- 0b79a77 Replace isTrapping with a new, similar method called by Eli Friedman · 15 years ago
- 50dead0 Revert yesterday's change by removing the LLVMContext parameter to AllocaInst and MallocInst. by Owen Anderson · 15 years ago
- 9adc0ab Move EVER MORE stuff over to LLVMContext. by Owen Anderson · 15 years ago
- 333c400 This started as a small change, I swear. Unfortunately, lots of things call the [I|F]CmpInst constructors. Who knew!? by Owen Anderson · 15 years ago
- 0a205a4 More LLVMContext-ification. by Owen Anderson · 15 years ago
- 1ff50b3 Second batch of passes using LLVMContext. by Owen Anderson · 15 years ago
- 0b8365c Revert r67798: it breaks llvm-gcc bootstrap on x86-64-linux, presumably due to by Duncan Sands · 15 years ago
- fafd98b While hoisting an instruction, update alias info set tracker. by Devang Patel · 15 years ago
- d7168dd reimplement AliasSetTracker in terms of DenseMap instead of hash_map, by Chris Lattner · 15 years ago
- d0ffa73 This code doesn't actually use the ExitingBlocks list. by Dan Gohman · 15 years ago
- 394f044 Change create*Pass factory functions to return Pass* instead of by Daniel Dunbar · 16 years ago
- ae73dc1 Tidy up several unbeseeming casts from pointer to intptr_t. by Dan Gohman · 16 years ago
- 3eee654 Put the LICM of constant GlobalVariables, introduced in r53945, under a by Dan Gohman · 16 years ago
- 967948b "Allow LICM to sink or lift loads from constant memory. Also add a test by Chris Lattner · 16 years ago
- 9b78763 Use Loop::block_iterator. by Dan Gohman · 16 years ago
- 02dea8b Tidy up BasicBlock::getFirstNonPHI, and change a bunch of places to by Dan Gohman · 16 years ago
- 19d9d43 rewrite the validity checking for memory promotion to be simpler, by Chris Lattner · 16 years ago
- 29d9293 Use 'continue' to reduce nesting in this loop. No functionality change. by Chris Lattner · 16 years ago
- 844731a Clean up the use of static and anonymous namespaces. This turned up by Dan Gohman · 16 years ago
- 4a3f6c8 Make several variable declarations static. by Dan Gohman · 16 years ago
- 3035959 Use empty() instead of comparing size() with zero. by Dan Gohman · 17 years ago
- 4ee451d Remove attribution from file headers, per discussion on llvmdev. by Chris Lattner · 17 years ago
- dff6710 Integrate the readonly/readnone logic more deeply by Duncan Sands · 17 years ago
- fd94dd5 Remove another leak. Due to some reason AliasSetTracker didn't had any dtor... by Anton Korobeynikov · 17 years ago
- 514ab34 Executive summary: getTypeSize -> getTypeStoreSize / getABITypeSize. by Duncan Sands · 17 years ago
- a24b294 Fix a typo in a comment. by Dan Gohman · 17 years ago
- cb406c2 Use empty() member functions when that's what's being tested for instead by Dan Gohman · 17 years ago
- bc2265a Relax unsafe use check. If there is one unconditional use inside the loop then it is safe to promote value even if there is another conditional use inside the loop. by Devang Patel · 17 years ago
- 798b4af Do not reserve DOM check for GetElementPtrInst. by Devang Patel · 17 years ago
- 128459b Do not promote null values because it may be unsafe to do so. by Devang Patel · 17 years ago
- f2038b1 Avoid unsafe promotion. by Devang Patel · 17 years ago
- e3bfb88 Fix PR1657 by Devang Patel · 17 years ago
- b7211a2 Use SmallVector instead of std::vector. by Devang Patel · 17 years ago
- af5cbc8 Typo. by Nick Lewycky · 17 years ago
- 8d246f0 Add note. by Devang Patel · 17 years ago
- 91d22c8 Implement Simple Analysis interfaces - cloneBasicBlockAnalysis and deleteAnalysisValue. by Devang Patel · 17 years ago
- 96b651c LICM preserves scalar evolution and dom frontier. by Devang Patel · 17 years ago
- 468502e Do not require ETForest. Now it is unused by LICM. by Devang Patel · 17 years ago
- 326821e Use DominatorTree instead of ETForest. by Devang Patel · 17 years ago
- d8af90c Allow insertelement, extractelement, and shufflevector to be hoisted/sunk by Dan Gohman · 17 years ago
- 2604242 s/llvm::DominatorTreeBase::DomTreeNode/llvm::DomTreeNode/g by Devang Patel · 17 years ago
- bec7647 s/DominatorTreeBase::Node/DominatorTreeBase:DomTreeNode/g by Devang Patel · 17 years ago
- e7ae1a9 Insert new instructions in AliasSet. by Devang Patel · 17 years ago
- f38ac5d Fix typo. by Devang Patel · 17 years ago
- ecd94c8 Fix typo in comment. by Nick Lewycky · 17 years ago
- 1997473 Drop 'const' by Devang Patel · 17 years ago
- 3e15bf3 Use 'static const char' instead of 'static const int'. by Devang Patel · 17 years ago
- 794fd75 Do not use typeinfo to identify pass in pass manager. by Devang Patel · 17 years ago
- 68c01b3 Mem2Reg does not need TargetData. by Devang Patel · 17 years ago
- 3a2b58f Rollback some changes that adversely affected performance. I'm currently rethinking by Owen Anderson · 17 years ago
- cdacc49 Fix a comment. by Owen Anderson · 17 years ago
- df07335 Move more passes to using ETForest instead of DominatorTree. by Owen Anderson · 17 years ago
- 325171e Switch more uses of DominatorTree over to ETForest. by Owen Anderson · 17 years ago
- 747603e Spell doFinalization right, so that it is a proper virtual override and by Dan Gohman · 17 years ago
- e34e9a2 fix long lines by Chris Lattner · 17 years ago
- ecb7a77 Change uses of Function::front to Function::getEntryBlock for readability. by Dan Gohman · 17 years ago
- 54959d6 Now LICM is a LoopPass. by Devang Patel · 17 years ago
- 9133fe2 Apply the VISIBILITY_HIDDEN field to the remaining anonymous classes in by Reid Spencer · 18 years ago
- 832254e Changes to support making the shift instructions be true BinaryOperators. by Reid Spencer · 18 years ago
- 79066fa Adjust #includes to match movement of constant folding code from transformutils to libanalysis. by Chris Lattner · 18 years ago
- e4d87aa For PR950: by Reid Spencer · 18 years ago
- 0e5f499 Switch over Transforms/Scalar to use the STATISTIC macro. For each statistic by Chris Lattner · 18 years ago
- ac0b6ae Detemplatize the Statistic class. The only type it is instantiated with by Chris Lattner · 18 years ago
- 3da59db For PR950: by Reid Spencer · 18 years ago
- b742703 Removed #include <iostream> and replaced with llvm_* streams. by Bill Wendling · 18 years ago
- 92f7365 An sinkable instruction may exist with uses, if those uses are in dead blocks. by Chris Lattner · 18 years ago
- 7f8897f eliminate RegisterOpt. It does the same thing as RegisterPass. by Chris Lattner · 18 years ago
- 3c80a51 random code cleanups, no functionality change by Chris Lattner · 18 years ago
- dac58ad Make iostream #inclusion explicit by Chris Lattner · 19 years ago
- 38b5807 prevent va_arg from being hoisted from a loop by Andrew Lenharth · 19 years ago
- 558bc88 core changes for varargs by Andrew Lenharth · 19 years ago
- fd93908 Remove trailing whitespace by Misha Brukman · 19 years ago
- 70ac2dc Fix a bug where LICM was not updating AA information properly when sinking by Chris Lattner · 19 years ago
- 1f309c1 enable -debug-only=licm by Chris Lattner · 19 years ago
- 68a9d3e Adjust to new interfaces by Chris Lattner · 20 years ago
- 8a91939 Allow hoisting loads of globals and alloca's in conditionals. by Chris Lattner · 20 years ago
- f3e1d69 Provide size information when checking to see if we can LICM a load, this by Chris Lattner · 20 years ago
- bdacd87 Fix a bug in the previous checkin that broke 255.vortex by Chris Lattner · 20 years ago
- fb3ee19 Make sure to update alias analysis information as we transform the function. by Chris Lattner · 20 years ago
- 551ccae Changes For Bug 352 by Reid Spencer · 20 years ago
- eaa2430 Fix hoisting of void typed values, e.g. calls by Chris Lattner · 20 years ago
- 56b7ee2 Now that dominator tree children are built in determinstic order, this horrible code by Chris Lattner · 20 years ago
- edb8433c Fix one source of nondeterminism in the -licm pass: the hoist pass by Chris Lattner · 20 years ago
- 802daf8 Fix typo in DEBUG printout. by Brian Gaeke · 20 years ago
- 2741c97 Adjust to the changes in the AliasSetTracker interface by Chris Lattner · 20 years ago
- 3280f7b Move isLoopInvariant to the Loop class by Chris Lattner · 20 years ago
- 5fa802f Loop exit sets are no longer explicitly held, they are dynamically computed on demand. by Chris Lattner · 20 years ago
- 118dd0c Implement LICM of calls in simple cases. This is sufficient to move around by Chris Lattner · 20 years ago
- 5294292 Update comment by Chris Lattner · 21 years ago
- 329c1c6 Improve encapsulation in the Loop and LoopInfo classes by eliminating the by Chris Lattner · 21 years ago
- a3df8a9 Remove the wierd "Operands" loop, by traversing basicblocks in reverse order by Chris Lattner · 21 years ago
- e4365b2 Implement LICM/sink_multiple.ll, by sinking all possible instructions in the by Chris Lattner · 21 years ago
- ba7df4c When we delete instructions from the loop, make sure to remove them from the by Chris Lattner · 21 years ago
- 00ad4a2 Do not promote volatile alias sets into registers by Chris Lattner · 21 years ago
- ea9403f Fix LICM/2003-12-11-SinkingToPHI.ll, and quite possibly all of the other known problems in the universe. by Chris Lattner · 21 years ago
- 7d3ced9 Fix bug: LICM/sink_multiple_exits.ll by Chris Lattner · 21 years ago
- df45bd3 Don't allow dead instructions to stop sinking early. by Chris Lattner · 21 years ago
- e3cfe8d Simplify code by Chris Lattner · 21 years ago
- f594a03 Avoid performing two identical lookups when one will suffice by Chris Lattner · 21 years ago
- 0ed2da9 Make LICM itself a bit more efficient, and make the generated code more efficient too: don't insert a store in every exit block, because a particular block may be exited to more than once by a loop by Chris Lattner · 21 years ago