llvm-6502

mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-17 18:31:04 +00:00

History

Arnold Schwaighofer 6bf4f67641 CostModel: Add parameter to instruction cost to further classify operand values On certain architectures we can support efficient vectorized version of instructions if the operand value is uniform (splat) or a constant scalar. An example of this is a vector shift on x86. We can efficiently support for (i = 0 ; i < ; i += 4) w[0:3] = v[0:3] << <2, 2, 2, 2> but not for (i = 0; i < ; i += 4) w[0:3] = v[0:3] << x[0:3] This patch adds a parameter to getArithmeticInstrCost to further qualify operand values as uniform or uniform constant. Targets can then choose to return a different cost for instructions with such operand values. A follow-up commit will test this feature on x86. radar://13576547 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178807 91177308-0d34-0410-b5e6-96231b3b80d8		2013-04-04 23:26:21 +00:00
..
IPA	ArrayRef ca accept one element. Simplify code a little bit, also it matches now	2013-03-07 20:01:19 +00:00
AliasAnalysis.cpp
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp	Support in AAEvaluator to print alias queries of loads/stores with TBAA tags.	2013-03-22 22:34:41 +00:00
AliasDebugger.cpp
AliasSetTracker.cpp
Analysis.cpp	Remove -print-dbginfo as it is unused & bitrotten.	2013-03-08 18:17:46 +00:00
BasicAliasAnalysis.cpp	BasicAA: Only query twice if the result of the more general query was MayAlias	2013-03-26 18:07:53 +00:00
BlockFrequencyInfo.cpp
BranchProbabilityInfo.cpp
CaptureTracking.cpp	Remove unneeded #includes. Use forward declarations instead.	2013-03-10 00:34:01 +00:00
CFGPrinter.cpp
CMakeLists.txt	Remove -print-dbginfo as it is unused & bitrotten.	2013-03-08 18:17:46 +00:00
CodeMetrics.cpp
ConstantFolding.cpp	Constant fold vector bitcasts of halves similarly to how floats and doubles are folded. Test case included.	2013-02-26 22:51:07 +00:00
CostModel.cpp	CostModel: Add parameter to instruction cost to further classify operand values	2013-04-04 23:26:21 +00:00
DependenceAnalysis.cpp
DominanceFrontier.cpp
DomPrinter.cpp
InstCount.cpp
InstructionSimplify.cpp
Interval.cpp
IntervalPartition.cpp
IVUsers.cpp
LazyValueInfo.cpp
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp
Lint.cpp
LLVMBuild.txt
Loads.cpp
LoopInfo.cpp
LoopPass.cpp
Makefile
MemDepPrinter.cpp
MemoryBuiltins.cpp	Early exit from getAllocationData() and isFreeCall() for intrinsics.	2013-03-08 21:15:00 +00:00
MemoryDependenceAnalysis.cpp	Fix loop style	2013-03-29 18:48:42 +00:00
ModuleDebugInfoPrinter.cpp
NoAliasAnalysis.cpp
PathNumbering.cpp
PathProfileInfo.cpp
PathProfileVerifier.cpp	Build fixes for STLPort + GCC	2013-03-29 18:48:45 +00:00
PHITransAddr.cpp
PostDominators.cpp
ProfileDataLoader.cpp
ProfileDataLoaderPass.cpp
ProfileEstimatorPass.cpp
ProfileInfo.cpp	Build fixes for STLPort + GCC	2013-03-29 18:48:45 +00:00
ProfileInfoLoader.cpp
ProfileInfoLoaderPass.cpp
ProfileVerifierPass.cpp
PtrUseVisitor.cpp
README.txt
RegionInfo.cpp
RegionPass.cpp
RegionPrinter.cpp
ScalarEvolution.cpp	Fix SCEV forgetMemoizedResults should search and destroy backedge exprs.	2013-03-26 03:14:53 +00:00
ScalarEvolutionAliasAnalysis.cpp
ScalarEvolutionExpander.cpp
ScalarEvolutionNormalization.cpp
SparsePropagation.cpp
TargetTransformInfo.cpp	CostModel: Add parameter to instruction cost to further classify operand values	2013-04-04 23:26:21 +00:00
Trace.cpp
TypeBasedAliasAnalysis.cpp
ValueTracking.cpp	Check whether a pointer is non-null (isKnownNonNull) in isKnownNonZero.	2013-03-18 21:23:25 +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))

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