2011-12-20 |
Jakub Staszak | Add some constantness to BranchProbabilityInfo and... |
tree | commitdiff |
2011-12-20 |
David Blaikie | Unweaken vtables as per llvm.org/docs/CodingStandards... |
tree | commitdiff |
2011-12-20 |
Andrew Trick | LSR: Fix another corner case in expansion of postinc... |
tree | commitdiff |
2011-12-18 |
Joerg Sonnenberger | Allow inlining of functions with returns_twice calls... |
tree | commitdiff |
2011-12-16 |
Eric Christopher | When recursing for the original size of a type, stop... |
tree | commitdiff |
2011-12-16 |
Devang Patel | In DICompositeType, referenced to derived type is eithe... |
tree | commitdiff |
2011-12-15 |
Devang Patel | Virtual table holder field is either metadata or null. |
tree | commitdiff |
2011-12-14 |
Dan Gohman | Move Instruction::isSafeToSpeculativelyExecute out... |
tree | commitdiff |
2011-12-14 |
Andrew Trick | LSR: Fold redundant bitcasts on-the-fly. |
tree | commitdiff |
2011-12-14 |
Eli Friedman | Fix a stupid typo in MemDepPrinter. |
tree | commitdiff |
2011-12-12 |
Daniel Dunbar | LLVMBuild: Introduce a common section which currently... |
tree | commitdiff |
2011-12-12 |
Daniel Dunbar | LLVMBuild: Remove trailing newline, which irked me. |
tree | commitdiff |
2011-12-12 |
Chandler Carruth | Switch llvm.cttz and llvm.ctlz to accept a second i1... |
tree | commitdiff |
2011-12-06 |
Chad Rosier | Probably not a good idea to convert a single vector... |
tree | commitdiff |
2011-12-05 |
Nadav Rotem | Add support for vectors of pointers. |
tree | commitdiff |
2011-12-03 |
Benjamin Kramer | Clear the new cache. |
tree | commitdiff |
2011-12-03 |
Benjamin Kramer | Add a "seen blocks" cache to LVI to avoid a linear... |
tree | commitdiff |
2011-12-03 |
Chad Rosier | Add support for constant folding the pow intrinsic. |
tree | commitdiff |
2011-12-02 |
Chad Rosier | Fix a few more places where TargetData/TargetLibraryInf... |
tree | commitdiff |
2011-12-01 |
Chad Rosier | Abuse of mass replace isn't warranted even when the... |
tree | commitdiff |
2011-12-01 |
Chad Rosier | Fix build by not assuming TLI is guaranteed. Will have... |
tree | commitdiff |
2011-12-01 |
Chad Rosier | Prevent library calls from being folded if -fno-builtin... |
tree | commitdiff |
2011-12-01 |
Chad Rosier | Last bit of TargetLibraryInfo propagation. Also fixed... |
tree | commitdiff |
2011-12-01 |
Chad Rosier | Propagate TargetLibraryInfo throughout ConstantFolding... |
tree | commitdiff |
2011-12-01 |
Nick Lewycky | Make use of "getScalarType()". No functionality change. |
tree | commitdiff |
2011-11-30 |
Andrew Trick | LSR: handle the expansion of phi operands that use... |
tree | commitdiff |
2011-11-29 |
Daniel Dunbar | build/CMake: Finish removal of add_llvm_library_depende... |
tree | commitdiff |
2011-11-29 |
Duncan Sands | Fix a theoretical problem (not seen in the wild): if... |
tree | commitdiff |
2011-11-29 |
Andrew Trick | SCEV fix. In general, Add/Mul expressions should not... |
tree | commitdiff |
2011-11-29 |
Andrew Trick | Make SCEV print <nsw><nuw> for Add/MulExpr. |
tree | commitdiff |
2011-11-28 |
Eli Friedman | Make SelectionDAG::InferPtrAlignment use llvm::ComputeM... |
tree | commitdiff |
2011-11-28 |
Andrew Trick | Remove the temporary flag -disable-unroll-scev and... |
tree | commitdiff |
2011-11-26 |
Benjamin Kramer | Move code into anonymous namespaces. |
tree | commitdiff |
2011-11-23 |
Benjamin Kramer | Validate the return type when checking if a function... |
tree | commitdiff |
2011-11-23 |
Duncan Sands | Fix a crash in which a multiplication was being reporte... |
tree | commitdiff |
2011-11-21 |
Nick Lewycky | Fix crasher in GVN due to my recent capture tracking... |
tree | commitdiff |
2011-11-21 |
Nick Lewycky | Add virtual destructor. Whoops! |
tree | commitdiff |
2011-11-20 |
Nick Lewycky | Less template, more virtual! Refactoring suggested... |
tree | commitdiff |
2011-11-20 |
Nick Lewycky | Refactor code to use new attribute getters on CallSite... |
tree | commitdiff |
2011-11-20 |
Benjamin Kramer | SCEV: Actually set overflow flags on add expressions. |
tree | commitdiff |
2011-11-18 |
Andrew Trick | Fix a corner case in updating LoopInfo after fully... |
tree | commitdiff |
2011-11-16 |
Andrew Trick | Fix SCEV overly optimistic back edge taken count for... |
tree | commitdiff |
2011-11-15 |
Benjamin Kramer | Missed some users of Value::getNameStr. |
tree | commitdiff |
2011-11-15 |
Benjamin Kramer | Remove all remaining uses of Value::getNameStr(). |
tree | commitdiff |
2011-11-15 |
Benjamin Kramer | Twinify GraphWriter a little bit. |
tree | commitdiff |
2011-11-14 |
Nick Lewycky | Refactor capture tracking (which already had a couple... |
tree | commitdiff |
2011-11-12 |
Nick Lewycky | Don't try to loop on iterators that are potentially... |
tree | commitdiff |
2011-11-09 |
Nick Lewycky | Fix typo in comment. |
tree | commitdiff |
2011-11-09 |
Nick Lewycky | Don't forget to check FlagNW when determining whether... |
tree | commitdiff |
2011-11-08 |
Eli Friedman | Fix code to match comment. Fixes PR11340, a regression... |
tree | commitdiff |
2011-11-04 |
Dan Gohman | Teach instsimplify to simplify calls to undef. |
tree | commitdiff |
2011-11-03 |
Daniel Dunbar | build: Add initial cut at LLVMBuild.txt files. |
tree | commitdiff |
2011-10-30 |
Duncan Sands | Reapply commit 143214 with a fix: m_ICmp doesn't match... |
tree | commitdiff |
2011-10-29 |
Eli Friedman | Revert r143214; it's breaking a bunch of stuff. |
tree | commitdiff |
2011-10-28 |
Duncan Sands | The expression icmp eq (select (icmp eq x, 0), 1, x... |
tree | commitdiff |
2011-10-28 |
Duncan Sands | A shift of a power of two is a power of two or zero. |
tree | commitdiff |
2011-10-28 |
Duncan Sands | Fold icmp ugt (udiv X, Y), X to false. Spotted by... |
tree | commitdiff |
2011-10-27 |
Duncan Sands | Reapply commit 143028 with a fix: the problem was casti... |
tree | commitdiff |
2011-10-27 |
Bob Wilson | Revert Duncan's r143028 expression folding which appear... |
tree | commitdiff |
2011-10-26 |
Duncan Sands | The maximum power of 2 dividing a power of 2 is itself... |
tree | commitdiff |
2011-10-26 |
Duncan Sands | My super-optimizer noticed that we weren't folding... |
tree | commitdiff |
2011-10-25 |
Duncan Sands | Restore commits 142790 and 142843 - they weren't breaki... |
tree | commitdiff |
2011-10-25 |
Chandler Carruth | Fix the API usage in loop probability heuristics. It... |
tree | commitdiff |
2011-10-25 |
Duncan Sands | Speculatively revert commits 142790 and 142843 to see... |
tree | commitdiff |
2011-10-24 |
Nick Lewycky | Now that we look at all the header PHIs, we need to... |
tree | commitdiff |
2011-10-24 |
Chandler Carruth | Remove return heuristics from the static branch probabi... |
tree | commitdiff |
2011-10-24 |
Nick Lewycky | Reapply r142781 with fix. Original message: |
tree | commitdiff |
2011-10-24 |
Nick Lewycky | PHI nodes not in the loop header aren't part of the... |
tree | commitdiff |
2011-10-24 |
Nick Lewycky | Speculatively revert r142781. Bots are showing |
tree | commitdiff |
2011-10-24 |
Chandler Carruth | Simplify the design of BranchProbabilityInfo by collaps... |
tree | commitdiff |
2011-10-23 |
Nick Lewycky | Enhance SCEV's brute force loop analysis to handle... |
tree | commitdiff |
2011-10-23 |
Chandler Carruth | Tidy up a loop to be more idiomatic for LLVM's codebase... |
tree | commitdiff |
2011-10-23 |
Chandler Carruth | Teach the BranchProbabilityInfo pass to print its resul... |
tree | commitdiff |
2011-10-23 |
Benjamin Kramer | Add compare operators to BranchProbability and use... |
tree | commitdiff |
2011-10-22 |
Nick Lewycky | Make SCEV's brute force analysis stronger in two ways... |
tree | commitdiff |
2011-10-21 |
Benjamin Kramer | Extend the floating point heuristic to consider NaN... |
tree | commitdiff |
2011-10-21 |
Benjamin Kramer | BranchProbabilityInfo: floating point equality is unlikely. |
tree | commitdiff |
2011-10-20 |
Eli Friedman | A FIXME about block addresses and indirectbr. |
tree | commitdiff |
2011-10-20 |
Eli Friedman | Simplify; no intended functional change. |
tree | commitdiff |
2011-10-20 |
Nick Lewycky | "@string = constant i8 0" is a value i8* string of... |
tree | commitdiff |
2011-10-19 |
Chandler Carruth | Generalize the reading of probability metadata to work... |
tree | commitdiff |
2011-10-19 |
Chandler Carruth | Teach the BranchProbabilityInfo analysis pass to read... |
tree | commitdiff |
2011-10-19 |
Chandler Carruth | Add pass printing support to BlockFrequencyInfo pass... |
tree | commitdiff |
2011-10-17 |
Devang Patel | Update DebugInfoFinder to match recent debug info encod... |
tree | commitdiff |
2011-10-17 |
Bill Wendling | Correct over-zealous removal of hack. |
tree | commitdiff |
2011-10-17 |
Bill Wendling | Now that we have the ReturnsTwice function attribute... |
tree | commitdiff |
2011-10-16 |
Chandler Carruth | Delete a dead member. Dunno if this was ever used,... |
tree | commitdiff |
2011-10-15 |
Andrew Trick | Fix SCEVExpander assert during LSR: "argument of incomp... |
tree | commitdiff |
2011-10-14 |
Nick Lewycky | An instruction's operands aren't necessarily instructio... |
tree | commitdiff |
2011-10-13 |
Eli Friedman | Enhance the memdep interface so that users can tell... |
tree | commitdiff |
2011-10-13 |
Andrew Trick | Reapply r141870, SCEV expansion of post-inc. |
tree | commitdiff |
2011-10-13 |
Andrew Trick | Fix memory corruption I introduced a few checkins ago. |
tree | commitdiff |
2011-10-13 |
Andrew Trick | Revert r141870. The test case crashes on linux with... |
tree | commitdiff |
2011-10-13 |
Andrew Trick | LSR: Reuse the post-inc expansion of expressions. |
tree | commitdiff |
2011-10-13 |
Andrew Trick | SCEV: Rewrite TrandformForPostIncUse to handle expressi... |
tree | commitdiff |
2011-10-13 |
Andrew Trick | Slightly more useful tracing. |
tree | commitdiff |
2011-10-11 |
Eric Christopher | Add a new wrapper node for a DILexicalBlock that encaps... |
tree | commitdiff |
2011-10-11 |
Andrew Trick | Move replaceCongruentIVs into SCEVExapander and bias... |
tree | commitdiff |
2011-10-08 |
Andrew Trick | Add an extra safety check in front of the optimization... |
tree | commitdiff |
2011-10-07 |
Andrew Trick | LSR should only reuse phis that match its formula. |
tree | commitdiff |
next |