RPCSX/llvm
1
0
Fork 0
mirror of https://github.com/RPCSX/llvm.git synced 2025-03-03 18:37:56 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

[SCEV] Be less conservative when extending bitwidths for computing ranges.

Browse Source 
Summary:
In getRangeForAffineAR we compute ranges for affine exprs E = A + B*C,
where ranges for A, B, and C are known. To avoid overflow, we need to
operate on a bigger bitwidth, and originally we chose 2*x+1 for this
(x being the original bitwidth). However, it is safe to use just 2*x:

A+B*C <= (2^x - 1) + (2^x - 1)*(2^x - 1) =
       =  2^x - 1 + 2^2x - 2^x - 2^x + 1 =
       = 2^2x - 2^x <= 2^2x - 1

Unnecessary extending of bitwidths results in noticeable slowdowns: ranges
perform arithmetic operations using APInt, which are much slower when bitwidths
are bigger than 64.

Reviewers: sanjoy, majnemer, chandlerc

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D27795

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@290211 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Michael Zolotukhin 2016-12-20 23:03:42 +00:00
parent 4138b67a0a
commit 2f3327339d
1 changed files with 6 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
lib/Analysis/ScalarEvolution.cpp
View File

@ -4679,19 +4679,18 @@ ConstantRange ScalarEvolution::getRangeForAffineAR(const SCEV *Start,
MaxBECount = getNoopOrZeroExtend(MaxBECount, Start->getType()); MaxBECount = getNoopOrZeroExtend(MaxBECount, Start->getType());
ConstantRange MaxBECountRange = getUnsignedRange(MaxBECount); ConstantRange MaxBECountRange = getUnsignedRange(MaxBECount);
ConstantRange ZExtMaxBECountRange = ConstantRange ZExtMaxBECountRange = MaxBECountRange.zextOrTrunc(BitWidth * 2);
MaxBECountRange.zextOrTrunc(BitWidth * 2 + 1);
ConstantRange StepSRange = getSignedRange(Step); ConstantRange StepSRange = getSignedRange(Step);
ConstantRange SExtStepSRange = StepSRange.sextOrTrunc(BitWidth * 2 + 1); ConstantRange SExtStepSRange = StepSRange.sextOrTrunc(BitWidth * 2);
ConstantRange StartURange = getUnsignedRange(Start); ConstantRange StartURange = getUnsignedRange(Start);
ConstantRange EndURange = ConstantRange EndURange =
StartURange.add(MaxBECountRange.multiply(StepSRange)); StartURange.add(MaxBECountRange.multiply(StepSRange));
// Check for unsigned overflow. // Check for unsigned overflow.
ConstantRange ZExtStartURange = StartURange.zextOrTrunc(BitWidth * 2 + 1); ConstantRange ZExtStartURange = StartURange.zextOrTrunc(BitWidth * 2);
ConstantRange ZExtEndURange = EndURange.zextOrTrunc(BitWidth * 2 + 1); ConstantRange ZExtEndURange = EndURange.zextOrTrunc(BitWidth * 2);
if (ZExtStartURange.add(ZExtMaxBECountRange.multiply(SExtStepSRange)) == if (ZExtStartURange.add(ZExtMaxBECountRange.multiply(SExtStepSRange)) ==
ZExtEndURange) { ZExtEndURange) {
APInt Min = APIntOps::umin(StartURange.getUnsignedMin(), APInt Min = APIntOps::umin(StartURange.getUnsignedMin(),
@ -4711,8 +4710,8 @@ ConstantRange ScalarEvolution::getRangeForAffineAR(const SCEV *Start,
// Check for signed overflow. This must be done with ConstantRange // Check for signed overflow. This must be done with ConstantRange
// arithmetic because we could be called from within the ScalarEvolution // arithmetic because we could be called from within the ScalarEvolution
// overflow checking code. // overflow checking code.
ConstantRange SExtStartSRange = StartSRange.sextOrTrunc(BitWidth * 2 + 1); ConstantRange SExtStartSRange = StartSRange.sextOrTrunc(BitWidth * 2);
ConstantRange SExtEndSRange = EndSRange.sextOrTrunc(BitWidth * 2 + 1); ConstantRange SExtEndSRange = EndSRange.sextOrTrunc(BitWidth * 2);
if (SExtStartSRange.add(ZExtMaxBECountRange.multiply(SExtStepSRange)) == if (SExtStartSRange.add(ZExtMaxBECountRange.multiply(SExtStepSRange)) ==
SExtEndSRange) { SExtEndSRange) {
APInt Min = APInt Min =