llvm/lib/Analysis
2010-03-25 23:25:28 +00:00
..
IPA Finally land the InvokeInst operand reordering. 2010-03-24 13:21:49 +00:00
AliasAnalysis.cpp
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp
AliasDebugger.cpp
AliasSetTracker.cpp
Analysis.cpp
BasicAliasAnalysis.cpp
CaptureTracking.cpp rename use_const_iterator to const_use_iterator for consistency's sake 2010-03-25 23:06:16 +00:00
CFGPrinter.cpp
CMakeLists.txt
ConstantFolding.cpp FP16 constfolding 2010-03-19 00:36:35 +00:00
DbgInfoPrinter.cpp
DebugInfo.cpp
DomPrinter.cpp
InlineCost.cpp Reapply r99451 with a fix to move the NoInline check to the cost functions 2010-03-25 04:49:10 +00:00
InstCount.cpp
InstructionSimplify.cpp
Interval.cpp
IntervalPartition.cpp
IVUsers.cpp
LazyValueInfo.cpp
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
LiveValues.cpp rename use_const_iterator to const_use_iterator for consistency's sake 2010-03-25 23:06:16 +00:00
LoopDependenceAnalysis.cpp
LoopInfo.cpp
LoopPass.cpp
Makefile
MemoryBuiltins.cpp rename use_const_iterator to const_use_iterator for consistency's sake 2010-03-25 23:06:16 +00:00
MemoryDependenceAnalysis.cpp
PHITransAddr.cpp
PointerTracking.cpp
PostDominators.cpp
ProfileEstimatorPass.cpp rename pred_const_iterator to const_pred_iterator for consistency's sake 2010-03-25 23:25:28 +00:00
ProfileInfo.cpp rename pred_const_iterator to const_pred_iterator for consistency's sake 2010-03-25 23:25:28 +00:00
ProfileInfoLoader.cpp
ProfileInfoLoaderPass.cpp rename pred_const_iterator to const_pred_iterator for consistency's sake 2010-03-25 23:25:28 +00:00
ProfileVerifierPass.cpp rename pred_const_iterator to const_pred_iterator for consistency's sake 2010-03-25 23:25:28 +00:00
README.txt
ScalarEvolution.cpp
ScalarEvolutionAliasAnalysis.cpp
ScalarEvolutionExpander.cpp Don't back past debug info intrinsics; SCEVExpander's strategy 2010-03-23 21:53:22 +00:00
SparsePropagation.cpp
Trace.cpp
ValueTracking.cpp

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.

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