llvm-mirror

mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-12-23 12:08:25 +00:00

History

Dan Gohman 198fbdadcf Revert the rule that considers comparisons between two pointers in the same object to be a non-capture; Duncan pointed out a way that such a comparison could be a capture. Make the rule that considers a comparison against null more specific, and only consider noalias return values compared against null. This still supports test/Transforms/GVN/nonescaping-malloc.ll, and is not susceptible to the problem Duncan pointed out with noalias arguments. llvm-svn: 89468		2009-11-20 17:50:21 +00:00
..
IPA	Fix DenseMap iterator constness.	2009-11-10 01:02:17 +00:00
AliasAnalysis.cpp
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp
AliasDebugger.cpp
AliasSetTracker.cpp	remove unneeded checks of isFreeCall	2009-11-03 05:35:19 +00:00
Analysis.cpp
BasicAliasAnalysis.cpp	Extend CaptureTracking to indicate when a value is never stored, even	2009-11-19 21:57:48 +00:00
CaptureTracking.cpp	Revert the rule that considers comparisons between two pointers in the	2009-11-20 17:50:21 +00:00
CFGPrinter.cpp
CMakeLists.txt	Stub out a new lazy value info pass, which will eventually	2009-11-11 00:22:30 +00:00
ConstantFolding.cpp	fix ConstantFoldCompareInstOperands to take the LHS/RHS as	2009-11-09 23:06:58 +00:00
DbgInfoPrinter.cpp
DebugInfo.cpp	Remove dead code.	2009-11-17 22:39:08 +00:00
DomPrinter.cpp
InlineCost.cpp	Add a missing closing parenthesis, and tweak to fit in 80	2009-11-01 19:12:43 +00:00
InstCount.cpp
InstructionSimplify.cpp	move some generally useful functions out of jump threading	2009-11-10 22:26:15 +00:00
Interval.cpp
IntervalPartition.cpp
IVUsers.cpp	Teach IVUsers to keep things simpler and track loop-invariant strides only	2009-11-19 02:05:44 +00:00
LazyValueInfo.cpp	typo spotted by duncan.	2009-11-16 03:51:42 +00:00
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
LiveValues.cpp	remove redundant foward declaration. This function is already in	2009-11-11 00:21:21 +00:00
LoopDependenceAnalysis.cpp
LoopInfo.cpp	Teach getSmallConstantTripMultiple about Shl operators.	2009-11-20 01:09:34 +00:00
LoopPass.cpp
Makefile
MemoryBuiltins.cpp	Simplify ComputeMultiple so that it doesn't depend on TargetData.	2009-11-18 00:58:27 +00:00
MemoryDependenceAnalysis.cpp
PointerTracking.cpp	remove a bunch of extraneous LLVMContext arguments	2009-11-06 04:27:31 +00:00
PostDominators.cpp
ProfileEstimatorPass.cpp
ProfileInfo.cpp
ProfileInfoLoader.cpp
ProfileInfoLoaderPass.cpp	Apply fix for PR5135, Credit to Andreas Neustifter.	2009-11-02 02:55:39 +00:00
ProfileVerifierPass.cpp	Apply fix for PR5135, Credit to Andreas Neustifter.	2009-11-02 02:55:39 +00:00
README.txt
ScalarEvolution.cpp	Pass the (optional) TargetData object to ConstantFoldInstOperands	2009-11-09 23:34:17 +00:00
ScalarEvolutionAliasAnalysis.cpp	Make ScalarEvolutionAliasAnalysis slightly more aggressive, by making an	2009-10-31 14:32:25 +00:00
ScalarEvolutionExpander.cpp
SparsePropagation.cpp
Trace.cpp
ValueTracking.cpp	Simplify ComputeMultiple so that it doesn't depend on TargetData.	2009-11-18 00:58:27 +00:00

README.txt

Analysis Opportunities:

//===---------------------------------------------------------------------===//

In test/Transforms/LoopStrengthReduce/quadradic-exit-value.ll, the
ScalarEvolution expression for %r is this:

  {1,+,3,+,2}<loop>

Outside the loop, this could be evaluated simply as (%n * %n), however
ScalarEvolution currently evaluates it as

  (-2 + (2 * (trunc i65 (((zext i64 (-2 + %n) to i65) * (zext i64 (-1 + %n) to i65)) /u 2) to i64)) + (3 * %n))

In addition to being much more complicated, it involves i65 arithmetic,
which is very inefficient when expanded into code.

//===---------------------------------------------------------------------===//