RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-19 11:02:59 +02:00
Code Issues Projects Releases Wiki Activity

[X86][AVX2] Begun generalizing lowering to VPERMD/VPERMPS in preparation for AVX512 support.

Browse Source 
llvm-svn: 284823
This commit is contained in:
Simon Pilgrim 2016-10-21 13:00:47 +00:00
parent ef32c05524
commit da9223b586
1 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
lib/Target/X86/X86ISelLowering.cpp
View File

@ -25528,16 +25528,17 @@ static bool combineX86ShuffleChain(ArrayRef<SDValue> Inputs, SDValue Root,
any_of(Mask, [](int M) { return M == SM_SentinelZero; }); any_of(Mask, [](int M) { return M == SM_SentinelZero; });
if (is128BitLaneCrossingShuffleMask(MaskVT, Mask)) { if (is128BitLaneCrossingShuffleMask(MaskVT, Mask)) {
// If we have a single input lane-crossing shuffle with 32-bit scalars then // If we have a single input lane-crossing shuffle then lower to VPERMV.
// lower to VPERMD/VPERMPS.
if (UnaryShuffle && (Depth >= 3 || HasVariableMask) && !MaskContainsZeros && if (UnaryShuffle && (Depth >= 3 || HasVariableMask) && !MaskContainsZeros &&
Subtarget.hasAVX2() && (MaskVT == MVT::v8f32 || MaskVT == MVT::v8i32)) { Subtarget.hasAVX2() && (MaskVT == MVT::v8f32 || MaskVT == MVT::v8i32)) {
SDValue VPermIdx[8]; MVT VPermMaskSVT = MVT::getIntegerVT(MaskEltSizeInBits);
for (int i = 0; i < 8; ++i) SmallVector<SDValue, 8> VPermIdx;
VPermIdx[i] = Mask[i] < 0 ? DAG.getUNDEF(MVT::i32) for (int M : Mask)
: DAG.getConstant(Mask[i], DL, MVT::i32); VPermIdx.push_back(M < 0 ? DAG.getUNDEF(VPermMaskSVT)
: DAG.getConstant(M, DL, VPermMaskSVT));
SDValue VPermMask = DAG.getBuildVector(MVT::v8i32, DL, VPermIdx); MVT VPermMaskVT = MVT::getVectorVT(VPermMaskSVT, NumMaskElts);
SDValue VPermMask = DAG.getBuildVector(VPermMaskVT, DL, VPermIdx);
DCI.AddToWorklist(VPermMask.getNode()); DCI.AddToWorklist(VPermMask.getNode());
Res = DAG.getBitcast(MaskVT, V1); Res = DAG.getBitcast(MaskVT, V1);
DCI.AddToWorklist(Res.getNode()); DCI.AddToWorklist(Res.getNode());