llvm-6502/lib/Analysis
2009-10-09 16:35:06 +00:00
..
IPA Extend the StartPassTimer and StopPassTimer functions so that the 2009-09-28 00:07:05 +00:00
AliasAnalysis.cpp Global Aliases are not identifiable objects. 2009-08-27 17:52:56 +00:00
AliasAnalysisCounter.cpp Do not try to override non-virtual methods, especially 2009-09-06 20:02:00 +00:00
AliasAnalysisEvaluator.cpp Fix a missing newline, now that Value's operator<< doesn't add one of its own. 2009-08-26 14:34:12 +00:00
AliasDebugger.cpp
AliasSetTracker.cpp eliminate the std::ostream form of WriteAsOperand and update clients. 2009-08-23 05:17:37 +00:00
Analysis.cpp
BasicAliasAnalysis.cpp Enhance analysis passes so that they apply the same analysis to malloc calls as to MallocInst. 2009-09-18 21:34:51 +00:00
CaptureTracking.cpp Push LLVMContexts through the IntegerType APIs. 2009-08-13 21:58:54 +00:00
CFGPrinter.cpp The flag "--dot-cfg-only" is at the moment equivalent to the flag "--dot-cfg". 2009-09-19 11:25:44 +00:00
CMakeLists.txt Update CMake build yet again after a source file was removed 2009-10-08 21:24:34 +00:00
ConstantFolding.cpp Extend ConstantFolding to understand signed overflow variants 2009-10-05 22:53:52 +00:00
DbgInfoPrinter.cpp Remove std::string uses from DebugInfo interface. 2009-09-29 18:40:58 +00:00
DebugInfo.cpp It's possible for a global variable to be optimized out of a metadata object. So 2009-10-08 20:52:51 +00:00
InstCount.cpp Change Pass::print to take a raw ostream instead of std::ostream, 2009-08-23 06:03:38 +00:00
Interval.cpp Change Pass::print to take a raw ostream instead of std::ostream, 2009-08-23 06:03:38 +00:00
IntervalPartition.cpp Change Pass::print to take a raw ostream instead of std::ostream, 2009-08-23 06:03:38 +00:00
IVUsers.cpp Instead of testing whether an instruction dominates the loop preheader, 2009-09-27 15:30:00 +00:00
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
LiveValues.cpp
LoopDependenceAnalysis.cpp remove a few dead insertion methods. 2009-08-24 02:39:26 +00:00
LoopInfo.cpp Move the dominator verification code out of special code embedded within 2009-09-28 00:27:48 +00:00
LoopPass.cpp Fix this debug output to handle the case where the loop has been deleted. 2009-09-28 15:40:01 +00:00
Makefile
MallocHelper.cpp Add PR to this FIXME, looks like I didn't commit this change after all. 2009-10-07 09:22:55 +00:00
MemoryDependenceAnalysis.cpp Revert r82404, it is causing a bootstrap miscompile. This is very very 2009-09-20 22:44:26 +00:00
PointerTracking.cpp Introduce and use convenience methods for getting pointer types 2009-10-06 15:40:36 +00:00
PostDominators.cpp Change Pass::print to take a raw ostream instead of std::ostream, 2009-08-23 06:03:38 +00:00
ProfileEstimatorPass.cpp Make ProfileEstimator even more robust on general CFGs. 2009-09-11 08:39:33 +00:00
ProfileInfo.cpp Add the first functions for updating ProfileInfo. 2009-09-09 17:52:57 +00:00
ProfileInfoLoader.cpp Reapplied r81355 with the problems fixed. 2009-09-16 11:35:50 +00:00
ProfileInfoLoaderPass.cpp Reapplied r81355 with the problems fixed. 2009-09-16 11:35:50 +00:00
ProfileVerifierPass.cpp Bugfix. Sorry. 2009-09-11 08:43:15 +00:00
README.txt Create a README.txt for lib/Analysis, and add an entry. 2009-09-28 18:38:53 +00:00
ScalarEvolution.cpp Revert r83606 and add comments explaining why it isn't safe. 2009-10-09 16:35:06 +00:00
ScalarEvolutionAliasAnalysis.cpp Add some comments. 2009-08-29 23:36:57 +00:00
ScalarEvolutionExpander.cpp Introduce and use convenience methods for getting pointer types 2009-10-06 15:40:36 +00:00
SparsePropagation.cpp Add a comment explaining why you would ever want to do this. 2009-09-19 19:00:06 +00:00
Trace.cpp eliminate the "Value" printing methods that print to a std::ostream. 2009-08-23 04:37:46 +00:00
ValueTracking.cpp Handle sqrt in CannotBeNegativeZero. absf and absl 2009-09-25 20:54:50 +00:00

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.

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