6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-09-19 04:26:05 +00:00
Code Issues Projects Releases Wiki Activity

This patch recognizes (+ (+ v0, v1) (+ v2, v3)), reorders them for bundling into vector of loads,

Browse Source 
and vectorizes it. 
 
 Test case :
 
       float hadd(float* a) {
           return (a[0] + a[1]) + (a[2] + a[3]);
        }
 
 
 AArch64 assembly before patch :
 
        ldp	s0, s1, [x0]
 	ldp	s2, s3, [x0, #8]
 	fadd	s0, s0, s1
 	fadd	s1, s2, s3
 	fadd	s0, s0, s1
 	ret
 
 AArch64 assembly after patch :
 
        ldp	d0, d1, [x0]
 	fadd	v0.2s, v0.2s, v1.2s
 	faddp	s0, v0.2s
 	ret

Reviewed Link : http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20141208/248531.html



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224119 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Suyog Sarda
2014-12-12 12:53:44 +00:00
parent c15d82e259
commit 1dea0dc279
2 changed files with 51 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
lib/Transforms/Vectorize/SLPVectorizer.cpp
View File

@@ -439,6 +439,13 @@ public:
/// \returns true if the memory operations A and B are consecutive.
bool isConsecutiveAccess(Value *A, Value *B);
/// For consecutive loads (+(+ v0, v1)(+ v2, v3)), Left had v0 and v2
/// while Right had v1 and v3, which prevented bundling them into
/// a vector of loads. Rorder them so that Left now has v0 and v1
/// while Right has v2 and v3 enabling their bundling into a vector.
void reorderIfConsecutiveLoads(SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right);
/// \brief Perform LICM and CSE on the newly generated gather sequences.
void optimizeGatherSequence();
@@ -1234,6 +1241,7 @@ void BoUpSLP::buildTree_rec(ArrayRef<Value *> VL, unsigned Depth) {
if (isa<BinaryOperator>(VL0) && VL0->isCommutative()) {
ValueList Left, Right;
reorderInputsAccordingToOpcode(VL, Left, Right);
reorderIfConsecutiveLoads (Left, Right);
buildTree_rec(Left, Depth + 1);
buildTree_rec(Right, Depth + 1);
return;
@@ -1818,6 +1826,19 @@ bool BoUpSLP::isConsecutiveAccess(Value *A, Value *B) {
return X == PtrSCEVB;
}
void BoUpSLP::reorderIfConsecutiveLoads(SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right) {
for (unsigned i = 0, e = Left.size(); i < e - 1; ++i) {
if (!isa<LoadInst>(Left[i]) || !isa<LoadInst>(Right[i]))
return;
if (!(isConsecutiveAccess(Left[i], Right[i]) &&
isConsecutiveAccess(Right[i], Left[i + 1])))
continue;
else
std::swap(Left[i + 1], Right[i]);
}
}
void BoUpSLP::setInsertPointAfterBundle(ArrayRef<Value *> VL) {
Instruction *VL0 = cast<Instruction>(VL[0]);
BasicBlock::iterator NextInst = VL0;
@@ -2048,9 +2069,10 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) {
case Instruction::Or:
case Instruction::Xor: {
ValueList LHSVL, RHSVL;
if (isa<BinaryOperator>(VL0) && VL0->isCommutative())
if (isa<BinaryOperator>(VL0) && VL0->isCommutative()) {
reorderInputsAccordingToOpcode(E->Scalars, LHSVL, RHSVL);
else
reorderIfConsecutiveLoads(LHSVL, RHSVL);
} else
for (int i = 0, e = E->Scalars.size(); i < e; ++i) {
LHSVL.push_back(cast<Instruction>(E->Scalars[i])->getOperand(0));
RHSVL.push_back(cast<Instruction>(E->Scalars[i])->getOperand(1));