mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-25 20:23:11 +01:00
[X86][SSE] combineVectorSignBitsTruncation - remove minimum vector length limitations
truncateVectorWithPACK has its own vector length controls, so we can rely on those directly. This helps some existing truncation to subvector tests, which were being combined later during shuffle lowering at which point the sign/zero bit detection had become obscured preventing lowerShuffleWithPACK working as well as it could.
This commit is contained in:
Simon Pilgrim
parent
abbd22c44e
commit
26620d6e83
@ -43489,13 +43489,17 @@ static SDValue combineVectorSignBitsTruncation(SDNode *N, const SDLoc &DL,
|
||||
MVT InSVT = InVT.getScalarType();
|
||||
|
||||
// Check we have a truncation suited for PACKSS/PACKUS.
|
||||
if (!VT.is128BitVector() && !VT.is256BitVector())
|
||||
if (!isPowerOf2_32(VT.getVectorNumElements()))
|
||||
return SDValue();
|
||||
if (SVT != MVT::i8 && SVT != MVT::i16 && SVT != MVT::i32)
|
||||
return SDValue();
|
||||
if (InSVT != MVT::i16 && InSVT != MVT::i32 && InSVT != MVT::i64)
|
||||
return SDValue();
|
||||
|
||||
// Truncation to sub-128bit vXi32 can be better handled with shuffles.
|
||||
if (SVT == MVT::i32 && VT.getSizeInBits() < 128)
|
||||
return SDValue();
|
||||
|
||||
// AVX512 has fast truncate, but if the input is already going to be split,
|
||||
// there's no harm in trying pack.
|
||||
if (Subtarget.hasAVX512() &&
|
||||
|
||||
@ -946,7 +946,7 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; SSE2-NEXT: pxor %xmm8, %xmm8
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [127,127]
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm11 = [2147483648,2147483648]
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm6
|
||||
; SSE2-NEXT: movdqa %xmm2, %xmm6
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm6
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [2147483775,2147483775]
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm7
|
||||
@ -957,23 +957,10 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; SSE2-NEXT: pand %xmm12, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm13 = xmm7[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm6, %xmm13
|
||||
; SSE2-NEXT: pand %xmm13, %xmm3
|
||||
; SSE2-NEXT: pand %xmm13, %xmm2
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm13
|
||||
; SSE2-NEXT: por %xmm3, %xmm13
|
||||
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm3
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm3, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm3
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm2
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm3
|
||||
; SSE2-NEXT: por %xmm2, %xmm3
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
||||
; SSE2-NEXT: por %xmm2, %xmm13
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm2
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm2
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm2, %xmm6
|
||||
@ -983,84 +970,92 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm2
|
||||
; SSE2-NEXT: pand %xmm2, %xmm1
|
||||
; SSE2-NEXT: pand %xmm2, %xmm3
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm2
|
||||
; SSE2-NEXT: por %xmm1, %xmm2
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm1
|
||||
; SSE2-NEXT: por %xmm3, %xmm2
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm3
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm1, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm3, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm3
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm0
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm3
|
||||
; SSE2-NEXT: por %xmm0, %xmm3
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm0
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm0, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm7, %xmm1
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm7, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm1, %xmm6
|
||||
; SSE2-NEXT: pand %xmm6, %xmm0
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm6
|
||||
; SSE2-NEXT: por %xmm0, %xmm6
|
||||
; SSE2-NEXT: pand %xmm6, %xmm1
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm6
|
||||
; SSE2-NEXT: por %xmm1, %xmm6
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [18446744073709551488,18446744073709551488]
|
||||
; SSE2-NEXT: movdqa %xmm6, %xmm0
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm0
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [18446744071562067840,18446744071562067840]
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm1[0,0,2,2]
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm7
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm7[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm1, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm7[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm0, %xmm1
|
||||
; SSE2-NEXT: pand %xmm1, %xmm6
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm1
|
||||
; SSE2-NEXT: por %xmm6, %xmm1
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm0
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm0
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm0
|
||||
; SSE2-NEXT: pand %xmm0, %xmm6
|
||||
; SSE2-NEXT: pand %xmm0, %xmm3
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm0
|
||||
; SSE2-NEXT: por %xmm6, %xmm0
|
||||
; SSE2-NEXT: por %xmm3, %xmm0
|
||||
; SSE2-NEXT: packssdw %xmm1, %xmm0
|
||||
; SSE2-NEXT: movdqa %xmm2, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm1
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm1
|
||||
; SSE2-NEXT: pand %xmm1, %xmm2
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm1
|
||||
; SSE2-NEXT: por %xmm2, %xmm1
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm2
|
||||
; SSE2-NEXT: pxor %xmm11, %xmm2
|
||||
; SSE2-NEXT: movdqa %xmm2, %xmm6
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm2
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm12, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm2
|
||||
; SSE2-NEXT: pand %xmm2, %xmm3
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm2
|
||||
; SSE2-NEXT: por %xmm3, %xmm2
|
||||
; SSE2-NEXT: pxor %xmm13, %xmm11
|
||||
; SSE2-NEXT: movdqa %xmm11, %xmm3
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm3
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm11
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm11[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm6, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm7, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm13
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm3
|
||||
; SSE2-NEXT: por %xmm13, %xmm3
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm6 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0]
|
||||
; SSE2-NEXT: pand %xmm6, %xmm3
|
||||
; SSE2-NEXT: pand %xmm6, %xmm2
|
||||
; SSE2-NEXT: packuswb %xmm3, %xmm2
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm6, %xmm1
|
||||
; SSE2-NEXT: pand %xmm6, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm1, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm2, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm1, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm2
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm3
|
||||
; SSE2-NEXT: por %xmm2, %xmm3
|
||||
; SSE2-NEXT: pxor %xmm13, %xmm11
|
||||
; SSE2-NEXT: movdqa %xmm11, %xmm1
|
||||
; SSE2-NEXT: pcmpgtd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm11
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm11[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm2, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm6, %xmm1
|
||||
; SSE2-NEXT: pand %xmm1, %xmm13
|
||||
; SSE2-NEXT: pandn %xmm9, %xmm1
|
||||
; SSE2-NEXT: por %xmm13, %xmm1
|
||||
; SSE2-NEXT: packssdw %xmm3, %xmm1
|
||||
; SSE2-NEXT: packssdw %xmm1, %xmm0
|
||||
; SSE2-NEXT: packsswb %xmm0, %xmm0
|
||||
; SSE2-NEXT: pcmpeqd %xmm8, %xmm5
|
||||
; SSE2-NEXT: pcmpeqd %xmm1, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm1, %xmm5
|
||||
@ -1133,45 +1128,40 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; SSE4-NEXT: pxor %xmm8, %xmm8
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm7 = [127,127]
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm10
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm10
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm2, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm3
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm2, %xmm3
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm10
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm2, %xmm10
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm2
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm2
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm2
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm9, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm9, %xmm7
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm3
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm9, %xmm3
|
||||
; SSE4-NEXT: movdqa %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm7
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm1 = [18446744073709551488,18446744073709551488]
|
||||
; SSE4-NEXT: movapd %xmm7, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm6
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm7, %xmm6
|
||||
; SSE4-NEXT: movapd %xmm2, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm7
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm2, %xmm7
|
||||
; SSE4-NEXT: movapd %xmm3, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm2
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm2
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm7
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm7
|
||||
; SSE4-NEXT: packssdw %xmm6, %xmm7
|
||||
; SSE4-NEXT: movapd %xmm2, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm3
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm2, %xmm3
|
||||
; SSE4-NEXT: movapd %xmm10, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm10, %xmm1
|
||||
; SSE4-NEXT: movapd {{.*#+}} xmm0 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0]
|
||||
; SSE4-NEXT: andpd %xmm0, %xmm1
|
||||
; SSE4-NEXT: andpd %xmm0, %xmm2
|
||||
; SSE4-NEXT: packusdw %xmm1, %xmm2
|
||||
; SSE4-NEXT: andpd %xmm0, %xmm7
|
||||
; SSE4-NEXT: andpd %xmm0, %xmm6
|
||||
; SSE4-NEXT: packusdw %xmm7, %xmm6
|
||||
; SSE4-NEXT: packusdw %xmm2, %xmm6
|
||||
; SSE4-NEXT: packuswb %xmm0, %xmm6
|
||||
; SSE4-NEXT: packssdw %xmm3, %xmm1
|
||||
; SSE4-NEXT: packssdw %xmm1, %xmm7
|
||||
; SSE4-NEXT: packsswb %xmm0, %xmm7
|
||||
; SSE4-NEXT: pcmpeqd %xmm8, %xmm5
|
||||
; SSE4-NEXT: pcmpeqd %xmm0, %xmm0
|
||||
; SSE4-NEXT: pxor %xmm0, %xmm5
|
||||
@ -1206,74 +1196,63 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; SSE4-NEXT: .LBB2_16: # %else14
|
||||
; SSE4-NEXT: retq
|
||||
; SSE4-NEXT: .LBB2_1: # %cond.store
|
||||
; SSE4-NEXT: pextrb $0, %xmm6, (%rdi)
|
||||
; SSE4-NEXT: pextrb $0, %xmm7, (%rdi)
|
||||
; SSE4-NEXT: testb $2, %al
|
||||
; SSE4-NEXT: je .LBB2_4
|
||||
; SSE4-NEXT: .LBB2_3: # %cond.store1
|
||||
; SSE4-NEXT: pextrb $1, %xmm6, 1(%rdi)
|
||||
; SSE4-NEXT: pextrb $1, %xmm7, 1(%rdi)
|
||||
; SSE4-NEXT: testb $4, %al
|
||||
; SSE4-NEXT: je .LBB2_6
|
||||
; SSE4-NEXT: .LBB2_5: # %cond.store3
|
||||
; SSE4-NEXT: pextrb $2, %xmm6, 2(%rdi)
|
||||
; SSE4-NEXT: pextrb $2, %xmm7, 2(%rdi)
|
||||
; SSE4-NEXT: testb $8, %al
|
||||
; SSE4-NEXT: je .LBB2_8
|
||||
; SSE4-NEXT: .LBB2_7: # %cond.store5
|
||||
; SSE4-NEXT: pextrb $3, %xmm6, 3(%rdi)
|
||||
; SSE4-NEXT: pextrb $3, %xmm7, 3(%rdi)
|
||||
; SSE4-NEXT: testb $16, %al
|
||||
; SSE4-NEXT: je .LBB2_10
|
||||
; SSE4-NEXT: .LBB2_9: # %cond.store7
|
||||
; SSE4-NEXT: pextrb $4, %xmm6, 4(%rdi)
|
||||
; SSE4-NEXT: pextrb $4, %xmm7, 4(%rdi)
|
||||
; SSE4-NEXT: testb $32, %al
|
||||
; SSE4-NEXT: je .LBB2_12
|
||||
; SSE4-NEXT: .LBB2_11: # %cond.store9
|
||||
; SSE4-NEXT: pextrb $5, %xmm6, 5(%rdi)
|
||||
; SSE4-NEXT: pextrb $5, %xmm7, 5(%rdi)
|
||||
; SSE4-NEXT: testb $64, %al
|
||||
; SSE4-NEXT: je .LBB2_14
|
||||
; SSE4-NEXT: .LBB2_13: # %cond.store11
|
||||
; SSE4-NEXT: pextrb $6, %xmm6, 6(%rdi)
|
||||
; SSE4-NEXT: pextrb $6, %xmm7, 6(%rdi)
|
||||
; SSE4-NEXT: testb $-128, %al
|
||||
; SSE4-NEXT: je .LBB2_16
|
||||
; SSE4-NEXT: .LBB2_15: # %cond.store13
|
||||
; SSE4-NEXT: pextrb $7, %xmm6, 7(%rdi)
|
||||
; SSE4-NEXT: pextrb $7, %xmm7, 7(%rdi)
|
||||
; SSE4-NEXT: retq
|
||||
;
|
||||
; AVX1-LABEL: truncstore_v8i64_v8i8:
|
||||
; AVX1: # %bb.0:
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} ymm9 = [127,127,127,127]
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm10
|
||||
; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [127,127]
|
||||
; AVX1-NEXT: vpcmpgtq %xmm10, %xmm5, %xmm6
|
||||
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm5, %xmm7
|
||||
; AVX1-NEXT: vinsertf128 $1, %xmm6, %ymm7, %ymm8
|
||||
; AVX1-NEXT: vblendvpd %ymm8, %ymm1, %ymm9, %ymm8
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
|
||||
; AVX1-NEXT: vpcmpgtq %xmm3, %xmm5, %xmm4
|
||||
; AVX1-NEXT: vpcmpgtq %xmm0, %xmm5, %xmm11
|
||||
; AVX1-NEXT: vinsertf128 $1, %xmm4, %ymm11, %ymm12
|
||||
; AVX1-NEXT: vblendvpd %ymm12, %ymm0, %ymm9, %ymm9
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} ymm12 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488]
|
||||
; AVX1-NEXT: vblendvpd %xmm4, %xmm3, %xmm5, %xmm3
|
||||
; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [18446744073709551488,18446744073709551488]
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm3, %xmm3
|
||||
; AVX1-NEXT: vblendvpd %xmm11, %xmm0, %xmm5, %xmm0
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm0, %ymm0
|
||||
; AVX1-NEXT: vblendvpd %ymm0, %ymm9, %ymm12, %ymm0
|
||||
; AVX1-NEXT: vblendvpd %xmm6, %xmm10, %xmm5, %xmm3
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm3, %xmm3
|
||||
; AVX1-NEXT: vblendvpd %xmm7, %xmm1, %xmm5, %xmm1
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm1
|
||||
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm1, %ymm1
|
||||
; AVX1-NEXT: vblendvpd %ymm1, %ymm8, %ymm12, %ymm1
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} ymm3 = [255,255,255,255]
|
||||
; AVX1-NEXT: vandpd %ymm3, %ymm1, %ymm1
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm4
|
||||
; AVX1-NEXT: vpackusdw %xmm4, %xmm1, %xmm1
|
||||
; AVX1-NEXT: vandpd %ymm3, %ymm0, %ymm0
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
|
||||
; AVX1-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
|
||||
; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [127,127]
|
||||
; AVX1-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm8
|
||||
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm4, %xmm9
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm7
|
||||
; AVX1-NEXT: vpcmpgtq %xmm7, %xmm4, %xmm5
|
||||
; AVX1-NEXT: vpcmpgtq %xmm0, %xmm4, %xmm6
|
||||
; AVX1-NEXT: vblendvpd %xmm6, %xmm0, %xmm4, %xmm0
|
||||
; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [18446744073709551488,18446744073709551488]
|
||||
; AVX1-NEXT: vpcmpgtq %xmm6, %xmm0, %xmm10
|
||||
; AVX1-NEXT: vblendvpd %xmm5, %xmm7, %xmm4, %xmm5
|
||||
; AVX1-NEXT: vpcmpgtq %xmm6, %xmm5, %xmm11
|
||||
; AVX1-NEXT: vblendvpd %xmm9, %xmm1, %xmm4, %xmm1
|
||||
; AVX1-NEXT: vpcmpgtq %xmm6, %xmm1, %xmm7
|
||||
; AVX1-NEXT: vblendvpd %xmm8, %xmm3, %xmm4, %xmm3
|
||||
; AVX1-NEXT: vpcmpgtq %xmm6, %xmm3, %xmm4
|
||||
; AVX1-NEXT: vblendvpd %xmm4, %xmm3, %xmm6, %xmm3
|
||||
; AVX1-NEXT: vblendvpd %xmm7, %xmm1, %xmm6, %xmm1
|
||||
; AVX1-NEXT: vpackssdw %xmm3, %xmm1, %xmm1
|
||||
; AVX1-NEXT: vblendvpd %xmm11, %xmm5, %xmm6, %xmm3
|
||||
; AVX1-NEXT: vblendvpd %xmm10, %xmm0, %xmm6, %xmm0
|
||||
; AVX1-NEXT: vpackssdw %xmm3, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpacksswb %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm1
|
||||
; AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
|
||||
; AVX1-NEXT: vpcmpeqd %xmm3, %xmm1, %xmm1
|
||||
@ -1344,26 +1323,20 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; AVX2: # %bb.0:
|
||||
; AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm4 = [127,127,127,127]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm4, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm4, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm4, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm4 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm0, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm1, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm4
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm5 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm4, %xmm4
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3]
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm4
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm5 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm4, %xmm4
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm0, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0
|
||||
; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpacksswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpcmpeqd %ymm3, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vmovmskps %ymm1, %eax
|
||||
; AVX2-NEXT: notl %eax
|
||||
@ -1756,7 +1729,7 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; SSE2-NEXT: pxor %xmm9, %xmm9
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [32767,32767]
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648]
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm5
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm5
|
||||
; SSE2-NEXT: pxor %xmm4, %xmm5
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [2147516415,2147516415]
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm7
|
||||
@ -1767,54 +1740,51 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; SSE2-NEXT: pand %xmm3, %xmm6
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm6, %xmm5
|
||||
; SSE2-NEXT: pand %xmm5, %xmm1
|
||||
; SSE2-NEXT: pand %xmm5, %xmm0
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm5
|
||||
; SSE2-NEXT: por %xmm1, %xmm5
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm4, %xmm1
|
||||
; SSE2-NEXT: por %xmm0, %xmm5
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE2-NEXT: pxor %xmm4, %xmm0
|
||||
; SSE2-NEXT: movdqa %xmm10, %xmm3
|
||||
; SSE2-NEXT: pcmpgtd %xmm1, %xmm3
|
||||
; SSE2-NEXT: pcmpgtd %xmm0, %xmm3
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm6, %xmm1
|
||||
; SSE2-NEXT: pcmpeqd %xmm10, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm6, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm1, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm0
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm3
|
||||
; SSE2-NEXT: por %xmm0, %xmm3
|
||||
; SSE2-NEXT: pand %xmm3, %xmm1
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm3
|
||||
; SSE2-NEXT: por %xmm1, %xmm3
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709518848,18446744073709518848]
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm4, %xmm1
|
||||
; SSE2-NEXT: movdqa %xmm3, %xmm0
|
||||
; SSE2-NEXT: pxor %xmm4, %xmm0
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm6 = [18446744071562035200,18446744071562035200]
|
||||
; SSE2-NEXT: movdqa %xmm1, %xmm7
|
||||
; SSE2-NEXT: movdqa %xmm0, %xmm7
|
||||
; SSE2-NEXT: pcmpgtd %xmm6, %xmm7
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm7[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm6, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm0, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm7[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm1, %xmm0
|
||||
; SSE2-NEXT: pand %xmm0, %xmm3
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm0
|
||||
; SSE2-NEXT: por %xmm3, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm7[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm6, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm1, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm7[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm0, %xmm1
|
||||
; SSE2-NEXT: pand %xmm1, %xmm3
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm1
|
||||
; SSE2-NEXT: por %xmm3, %xmm1
|
||||
; SSE2-NEXT: pxor %xmm5, %xmm4
|
||||
; SSE2-NEXT: movdqa %xmm4, %xmm1
|
||||
; SSE2-NEXT: pcmpgtd %xmm6, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[0,0,2,2]
|
||||
; SSE2-NEXT: movdqa %xmm4, %xmm0
|
||||
; SSE2-NEXT: pcmpgtd %xmm6, %xmm0
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[0,0,2,2]
|
||||
; SSE2-NEXT: pcmpeqd %xmm6, %xmm4
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
|
||||
; SSE2-NEXT: pand %xmm3, %xmm4
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm4, %xmm1
|
||||
; SSE2-NEXT: pand %xmm1, %xmm5
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm1
|
||||
; SSE2-NEXT: por %xmm5, %xmm1
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
|
||||
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
|
||||
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
|
||||
; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
|
||||
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
||||
; SSE2-NEXT: por %xmm4, %xmm0
|
||||
; SSE2-NEXT: pand %xmm0, %xmm5
|
||||
; SSE2-NEXT: pandn %xmm8, %xmm0
|
||||
; SSE2-NEXT: por %xmm5, %xmm0
|
||||
; SSE2-NEXT: packssdw %xmm1, %xmm0
|
||||
; SSE2-NEXT: packssdw %xmm0, %xmm0
|
||||
; SSE2-NEXT: pcmpeqd %xmm2, %xmm9
|
||||
; SSE2-NEXT: movmskps %xmm9, %eax
|
||||
; SSE2-NEXT: xorl $15, %eax
|
||||
@ -1857,26 +1827,24 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; SSE4-NEXT: pxor %xmm4, %xmm4
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm5 = [32767,32767]
|
||||
; SSE4-NEXT: movdqa %xmm5, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm5, %xmm6
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm6
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm6
|
||||
; SSE4-NEXT: movdqa %xmm5, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm5
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm3 = [18446744073709518848,18446744073709518848]
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm5
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm1 = [18446744073709518848,18446744073709518848]
|
||||
; SSE4-NEXT: movapd %xmm5, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm3, %xmm1
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm5, %xmm1
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm3
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm5, %xmm3
|
||||
; SSE4-NEXT: movapd %xmm6, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm6, %xmm3
|
||||
; SSE4-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm4[1,2,3],xmm3[4],xmm4[5,6,7]
|
||||
; SSE4-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
|
||||
; SSE4-NEXT: packusdw %xmm3, %xmm1
|
||||
; SSE4-NEXT: packusdw %xmm1, %xmm1
|
||||
; SSE4-NEXT: pcmpeqd %xmm4, %xmm2
|
||||
; SSE4-NEXT: movmskps %xmm2, %eax
|
||||
; SSE4-NEXT: pcmpgtq %xmm1, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm6, %xmm1
|
||||
; SSE4-NEXT: packssdw %xmm3, %xmm1
|
||||
; SSE4-NEXT: packssdw %xmm0, %xmm1
|
||||
; SSE4-NEXT: pcmpeqd %xmm2, %xmm4
|
||||
; SSE4-NEXT: movmskps %xmm4, %eax
|
||||
; SSE4-NEXT: xorl $15, %eax
|
||||
; SSE4-NEXT: testb $1, %al
|
||||
; SSE4-NEXT: jne .LBB4_1
|
||||
@ -1920,11 +1888,9 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; AVX1-NEXT: vblendvpd %xmm5, %xmm3, %xmm4, %xmm3
|
||||
; AVX1-NEXT: vpcmpgtq %xmm6, %xmm3, %xmm4
|
||||
; AVX1-NEXT: vblendvpd %xmm4, %xmm3, %xmm6, %xmm3
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0],xmm2[1,2,3],xmm3[4],xmm2[5,6,7]
|
||||
; AVX1-NEXT: vblendvpd %xmm7, %xmm0, %xmm6, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm2[1,2,3],xmm0[4],xmm2[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackssdw %xmm3, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackssdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpcmpeqd %xmm2, %xmm1, %xmm1
|
||||
; AVX1-NEXT: vmovmskps %xmm1, %eax
|
||||
; AVX1-NEXT: xorl $15, %eax
|
||||
@ -1969,10 +1935,8 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; AVX2-NEXT: vpcmpgtq %ymm3, %ymm0, %ymm4
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm0, %ymm3, %ymm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm3
|
||||
; AVX2-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0],xmm2[1,2,3],xmm3[4],xmm2[5,6,7]
|
||||
; AVX2-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm2[1,2,3],xmm0[4],xmm2[5,6,7]
|
||||
; AVX2-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackssdw %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackssdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpcmpeqd %xmm2, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vmovmskps %xmm1, %eax
|
||||
; AVX2-NEXT: xorl $15, %eax
|
||||
|
||||
@ -1129,24 +1129,18 @@ define void @truncstore_v8i64_v8i8(<8 x i64> %x, <8 x i8>* %p, <8 x i32> %mask)
|
||||
; AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3
|
||||
; AVX2-NEXT: vbroadcastsd {{.*#+}} ymm4 = [255,255,255,255]
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm5 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
|
||||
; AVX2-NEXT: vpxor %ymm5, %ymm0, %ymm6
|
||||
; AVX2-NEXT: vpxor %ymm5, %ymm1, %ymm6
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm7 = [9223372036854776063,9223372036854776063,9223372036854776063,9223372036854776063]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm6, %ymm7, %ymm6
|
||||
; AVX2-NEXT: vblendvpd %ymm6, %ymm0, %ymm4, %ymm0
|
||||
; AVX2-NEXT: vpxor %ymm5, %ymm1, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm6, %ymm1, %ymm4, %ymm1
|
||||
; AVX2-NEXT: vpxor %ymm5, %ymm0, %ymm5
|
||||
; AVX2-NEXT: vpcmpgtq %ymm5, %ymm7, %ymm5
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm4
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm5 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm4, %xmm4
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3]
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm4
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm5 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm4, %xmm4
|
||||
; AVX2-NEXT: vpshufb %xmm5, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0
|
||||
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpcmpeqd %ymm3, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vmovmskps %ymm1, %eax
|
||||
; AVX2-NEXT: notl %eax
|
||||
@ -1569,27 +1563,25 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
;
|
||||
; SSE4-LABEL: truncstore_v4i64_v4i16:
|
||||
; SSE4: # %bb.0:
|
||||
; SSE4-NEXT: movdqa %xmm0, %xmm3
|
||||
; SSE4-NEXT: pxor %xmm8, %xmm8
|
||||
; SSE4-NEXT: movapd {{.*#+}} xmm7 = [65535,65535]
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm6 = [9223372036854775808,9223372036854775808]
|
||||
; SSE4-NEXT: movdqa %xmm0, %xmm5
|
||||
; SSE4-NEXT: pxor %xmm6, %xmm5
|
||||
; SSE4-NEXT: movdqa %xmm0, %xmm8
|
||||
; SSE4-NEXT: pxor %xmm6, %xmm6
|
||||
; SSE4-NEXT: movapd {{.*#+}} xmm5 = [65535,65535]
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm7 = [9223372036854775808,9223372036854775808]
|
||||
; SSE4-NEXT: movdqa %xmm1, %xmm3
|
||||
; SSE4-NEXT: pxor %xmm7, %xmm3
|
||||
; SSE4-NEXT: movdqa {{.*#+}} xmm4 = [9223372036854841343,9223372036854841343]
|
||||
; SSE4-NEXT: movdqa %xmm4, %xmm0
|
||||
; SSE4-NEXT: pcmpgtq %xmm5, %xmm0
|
||||
; SSE4-NEXT: movapd %xmm7, %xmm5
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm3, %xmm5
|
||||
; SSE4-NEXT: pxor %xmm1, %xmm6
|
||||
; SSE4-NEXT: pcmpgtq %xmm6, %xmm4
|
||||
; SSE4-NEXT: pcmpgtq %xmm3, %xmm0
|
||||
; SSE4-NEXT: movapd %xmm5, %xmm3
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm3
|
||||
; SSE4-NEXT: pxor %xmm8, %xmm7
|
||||
; SSE4-NEXT: pcmpgtq %xmm7, %xmm4
|
||||
; SSE4-NEXT: movdqa %xmm4, %xmm0
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm1, %xmm7
|
||||
; SSE4-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0],xmm8[1,2,3],xmm7[4],xmm8[5,6,7]
|
||||
; SSE4-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0],xmm8[1,2,3],xmm5[4],xmm8[5,6,7]
|
||||
; SSE4-NEXT: packusdw %xmm7, %xmm5
|
||||
; SSE4-NEXT: packusdw %xmm5, %xmm5
|
||||
; SSE4-NEXT: pcmpeqd %xmm8, %xmm2
|
||||
; SSE4-NEXT: movmskps %xmm2, %eax
|
||||
; SSE4-NEXT: blendvpd %xmm0, %xmm8, %xmm5
|
||||
; SSE4-NEXT: packusdw %xmm3, %xmm5
|
||||
; SSE4-NEXT: packusdw %xmm0, %xmm5
|
||||
; SSE4-NEXT: pcmpeqd %xmm2, %xmm6
|
||||
; SSE4-NEXT: movmskps %xmm6, %eax
|
||||
; SSE4-NEXT: xorl $15, %eax
|
||||
; SSE4-NEXT: testb $1, %al
|
||||
; SSE4-NEXT: jne .LBB4_1
|
||||
@ -1632,9 +1624,7 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; AVX1-NEXT: vpcmpgtq %xmm3, %xmm5, %xmm3
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} xmm5 = [65535,65535]
|
||||
; AVX1-NEXT: vblendvpd %xmm3, %xmm6, %xmm5, %xmm3
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0],xmm2[1,2,3],xmm3[4],xmm2[5,6,7]
|
||||
; AVX1-NEXT: vblendvpd %xmm4, %xmm0, %xmm5, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm2[1,2,3],xmm0[4],xmm2[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpcmpeqd %xmm2, %xmm1, %xmm1
|
||||
@ -1680,10 +1670,8 @@ define void @truncstore_v4i64_v4i16(<4 x i64> %x, <4 x i16>* %p, <4 x i32> %mask
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm5 = [9223372036854841343,9223372036854841343,9223372036854841343,9223372036854841343]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm5, %ymm4
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm0, %ymm3, %ymm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm4
|
||||
; AVX2-NEXT: vandpd %xmm3, %xmm4, %xmm4
|
||||
; AVX2-NEXT: vandpd %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm4, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm3
|
||||
; AVX2-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpcmpeqd %xmm2, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vmovmskps %xmm1, %eax
|
||||
@ -7061,9 +7049,10 @@ define void @truncstore_v8i16_v8i8(<8 x i16> %x, <8 x i8>* %p, <8 x i16> %mask)
|
||||
; SSE2-LABEL: truncstore_v8i16_v8i8:
|
||||
; SSE2: # %bb.0:
|
||||
; SSE2-NEXT: pxor %xmm2, %xmm2
|
||||
; SSE2-NEXT: pxor {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [32768,32768,32768,32768,32768,32768,32768,32768]
|
||||
; SSE2-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE2-NEXT: pminsw {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSE2-NEXT: pcmpeqw %xmm1, %xmm2
|
||||
; SSE2-NEXT: pcmpeqd %xmm1, %xmm1
|
||||
|
||||
@ -1471,7 +1471,6 @@ define <4 x i16> @trunc_packus_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
||||
; SSE41-NEXT: movapd {{.*#+}} xmm4 = [65535,65535]
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [2147483648,2147483648]
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [2147549183,2147549183]
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm5
|
||||
@ -1482,27 +1481,27 @@ define <4 x i16> @trunc_packus_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; SSE41-NEXT: pand %xmm5, %xmm0
|
||||
; SSE41-NEXT: por %xmm7, %xmm0
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm5
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm5
|
||||
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm5
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm1
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm1
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm2
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm2
|
||||
; SSE41-NEXT: pcmpgtd %xmm0, %xmm6
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm1, %xmm0
|
||||
; SSE41-NEXT: pand %xmm2, %xmm0
|
||||
; SSE41-NEXT: por %xmm6, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm4
|
||||
; SSE41-NEXT: pxor %xmm2, %xmm2
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm1
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm1
|
||||
; SSE41-NEXT: movapd %xmm1, %xmm6
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm4
|
||||
; SSE41-NEXT: xorpd %xmm1, %xmm1
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm2
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm2
|
||||
; SSE41-NEXT: movapd %xmm2, %xmm6
|
||||
; SSE41-NEXT: pcmpeqd %xmm3, %xmm6
|
||||
; SSE41-NEXT: pcmpgtd %xmm3, %xmm1
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2]
|
||||
; SSE41-NEXT: pcmpgtd %xmm3, %xmm2
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm6, %xmm0
|
||||
; SSE41-NEXT: por %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm1, %xmm1
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm4, %xmm1
|
||||
; SSE41-NEXT: por %xmm2, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm2, %xmm2
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm4, %xmm2
|
||||
; SSE41-NEXT: movapd %xmm5, %xmm4
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm4
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm6
|
||||
@ -1511,12 +1510,9 @@ define <4 x i16> @trunc_packus_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm6, %xmm0
|
||||
; SSE41-NEXT: por %xmm4, %xmm0
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm3
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm3
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm2[1,2,3],xmm3[4],xmm2[5,6,7]
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0],xmm2[1,2,3],xmm1[4],xmm2[5,6,7]
|
||||
; SSE41-NEXT: packusdw %xmm3, %xmm1
|
||||
; SSE41-NEXT: packusdw %xmm1, %xmm1
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm1
|
||||
; SSE41-NEXT: packusdw %xmm2, %xmm1
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm1
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: retq
|
||||
;
|
||||
@ -1532,9 +1528,7 @@ define <4 x i16> @trunc_packus_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; AVX1-NEXT: vblendvpd %xmm3, %xmm1, %xmm2, %xmm1
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm2
|
||||
; AVX1-NEXT: vpand %xmm1, %xmm2, %xmm1
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpand %xmm0, %xmm5, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1,2,3],xmm0[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vzeroupper
|
||||
@ -1545,13 +1539,11 @@ define <4 x i16> @trunc_packus_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm1 = [65535,65535,65535,65535]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm1, %ymm2
|
||||
; AVX2-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vpand %xmm1, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpand %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm0, %ymm1
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
@ -1733,7 +1725,6 @@ define void @trunc_packus_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
||||
; SSE41-NEXT: movapd {{.*#+}} xmm4 = [65535,65535]
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [2147483648,2147483648]
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [2147549183,2147549183]
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm5
|
||||
@ -1744,17 +1735,17 @@ define void @trunc_packus_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; SSE41-NEXT: pand %xmm5, %xmm0
|
||||
; SSE41-NEXT: por %xmm7, %xmm0
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm5
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm5
|
||||
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm5
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm3, %xmm0
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm1
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm1
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm2
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm2
|
||||
; SSE41-NEXT: pcmpgtd %xmm0, %xmm6
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm1, %xmm0
|
||||
; SSE41-NEXT: pand %xmm2, %xmm0
|
||||
; SSE41-NEXT: por %xmm6, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm4
|
||||
; SSE41-NEXT: pxor %xmm1, %xmm1
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm4
|
||||
; SSE41-NEXT: xorpd %xmm1, %xmm1
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm2
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm2
|
||||
; SSE41-NEXT: movapd %xmm2, %xmm6
|
||||
@ -1773,13 +1764,10 @@ define void @trunc_packus_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm6, %xmm0
|
||||
; SSE41-NEXT: por %xmm4, %xmm0
|
||||
; SSE41-NEXT: xorpd %xmm3, %xmm3
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm3
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm1[1,2,3],xmm3[4],xmm1[5,6,7]
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0],xmm1[1,2,3],xmm2[4],xmm1[5,6,7]
|
||||
; SSE41-NEXT: packusdw %xmm3, %xmm2
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm2
|
||||
; SSE41-NEXT: movq %xmm2, (%rdi)
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm1
|
||||
; SSE41-NEXT: packusdw %xmm2, %xmm1
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm1
|
||||
; SSE41-NEXT: movq %xmm1, (%rdi)
|
||||
; SSE41-NEXT: retq
|
||||
;
|
||||
; AVX1-LABEL: trunc_packus_v4i64_v4i16_store:
|
||||
@ -1794,9 +1782,7 @@ define void @trunc_packus_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; AVX1-NEXT: vblendvpd %xmm3, %xmm1, %xmm2, %xmm1
|
||||
; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm2
|
||||
; AVX1-NEXT: vpand %xmm1, %xmm2, %xmm1
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpand %xmm0, %xmm5, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1,2,3],xmm0[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vmovq %xmm0, (%rdi)
|
||||
@ -1808,13 +1794,11 @@ define void @trunc_packus_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm1 = [65535,65535,65535,65535]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm1, %ymm2
|
||||
; AVX2-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vpand %xmm1, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpand %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm0, %ymm1
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vmovq %xmm0, (%rdi)
|
||||
; AVX2-NEXT: vzeroupper
|
||||
@ -3893,26 +3877,20 @@ define <8 x i8> @trunc_packus_v8i64_v8i8(<8 x i64>* %p0) "min-legal-vector-width
|
||||
; AVX2-NEXT: vmovdqa (%rdi), %ymm0
|
||||
; AVX2-NEXT: vmovdqa 32(%rdi), %ymm1
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [255,255,255,255]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm3, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3
|
||||
; AVX2-NEXT: vpand %ymm1, %ymm3, %ymm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
;
|
||||
@ -4302,26 +4280,20 @@ define void @trunc_packus_v8i64_v8i8_store(<8 x i64>* %p0, <8 x i8> *%p1) "min-l
|
||||
; AVX2-NEXT: vmovdqa (%rdi), %ymm0
|
||||
; AVX2-NEXT: vmovdqa 32(%rdi), %ymm1
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [255,255,255,255]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm3, %ymm0
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3
|
||||
; AVX2-NEXT: vpand %ymm1, %ymm3, %ymm1
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm2
|
||||
; AVX2-NEXT: vpand %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vmovq %xmm0, (%rsi)
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@ -1022,34 +1022,32 @@ define <4 x i16> @trunc_usat_v4i64_v4i16(<4 x i64> %a0) {
|
||||
;
|
||||
; SSE41-LABEL: trunc_usat_v4i64_v4i16:
|
||||
; SSE41: # %bb.0:
|
||||
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
||||
; SSE41-NEXT: movapd {{.*#+}} xmm4 = [65535,65535]
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [9223372039002259456,9223372039002259456]
|
||||
; SSE41-NEXT: pxor %xmm5, %xmm0
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [9223372039002324991,9223372039002324991]
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm3
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm3
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm7
|
||||
; SSE41-NEXT: movdqa %xmm0, %xmm3
|
||||
; SSE41-NEXT: movapd {{.*#+}} xmm2 = [65535,65535]
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [9223372039002259456,9223372039002259456]
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm4, %xmm0
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [9223372039002324991,9223372039002324991]
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm6
|
||||
; SSE41-NEXT: pcmpeqd %xmm0, %xmm6
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm7
|
||||
; SSE41-NEXT: pcmpgtd %xmm0, %xmm7
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm7[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm3, %xmm0
|
||||
; SSE41-NEXT: pand %xmm6, %xmm0
|
||||
; SSE41-NEXT: por %xmm7, %xmm0
|
||||
; SSE41-NEXT: movapd %xmm4, %xmm3
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm3
|
||||
; SSE41-NEXT: pxor %xmm1, %xmm5
|
||||
; SSE41-NEXT: movdqa %xmm6, %xmm2
|
||||
; SSE41-NEXT: pcmpeqd %xmm5, %xmm2
|
||||
; SSE41-NEXT: pcmpgtd %xmm5, %xmm6
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm2, %xmm0
|
||||
; SSE41-NEXT: por %xmm6, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm4
|
||||
; SSE41-NEXT: pxor %xmm0, %xmm0
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0],xmm0[1,2,3],xmm4[4],xmm0[5,6,7]
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm0[1,2,3],xmm3[4],xmm0[5,6,7]
|
||||
; SSE41-NEXT: packusdw %xmm4, %xmm3
|
||||
; SSE41-NEXT: packusdw %xmm3, %xmm3
|
||||
; SSE41-NEXT: movdqa %xmm3, %xmm0
|
||||
; SSE41-NEXT: movapd %xmm2, %xmm6
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm6
|
||||
; SSE41-NEXT: pxor %xmm3, %xmm4
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm1
|
||||
; SSE41-NEXT: pcmpeqd %xmm4, %xmm1
|
||||
; SSE41-NEXT: pcmpgtd %xmm4, %xmm5
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm1, %xmm0
|
||||
; SSE41-NEXT: por %xmm5, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm2
|
||||
; SSE41-NEXT: packusdw %xmm6, %xmm2
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm2
|
||||
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
||||
; SSE41-NEXT: retq
|
||||
;
|
||||
; AVX1-LABEL: trunc_usat_v4i64_v4i16:
|
||||
@ -1063,10 +1061,7 @@ define <4 x i16> @trunc_usat_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm3, %xmm1
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} xmm3 = [65535,65535]
|
||||
; AVX1-NEXT: vblendvpd %xmm1, %xmm4, %xmm3, %xmm1
|
||||
; AVX1-NEXT: vpxor %xmm4, %xmm4, %xmm4
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm3, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1,2,3],xmm0[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vzeroupper
|
||||
@ -1080,10 +1075,8 @@ define <4 x i16> @trunc_usat_v4i64_v4i16(<4 x i64> %a0) {
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm3 = [9223372036854841343,9223372036854841343,9223372036854841343,9223372036854841343]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm3, %ymm2
|
||||
; AVX2-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vandpd %xmm1, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vandpd %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
@ -1207,6 +1200,7 @@ define void @trunc_usat_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
||||
; SSE41-NEXT: movapd {{.*#+}} xmm3 = [65535,65535]
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [9223372039002259456,9223372039002259456]
|
||||
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
||||
; SSE41-NEXT: pxor %xmm4, %xmm0
|
||||
; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [9223372039002324991,9223372039002324991]
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm6
|
||||
@ -1217,21 +1211,18 @@ define void @trunc_usat_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; SSE41-NEXT: pand %xmm6, %xmm0
|
||||
; SSE41-NEXT: por %xmm7, %xmm0
|
||||
; SSE41-NEXT: movapd %xmm3, %xmm6
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm6
|
||||
; SSE41-NEXT: pxor %xmm1, %xmm4
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm2
|
||||
; SSE41-NEXT: pcmpeqd %xmm4, %xmm2
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm6
|
||||
; SSE41-NEXT: pxor %xmm2, %xmm4
|
||||
; SSE41-NEXT: movdqa %xmm5, %xmm1
|
||||
; SSE41-NEXT: pcmpeqd %xmm4, %xmm1
|
||||
; SSE41-NEXT: pcmpgtd %xmm4, %xmm5
|
||||
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2]
|
||||
; SSE41-NEXT: pand %xmm2, %xmm0
|
||||
; SSE41-NEXT: pand %xmm1, %xmm0
|
||||
; SSE41-NEXT: por %xmm5, %xmm0
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm3
|
||||
; SSE41-NEXT: pxor %xmm0, %xmm0
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm0[1,2,3],xmm3[4],xmm0[5,6,7]
|
||||
; SSE41-NEXT: pblendw {{.*#+}} xmm6 = xmm6[0],xmm0[1,2,3],xmm6[4],xmm0[5,6,7]
|
||||
; SSE41-NEXT: packusdw %xmm3, %xmm6
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm6
|
||||
; SSE41-NEXT: movq %xmm6, (%rdi)
|
||||
; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm3
|
||||
; SSE41-NEXT: packusdw %xmm6, %xmm3
|
||||
; SSE41-NEXT: packusdw %xmm0, %xmm3
|
||||
; SSE41-NEXT: movq %xmm3, (%rdi)
|
||||
; SSE41-NEXT: retq
|
||||
;
|
||||
; AVX1-LABEL: trunc_usat_v4i64_v4i16_store:
|
||||
@ -1245,10 +1236,7 @@ define void @trunc_usat_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm3, %xmm1
|
||||
; AVX1-NEXT: vmovapd {{.*#+}} xmm3 = [65535,65535]
|
||||
; AVX1-NEXT: vblendvpd %xmm1, %xmm4, %xmm3, %xmm1
|
||||
; AVX1-NEXT: vpxor %xmm4, %xmm4, %xmm4
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm3, %xmm0
|
||||
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1,2,3],xmm0[4],xmm4[5,6,7]
|
||||
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX1-NEXT: vmovq %xmm0, (%rdi)
|
||||
@ -1263,10 +1251,8 @@ define void @trunc_usat_v4i64_v4i16_store(<4 x i64> %a0, <4 x i16> *%p1) {
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm3 = [9223372036854841343,9223372036854841343,9223372036854841343,9223372036854841343]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm2, %ymm3, %ymm2
|
||||
; AVX2-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vandpd %xmm1, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vandpd %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackusdw %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vmovq %xmm0, (%rdi)
|
||||
; AVX2-NEXT: vzeroupper
|
||||
@ -2818,24 +2804,18 @@ define <8 x i8> @trunc_usat_v8i64_v8i8(<8 x i64>* %p0) {
|
||||
; AVX2-NEXT: vmovdqa 32(%rdi), %ymm1
|
||||
; AVX2-NEXT: vbroadcastsd {{.*#+}} ymm2 = [255,255,255,255]
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm3 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm0, %ymm4
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm1, %ymm4
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm5 = [9223372036854776063,9223372036854776063,9223372036854776063,9223372036854776063]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm5, %ymm4
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm1, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm0, %ymm3
|
||||
; AVX2-NEXT: vpcmpgtq %ymm3, %ymm5, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
;
|
||||
@ -3085,24 +3065,18 @@ define void @trunc_usat_v8i64_v8i8_store(<8 x i64>* %p0, <8 x i8> *%p1) {
|
||||
; AVX2-NEXT: vmovdqa 32(%rdi), %ymm1
|
||||
; AVX2-NEXT: vbroadcastsd {{.*#+}} ymm2 = [255,255,255,255]
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm3 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm0, %ymm4
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm1, %ymm4
|
||||
; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm5 = [9223372036854776063,9223372036854776063,9223372036854776063,9223372036854776063]
|
||||
; AVX2-NEXT: vpcmpgtq %ymm4, %ymm5, %ymm4
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm1, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm4, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vpxor %ymm3, %ymm0, %ymm3
|
||||
; AVX2-NEXT: vpcmpgtq %ymm3, %ymm5, %ymm3
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm1, %xmm1
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
|
||||
; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm2
|
||||
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
|
||||
; AVX2-NEXT: vpshufb %xmm3, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
|
||||
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
|
||||
; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0
|
||||
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
|
||||
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
|
||||
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
|
||||
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
|
||||
; AVX2-NEXT: vmovq %xmm0, (%rsi)
|
||||
; AVX2-NEXT: vzeroupper
|
||||
; AVX2-NEXT: retq
|
||||
@ -3830,23 +3804,21 @@ define <8 x i8> @trunc_usat_v8i32_v8i8(<8 x i32> %a0) {
|
||||
; SSSE3: # %bb.0:
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255]
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [2147483648,2147483648,2147483648,2147483648]
|
||||
; SSSE3-NEXT: movdqa %xmm0, %xmm4
|
||||
; SSSE3-NEXT: movdqa %xmm1, %xmm4
|
||||
; SSSE3-NEXT: pxor %xmm3, %xmm4
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [2147483903,2147483903,2147483903,2147483903]
|
||||
; SSSE3-NEXT: movdqa %xmm5, %xmm6
|
||||
; SSSE3-NEXT: pcmpgtd %xmm4, %xmm6
|
||||
; SSSE3-NEXT: pand %xmm6, %xmm0
|
||||
; SSSE3-NEXT: pand %xmm6, %xmm1
|
||||
; SSSE3-NEXT: pandn %xmm2, %xmm6
|
||||
; SSSE3-NEXT: por %xmm6, %xmm0
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm3
|
||||
; SSSE3-NEXT: por %xmm1, %xmm6
|
||||
; SSSE3-NEXT: pxor %xmm0, %xmm3
|
||||
; SSSE3-NEXT: pcmpgtd %xmm3, %xmm5
|
||||
; SSSE3-NEXT: pand %xmm5, %xmm1
|
||||
; SSSE3-NEXT: pand %xmm5, %xmm0
|
||||
; SSSE3-NEXT: pandn %xmm2, %xmm5
|
||||
; SSSE3-NEXT: por %xmm1, %xmm5
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; SSSE3-NEXT: pshufb %xmm1, %xmm5
|
||||
; SSSE3-NEXT: pshufb %xmm1, %xmm0
|
||||
; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1]
|
||||
; SSSE3-NEXT: por %xmm5, %xmm0
|
||||
; SSSE3-NEXT: packuswb %xmm6, %xmm0
|
||||
; SSSE3-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSSE3-NEXT: retq
|
||||
;
|
||||
; SSE41-LABEL: trunc_usat_v8i32_v8i8:
|
||||
@ -3943,24 +3915,22 @@ define void @trunc_usat_v8i32_v8i8_store(<8 x i32> %a0, <8 x i8> *%p1) {
|
||||
; SSSE3: # %bb.0:
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255]
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [2147483648,2147483648,2147483648,2147483648]
|
||||
; SSSE3-NEXT: movdqa %xmm0, %xmm4
|
||||
; SSSE3-NEXT: movdqa %xmm1, %xmm4
|
||||
; SSSE3-NEXT: pxor %xmm3, %xmm4
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [2147483903,2147483903,2147483903,2147483903]
|
||||
; SSSE3-NEXT: movdqa %xmm5, %xmm6
|
||||
; SSSE3-NEXT: pcmpgtd %xmm4, %xmm6
|
||||
; SSSE3-NEXT: pand %xmm6, %xmm0
|
||||
; SSSE3-NEXT: pand %xmm6, %xmm1
|
||||
; SSSE3-NEXT: pandn %xmm2, %xmm6
|
||||
; SSSE3-NEXT: por %xmm0, %xmm6
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm3
|
||||
; SSSE3-NEXT: por %xmm1, %xmm6
|
||||
; SSSE3-NEXT: pxor %xmm0, %xmm3
|
||||
; SSSE3-NEXT: pcmpgtd %xmm3, %xmm5
|
||||
; SSSE3-NEXT: pand %xmm5, %xmm1
|
||||
; SSSE3-NEXT: pand %xmm5, %xmm0
|
||||
; SSSE3-NEXT: pandn %xmm2, %xmm5
|
||||
; SSSE3-NEXT: por %xmm1, %xmm5
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm0 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
|
||||
; SSSE3-NEXT: pshufb %xmm0, %xmm5
|
||||
; SSSE3-NEXT: pshufb %xmm0, %xmm6
|
||||
; SSSE3-NEXT: punpckldq {{.*#+}} xmm6 = xmm6[0],xmm5[0],xmm6[1],xmm5[1]
|
||||
; SSSE3-NEXT: movq %xmm6, (%rdi)
|
||||
; SSSE3-NEXT: por %xmm0, %xmm5
|
||||
; SSSE3-NEXT: packuswb %xmm6, %xmm5
|
||||
; SSSE3-NEXT: packuswb %xmm0, %xmm5
|
||||
; SSSE3-NEXT: movq %xmm5, (%rdi)
|
||||
; SSSE3-NEXT: retq
|
||||
;
|
||||
; SSE41-LABEL: trunc_usat_v8i32_v8i8_store:
|
||||
@ -4325,17 +4295,20 @@ define void @trunc_usat_v16i32_v16i8_store(<16 x i32>* %p0, <16 x i8>* %p1) {
|
||||
define <8 x i8> @trunc_usat_v8i16_v8i8(<8 x i16> %a0) {
|
||||
; SSE2-LABEL: trunc_usat_v8i16_v8i8:
|
||||
; SSE2: # %bb.0:
|
||||
; SSE2-NEXT: pxor {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [32768,32768,32768,32768,32768,32768,32768,32768]
|
||||
; SSE2-NEXT: pxor %xmm1, %xmm0
|
||||
; SSE2-NEXT: pminsw {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pxor %xmm1, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSE2-NEXT: retq
|
||||
;
|
||||
; SSSE3-LABEL: trunc_usat_v8i16_v8i8:
|
||||
; SSSE3: # %bb.0:
|
||||
; SSSE3-NEXT: pxor {{.*}}(%rip), %xmm0
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [32768,32768,32768,32768,32768,32768,32768,32768]
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm0
|
||||
; SSSE3-NEXT: pminsw {{.*}}(%rip), %xmm0
|
||||
; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm0
|
||||
; SSSE3-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSSE3-NEXT: retq
|
||||
;
|
||||
; SSE41-LABEL: trunc_usat_v8i16_v8i8:
|
||||
@ -4388,18 +4361,21 @@ define <8 x i8> @trunc_usat_v8i16_v8i8(<8 x i16> %a0) {
|
||||
define void @trunc_usat_v8i16_v8i8_store(<8 x i16> %a0, <8 x i8> *%p1) {
|
||||
; SSE2-LABEL: trunc_usat_v8i16_v8i8_store:
|
||||
; SSE2: # %bb.0:
|
||||
; SSE2-NEXT: pxor {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [32768,32768,32768,32768,32768,32768,32768,32768]
|
||||
; SSE2-NEXT: pxor %xmm1, %xmm0
|
||||
; SSE2-NEXT: pminsw {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
||||
; SSE2-NEXT: pxor %xmm1, %xmm0
|
||||
; SSE2-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSE2-NEXT: movq %xmm0, (%rdi)
|
||||
; SSE2-NEXT: retq
|
||||
;
|
||||
; SSSE3-LABEL: trunc_usat_v8i16_v8i8_store:
|
||||
; SSSE3: # %bb.0:
|
||||
; SSSE3-NEXT: pxor {{.*}}(%rip), %xmm0
|
||||
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [32768,32768,32768,32768,32768,32768,32768,32768]
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm0
|
||||
; SSSE3-NEXT: pminsw {{.*}}(%rip), %xmm0
|
||||
; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
|
||||
; SSSE3-NEXT: pxor %xmm1, %xmm0
|
||||
; SSSE3-NEXT: packuswb %xmm0, %xmm0
|
||||
; SSSE3-NEXT: movq %xmm0, (%rdi)
|
||||
; SSSE3-NEXT: retq
|
||||
;
|
||||
|
||||
Loading…
Reference in New Issue
Block a user