llvm-6502

mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-11-10 17:07:06 +00:00

History

Andrew Trick 762797d1af An algorithm for incrementally updating LoopInfo within a LoopPassManager. The incremental update should be extremely cheap in most cases and can be used in places where it's not feasible to regenerate the entire loop forest. - "Unloop" is a node in the loop tree whose last backedge has been removed. - Perform reverse dataflow on the block inside Unloop to propagate the nearest loop from the block's successors. - For reducible CFG, each block in unloop is visited exactly once. This is because unloop no longer has a backedge and blocks within subloops don't change parents. - Immediate subloops are summarized by the nearest loop reachable from their exits or exits within nested subloops. - At completion the unloop blocks each have a new parent loop, and each immediate subloop has a new parent. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@136844 91177308-0d34-0410-b5e6-96231b3b80d8		2011-08-03 23:50:25 +00:00
..
IPA	Rewrite the CMake build to use explicit dependencies between libraries,	2011-07-29 00:14:25 +00:00
AliasAnalysis.cpp
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp
AliasDebugger.cpp
AliasSetTracker.cpp	Minor simplification.	2011-07-27 01:02:25 +00:00
Analysis.cpp	Rename BlockFrequency to BlockFrequencyInfo and MachineBlockFrequency to	2011-07-25 19:25:40 +00:00
BasicAliasAnalysis.cpp	Explicitly cast narrowing conversions inside {}s that will become errors in	2011-07-27 06:22:51 +00:00
BlockFrequencyInfo.cpp	Add more constantness in BlockFrequencyInfo.	2011-08-03 21:30:57 +00:00
BranchProbabilityInfo.cpp	Change SmallVector to SmallPtrSet in BranchProbabilityInfo. Handle cases where	2011-08-01 19:16:26 +00:00
CaptureTracking.cpp
CFGPrinter.cpp
CMakeLists.txt	Rewrite the CMake build to use explicit dependencies between libraries,	2011-07-29 00:14:25 +00:00
ConstantFolding.cpp	Convert GetElementPtrInst to use ArrayRef.	2011-07-25 09:48:08 +00:00
DbgInfoPrinter.cpp
DebugInfo.cpp
DIBuilder.cpp
DominanceFrontier.cpp
DomPrinter.cpp
InlineCost.cpp
InstCount.cpp
InstructionSimplify.cpp	Add helper function for getting true/false constants in a uniform	2011-07-26 15:03:53 +00:00
Interval.cpp
IntervalPartition.cpp
IVUsers.cpp
LazyValueInfo.cpp
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
Lint.cpp
Loads.cpp
LoopDependenceAnalysis.cpp
LoopInfo.cpp	An algorithm for incrementally updating LoopInfo within a	2011-08-03 23:50:25 +00:00
LoopPass.cpp	An algorithm for incrementally updating LoopInfo within a	2011-08-03 23:50:25 +00:00
Makefile
MemDepPrinter.cpp
MemoryBuiltins.cpp
MemoryDependenceAnalysis.cpp
ModuleDebugInfoPrinter.cpp
NoAliasAnalysis.cpp
PathNumbering.cpp	Add this back in for now. There are still a few passes which create unwind instructions at the moment.	2011-08-03 01:07:57 +00:00
PathProfileInfo.cpp
PathProfileVerifier.cpp
PHITransAddr.cpp	Shorten some expressions by using ArrayRef::slice().	2011-07-25 15:13:01 +00:00
PostDominators.cpp
ProfileEstimatorPass.cpp
ProfileInfo.cpp
ProfileInfoLoader.cpp
ProfileInfoLoaderPass.cpp
ProfileVerifierPass.cpp
README.txt
RegionInfo.cpp
RegionPass.cpp
RegionPrinter.cpp
ScalarEvolution.cpp	Use consistent terminology for loop exit/exiting blocks. Name change only.	2011-08-02 04:23:35 +00:00
ScalarEvolutionAliasAnalysis.cpp
ScalarEvolutionExpander.cpp	Convert IRBuilder::CreateGEP and IRBuilder::CreateInBoundsGEP to use	2011-07-22 08:16:57 +00:00
ScalarEvolutionNormalization.cpp
SparsePropagation.cpp
Trace.cpp
TypeBasedAliasAnalysis.cpp
ValueTracking.cpp

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.

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

In formatValue in test/CodeGen/X86/lsr-delayed-fold.ll,

ScalarEvolution is forming this expression:

((trunc i64 (-1 * %arg5) to i32) + (trunc i64 %arg5 to i32) + (-1 * (trunc i64 undef to i32)))

This could be folded to

(-1 * (trunc i64 undef to i32))

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