Optimize some 64-bit multiplication by constants into two lea's or one lea + shl since imulq is slow (latency 5). e.g.
x * 40
=>
shlq $3, %rdi
leaq (%rdi,%rdi,4), %rax
This has the added benefit of allowing more multiply to be folded into addressing mode. e.g.
a * 24 + b
=>
leaq (%rdi,%rdi,2), %rax
leaq (%rsi,%rax,8), %rax
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@67917 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 4856059..29174b9 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -826,6 +826,8 @@
setTargetDAGCombine(ISD::SRA);
setTargetDAGCombine(ISD::SRL);
setTargetDAGCombine(ISD::STORE);
+ if (Subtarget->is64Bit())
+ setTargetDAGCombine(ISD::MUL);
computeRegisterProperties();
@@ -8407,6 +8409,74 @@
}
+/// PerformMulCombine - Optimize a single multiply with constant into two
+/// in order to implement it with two cheaper instructions, e.g.
+/// LEA + SHL, LEA + LEA.
+static SDValue PerformMulCombine(SDNode *N, SelectionDAG &DAG,
+ TargetLowering::DAGCombinerInfo &DCI) {
+ if (DAG.getMachineFunction().
+ getFunction()->hasFnAttr(Attribute::OptimizeForSize))
+ return SDValue();
+
+ if (DCI.isBeforeLegalize() || DCI.isCalledByLegalizer())
+ return SDValue();
+
+ MVT VT = N->getValueType(0);
+ if (VT != MVT::i64)
+ return SDValue();
+
+ ConstantSDNode *C = dyn_cast<ConstantSDNode>(N->getOperand(1));
+ if (!C)
+ return SDValue();
+ uint64_t MulAmt = C->getZExtValue();
+ if (isPowerOf2_64(MulAmt) || MulAmt == 3 || MulAmt == 5 || MulAmt == 9)
+ return SDValue();
+
+ uint64_t MulAmt1 = 0;
+ uint64_t MulAmt2 = 0;
+ if ((MulAmt % 9) == 0) {
+ MulAmt1 = 9;
+ MulAmt2 = MulAmt / 9;
+ } else if ((MulAmt % 5) == 0) {
+ MulAmt1 = 5;
+ MulAmt2 = MulAmt / 5;
+ } else if ((MulAmt % 3) == 0) {
+ MulAmt1 = 3;
+ MulAmt2 = MulAmt / 3;
+ }
+ if (MulAmt2 &&
+ (isPowerOf2_64(MulAmt2) || MulAmt2 == 3 || MulAmt2 == 5 || MulAmt2 == 9)){
+ DebugLoc DL = N->getDebugLoc();
+
+ if (isPowerOf2_64(MulAmt2) &&
+ !(N->hasOneUse() && N->use_begin()->getOpcode() == ISD::ADD))
+ // If second multiplifer is pow2, issue it first. We want the multiply by
+ // 3, 5, or 9 to be folded into the addressing mode unless the lone use
+ // is an add.
+ std::swap(MulAmt1, MulAmt2);
+
+ SDValue NewMul;
+ if (isPowerOf2_64(MulAmt1))
+ NewMul = DAG.getNode(ISD::SHL, DL, VT, N->getOperand(0),
+ DAG.getConstant(Log2_64(MulAmt1), MVT::i8));
+ else
+ NewMul = DAG.getNode(ISD::MUL, DL, VT, N->getOperand(0),
+ DAG.getConstant(MulAmt1, VT));
+
+ if (isPowerOf2_64(MulAmt2))
+ NewMul = DAG.getNode(ISD::SHL, DL, VT, NewMul,
+ DAG.getConstant(Log2_64(MulAmt2), MVT::i8));
+ else
+ NewMul = DAG.getNode(ISD::MUL, DL, VT, NewMul,
+ DAG.getConstant(MulAmt2, VT));
+
+ // Do not add new nodes to DAG combiner worklist.
+ DCI.CombineTo(N, NewMul, false);
+ }
+ return SDValue();
+}
+
+
/// PerformShiftCombine - Transforms vector shift nodes to use vector shifts
/// when possible.
static SDValue PerformShiftCombine(SDNode* N, SelectionDAG &DAG,
@@ -8668,6 +8738,7 @@
return PerformBuildVectorCombine(N, DAG, DCI, Subtarget, *this);
case ISD::SELECT: return PerformSELECTCombine(N, DAG, Subtarget);
case X86ISD::CMOV: return PerformCMOVCombine(N, DAG, DCI);
+ case ISD::MUL: return PerformMulCombine(N, DAG, DCI);
case ISD::SHL:
case ISD::SRA:
case ISD::SRL: return PerformShiftCombine(N, DAG, Subtarget);