summaryrefslogtreecommitdiffstats
path: root/llvm/lib
diff options
context:
space:
mode:
authorCraig Topper <craig.topper@gmail.com>2015-03-05 06:38:42 +0000
committerCraig Topper <craig.topper@gmail.com>2015-03-05 06:38:42 +0000
commit0ee8470a436f696bace4514d56eb51960fc419cf (patch)
treea242c77ed771402291a32ffca92f255fe761a7d9 /llvm/lib
parent6e56345dbcca8f8aa29865c17fae46f61384a4f0 (diff)
downloadbcm5719-llvm-0ee8470a436f696bace4514d56eb51960fc419cf.tar.gz
bcm5719-llvm-0ee8470a436f696bace4514d56eb51960fc419cf.zip
[X86] Use vmovss to handle inserting an element into index 0 of a v8f32 vector of zeros.
llvm-svn: 231354
Diffstat (limited to 'llvm/lib')
-rw-r--r--llvm/lib/Target/X86/X86ISelLowering.cpp14
1 files changed, 12 insertions, 2 deletions
diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index e5af7293b0e..e51149210bf 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -5634,12 +5634,13 @@ X86TargetLowering::LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG) const {
if (ExtVT == MVT::i32 || ExtVT == MVT::f32 || ExtVT == MVT::f64 ||
(ExtVT == MVT::i64 && Subtarget->is64Bit())) {
- if (VT.is256BitVector() || VT.is512BitVector()) {
+ if (VT.is512BitVector()) {
SDValue ZeroVec = getZeroVector(VT, Subtarget, DAG, dl);
return DAG.getNode(ISD::INSERT_VECTOR_ELT, dl, VT, ZeroVec,
Item, DAG.getIntPtrConstant(0));
}
- assert(VT.is128BitVector() && "Expected an SSE value type!");
+ assert((VT.is128BitVector() || VT.is256BitVector()) &&
+ "Expected an SSE value type!");
Item = DAG.getNode(ISD::SCALAR_TO_VECTOR, dl, VT, Item);
// Turn it into a MOVL (i.e. movss, movsd, or movd) to a zero vector.
return getShuffleVectorZeroOrUndef(Item, 0, true, Subtarget, DAG);
@@ -9333,6 +9334,15 @@ static SDValue lowerV8F32VectorShuffle(SDValue Op, SDValue V1, SDValue V2,
ArrayRef<int> Mask = SVOp->getMask();
assert(Mask.size() == 8 && "Unexpected mask size for v8 shuffle!");
+ // If we have a single input to the zero element, insert that into V1 if we
+ // can do so cheaply.
+ int NumV2Elements =
+ std::count_if(Mask.begin(), Mask.end(), [](int M) { return M >= 8; });
+ if (NumV2Elements == 1 && Mask[0] >= 8)
+ if (SDValue Insertion = lowerVectorShuffleAsElementInsertion(
+ DL, MVT::v8f32, V1, V2, Mask, Subtarget, DAG))
+ return Insertion;
+
if (SDValue Blend = lowerVectorShuffleAsBlend(DL, MVT::v8f32, V1, V2, Mask,
Subtarget, DAG))
return Blend;
OpenPOWER on IntegriCloud