RPCSX/llvm
1
0
Fork 0
mirror of https://github.com/RPCSX/llvm.git synced 2024-11-24 20:29:53 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

[LibCallSimplifier] don't allow sqrt transform unless all ops are unsafe

Browse Source 
Fix the FIXME added with:
http://reviews.llvm.org/rL257400



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257404 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Sanjay Patel 2016-01-11 22:50:36 +00:00
parent 2ab07ac8fe
commit be014a6bcf
2 changed files with 17 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lib/Transforms/Utils/SimplifyLibCalls.cpp
View File

@ -1422,10 +1422,10 @@ Value *LibCallSimplifier::optimizeSqrt(CallInst *CI, IRBuilder<> &B) {
// variations of this pattern because instcombine's visitFMUL and/or the
// reassociation pass should give us this form.
Value *OtherMul0, *OtherMul1;
// FIXME: This multiply must be unsafe to allow this transform.
if (match(Op0, m_FMul(m_Value(OtherMul0), m_Value(OtherMul1)))) {
// Pattern: sqrt((x * y) * z)
if (OtherMul0 == OtherMul1) {
if (OtherMul0 == OtherMul1 &&
cast<Instruction>(Op0)->hasUnsafeAlgebra()) {
// Matched: sqrt((x * x) * z)
RepeatOp = OtherMul0;
OtherOp = Op1;

15
test/Transforms/InstCombine/fast-math.ll
View File

@ -649,6 +649,21 @@ define double @sqrt_intrinsic_three_args6(double %x, double %y) {
; CHECK-NEXT: ret double %1
}
; If any operation is not 'fast', we can't simplify.
define double @sqrt_intrinsic_not_so_fast(double %x, double %y) {
%mul = fmul double %x, %x
%mul2 = fmul fast double %mul, %y
%sqrt = call fast double @llvm.sqrt.f64(double %mul2)
ret double %sqrt
; CHECK-LABEL: sqrt_intrinsic_not_so_fast(
; CHECK-NEXT: %mul = fmul double %x, %x
; CHECK-NEXT: %mul2 = fmul fast double %mul, %y
; CHECK-NEXT: %sqrt = call fast double @llvm.sqrt.f64(double %mul2)
; CHECK-NEXT: ret double %sqrt
}
define double @sqrt_intrinsic_arg_4th(double %x) {
%mul = fmul fast double %x, %x
%mul2 = fmul fast double %mul, %mul