Reapply r68211, with the miscompilations it caused fixed.
[oota-llvm.git] / lib / Transforms / Scalar /
2009-04-01 Owen AndersonReapply r68211, with the miscompilations it caused...
2009-04-01 Dan GohmanRevert r68172. It caused regressions in
2009-04-01 Owen AndersonEnhance GVN to propagate simple conditionals. This...
2009-03-31 Chris LattnerMake the key of ValueRankMap an AssertingVH, so that...
2009-03-31 Evan ChengThrottle back "fold select into operand" transformation...
2009-03-30 Devang PatelLoop Index Split can eliminate a loop if it can determi...
2009-03-27 Duncan SandsRevert r67798: it breaks llvm-gcc bootstrap on x86...
2009-03-27 Dale JohannesenOne more place to skip debug info.
2009-03-26 Devang PatelWhile hoisting an instruction, update alias info set...
2009-03-26 Dale JohannesenSkip debug info one more place. (This one gets
2009-03-25 Devang PatelBefore deleting a basic block, give other loop passes...
2009-03-25 Chris LattnerFix PR3874 by restoring a condition I removed, but...
2009-03-24 Chris Lattneroops, I intended to remove this, not comment it out...
2009-03-24 Chris Lattnercanonicalize inttoptr and ptrtoint instructions which...
2009-03-24 Chris Lattnertwo changes:
2009-03-21 Duncan SandsFactorize out a concept - no functionality change.
2009-03-20 Chris LattnerFix instcombine to not introduce undefined shifts when...
2009-03-20 Duncan SandsDon't load values out of global constants with weak
2009-03-19 Dale JohannesenThis pass keeps a map of Instructions to Rank numbers,
2009-03-19 Nick LewyckyRemove strange extra semicolons.
2009-03-18 Chris LattnerFix PR3826 - InstComb assert with vector shift, by...
2009-03-18 Zhou ShengExplicitly check for StoreInst, do not lose the chance...
2009-03-18 Zhou ShengRevert my previous change on Local.cpp, instead, fix...
2009-03-17 Chris LattnerLSR shouldn't ever try to hack on integer IV's larger...
2009-03-17 Chris LattnerRemove a condition which is always true.
2009-03-17 Dale JohannesenFix a debug info dependency in jump threading.
2009-03-13 Evan ChengFix PR3784: If the source of a phi comes from a bb...
2009-03-13 Bill WendlingRevert r66920. It was causing failures in the self...
2009-03-13 Dale JohannesenOne more place where debug info affects codegen.
2009-03-13 Gabor GreifSecond installment of "BasicBlock operands to the back"
2009-03-13 Bill WendlingOops...I committed too much.
2009-03-13 Bill WendlingTemporarily XFAIL this test.
2009-03-12 Duncan SandsRevert commit 66140 since it caused several failures
2009-03-11 Dale JohannesenSkip interleaved debug info when fast-forwarding through
2009-03-10 Dale JohannesenRemoving a dead debug intrinsic shouldn't trigger
2009-03-10 John CriswellDo not attempt to do parial redundancy elimination...
2009-03-09 Dan GohmanDon't record the increment instruction; just recompute...
2009-03-09 Dan GohmanFix a few more places where induction variable types...
2009-03-09 Dan GohmanUse ReplacedTy instead of recomputing the same value.
2009-03-09 Dan GohmanUse LoopInfo's getLoopLatch() instead of doing what...
2009-03-09 Dan GohmanDon't use an induction variable type as a memory access...
2009-03-09 Dan GohmanFactor out the code that determines the memory access...
2009-03-09 Dan GohmanMove the sorting of the StrideOrder array earlier so...
2009-03-09 Dan GohmanDelete the isOnlyStride argument, which is unused.
2009-03-09 Dan GohmanTidy some LSR debug output: announce the loop it's...
2009-03-09 Chris Lattnerreimplement AliasSetTracker in terms of DenseMap instea...
2009-03-08 Chris Lattnerteach SROA to handle promoting vector allocas with...
2009-03-08 Chris LattnerEnhance SROA to "promote to scalar" allocas which are
2009-03-08 Chris Lattnerchange the MemIntrinsic get/setAlignment method to...
2009-03-08 Chris LattnerIntroduce a new MemTransferInst pseudo class, which...
2009-03-06 Chris Lattneradd a bunch more passes to the C bindings (PR3734)...
2009-03-06 Devang PatelWhile converting an aggregate to scalare, ignore and...
2009-03-06 Chris Lattnerthis wasn't intended to go in.
2009-03-06 Chris LattnerChange various llvm utilities to use PrettyStackTracePr...
2009-03-06 Devang PatelDo not count DbgInfoIntrinsic while estimating loop...
2009-03-06 Devang PatelSkip DbgInfoIntrinsic.
2009-03-06 Dale JohannesenDon't assign rank numbers to debug intrinsic "calls".
2009-03-06 Evan ChengSRThreshold is meant to be inclusive.
2009-03-05 Evan ChengDo not split edges to EH landing pads. It will cause...
2009-03-05 Dale JohannesenFix another case where debug info was affecting
2009-03-05 Bill WendlingAdd comment to emphasize that the while body is empty.
2009-03-05 Dale JohannesenFix another case where a dbg.declare meant something
2009-03-04 Dale JohannesenRe-commit 65975 and a fix for the problem that
2009-03-04 Dan GohmanFix this comment.
2009-03-04 Dan GohmanAdd an assertion for a condition that's always true...
2009-03-04 Chris Lattnercomplete comment.
2009-03-04 Chris Lattnerthis wasn't intended to be committed.
2009-03-04 Chris LattnerFix PR3720 by properly propagating alignment informatio...
2009-03-04 Dale JohannesenAlways skip ptr-to-ptr bitcasts when counting,
2009-03-04 Dale JohannesenMake my earlier patch to skip debug intrinsics
2009-03-03 Dale JohannesenInstruction counters must skip the bitcasts that
2009-03-03 Dale JohannesenWhen removing a store to an alloca that has only one
2009-03-03 Bill WendlingUse > instead of >=. We want to promote aggregates...
2009-03-03 Bill WendlingReapply r65755, but reversing "<" to ">=".
2009-03-03 Dan GohmanFix a bunch of Doxygen syntax issues. Escape special...
2009-03-03 Dale JohannesenDon't count DebugInfo instructions in another limit
2009-03-03 Dale JohannesenWhen sinking an insn in InstCombine bring its debug
2009-03-02 Devang PatelIf branch conditions' one successor is dominating anoth...
2009-03-02 Duncan SandsFix PR3694: add an instcombine micro-optimization that...
2009-03-01 Bill WendlingTemporarily revert r65755. It was causing failures...
2009-03-01 Chris Lattnerhoist the check for alloca size up so that it controls...
2009-02-27 Nick LewyckySilence compiler warning about use of uninitialized...
2009-02-25 Chris LattnerFix PR3667
2009-02-24 Dan GohmanRename ScalarEvolution's getIterationCount to getBacked...
2009-02-24 Dan GohmanFix a ValueTracking rule: RHS means operand 1, not...
2009-02-24 Dan GohmanGeneralize the ChangeCompareStride code, in preparation for
2009-02-24 Dan GohmanPreserve the DominanceFrontier analysis in the LoopDele...
2009-02-23 Dan GohmanBack out the change in 64918 that used sign-extensions...
2009-02-23 Dan GohmanLoopDeletion needs to inform ScalarEvolution when a...
2009-02-23 Dan GohmanIndVarSimplify preserves ScalarEvolution. In the
2009-02-23 Zhou ShengShould reset DBI_Prev if DBI_Next == 0.
2009-02-23 Chris Lattnerfix some typos that Duncan noticed
2009-02-22 Dan GohmanRevert the part of 64623 that attempted to align the...
2009-02-22 Dan GohmanProperly parenthesize this expression, fixing a real...
2009-02-22 Evan ChengOnly try to sink immediate when TLI is not null. It...
2009-02-21 Nick LewyckyDon't sign extend the char when expanding char -> int...
2009-02-21 Evan ChengTeach LSR sink to sink the immediate portion of the...
2009-02-21 Chris Lattnerrename a function to indicate that it checks for profit...
2009-02-20 Evan ChengFix strange logic in CollectIVUsers used to determine...
2009-02-20 Dan GohmanSimplify code and reduce indentation. No functionality...
next