[InstCombine] Expand the simplification of pow(x, 0.5) to sqrt(x)
Expand the number of cases when `pow(x, 0.5)` is simplified into `sqrt(x)`
by considering the math semantics with more granularity.
Differential revision: https://reviews.llvm.org/D50036
llvm-svn: 339887
diff --git a/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp b/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
index a2e5f8d..a0f1f81 100644
--- a/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
+++ b/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
@@ -1181,12 +1181,9 @@
/// Use square root in place of pow(x, +/-0.5).
Value *LibCallSimplifier::replacePowWithSqrt(CallInst *Pow, IRBuilder<> &B) {
- // TODO: There is some subset of 'fast' under which these transforms should
- // be allowed.
- if (!Pow->isFast())
- return nullptr;
-
Value *Sqrt, *Base = Pow->getArgOperand(0), *Expo = Pow->getArgOperand(1);
+ AttributeList Attrs = Pow->getCalledFunction()->getAttributes();
+ Module *Mod = Pow->getModule();
Type *Ty = Pow->getType();
const APFloat *ExpoF;
@@ -1198,18 +1195,32 @@
if (Pow->hasFnAttr(Attribute::ReadNone)) {
Function *SqrtFn = Intrinsic::getDeclaration(Pow->getModule(),
Intrinsic::sqrt, Ty);
- Sqrt = B.CreateCall(SqrtFn, Base);
+ Sqrt = B.CreateCall(SqrtFn, Base, "sqrt");
}
// Otherwise, use the libcall for sqrt().
else if (hasUnaryFloatFn(TLI, Ty, LibFunc_sqrt, LibFunc_sqrtf, LibFunc_sqrtl))
// TODO: We also should check that the target can in fact lower the sqrt()
// libcall. We currently have no way to ask this question, so we ask if
// the target has a sqrt() libcall, which is not exactly the same.
- Sqrt = emitUnaryFloatFnCall(Base, TLI->getName(LibFunc_sqrt), B,
- Pow->getCalledFunction()->getAttributes());
+ Sqrt = emitUnaryFloatFnCall(Base, TLI->getName(LibFunc_sqrt), B, Attrs);
else
return nullptr;
+ // Handle signed zero base by expanding to fabs(sqrt(x)).
+ if (!Pow->hasNoSignedZeros()) {
+ Function *FAbsFn = Intrinsic::getDeclaration(Mod, Intrinsic::fabs, Ty);
+ Sqrt = B.CreateCall(FAbsFn, Sqrt, "abs");
+ }
+
+ // Handle non finite base by expanding to
+ // (x == -infinity ? +infinity : sqrt(x)).
+ if (!Pow->hasNoInfs()) {
+ Value *PosInf = ConstantFP::getInfinity(Ty),
+ *NegInf = ConstantFP::getInfinity(Ty, true);
+ Value *FCmp = B.CreateFCmpOEQ(Base, NegInf, "isinf");
+ Sqrt = B.CreateSelect(FCmp, PosInf, Sqrt);
+ }
+
// If the exponent is negative, then get the reciprocal.
if (ExpoF->isNegative())
Sqrt = B.CreateFDiv(ConstantFP::get(Ty, 1.0), Sqrt, "reciprocal");
@@ -1265,7 +1276,7 @@
// We enable these only with fast-math. Besides rounding differences, the
// transformation changes overflow and underflow behavior quite dramatically.
// Example: x = 1000, y = 0.001.
- // pow(exp(x), y) = pow(inf, 0.001) = inf, whereas exp(x*y) = exp(1).
+ // pow(exp(x), y) = pow(inf, 0.001) = inf, whereas exp(x * y) = exp(1).
auto *BaseFn = dyn_cast<CallInst>(Base);
if (BaseFn && BaseFn->isFast() && Pow->isFast()) {
LibFunc LibFn;
@@ -1299,28 +1310,6 @@
if (Value *Sqrt = replacePowWithSqrt(Pow, B))
return Sqrt;
- // FIXME: Correct the transforms and pull this into replacePowWithSqrt().
- ConstantFP *ExpoC = dyn_cast<ConstantFP>(Expo);
- if (ExpoC && ExpoC->isExactlyValue(0.5) &&
- hasUnaryFloatFn(TLI, Ty, LibFunc_sqrt, LibFunc_sqrtf, LibFunc_sqrtl)) {
- // Expand pow(x, 0.5) to (x == -infinity ? +infinity : fabs(sqrt(x))).
- // This is faster than calling pow(), and still handles -0.0 and
- // negative infinity correctly.
- // TODO: In finite-only mode, this could be just fabs(sqrt(x)).
- Value *PosInf = ConstantFP::getInfinity(Ty);
- Value *NegInf = ConstantFP::getInfinity(Ty, true);
-
- // TODO: As above, we should lower to the sqrt() intrinsic if the pow() is
- // an intrinsic, to match errno semantics.
- Value *Sqrt = emitUnaryFloatFnCall(Base, TLI->getName(LibFunc_sqrt),
- B, Attrs);
- Function *FAbsFn = Intrinsic::getDeclaration(Module, Intrinsic::fabs, Ty);
- Value *FAbs = B.CreateCall(FAbsFn, Sqrt, "abs");
- Value *FCmp = B.CreateFCmpOEQ(Base, NegInf, "isinf");
- Sqrt = B.CreateSelect(FCmp, PosInf, FAbs);
- return Sqrt;
- }
-
// pow(x, n) -> x * x * x * ...
const APFloat *ExpoF;
if (Pow->isFast() && match(Expo, m_APFloat(ExpoF))) {