Don't try to match 'unpackl/h v, v' for 32xi8 and 16xi16 when only AVX1 is supported. Fix 'unpackh v, v' for 256-bit types to understand 128-bit lanes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146726 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 03727a2..a7b38a5 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -3472,7 +3472,7 @@
/// specifies a shuffle of elements that is suitable for input to UNPCKL.
static bool isUNPCKLMask(const SmallVectorImpl<int> &Mask, EVT VT,
bool HasAVX2, bool V2IsSplat = false) {
- int NumElts = VT.getVectorNumElements();
+ unsigned NumElts = VT.getVectorNumElements();
assert((VT.is128BitVector() || VT.is256BitVector()) &&
"Unsupported vector type for unpckh");
@@ -3486,11 +3486,9 @@
unsigned NumLanes = VT.getSizeInBits()/128;
unsigned NumLaneElts = NumElts/NumLanes;
- unsigned Start = 0;
- unsigned End = NumLaneElts;
- for (unsigned s = 0; s < NumLanes; ++s) {
- for (unsigned i = Start, j = s * NumLaneElts;
- i != End;
+ for (unsigned l = 0; l != NumLanes; ++l) {
+ for (unsigned i = l*NumLaneElts, j = l*NumLaneElts;
+ i != (l+1)*NumLaneElts;
i += 2, ++j) {
int BitI = Mask[i];
int BitI1 = Mask[i+1];
@@ -3504,9 +3502,6 @@
return false;
}
}
- // Process the next 128 bits.
- Start += NumLaneElts;
- End += NumLaneElts;
}
return true;
@@ -3522,7 +3517,7 @@
/// specifies a shuffle of elements that is suitable for input to UNPCKH.
static bool isUNPCKHMask(const SmallVectorImpl<int> &Mask, EVT VT,
bool HasAVX2, bool V2IsSplat = false) {
- int NumElts = VT.getVectorNumElements();
+ unsigned NumElts = VT.getVectorNumElements();
assert((VT.is128BitVector() || VT.is256BitVector()) &&
"Unsupported vector type for unpckh");
@@ -3536,11 +3531,9 @@
unsigned NumLanes = VT.getSizeInBits()/128;
unsigned NumLaneElts = NumElts/NumLanes;
- unsigned Start = 0;
- unsigned End = NumLaneElts;
for (unsigned l = 0; l != NumLanes; ++l) {
- for (unsigned i = Start, j = (l*NumLaneElts)+NumLaneElts/2;
- i != End; i += 2, ++j) {
+ for (unsigned i = l*NumLaneElts, j = (l*NumLaneElts)+NumLaneElts/2;
+ i != (l+1)*NumLaneElts; i += 2, ++j) {
int BitI = Mask[i];
int BitI1 = Mask[i+1];
if (!isUndefOrEqual(BitI, j))
@@ -3553,9 +3546,6 @@
return false;
}
}
- // Process the next 128 bits.
- Start += NumLaneElts;
- End += NumLaneElts;
}
return true;
}
@@ -3569,26 +3559,32 @@
/// isUNPCKL_v_undef_Mask - Special case of isUNPCKLMask for canonical form
/// of vector_shuffle v, v, <0, 4, 1, 5>, i.e. vector_shuffle v, undef,
/// <0, 0, 1, 1>
-static bool isUNPCKL_v_undef_Mask(const SmallVectorImpl<int> &Mask, EVT VT) {
- int NumElems = VT.getVectorNumElements();
- if (NumElems != 2 && NumElems != 4 && NumElems != 8 && NumElems != 16)
+static bool isUNPCKL_v_undef_Mask(const SmallVectorImpl<int> &Mask, EVT VT,
+ bool HasAVX2) {
+ unsigned NumElts = VT.getVectorNumElements();
+
+ assert((VT.is128BitVector() || VT.is256BitVector()) &&
+ "Unsupported vector type for unpckh");
+
+ if (VT.getSizeInBits() == 256 && NumElts != 4 && NumElts != 8 &&
+ (!HasAVX2 || (NumElts != 16 && NumElts != 32)))
return false;
// For 256-bit i64/f64, use MOVDDUPY instead, so reject the matching pattern
// FIXME: Need a better way to get rid of this, there's no latency difference
// between UNPCKLPD and MOVDDUP, the later should always be checked first and
// the former later. We should also remove the "_undef" special mask.
- if (NumElems == 4 && VT.getSizeInBits() == 256)
+ if (NumElts == 4 && VT.getSizeInBits() == 256)
return false;
// Handle 128 and 256-bit vector lengths. AVX defines UNPCK* to operate
// independently on 128-bit lanes.
- unsigned NumLanes = VT.getSizeInBits() / 128;
- unsigned NumLaneElts = NumElems / NumLanes;
+ unsigned NumLanes = VT.getSizeInBits()/128;
+ unsigned NumLaneElts = NumElts/NumLanes;
- for (unsigned s = 0; s < NumLanes; ++s) {
- for (unsigned i = s * NumLaneElts, j = s * NumLaneElts;
- i != NumLaneElts * (s + 1);
+ for (unsigned l = 0; l != NumLanes; ++l) {
+ for (unsigned i = l*NumLaneElts, j = l*NumLaneElts;
+ i != (l+1)*NumLaneElts;
i += 2, ++j) {
int BitI = Mask[i];
int BitI1 = Mask[i+1];
@@ -3603,35 +3599,49 @@
return true;
}
-bool X86::isUNPCKL_v_undef_Mask(ShuffleVectorSDNode *N) {
+bool X86::isUNPCKL_v_undef_Mask(ShuffleVectorSDNode *N, bool HasAVX2) {
SmallVector<int, 8> M;
N->getMask(M);
- return ::isUNPCKL_v_undef_Mask(M, N->getValueType(0));
+ return ::isUNPCKL_v_undef_Mask(M, N->getValueType(0), HasAVX2);
}
/// isUNPCKH_v_undef_Mask - Special case of isUNPCKHMask for canonical form
/// of vector_shuffle v, v, <2, 6, 3, 7>, i.e. vector_shuffle v, undef,
/// <2, 2, 3, 3>
-static bool isUNPCKH_v_undef_Mask(const SmallVectorImpl<int> &Mask, EVT VT) {
- int NumElems = VT.getVectorNumElements();
- if (NumElems != 2 && NumElems != 4 && NumElems != 8 && NumElems != 16)
+static bool isUNPCKH_v_undef_Mask(const SmallVectorImpl<int> &Mask, EVT VT,
+ bool HasAVX2) {
+ unsigned NumElts = VT.getVectorNumElements();
+
+ assert((VT.is128BitVector() || VT.is256BitVector()) &&
+ "Unsupported vector type for unpckh");
+
+ if (VT.getSizeInBits() == 256 && NumElts != 4 && NumElts != 8 &&
+ (!HasAVX2 || (NumElts != 16 && NumElts != 32)))
return false;
- for (int i = 0, j = NumElems / 2; i != NumElems; i += 2, ++j) {
- int BitI = Mask[i];
- int BitI1 = Mask[i+1];
- if (!isUndefOrEqual(BitI, j))
- return false;
- if (!isUndefOrEqual(BitI1, j))
- return false;
+ // Handle 128 and 256-bit vector lengths. AVX defines UNPCK* to operate
+ // independently on 128-bit lanes.
+ unsigned NumLanes = VT.getSizeInBits()/128;
+ unsigned NumLaneElts = NumElts/NumLanes;
+
+ for (unsigned l = 0; l != NumLanes; ++l) {
+ for (unsigned i = l*NumLaneElts, j = (l*NumLaneElts)+NumLaneElts/2;
+ i != (l+1)*NumLaneElts; i += 2, ++j) {
+ int BitI = Mask[i];
+ int BitI1 = Mask[i+1];
+ if (!isUndefOrEqual(BitI, j))
+ return false;
+ if (!isUndefOrEqual(BitI1, j))
+ return false;
+ }
}
return true;
}
-bool X86::isUNPCKH_v_undef_Mask(ShuffleVectorSDNode *N) {
+bool X86::isUNPCKH_v_undef_Mask(ShuffleVectorSDNode *N, bool HasAVX2) {
SmallVector<int, 8> M;
N->getMask(M);
- return ::isUNPCKH_v_undef_Mask(M, N->getValueType(0));
+ return ::isUNPCKH_v_undef_Mask(M, N->getValueType(0), HasAVX2);
}
/// isMOVLMask - Return true if the specified VECTOR_SHUFFLE operand
@@ -6481,9 +6491,9 @@
// NOTE: isPSHUFDMask can also match both masks below (unpckl_undef and
// unpckh_undef). Only use pshufd if speed is more important than size.
- if (OptForSize && X86::isUNPCKL_v_undef_Mask(SVOp))
+ if (OptForSize && X86::isUNPCKL_v_undef_Mask(SVOp, HasAVX2))
return getTargetShuffleNode(X86ISD::UNPCKL, dl, VT, V1, V1, DAG);
- if (OptForSize && X86::isUNPCKH_v_undef_Mask(SVOp))
+ if (OptForSize && X86::isUNPCKH_v_undef_Mask(SVOp, HasAVX2))
return getTargetShuffleNode(X86ISD::UNPCKH, dl, VT, V1, V1, DAG);
if (X86::isMOVDDUPMask(SVOp) && Subtarget->hasSSE3orAVX() &&
@@ -6663,9 +6673,9 @@
return getTargetShuffleNode(getSHUFPOpcode(VT), dl, VT, V1, V2,
X86::getShuffleSHUFImmediate(SVOp), DAG);
- if (isUNPCKL_v_undef_Mask(M, VT))
+ if (isUNPCKL_v_undef_Mask(M, VT, HasAVX2))
return getTargetShuffleNode(X86ISD::UNPCKL, dl, VT, V1, V1, DAG);
- if (isUNPCKH_v_undef_Mask(M, VT))
+ if (isUNPCKH_v_undef_Mask(M, VT, HasAVX2))
return getTargetShuffleNode(X86ISD::UNPCKH, dl, VT, V1, V1, DAG);
//===--------------------------------------------------------------------===//
@@ -11100,8 +11110,8 @@
isPALIGNRMask(M, VT, Subtarget->hasSSSE3orAVX()) ||
isUNPCKLMask(M, VT, Subtarget->hasAVX2()) ||
isUNPCKHMask(M, VT, Subtarget->hasAVX2()) ||
- isUNPCKL_v_undef_Mask(M, VT) ||
- isUNPCKH_v_undef_Mask(M, VT));
+ isUNPCKL_v_undef_Mask(M, VT, Subtarget->hasAVX2()) ||
+ isUNPCKH_v_undef_Mask(M, VT, Subtarget->hasAVX2()));
}
bool