[llvm] r245382 - use TLI.allowsMemoryAccess() to check if memory accesses are fast; NFCI
Sanjay Patel via llvm-commits
llvm-commits at lists.llvm.org
Tue Aug 18 15:48:12 PDT 2015
Author: spatel
Date: Tue Aug 18 17:48:12 2015
New Revision: 245382
URL: http://llvm.org/viewvc/llvm-project?rev=245382&view=rev
Log:
use TLI.allowsMemoryAccess() to check if memory accesses are fast; NFCI
This consolidates use of isUnalignedMem32Slow() in one place.
There is a slight change in logic although I'm not sure that it would ever
come up in the real world: we were assuming that an alignment of the type
size is always fast; now, we actually check the data layout to confirm that.
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=245382&r1=245381&r2=245382&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Tue Aug 18 17:48:12 2015
@@ -24101,10 +24101,13 @@ static SDValue PerformLOADCombine(SDNode
// For chips with slow 32-byte unaligned loads, break the 32-byte operation
// into two 16-byte operations.
ISD::LoadExtType Ext = Ld->getExtensionType();
+ bool Fast;
+ unsigned AddressSpace = Ld->getAddressSpace();
unsigned Alignment = Ld->getAlignment();
- bool IsAligned = Alignment == 0 || Alignment >= MemVT.getSizeInBits()/8;
- if (RegVT.is256BitVector() && Subtarget->isUnalignedMem32Slow() &&
- !DCI.isBeforeLegalizeOps() && !IsAligned && Ext == ISD::NON_EXTLOAD) {
+ if (RegVT.is256BitVector() && !DCI.isBeforeLegalizeOps() &&
+ Ext == ISD::NON_EXTLOAD &&
+ TLI.allowsMemoryAccess(*DAG.getContext(), DAG.getDataLayout(), RegVT,
+ AddressSpace, Alignment, &Fast) && !Fast) {
unsigned NumElems = RegVT.getVectorNumElements();
if (NumElems < 2)
return SDValue();
@@ -24318,10 +24321,12 @@ static SDValue PerformSTORECombine(SDNod
// If we are saving a concatenation of two XMM registers and 32-byte stores
// are slow, such as on Sandy Bridge, perform two 16-byte stores.
+ bool Fast;
+ unsigned AddressSpace = St->getAddressSpace();
unsigned Alignment = St->getAlignment();
- bool IsAligned = Alignment == 0 || Alignment >= VT.getSizeInBits()/8;
- if (VT.is256BitVector() && Subtarget->isUnalignedMem32Slow() &&
- StVT == VT && !IsAligned) {
+ if (VT.is256BitVector() && StVT == VT &&
+ TLI.allowsMemoryAccess(*DAG.getContext(), DAG.getDataLayout(), VT,
+ AddressSpace, Alignment, &Fast) && !Fast) {
unsigned NumElems = VT.getVectorNumElements();
if (NumElems < 2)
return SDValue();
More information about the llvm-commits
mailing list