llvm-6502

mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-13 20:32:21 +00:00

History

Bob Wilson 28f872f8a1 Clean up handling of always-inline functions in the inliner. This patch moves the isInlineViable function from the InlineAlways pass into the InlineCostAnalyzer and then changes the InlineCost computation to use that simple check for always-inline functions. All the special-case checks for AlwaysInline in the CallAnalyzer can then go away. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168300 91177308-0d34-0410-b5e6-96231b3b80d8		2012-11-19 07:04:35 +00:00
..
IPA	Add doInitialization and doFinalization methods to ModulePass's, to allow them to be re-initialized and reused on multiple Module's.	2012-11-15 00:14:15 +00:00
AliasAnalysis.cpp	Use the attribute enums to query if a parameter has an attribute.	2012-10-09 21:38:14 +00:00
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp
AliasDebugger.cpp
AliasSetTracker.cpp	Move TargetData to DataLayout.	2012-10-08 16:38:25 +00:00
Analysis.cpp	Add a cost model analysis that allows us to estimate the cost of IR-level instructions.	2012-11-02 21:48:17 +00:00
BasicAliasAnalysis.cpp	Phi speculation improvement for BasicAA	2012-11-17 02:33:15 +00:00
BlockFrequencyInfo.cpp
BranchProbabilityInfo.cpp	BranchProb: modify the definition of an edge in BranchProbabilityInfo to handle	2012-08-24 18:14:27 +00:00
CaptureTracking.cpp	Give CaptureTracker::shouldExplore a base implementation. Most users want to do	2012-10-08 22:12:48 +00:00
CFGPrinter.cpp
CMakeLists.txt	Add a cost model analysis that allows us to estimate the cost of IR-level instructions.	2012-11-02 21:48:17 +00:00
CodeMetrics.cpp	Revert the majority of the next patch in the address space series:	2012-11-01 09:14:31 +00:00
ConstantFolding.cpp	llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware.	2012-11-08 20:34:25 +00:00
CostModel.cpp	CostModel: add support for Vector Insert and Extract.	2012-11-02 22:31:56 +00:00
DbgInfoPrinter.cpp	Move lib/Analysis/DebugInfo.cpp to lib/VMCore/DebugInfo.cpp and	2012-06-28 00:05:13 +00:00
DependenceAnalysis.cpp	DependenceAnalysis: Print all dependency pairs when dumping. Update all testcases.	2012-11-13 12:12:02 +00:00
DominanceFrontier.cpp	Release build: guard dump functions with	2012-09-12 05:06:18 +00:00
DomPrinter.cpp
InlineCost.cpp	Clean up handling of always-inline functions in the inliner.	2012-11-19 07:04:35 +00:00
InstCount.cpp
InstructionSimplify.cpp	Make this easier to understand, as suggested by Chandler.	2012-11-16 20:53:08 +00:00
Interval.cpp	Revert 'Fix a typo 'iff' => 'if''. iff is an abreviation of if and only if. See: http://en.wikipedia.org/wiki/If_and_only_if Commit 164767	2012-09-27 10:14:43 +00:00
IntervalPartition.cpp
IVUsers.cpp	Move TargetData to DataLayout.	2012-10-08 16:38:25 +00:00
LazyValueInfo.cpp	Hoist out some work done inside a loop doing a linear scan over all	2012-10-26 04:43:47 +00:00
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
Lint.cpp	Revert the series of commits starting with r166578 which introduced the	2012-11-01 08:07:29 +00:00
LLVMBuild.txt
Loads.cpp	Move TargetData to DataLayout.	2012-10-08 16:38:25 +00:00
LoopInfo.cpp	Release build: guard dump functions with	2012-09-12 05:06:18 +00:00
LoopPass.cpp	Enable the new LoopInfo algorithm by default.	2012-06-26 04:11:38 +00:00
Makefile
MemDepPrinter.cpp
MemoryBuiltins.cpp	Revert the series of commits starting with r166578 which introduced the	2012-11-01 08:07:29 +00:00
MemoryDependenceAnalysis.cpp	Ignore unreachable blocks when doing memory dependence analysis on non-local	2012-10-23 18:37:11 +00:00
ModuleDebugInfoPrinter.cpp	Move lib/Analysis/DebugInfo.cpp to lib/VMCore/DebugInfo.cpp and	2012-06-28 00:05:13 +00:00
NoAliasAnalysis.cpp	Move TargetData to DataLayout.	2012-10-08 16:38:25 +00:00
PathNumbering.cpp	Move llvm/Support/TypeBuilder.h -> llvm/TypeBuilder.h. This completes	2012-07-15 23:45:24 +00:00
PathProfileInfo.cpp
PathProfileVerifier.cpp
PHITransAddr.cpp	Release build: guard dump functions with	2012-09-12 05:06:18 +00:00
PostDominators.cpp
ProfileDataLoader.cpp	Remove code to saturate profile counts.	2012-10-29 17:27:39 +00:00
ProfileDataLoaderPass.cpp	Cleanups due to feedback. No functionality change. Patch by Alistair.	2012-08-31 05:18:31 +00:00
ProfileEstimatorPass.cpp	Fix floating-point divide by zero, in a case where the value was not going to be used anyway.	2012-08-24 00:31:45 +00:00
ProfileInfo.cpp	Profile: set branch weight metadata with data generated from profiling.	2012-08-28 22:21:25 +00:00
ProfileInfoLoader.cpp	Remove unused private member variables uncovered by the recent changes to clang's -Wunused-private-field.	2012-07-20 22:05:57 +00:00
ProfileInfoLoaderPass.cpp
ProfileVerifierPass.cpp
README.txt
RegionInfo.cpp	Release build: guard dump functions with	2012-09-12 05:06:18 +00:00
RegionPass.cpp	Remove the the block_node_iterator of Region, replace it by the block_iterator.	2012-08-27 13:49:24 +00:00
RegionPrinter.cpp
ScalarEvolution.cpp	Revert the series of commits starting with r166578 which introduced the	2012-11-01 08:07:29 +00:00
ScalarEvolutionAliasAnalysis.cpp
ScalarEvolutionExpander.cpp	Revert the series of commits starting with r166578 which introduced the	2012-11-01 08:07:29 +00:00
ScalarEvolutionNormalization.cpp
SparsePropagation.cpp
Trace.cpp	Release build: guard dump functions with	2012-09-12 05:06:18 +00:00
TypeBasedAliasAnalysis.cpp
ValueTracking.cpp	Revert the majority of the next patch in the address space series:	2012-11-01 09:14:31 +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.

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

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))

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