ unsigned AllowedUses = InVec.getNumOperands() > 1 &&
+ InVec.getOperand(0) == InVec.getOperand(1) ? 2 : 1;
if (LdNode.getOpcode() == ISD::BITCAST) {
// Don't duplicate a load with other uses.
(VMOVHPSrm VR128:$src1, addr:$src2)>;
// VMOVHPD patterns
-
+
// FIXME: Instead of X86Unpckl, there should be a X86Movlhpd here, the problem
// is during lowering, where it's not possible to recognize the load fold
// cause it has two uses through a bitcast. One use disappears at isel time
(VUNPCKHPDYrr VR256:$src1, VR256:$src2)>;
}
-let Predicates = [HasAVX] in {
- // FIXME: Instead of X86Movddup, there should be a X86Unpckl here, the
- // problem is during lowering, where it's not possible to recognize the load
- // fold cause it has two uses through a bitcast. One use disappears at isel
- // time and the fold opportunity reappears.
- def : Pat<(v2f64 (X86Movddup VR128:$src)),
- (VUNPCKLPDrr VR128:$src, VR128:$src)>;
-}
-
-let Predicates = [UseSSE2] in {
- // FIXME: Instead of X86Movddup, there should be a X86Unpckl here, the
- // problem is during lowering, where it's not possible to recognize the load
- // fold cause it has two uses through a bitcast. One use disappears at isel
- // time and the fold opportunity reappears.
- def : Pat<(v2f64 (X86Movddup VR128:$src)),
- (UNPCKLPDrr VR128:$src, VR128:$src)>;
-}
-
//===----------------------------------------------------------------------===//
// SSE 1 & 2 - Extract Floating-Point Sign mask
//===----------------------------------------------------------------------===//
//===---------------------------------------------------------------------===//
multiclass sse3_replicate_dfp<string OpcodeStr> {
-let hasSideEffects = 0 in
def rr : S3DI<0x12, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
!strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
- [], IIC_SSE_MOV_LH>, Sched<[WriteFShuffle]>;
+ [(set VR128:$dst, (v2f64 (X86Movddup VR128:$src)))],
+ IIC_SSE_MOV_LH>, Sched<[WriteFShuffle]>;
def rm : S3DI<0x12, MRMSrcMem, (outs VR128:$dst), (ins f64mem:$src),
!strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
[(set VR128:$dst,
(loadv2f64 (add addr:$src, (iPTR 16))),
(iPTR 2)),
(VMOVUPDYrm addr:$src)>;
-
+
def : Pat<(insert_subvector
(v32i8 (insert_subvector
undef, (bc_v16i8 (loadv2i64 addr:$src)), (iPTR 0))),
(bc_v16i8 (loadv2i64 (add addr:$src, (iPTR 16)))),
(iPTR 16)),
(VMOVDQUYrm addr:$src)>;
-
+
def : Pat<(insert_subvector
(v16i16 (insert_subvector
undef, (bc_v8i16 (loadv2i64 addr:$src)), (iPTR 0))),
(bc_v8i16 (loadv2i64 (add addr:$src, (iPTR 16)))),
(iPTR 8)),
(VMOVDQUYrm addr:$src)>;
-
+
def : Pat<(insert_subvector
(v8i32 (insert_subvector
undef, (bc_v4i32 (loadv2i64 addr:$src)), (iPTR 0))),
entry:
%shuffle = shufflevector <16 x i16> %a, <16 x i16> undef, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, 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 i16> %shuffle
-}
-
-; CHECK: vmovq
-; CHECK-NEXT: vunpcklpd %xmm
-; CHECK-NEXT: vinsertf128 $1
-define <4 x i64> @funcC(i64 %q) nounwind uwtable readnone ssp {
-entry:
+}\r
+\r
+; CHECK: vmovq\r
+; CHECK-NEXT: vmovddup %xmm\r
+; CHECK-NEXT: vinsertf128 $1\r
+define <4 x i64> @funcC(i64 %q) nounwind uwtable readnone ssp {\r
+entry:\r
%vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
%vecinit2.i = insertelement <4 x i64> %vecinit.i, i64 %q, i32 1
%vecinit4.i = insertelement <4 x i64> %vecinit2.i, i64 %q, i32 2
%vecinit6.i = insertelement <4 x i64> %vecinit4.i, i64 %q, i32 3
- ret <4 x i64> %vecinit6.i
-}
-
-; CHECK: vunpcklpd %xmm
-; CHECK-NEXT: vinsertf128 $1
-define <4 x double> @funcD(double %q) nounwind uwtable readnone ssp {
-entry:
+ ret <4 x i64> %vecinit6.i\r
+}\r
+\r
+; CHECK: vmovddup %xmm\r
+; CHECK-NEXT: vinsertf128 $1\r
+define <4 x double> @funcD(double %q) nounwind uwtable readnone ssp {\r
+entry:\r
%vecinit.i = insertelement <4 x double> undef, double %q, i32 0
%vecinit2.i = insertelement <4 x double> %vecinit.i, double %q, i32 1
%vecinit4.i = insertelement <4 x double> %vecinit2.i, double %q, i32 2
%vecinit6.i = insertelement <4 x double> %vecinit4.i, double %q, i32 3
ret <4 x double> %vecinit6.i
}
-
-; Test this turns into a broadcast:
-; shuffle (scalar_to_vector (load (ptr + 4))), undef, <0, 0, 0, 0>
-;
-; CHECK: vbroadcastss
-define <8 x float> @funcE() nounwind {
-allocas:
+\r
+; Test this turns into a broadcast:\r
+; shuffle (scalar_to_vector (load (ptr + 4))), undef, <0, 0, 0, 0>\r
+;\r
+; CHECK: vbroadcastss\r
+define <8 x float> @funcE() nounwind {\r
+allocas:\r
%udx495 = alloca [18 x [18 x float]], align 32
br label %for_test505.preheader
define <4 x double> @_inreg4xdouble(<4 x double> %a) {
%b = shufflevector <4 x double> %a, <4 x double> undef, <4 x i32> zeroinitializer
ret <4 x double> %b
-}
-
-;CHECK-LABEL: _inreg2xdouble:
-;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
+}\r
+\r
+;CHECK-LABEL: _inreg2xdouble:\r
+;CHECK: vmovddup\r
+;CHECK: ret\r
+define <2 x double> @_inreg2xdouble(<2 x double> %a) {\r
+ %b = shufflevector <2 x double> %a, <2 x double> undef, <2 x i32> zeroinitializer\r
ret <2 x double> %b
}
; GNU_SINCOS: callq sincosf
; GNU_SINCOS: movss 4(%rsp), %xmm0
; GNU_SINCOS: addss (%rsp), %xmm0
-
-; OSX_SINCOS-LABEL: test1:
-; OSX_SINCOS: callq ___sincosf_stret
-; OSX_SINCOS: movaps %xmm0, %xmm1
-; OSX_SINCOS: shufps {{.*}} ## xmm1 = xmm1[1,1,2,3]
-; OSX_SINCOS: addss %xmm0, %xmm1
-
-; OSX_NOOPT: test1
-; OSX_NOOPT: callq _sinf
+\r
+; OSX_SINCOS-LABEL: test1:\r
+; OSX_SINCOS: callq ___sincosf_stret\r
+; OSX_SINCOS: movshdup {{.*}} xmm1 = xmm0[1,1,3,3]\r
+; OSX_SINCOS: addss %xmm1, %xmm0\r
+\r
+; OSX_NOOPT: test1\r
+; OSX_NOOPT: callq _sinf\r
; OSX_NOOPT: callq _cosf
%call = tail call float @sinf(float %x) nounwind readnone
%call1 = tail call float @cosf(float %x) nounwind readnone
; This used to compile to insertps $0 + insertps $16. insertps $0 is always
; pointless.
-define <2 x float> @buildvector(<2 x float> %A, <2 x float> %B) nounwind {
-; X32-LABEL: buildvector:
-; X32: ## BB#0: ## %entry
-; X32-NEXT: movaps %xmm0, %xmm2
-; X32-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,1,2,3]
-; X32-NEXT: addss %xmm1, %xmm0
-; 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: movaps %xmm0, %xmm2
-; X64-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,1,2,3]
-; X64-NEXT: addss %xmm1, %xmm0
-; 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
+define <2 x float> @buildvector(<2 x float> %A, <2 x float> %B) nounwind {\r
+; X32-LABEL: buildvector:\r
+; X32: ## BB#0: ## %entry\r
+; X32-NEXT: movshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]\r
+; X32-NEXT: movshdup {{.*#+}} xmm3 = xmm1[1,1,3,3]\r
+; X32-NEXT: addss %xmm1, %xmm0\r
+; X32-NEXT: addss %xmm2, %xmm3\r
+; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[2,3]\r
+; X32-NEXT: retl\r
+;\r
+; X64-LABEL: buildvector:\r
+; X64: ## BB#0: ## %entry\r
+; X64-NEXT: movshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]\r
+; X64-NEXT: movshdup {{.*#+}} xmm3 = xmm1[1,1,3,3]\r
+; X64-NEXT: addss %xmm1, %xmm0\r
+; X64-NEXT: addss %xmm2, %xmm3\r
+; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[2,3]\r
+; X64-NEXT: retq\r
+entry:\r
+ %tmp7 = extractelement <2 x float> %A, i32 0\r
%tmp5 = extractelement <2 x float> %A, i32 1
%tmp3 = extractelement <2 x float> %B, i32 0
%tmp1 = extractelement <2 x float> %B, i32 1
define void @rsqrtss(<4 x float> %a) nounwind uwtable ssp {
entry:
-; CHECK-LABEL: rsqrtss:
-; CHECK: rsqrtss %xmm0, %xmm0
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: shufps
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: movap
-; CHECK-NEXT: jmp
+; CHECK-LABEL: rsqrtss:\r
+; CHECK: rsqrtss %xmm0, %xmm0\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movshdup\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movap\r
+; CHECK-NEXT: jmp\r
%0 = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %a) nounwind
%a.addr.0.extract = extractelement <4 x float> %0, i32 0
define void @rcpss(<4 x float> %a) nounwind uwtable ssp {
entry:
-; CHECK-LABEL: rcpss:
-; CHECK: rcpss %xmm0, %xmm0
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: shufps
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: movap
-; CHECK-NEXT: jmp
+; CHECK-LABEL: rcpss:\r
+; CHECK: rcpss %xmm0, %xmm0\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movshdup\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movap\r
+; CHECK-NEXT: jmp\r
%0 = tail call <4 x float> @llvm.x86.sse.rcp.ss(<4 x float> %a) nounwind
%a.addr.0.extract = extractelement <4 x float> %0, i32 0
define void @sqrtss(<4 x float> %a) nounwind uwtable ssp {
entry:
-; CHECK-LABEL: sqrtss:
-; CHECK: sqrtss %xmm0, %xmm0
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: shufps
-; CHECK-NEXT: cvtss2sd %xmm0
-; CHECK-NEXT: movap
-; CHECK-NEXT: jmp
+; CHECK-LABEL: sqrtss:\r
+; CHECK: sqrtss %xmm0, %xmm0\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movshdup\r
+; CHECK-NEXT: cvtss2sd %xmm0\r
+; CHECK-NEXT: movap\r
+; CHECK-NEXT: jmp\r
%0 = tail call <4 x float> @llvm.x86.sse.sqrt.ss(<4 x float> %a) nounwind
%a.addr.0.extract = extractelement <4 x float> %0, i32 0
; RUN: llc < %s -mcpu=yonah -march=x86 -mtriple=i386-linux-gnu -o - | FileCheck %s --check-prefix=X32
; PR7518
-define void @test1(<2 x float> %Q, float *%P2) nounwind {
-; X64-LABEL: test1:
-; X64: # BB#0:
-; X64-NEXT: movaps %xmm0, %xmm1
-; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[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: movaps %xmm0, %xmm1
-; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1,2,3]
-; X32-NEXT: addss %xmm0, %xmm1
-; X32-NEXT: movss %xmm1, (%eax)
-; X32-NEXT: retl
+define void @test1(<2 x float> %Q, float *%P2) nounwind {\r
+; X64-LABEL: test1:\r
+; X64: # BB#0:\r
+; X64-NEXT: movshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]\r
+; X64-NEXT: addss %xmm0, %xmm1\r
+; X64-NEXT: movss %xmm1, (%rdi)\r
+; X64-NEXT: retq\r
+;\r
+; X32-LABEL: test1:\r
+; X32: # BB#0:\r
+; X32-NEXT: movl {{[0-9]+}}(%esp), %eax\r
+; X32-NEXT: movshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]\r
+; X32-NEXT: addss %xmm0, %xmm1\r
+; X32-NEXT: movss %xmm1, (%eax)\r
+; X32-NEXT: retl\r
%a = extractelement <2 x float> %Q, i32 0
%b = extractelement <2 x float> %Q, i32 1
%c = fadd float %a, %b
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
; 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: 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: movzbl %dl, %edx
-; CHECK-WIDE-NEXT: orl %eax, %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: 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: vcvttss2si %xmm1, %ecx\r
+; CHECK-WIDE-NEXT: movzbl %cl, %ecx\r
+; CHECK-WIDE-NEXT: orl %eax, %ecx\r
+; CHECK-WIDE-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax\r
+; CHECK-WIDE-NEXT: shll $8, %eax\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %edx\r
+; CHECK-WIDE-NEXT: movzbl %dl, %edx\r
+; CHECK-WIDE-NEXT: orl %eax, %edx\r
+; CHECK-WIDE-NEXT: vpinsrw $0, %edx, %xmm0, %xmm1\r
+; CHECK-WIDE-NEXT: vpinsrw $1, %ecx, %xmm1, %xmm1\r
+; CHECK-WIDE-NEXT: vextractf128 $1, %ymm0, %xmm0\r
+; CHECK-WIDE-NEXT: vmovshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm2, %eax\r
+; CHECK-WIDE-NEXT: shll $8, %eax\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %ecx\r
; CHECK-WIDE-NEXT: movzbl %cl, %ecx
; CHECK-WIDE-NEXT: orl %eax, %ecx
; CHECK-WIDE-NEXT: vpinsrw $2, %ecx, %xmm1, %xmm1
; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax
; CHECK-WIDE-NEXT: shll $8, %eax
; 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: 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: vcvttss2si %xmm1, %ecx\r
+; CHECK-WIDE-NEXT: movzbl %cl, %ecx\r
+; CHECK-WIDE-NEXT: orl %eax, %ecx\r
+; CHECK-WIDE-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm1, %eax\r
+; CHECK-WIDE-NEXT: shll $8, %eax\r
+; CHECK-WIDE-NEXT: vcvttss2si %xmm0, %edx\r
; CHECK-WIDE-NEXT: movzbl %dl, %edx
; CHECK-WIDE-NEXT: orl %eax, %edx
; CHECK-WIDE-NEXT: vpinsrw $0, %edx, %xmm0, %xmm0
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSE3-LABEL: shuffle_v2f64_00:
-; SSE3: # BB#0:
-; SSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE3-NEXT: retq
-;
-; SSSE3-LABEL: shuffle_v2f64_00:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: shuffle_v2f64_00:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: shuffle_v2f64_00:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX-NEXT: retq
- %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 0, i32 0>
- ret <2 x double> %shuffle
+;\r
+; SSE3-LABEL: shuffle_v2f64_00:\r
+; SSE3: # BB#0:\r
+; SSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE3-NEXT: retq\r
+;\r
+; SSSE3-LABEL: shuffle_v2f64_00:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: shuffle_v2f64_00:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: shuffle_v2f64_00:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 0, i32 0>\r
+ ret <2 x double> %shuffle\r
}
define <2 x double> @shuffle_v2f64_10(<2 x double> %a, <2 x double> %b) {
; SSE-LABEL: shuffle_v2f64_10:
; SSE2-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0,0]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
-;
-; SSE3-LABEL: shuffle_v2f64_22:
-; SSE3: # BB#0:
-; SSE3-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; SSE3-NEXT: movapd %xmm1, %xmm0
-; SSE3-NEXT: retq
-;
-; SSSE3-LABEL: shuffle_v2f64_22:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; SSSE3-NEXT: movapd %xmm1, %xmm0
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: shuffle_v2f64_22:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; SSE41-NEXT: movapd %xmm1, %xmm0
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: shuffle_v2f64_22:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0,0]
-; AVX-NEXT: retq
- %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 2, i32 2>
- ret <2 x double> %shuffle
+;\r
+; SSE3-LABEL: shuffle_v2f64_22:\r
+; SSE3: # BB#0:\r
+; SSE3-NEXT: movddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; SSE3-NEXT: retq\r
+;\r
+; SSSE3-LABEL: shuffle_v2f64_22:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: shuffle_v2f64_22:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: shuffle_v2f64_22:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; AVX-NEXT: retq\r
+ %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 2, i32 2>\r
+ ret <2 x double> %shuffle\r
}
define <2 x double> @shuffle_v2f64_32(<2 x double> %a, <2 x double> %b) {
; SSE-LABEL: shuffle_v2f64_32:
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSE3-LABEL: insert_dup_reg_v2f64:
-; SSE3: # BB#0:
-; SSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE3-NEXT: retq
-;
-; SSSE3-LABEL: insert_dup_reg_v2f64:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: insert_dup_reg_v2f64:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: insert_dup_reg_v2f64:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX-NEXT: retq
- %v = insertelement <2 x double> undef, double %a, i32 0
- %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <2 x i32> <i32 0, i32 0>
+;\r
+; SSE3-LABEL: insert_dup_reg_v2f64:\r
+; SSE3: # BB#0:\r
+; SSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE3-NEXT: retq\r
+;\r
+; SSSE3-LABEL: insert_dup_reg_v2f64:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: insert_dup_reg_v2f64:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: insert_dup_reg_v2f64:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %v = insertelement <2 x double> undef, double %a, i32 0\r
+ %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <2 x i32> <i32 0, i32 0>\r
ret <2 x double> %shuffle
}
define <2 x double> @insert_dup_mem_v2f64(double* %ptr) {
target triple = "x86_64-unknown-unknown"
-define <4 x double> @shuffle_v4f64_0000(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_0000:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
+define <4 x double> @shuffle_v4f64_0000(<4 x double> %a, <4 x double> %b) {\r
+; AVX1-LABEL: shuffle_v4f64_0000:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4f64_0000:
; AVX2: # BB#0:
; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
ret <4 x double> %shuffle
}
-define <4 x double> @shuffle_v4f64_0001(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_0001:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX1-NEXT: retq
-;
+define <4 x double> @shuffle_v4f64_0001(<4 x double> %a, <4 x double> %b) {\r
+; AVX1-LABEL: shuffle_v4f64_0001:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4f64_0001:
; AVX2: # BB#0:
; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
define <4 x double> @shuffle_v4f64_0020(<4 x double> %a, <4 x double> %b) {
; AVX1-LABEL: shuffle_v4f64_0020:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1\r
+; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4f64_0020:
; AVX2: # BB#0:
; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,2,0]
}
define <4 x double> @shuffle_v4f64_1000(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_1000:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX1-NEXT: retq
-;
+; AVX1-LABEL: shuffle_v4f64_1000:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4f64_1000:
; AVX2: # BB#0:
; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[1,0,0,0]
}
define <4 x double> @shuffle_v4f64_2200(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_2200:
-; AVX1: # BB#0:
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
-; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v4f64_2200:
+; AVX1-LABEL: shuffle_v4f64_2200:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]\r
+; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; AVX1-NEXT: retq\r
+;\r
+; AVX2-LABEL: shuffle_v4f64_2200:\r
; AVX2: # BB#0:
; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]
; AVX2-NEXT: retq
ret <4 x double> %shuffle
}
-define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) {
-; ALL-LABEL: shuffle_v4f64_0022:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; ALL-NEXT: retq
- %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
- ret <4 x double> %shuffle
+define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) {\r
+; ALL-LABEL: shuffle_v4f64_0022:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 2>\r
+ ret <4 x double> %shuffle\r
}
define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) {
ret <4 x double> %shuffle
}
-define <4 x double> @shuffle_v4f64_0423(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_0423:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,2,2]
-; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
-; AVX1-NEXT: retq
-;
+define <4 x double> @shuffle_v4f64_0423(<4 x double> %a, <4 x double> %b) {\r
+; AVX1-LABEL: shuffle_v4f64_0423:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]\r
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4f64_0423:
; AVX2: # BB#0:
; AVX2-NEXT: vbroadcastsd %xmm1, %ymm1
ret <4 x double> %shuffle
}
-define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) {
-; ALL-LABEL: shuffle_v4f64_0462:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,2,2]
-; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3]
-; ALL-NEXT: retq
- %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 6, i32 2>
+define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) {\r
+; ALL-LABEL: shuffle_v4f64_0462:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3]\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 6, i32 2>\r
ret <4 x double> %shuffle
}
ret <4 x double> %shuffle
}
-define <4 x i64> @shuffle_v4i64_0000(<4 x i64> %a, <4 x i64> %b) {
-; AVX1-LABEL: shuffle_v4i64_0000:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
+define <4 x i64> @shuffle_v4i64_0000(<4 x i64> %a, <4 x i64> %b) {\r
+; AVX1-LABEL: shuffle_v4i64_0000:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4i64_0000:
; AVX2: # BB#0:
; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
ret <4 x i64> %shuffle
}
-define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {
-; AVX1-LABEL: shuffle_v4i64_0001:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX1-NEXT: retq
-;
+define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {\r
+; AVX1-LABEL: shuffle_v4i64_0001:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4i64_0001:
; AVX2: # BB#0:
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
define <4 x i64> @shuffle_v4i64_0020(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_0020:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1\r
+; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4i64_0020:
; AVX2: # BB#0:
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
}
define <4 x i64> @shuffle_v4i64_1000(<4 x i64> %a, <4 x i64> %b) {
-; AVX1-LABEL: shuffle_v4i64_1000:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; AVX1-NEXT: retq
-;
+; AVX1-LABEL: shuffle_v4i64_1000:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4i64_1000:
; AVX2: # BB#0:
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
}
define <4 x i64> @shuffle_v4i64_2200(<4 x i64> %a, <4 x i64> %b) {
-; AVX1-LABEL: shuffle_v4i64_2200:
-; AVX1: # BB#0:
-; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
-; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v4i64_2200:
+; AVX1-LABEL: shuffle_v4i64_2200:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]\r
+; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; AVX1-NEXT: retq\r
+;\r
+; AVX2-LABEL: shuffle_v4i64_2200:\r
; AVX2: # BB#0:
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,2,0,0]
; AVX2-NEXT: retq
ret <4 x i64> %shuffle
}
-define <4 x i64> @shuffle_v4i64_0124(<4 x i64> %a, <4 x i64> %b) {
-; AVX1-LABEL: shuffle_v4i64_0124:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
-; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],ymm1[3]
-; AVX1-NEXT: retq
+define <4 x i64> @shuffle_v4i64_0124(<4 x i64> %a, <4 x i64> %b) {\r
+; AVX1-LABEL: shuffle_v4i64_0124:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm1[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1\r
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],ymm1[3]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v4i64_0124:
; AVX2: # BB#0:
define <4 x i64> @shuffle_v4i64_0412(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_0412:
; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,2,2]
-; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
-; AVX1-NEXT: retq
-;
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2\r
+; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0\r
+; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]\r
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: shuffle_v4i64_0412:
; AVX2: # BB#0:
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) {
; AVX1-LABEL: shuffle_v4i64_4012:
-; AVX1: # BB#0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
-; AVX1-NEXT: retq
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2\r
+; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0\r
+; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v4i64_4012:
; AVX2: # BB#0:
ret <4 x double> %3
}
-define <4 x double> @splat_v4f64(<2 x double> %r) {
-; AVX1-LABEL: splat_v4f64:
-; AVX1: # BB#0:
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX1-NEXT: retq
-;
+define <4 x double> @splat_v4f64(<2 x double> %r) {\r
+; AVX1-LABEL: splat_v4f64:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0\r
+; AVX1-NEXT: retq\r
+;\r
; AVX2-LABEL: splat_v4f64:
; AVX2: # BB#0:
; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
ret <8 x float> %shuffle
}
-define <8 x float> @shuffle_v8f32_01014545(<8 x float> %a, <8 x float> %b) {
-; ALL-LABEL: shuffle_v8f32_01014545:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; ALL-NEXT: retq
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>
- ret <8 x float> %shuffle
+define <8 x float> @shuffle_v8f32_01014545(<8 x float> %a, <8 x float> %b) {\r
+; ALL-LABEL: shuffle_v8f32_01014545:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>\r
+ ret <8 x float> %shuffle\r
}
define <8 x float> @shuffle_v8f32_00112233(<8 x float> %a, <8 x float> %b) {
define <8 x float> @shuffle_v8f32_08080808(<8 x float> %a, <8 x float> %b) {
; AVX1-LABEL: shuffle_v8f32_08080808:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[0,0,2,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
-; AVX1-NEXT: retq
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[0,0,2,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0\r
+; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v8f32_08080808:
; AVX2: # BB#0:
ret <8 x float> %shuffle
}
-define <8 x float> @shuffle_v8f32_09ab1def(<8 x float> %a, <8 x float> %b) {
-; AVX1-LABEL: shuffle_v8f32_09ab1def:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[1,1,2,3]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
-; AVX1-NEXT: retq
+define <8 x float> @shuffle_v8f32_09ab1def(<8 x float> %a, <8 x float> %b) {\r
+; AVX1-LABEL: shuffle_v8f32_09ab1def:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0\r
+; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v8f32_09ab1def:
; AVX2: # BB#0:
ret <8 x float> %shuffle
}
-define <8 x float> @shuffle_v8f32_00224466(<8 x float> %a, <8 x float> %b) {
-; ALL-LABEL: shuffle_v8f32_00224466:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
-; ALL-NEXT: retq
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
- ret <8 x float> %shuffle
+define <8 x float> @shuffle_v8f32_00224466(<8 x float> %a, <8 x float> %b) {\r
+; ALL-LABEL: shuffle_v8f32_00224466:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>\r
+ ret <8 x float> %shuffle\r
}
define <8 x float> @shuffle_v8f32_10325476(<8 x float> %a, <8 x float> %b) {
ret <8 x float> %shuffle
}
-define <8 x float> @shuffle_v8f32_11335577(<8 x float> %a, <8 x float> %b) {
-; ALL-LABEL: shuffle_v8f32_11335577:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
-; ALL-NEXT: retq
- %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
- ret <8 x float> %shuffle
+define <8 x float> @shuffle_v8f32_11335577(<8 x float> %a, <8 x float> %b) {\r
+; ALL-LABEL: shuffle_v8f32_11335577:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>\r
+ ret <8 x float> %shuffle\r
}
define <8 x float> @shuffle_v8f32_10235467(<8 x float> %a, <8 x float> %b) {
ret <8 x i32> %shuffle
}
-define <8 x i32> @shuffle_v8i32_01014545(<8 x i32> %a, <8 x i32> %b) {
-; AVX1-LABEL: shuffle_v8i32_01014545:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8i32_01014545:
+define <8 x i32> @shuffle_v8i32_01014545(<8 x i32> %a, <8 x i32> %b) {\r
+; AVX1-LABEL: shuffle_v8i32_01014545:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; AVX1-NEXT: retq\r
+;\r
+; AVX2-LABEL: shuffle_v8i32_01014545:\r
; AVX2: # BB#0:
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[0,1,0,1,4,5,4,5]
; AVX2-NEXT: retq
define <8 x i32> @shuffle_v8i32_08080808(<8 x i32> %a, <8 x i32> %b) {
; AVX1-LABEL: shuffle_v8i32_08080808:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[0,0,2,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1
-; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
-; AVX1-NEXT: retq
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm1[0,0,2,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1\r
+; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0\r
+; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v8i32_08080808:
; AVX2: # BB#0:
ret <8 x i32> %shuffle
}
-define <8 x i32> @shuffle_v8i32_09ab1def(<8 x i32> %a, <8 x i32> %b) {
-; AVX1-LABEL: shuffle_v8i32_09ab1def:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[1,1,2,3]
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
-; AVX1-NEXT: retq
+define <8 x i32> @shuffle_v8i32_09ab1def(<8 x i32> %a, <8 x i32> %b) {\r
+; AVX1-LABEL: shuffle_v8i32_09ab1def:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]\r
+; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0\r
+; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]\r
+; AVX1-NEXT: retq\r
;
; AVX2-LABEL: shuffle_v8i32_09ab1def:
; AVX2: # BB#0:
ret <8 x i32> %shuffle
}
-define <8 x i32> @shuffle_v8i32_00224466(<8 x i32> %a, <8 x i32> %b) {
-; AVX1-LABEL: shuffle_v8i32_00224466:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8i32_00224466:
+define <8 x i32> @shuffle_v8i32_00224466(<8 x i32> %a, <8 x i32> %b) {\r
+; AVX1-LABEL: shuffle_v8i32_00224466:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]\r
+; AVX1-NEXT: retq\r
+;\r
+; AVX2-LABEL: shuffle_v8i32_00224466:\r
; AVX2: # BB#0:
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
; AVX2-NEXT: retq
ret <8 x i32> %shuffle
}
-define <8 x i32> @shuffle_v8i32_11335577(<8 x i32> %a, <8 x i32> %b) {
-; AVX1-LABEL: shuffle_v8i32_11335577:
-; AVX1: # BB#0:
-; AVX1-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
-; AVX1-NEXT: retq
-;
-; AVX2-LABEL: shuffle_v8i32_11335577:
+define <8 x i32> @shuffle_v8i32_11335577(<8 x i32> %a, <8 x i32> %b) {\r
+; AVX1-LABEL: shuffle_v8i32_11335577:\r
+; AVX1: # BB#0:\r
+; AVX1-NEXT: vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]\r
+; AVX1-NEXT: retq\r
+;\r
+; AVX2-LABEL: shuffle_v8i32_11335577:\r
; AVX2: # BB#0:
; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
; AVX2-NEXT: retq
ret <8 x double> %shuffle
}
-define <8 x double> @shuffle_v8f64_00224466(<8 x double> %a, <8 x double> %b) {
-; ALL-LABEL: shuffle_v8f64_00224466:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm0[0,0,2,2]
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0
-; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,2]
-; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
-; ALL-NEXT: retq
- %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
+define <8 x double> @shuffle_v8f64_00224466(<8 x double> %a, <8 x double> %b) {\r
+; ALL-LABEL: shuffle_v8f64_00224466:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm0[0,0,2,2]\r
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0 \r
+; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]\r
+; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>\r
ret <8 x double> %shuffle
}
ret <8 x double> %shuffle
}
-define <8 x double> @shuffle_v8f64_00226644(<8 x double> %a, <8 x double> %b) {
-; ALL-LABEL: shuffle_v8f64_00226644:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm0[0,0,2,2]
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0
-; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]
-; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+define <8 x double> @shuffle_v8f64_00226644(<8 x double> %a, <8 x double> %b) {\r
+; ALL-LABEL: shuffle_v8f64_00226644:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm0[0,0,2,2]\r
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0 \r
+; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]\r
+; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0\r
; ALL-NEXT: retq
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 6, i32 6, i32 4, i32 4>
ret <8 x double> %shuffle
ret <8 x double> %shuffle
}
-define <8 x double> @shuffle_v8f64_002u6u44(<8 x double> %a, <8 x double> %b) {
-; ALL-LABEL: shuffle_v8f64_002u6u44:
-; ALL: # BB#0:
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm0[0,0,2,2]
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0
-; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,1,0,0]
-; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0
+define <8 x double> @shuffle_v8f64_002u6u44(<8 x double> %a, <8 x double> %b) {\r
+; ALL-LABEL: shuffle_v8f64_002u6u44:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm0[0,0,2,2]\r
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0 \r
+; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,1,0,0]\r
+; ALL-NEXT: vinsertf64x4 $1, %ymm0, %zmm1, %zmm0\r
; ALL-NEXT: retq
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 undef, i32 6, i32 undef, i32 4, i32 4>
ret <8 x double> %shuffle
}
define <8 x double> @shuffle_v8f64_uuu3uu66(<8 x double> %a, <8 x double> %b) {
-; ALL-LABEL: shuffle_v8f64_uuu3uu66:
-; ALL: # BB#0:
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm1
-; ALL-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,2,2]
-; ALL-NEXT: vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
-; ALL-NEXT: retq
- %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 undef, i32 undef, i32 undef, i32 3, i32 undef, i32 undef, i32 6, i32 6>
+; ALL-LABEL: shuffle_v8f64_uuu3uu66:\r
+; ALL: # BB#0:\r
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm1 \r
+; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]\r
+; ALL-NEXT: vinsertf64x4 $1, %ymm1, %zmm0, %zmm0\r
+; ALL-NEXT: retq\r
+ %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 undef, i32 undef, i32 undef, i32 3, i32 undef, i32 undef, i32 6, i32 6>\r
ret <8 x double> %shuffle
}
; ALL-LABEL: shuffle_v8f64_f511235a:
; ALL: # BB#0:
; ALL-NEXT: vperm2f128 {{.*#+}} ymm2 = ymm0[2,3,0,1]
-; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm3
-; ALL-NEXT: vpermpd {{.*#+}} ymm4 = ymm3[0,1,1,3]
-; ALL-NEXT: vblendpd {{.*#+}} ymm2 = ymm2[0,1],ymm4[2],ymm2[3]
-; ALL-NEXT: vpermilpd {{.*#+}} ymm4 = ymm1[0,0,2,2]
-; ALL-NEXT: vblendpd {{.*#+}} ymm2 = ymm2[0,1,2],ymm4[3]
-; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,1]
-; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm3[1],ymm0[2,3]
+; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm3 \r
+; ALL-NEXT: vpermpd {{.*#+}} ymm4 = ymm3[0,1,1,3]\r
+; ALL-NEXT: vblendpd {{.*#+}} ymm2 = ymm2[0,1],ymm4[2],ymm2[3]\r
+; ALL-NEXT: vmovddup {{.*#+}} ymm4 = ymm1[0,0,2,2]\r
+; ALL-NEXT: vblendpd {{.*#+}} ymm2 = ymm2[0,1,2],ymm4[3]\r
+; ALL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,1]\r
+; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm3[1],ymm0[2,3]\r
; ALL-NEXT: vextractf64x4 $1, %zmm1, %ymm1
; ALL-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[3,1,2,3]
; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
; SSE2-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0,0]
; SSE2-NEXT: movaps %xmm1, %xmm0
; SSE2-NEXT: retq
-;
-; SSSE3-LABEL: combine_test2b:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; SSSE3-NEXT: movapd %xmm1, %xmm0
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: combine_test2b:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm1 = xmm1[0,0]
-; SSE41-NEXT: movapd %xmm1, %xmm0
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: combine_test2b:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0,0]
-; 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>
+;\r
+; SSSE3-LABEL: combine_test2b:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: combine_test2b:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: combine_test2b:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm1[0,0]\r
+; AVX-NEXT: retq\r
+ %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>\r
+ %2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 0, i32 5>\r
ret <4 x float> %2
}
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSSE3-LABEL: combine_undef_input_test7:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: combine_undef_input_test7:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: combine_undef_input_test7:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; 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> %a, <4 x i32> <i32 1, i32 2, i32 4, i32 5>
+;\r
+; SSSE3-LABEL: combine_undef_input_test7:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: combine_undef_input_test7:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: combine_undef_input_test7:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 6, i32 0, i32 1, i32 7>\r
+ %2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 1, i32 2, i32 4, i32 5>\r
ret <4 x float> %2
}
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSSE3-LABEL: combine_undef_input_test8:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: combine_undef_input_test8:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: combine_undef_input_test8:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; 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> %a, <4 x i32> <i32 0, i32 2, i32 4, i32 1>
+;\r
+; SSSE3-LABEL: combine_undef_input_test8:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: combine_undef_input_test8:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: combine_undef_input_test8:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 5, i32 1, i32 7>\r
+ %2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 0, i32 2, i32 4, i32 1>\r
ret <4 x float> %2
}
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSSE3-LABEL: combine_undef_input_test17:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: combine_undef_input_test17:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: combine_undef_input_test17:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; 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> %a, <4 x float> %1, <4 x i32> <i32 5, i32 6, i32 0, i32 1>
+;\r
+; SSSE3-LABEL: combine_undef_input_test17:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: combine_undef_input_test17:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: combine_undef_input_test17:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 6, i32 0, i32 1, i32 7>\r
+ %2 = shufflevector <4 x float> %a, <4 x float> %1, <4 x i32> <i32 5, i32 6, i32 0, i32 1>\r
ret <4 x float> %2
}
; SSE2: # BB#0:
; SSE2-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0,0]
; SSE2-NEXT: retq
-;
-; SSSE3-LABEL: combine_undef_input_test18:
-; SSSE3: # BB#0:
-; SSSE3-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSSE3-NEXT: retq
-;
-; SSE41-LABEL: combine_undef_input_test18:
-; SSE41: # BB#0:
-; SSE41-NEXT: unpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; SSE41-NEXT: retq
-;
-; AVX-LABEL: combine_undef_input_test18:
-; AVX: # BB#0:
-; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0,0]
-; 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> %a, <4 x float> %1, <4 x i32> <i32 4, i32 6, i32 0, i32 5>
+;\r
+; SSSE3-LABEL: combine_undef_input_test18:\r
+; SSSE3: # BB#0:\r
+; SSSE3-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSSE3-NEXT: retq\r
+;\r
+; SSE41-LABEL: combine_undef_input_test18:\r
+; SSE41: # BB#0:\r
+; SSE41-NEXT: movddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; SSE41-NEXT: retq\r
+;\r
+; AVX-LABEL: combine_undef_input_test18:\r
+; AVX: # BB#0:\r
+; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]\r
+; AVX-NEXT: retq\r
+ %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 0, i32 5, i32 1, i32 7>\r
+ %2 = shufflevector <4 x float> %a, <4 x float> %1, <4 x i32> <i32 4, i32 6, i32 0, i32 5>\r
ret <4 x float> %2
}
projects / oota-llvm.git / commitdiff