- 0ce4b8e Allow DOUT to be used outside of the llvm namespace. by Dan Gohman · 17 years ago
- 89b20c0 Make the comment for ScalarizeVectorOp mention that it is only for use by Dan Gohman · 17 years ago
- 0d7d365 Fix a bug. by Zhou Sheng · 17 years ago
- df82c93 ok, this is something of a dirty hack, but it seems to work. (fixes e.g. by Duraid Madina · 17 years ago
- b2efabd ok, this much doesn't seem to bork anything by Duraid Madina · 17 years ago
- 4e378c6 revert evan's fixes (and my doofusness) since they had a huge code by Duraid Madina · 17 years ago
- 669f738 pull evan's fixes - should help the nightly tester (but there are still by Duraid Madina · 17 years ago
- 0410407 Replace std::set with SmallPtrSet. by Evan Cheng · 17 years ago
- 62cf8ba Add support for performing GVNPRE on the three vector-specific operations. by Owen Anderson · 17 years ago
- cc389e0 Add tests for performing GVNPRE on the three vector-specific instructions. by Owen Anderson · 17 years ago
- 75611fb Fix an obvious bug. Old code only worked for the entry block. by Evan Cheng · 17 years ago
- 6032a5b 1. Correct some comments and clean up some dead code. by Owen Anderson · 17 years ago
- c33aa47 Add comment. by Evan Cheng · 17 years ago
- 7ac19af Correctly handle implcit def / use operands. by Evan Cheng · 17 years ago
- 4efe741 Properly handle kills of a physical register which has sub-registers that are read by later instructions. by Evan Cheng · 17 years ago
- c4f2fe0 Add immediate sub-registers. by Evan Cheng · 17 years ago
- e2446c6 Silence a warning. by Evan Cheng · 17 years ago
- c9a15d5 Update for Subversion conversion: by Reid Spencer · 17 years ago
- 4d2a0f5 Replace ?: with if statements, for clarity. by Dan Gohman · 17 years ago
- 25aceac Simplify the expression for MVT::isExtendedValueType. by Dan Gohman · 17 years ago
- 71d7794 Simplify the expression for TargetLowering::isTypeLegal. by Dan Gohman · 17 years ago
- acaf32e Use utostr from StringExtras.h instead of ostringstream from <sstream>. by Dan Gohman · 17 years ago
- 798cb2a Renumber the SimpleValueType values to fill in the hole left by by Dan Gohman · 17 years ago
- d45eddd Revert the earlier change that removed the M_REMATERIALIZABLE machine by Dan Gohman · 17 years ago
- 9a0930d Fix a typo in a comment. by Dan Gohman · 17 years ago
- bc9d89e Add a convenince member function for appending strings to a module's by Dan Gohman · 17 years ago
- 837a600 tidy this file up a bit by Duraid Madina · 17 years ago
- 2e0930c A bunch of fixes to the BigBlock allocator improve compile-time by ~20% by Duraid Madina · 17 years ago
- ba6801e fix Transforms/Inline/2007-06-25-WeakInline.ll by not inlining functions by Chris Lattner · 17 years ago
- 4939deb new testcase, the inliner shouldn't inline this. by Chris Lattner · 17 years ago
- 9030d38 Use the built-in postorder iterators rather than computing a postorder walk by hand. by Owen Anderson · 17 years ago
- c38adbb Fix PR1525: by Reid Spencer · 17 years ago
- 7f32156 Generalize MVT::ValueType and associated functions to be able to represent by Dan Gohman · 17 years ago
- 32791e0 Make minor adjustments to whitespace and comments to reduce differences by Dan Gohman · 17 years ago
- 0197630 Fix loadv2i32 to be loadv4i32, though it isn't actually used anywhere yet. by Dan Gohman · 17 years ago
- 8bc49c2 Say AT&T instead of Intel in the comments for AT&T support. by Dan Gohman · 17 years ago
- 5248896 Provide hook for alloca on VCPP. Patch by Scott Graham by Anton Korobeynikov · 17 years ago
- f62c44a 1) Fix an issue with non-deterministic iteration order in phi_translate by Owen Anderson · 17 years ago
- 4390feb Fix value ranges. by Nick Lewycky · 17 years ago
- 79a6370 Fix a silly mistake that was causing failures. by Owen Anderson · 17 years ago
- f107bef Rename variables to expose the fact that this test is failing. by Owen Anderson · 17 years ago
- dea2526 Remove tabs. by Nick Lewycky · 17 years ago
- 984504b Remove use of ETForest. Also cleaned up issues around unreachable basic by Nick Lewycky · 17 years ago
- 6475804 Rework topo_sort so eliminate some behavior that scaled terribly. This reduces the time to optimize 403.gcc from 18.2s to 17.5s, by Owen Anderson · 17 years ago
- 2106f61 Perform fewer set insertions while calculating ANTIC_IN. This reduces the amount of time to optimize 403.gcc from 21.9s to 18.2s. by Owen Anderson · 17 years ago
- b3b3734 Remove some code that I was using for collecting performance information that should not have been committed. by Owen Anderson · 17 years ago
- 0819a9d Fix the build. by Owen Anderson · 17 years ago
- ea859be Move ComputeMaskedBits, MaskedValueIsZero, and ComputeNumSignBits from by Dan Gohman · 17 years ago
- a8c7682 check in the BigBlock local register allocator by Duraid Madina · 17 years ago
- 1240846 Avoid excessive calls to find_leader when calculating AVAIL_OUT. This reduces the time to optimize 403.gcc from 23.5s to 21.9s. by Owen Anderson · 17 years ago
- 0b2ce1f std::set is really really terrible. Switch to SmallPtrSet to reduce compile time. For Duraid's example. The overall isel time is reduced from 0.6255 sec to 0.1876 sec. by Evan Cheng · 17 years ago
- 5411835 Quote complex names for Darwin X86 and ARM. by Dale Johannesen · 17 years ago
- e8138ff Reserve space in vectors before topologically sorting into them. This improves the time to optimize 403.gcc from 28s to 23.5s. by Owen Anderson · 17 years ago
- 82575d8 Make a bunch of optimizations for compile time to GVNPRE, including smarter set unions, deferring blocks rather than computing maximal sets, and smarter use of sets. With these enhancements, the time to optimize 273.perlbmk goes from 5.3s to 2.7s. by Owen Anderson · 17 years ago
- 6394e5e Fix a bug in SmallPtrSet that was causing GVNPRE to enter an infinite loop. by Owen Anderson · 17 years ago
- 61766ca Two changes: by Chris Lattner · 17 years ago
- 25e681a CallGraphSCCPass manager may require other passes. by Devang Patel · 17 years ago
- d3fb671 Have internal df_iterator's use SmallPtrSet instead of std::set. This provides compile time speedups to any pass using df_iterator. by Owen Anderson · 17 years ago
- 52974eb Fix tests. by Evan Cheng · 17 years ago
- 1dbfd48 Significantly improve the documentation of the instcombine divide/compare by Chris Lattner · 17 years ago
- e47f30f new testcase miscompiled by instcombine, reduced from perl by Chris Lattner · 17 years ago
- 66fd906 Change lots of sets from std::set to SmallPtrSet. This reduces the time required to optimize 253.perlbmk from 10.9s to 5.3s. by Owen Anderson · 17 years ago
- 0e7f728 Move code to update dominator information after basic block is split by Devang Patel · 17 years ago
- 2d74a31 Tidy up ValueType names in comments. by Dan Gohman · 17 years ago
- b9f1019 Rename TargetLowering::getNumElements and friends to by Dan Gohman · 17 years ago
- 53855ca New tests. by Evan Cheng · 17 years ago
- b13cdbd Xforms: by Evan Cheng · 17 years ago
- 58e087b Eliminate a redundant check. This speeds up optimization of 253.perlbmk from 13.5 seconds to 10.9 seconds. by Owen Anderson · 17 years ago
- 9cb5601 Comment-ize the functions in GVNPRE. by Owen Anderson · 17 years ago
- 562ef78 refactor a bunch of code out of visitICmpInstWithInstAndIntCst into its own by Chris Lattner · 17 years ago
- 3eaca71 Split runOnFunction into many smaller functions. This make it easier to get accurate performance analysis of GVNPRE. by Owen Anderson · 17 years ago
- c5dfcdb Modify deleting global variable with an even easier way. by Tanya Lattner · 17 years ago
- 2397f8d Update AnalysisGroup documentation to document restriction that allows by Devang Patel · 17 years ago
- b011c66 Add blurb on deleting global variables. by Tanya Lattner · 17 years ago
- 0304b2b Make GVNPRE accurate report whether it modified the function or not. by Owen Anderson · 17 years ago
- 60c916b Added some if-conversion tests. by Evan Cheng · 17 years ago
- 5dbea73 Get rid of an unneeded helper function. by Owen Anderson · 17 years ago
- 97e604e Be more conservative of duplicating blocks. by Evan Cheng · 17 years ago
- 4f703ec Use a DenseMap instead of an std::map for the value numbering. This reduces the time to optimize lencod on a PPC Debug build from ~300s to ~140s. by Owen Anderson · 17 years ago
- 52471b1 Make dependsOnInvoke much more specific in what it tests, which in turn make it much faster to run. This reduces the time to optimize lencondwith a debug build on PPC from ~450s to ~300s. by Owen Anderson · 17 years ago
- c91c56c Moved Inliner.h to include/llvm/Transforms/IPO/InlinerPass.h by Tanya Lattner · 17 years ago
- 6f7426e Inliner pass header file was moved. by Tanya Lattner · 17 years ago
- 9933e4b Move inliner pass header file. by Tanya Lattner · 17 years ago
- 9ffd340 Avoid if-converting simple block that ends with unconditional branch or fallthrough unless it branches / falls to the 'false' block. Not profitable, may end up increasing code size. by Evan Cheng · 17 years ago
- 277f074 Allow predicated immediate ARM to ARM calls. by Evan Cheng · 17 years ago
- 3f931b8 Fix pr1448 by Chris Lattner · 17 years ago
- 5cec4db Rename ScalarEvolution::deleteInstructionFromRecords to by Dan Gohman · 17 years ago
- b5bec2b Pass a SelectionDAG into SDNode::dump everywhere it's used, in prepration by Dan Gohman · 17 years ago
- d095410 Handle constants in phi nodes properly. This fixes test/Transforms/GVNPRE/2007-06-18-ConstantInPhi.ll by Owen Anderson · 17 years ago
- 20c2b35 silence warning when assertions are disabled. by Chris Lattner · 17 years ago
- 831dab3 Testcase for instances where a constant only occurs as an operand to a phi node. by Owen Anderson · 17 years ago
- defb9fb don't use binutils 2.17 by Chris Lattner · 17 years ago
- 85ef254 document and hide two options. by Chris Lattner · 17 years ago
- 3ee7740 describe an argument, hide it. by Chris Lattner · 17 years ago
- 2384d7b silence a bogus warning Duraid ran into. by Chris Lattner · 17 years ago
- 5b689c3 Add a new testcase for memory corruption issues. by Owen Anderson · 17 years ago
- 20cb51f Be careful to erase values from all of the appropriate sets when they're not needed anymore. This fixes a few more memory-related issues. by Owen Anderson · 17 years ago
- 6cae740 Remember to clear the maximal sets between functions. by Owen Anderson · 17 years ago
- 086f5f3 Refactor GVNPRE to use a much smart method of uniquing value sets, and centralize a lot of the value numbering information. No functionality change. by Owen Anderson · 17 years ago
- 82a87a0 Replace M_REMATERIALIZIBLE and the newly-added isOtherReMaterializableLoad by Dan Gohman · 17 years ago