1. 7f93dc8 Switch obvious clients to Twine instead of utostr (when they were already using by Daniel Dunbar · 15 years ago
  2. e922c02 Get rid of the Pass+Context magic. by Owen Anderson · 15 years ago
  3. 07cf79e "LLVMContext* " --> "LLVMContext *" by Owen Anderson · 15 years ago
  4. 0a205a4 More LLVMContext-ification. by Owen Anderson · 15 years ago
  5. e9d87f4 Simplify code by using SmallVector's pop_back_val() instead of by Dan Gohman · 15 years ago
  6. bde6fda Fix for PR3944: make mem2reg O(N) instead of O(N^2) in the number of by Eli Friedman · 15 years ago
  7. 2511abf Tweak the check for promotable alloca's to handle debug intrinsics correctly. by Dale Johannesen · 15 years ago
  8. b8c564f Let AnalyzeAlloca() remove debug intrinsics. by Devang Patel · 16 years ago
  9. 746867c Rework r58829, allowing removal of dbg info intrinsics during alloca by Daniel Dunbar · 16 years ago
  10. 068a795 BCUI + 1 doesn't work. Use next instead. by Bill Wendling · 16 years ago
  11. 180ffae Handle (delete) dbg intrinsics while promoting alloca. by Devang Patel · 16 years ago
  12. 0fd77a5 Rewrite all the 'PromoteLocallyUsedAlloca[s]' logic. With the power of by Chris Lattner · 16 years ago
  13. 3321060 Add a new LargeBlockInfo helper, which is just a wrapper around by Chris Lattner · 16 years ago
  14. 051a950 API changes for class Use size reduction, wave 1. by Gabor Greif · 16 years ago
  15. 6e7aeb1 Update -mem2reg to use succ_iterator instead of iterating across TerminatorInst by Nick Lewycky · 16 years ago
  16. 2663ffe Make RenamePass faster by making the 'is this a new phi node' by Chris Lattner · 17 years ago
  17. 4ee451d Remove attribution from file headers, per discussion on llvmdev. by Chris Lattner · 17 years ago
  18. c86b677 Finishing initial docs for all transformations in Passes.html. by Gordon Henriksen · 17 years ago
  19. 76c1b97 Merge DenseMapKeyInfo & DenseMapValueInfo into DenseMapInfo by Chris Lattner · 17 years ago
  20. 9f528e6 Don't promote volatile loads/stores. This is needed (for example) to handle setjmp/longjmp properly. by Anton Korobeynikov · 17 years ago
  21. f12f8de rewrite the code used to construct pruned SSA form with the IDF method. by Chris Lattner · 17 years ago
  22. 0ec8df3 Factor out a whole bunch of code into it's own method. by Chris Lattner · 17 years ago
  23. 127ed3c Use getNumPreds(BB) instead of computing them manually. This is a very small but by Chris Lattner · 17 years ago
  24. 1e76af3 Change the rename pass to be "tail recursive", only adding N-1 successors by Chris Lattner · 17 years ago
  25. e7b653d cache computation of #preds for a BB. This speeds up by Chris Lattner · 17 years ago
  26. 384c7e0 reserve operand space for phi nodes when we insert them. by Chris Lattner · 17 years ago
  27. 6c81213 use continue to avoid nesting, no functionality change. by Chris Lattner · 17 years ago
  28. fb312c7 Promoting allocas with the 'single store' fastpath is by Chris Lattner · 17 years ago
  29. 7a5745b When PromoteLocallyUsedAllocas promoted allocas, it didn't remember by Chris Lattner · 17 years ago
  30. 59a2837 std::map -> DenseMap by Chris Lattner · 17 years ago
  31. c69e491 fix a logic bug where we wouldn't promote single store allocas if the by Chris Lattner · 17 years ago
  32. 4f63e76 When we do the single-store optimization, delete both the store by Chris Lattner · 17 years ago
  33. d0458e5 Three improvements: by Chris Lattner · 17 years ago
  34. c3f6ea8 switch from using a std::set to using a SmallPtrSet. This speeds up the by Chris Lattner · 17 years ago
  35. b776a33 In mem2reg, when handling the single-store case, make sure to remove by Chris Lattner · 17 years ago
  36. 5dd75b4 split rewriting of single-store allocas into its own method. by Chris Lattner · 17 years ago
  37. bbe1040 refactor some code to shrink PromoteMem2Reg::run a bit by Chris Lattner · 17 years ago
  38. 483ce14 add a typedef, no other change. by Chris Lattner · 17 years ago
  39. 63cdcaa avoid an unneeded vector copy. This speeds up mem2reg on the testcase by Chris Lattner · 17 years ago
  40. ac4aa4b make RenamePassWorkList a local var instead of an ivar. by Chris Lattner · 17 years ago
  41. 326821e Use DominatorTree instead of ETForest. by Devang Patel · 17 years ago
  42. 68c01b3 Mem2Reg does not need TargetData. by Devang Patel · 17 years ago
  43. 41968df Remove unused function argument. by Devang Patel · 17 years ago
  44. cdacc49 Fix a comment. by Owen Anderson · 17 years ago
  45. df07335 Move more passes to using ETForest instead of DominatorTree. by Owen Anderson · 17 years ago
  46. d64d3a1 Reduce malloc/free traffic. by Devang Patel · 17 years ago
  47. f7543fe Remove dead comments. by Devang Patel · 17 years ago
  48. a5b7dc5 Avoid recursion. Use iterative algorithm for RenamePass(). by Devang Patel · 17 years ago
  49. dfb22c3 redesign the primary datastructure used by mem2reg to eliminate an by Chris Lattner · 18 years ago
  50. d387404 With the last change, we no longer need both directions of mapping from by Chris Lattner · 18 years ago
  51. a0d1548 Simplify use of DFBlocks, this makes no noticable performance difference, by Chris Lattner · 18 years ago
  52. 6a1a28d Switch InsertedPHINodes back to SmallPtrSet now that the SmallPtrSet::erase by Chris Lattner · 18 years ago
  53. fc7d13d switch a SmallPtrSet back to an std::set for now, this caused problems. by Chris Lattner · 18 years ago
  54. c670f3d switch an std::set over to a SmallPtrSet, speeding up mem2reg 6% on 176.gcc. by Chris Lattner · 18 years ago
  55. c837615 switch an std::set over to SmallPtrSet, speeding up mem2reg 3.4% on 176.gcc. by Chris Lattner · 18 years ago
  56. 40b6555 eliminate some malloc traffic, this speeds up mem2reg by 3.4%. by Chris Lattner · 18 years ago
  57. a4f0b3a s|llvm/Support/Visibility.h|llvm/Support/Compiler.h| by Chris Lattner · 18 years ago
  58. 9525528 Use hidden visibility to make symbols in an anonymous namespace get by Chris Lattner · 18 years ago
  59. 634c76c Fix some nondeterminstic behavior in the mem2reg pass that (in addition to by Chris Lattner · 18 years ago
  60. 36ba500 Implement a refinement to the mem2reg algorithm for cases where an alloca by Chris Lattner · 19 years ago
  61. fed40df This needs proper dominance by Chris Lattner · 19 years ago
  62. 5e1b231 This code can handle non-dominating instructions by Chris Lattner · 19 years ago
  63. a83ba0f Fix a fixme in CondPropagate.cpp by moving a PhiNode optimization into by Nate Begeman · 19 years ago
  64. 00b16889 Eliminate all remaining tabs and trailing spaces. by Jeff Cohen · 19 years ago
  65. 6cfd1eb Fix PR590 and Transforms/Mem2Reg/2005-06-30-ReadBeforeWrite.ll. by Chris Lattner · 19 years ago
  66. fd93908 Remove trailing whitespace by Misha Brukman · 19 years ago
  67. 5a85d9c Fix test/Regression/Transforms/LICM/2004-09-14-AliasAnalysisInvalidate.llx by Chris Lattner · 20 years ago
  68. 17e6e44 Correction to allow compilation with Visual C++. by Reid Spencer · 20 years ago
  69. 28e792c Fix a bug that occurs when the constant value is the result of an invoke. In by Chris Lattner · 20 years ago
  70. 7e40f63 When inserting PHI nodes, don't insert any phi nodes that are obviously by Chris Lattner · 20 years ago
  71. b20724d When promoting mem2reg, make uninitialized values become undef isntead of 0. by Chris Lattner · 20 years ago
  72. a744b77 Remove a whole bunch of horrible hacky code that was used to promote allocas by Chris Lattner · 20 years ago
  73. afbb1cc Make sure to remove the Select instruction as well by Chris Lattner · 20 years ago
  74. 62e29b5 If given an AliasSetTracker object to update, update it. by Chris Lattner · 20 years ago
  75. 20aa474 Fixes to make LLVM compile with vc7.1. by Alkis Evlogimenos · 20 years ago
  76. 551ccae Changes For Bug 352 by Reid Spencer · 20 years ago
  77. 47b14a4 Fix #includes of i*.h => Instructions.h as per PR403. by Misha Brukman · 20 years ago
  78. abc35bc Change to use the StableBasicBlockNumbering class by Chris Lattner · 20 years ago
  79. 63168d2 Do not let the numbering of PHI nodes placed in the function depend on by Chris Lattner · 20 years ago
  80. 8ba725b Implement ScalarRepl/select_promote.ll by Chris Lattner · 20 years ago
  81. e47f78e Bunch up all locally used allocas by the block they are allocated in, and by Chris Lattner · 21 years ago
  82. 7fecc2e Handle extremely trivial cases extremely efficiently. This speeds up by Chris Lattner · 21 years ago
  83. 2d11f16 Implement Transforms/ScalarRepl/phinodepromote.ll, which is an important by Chris Lattner · 21 years ago
  84. f7703df Finegrainify namespacification by Chris Lattner · 21 years ago
  85. d0fde30 Put all LLVM code into the llvm namespace, as per bug 109. by Brian Gaeke · 21 years ago
  86. b576c94 Added LLVM project notice to the top of every C++ source file. by John Criswell · 21 years ago
  87. 3c4a34e Fix PR#50 by Chris Lattner · 21 years ago
  88. 69091be This changes the PromoteMemToReg function to create "pruned" SSA form, not by Chris Lattner · 21 years ago
  89. 43f820d Change the interface to PromoteMemToReg to also take a DominatorTree by Chris Lattner · 21 years ago
  90. 24011be Speed up the mem2reg transform for allocas which are only read/written in a single by Chris Lattner · 21 years ago
  91. 92e4975 The first PHI node may be null, scan for the first non-null one by Chris Lattner · 21 years ago
  92. 3c881cb The VersionNumbers vector is only used during PHI placement. Turn it into an argument, allowing us to get rid of the vector. by Chris Lattner · 21 years ago
  93. afa060e * Update file header comment by Chris Lattner · 21 years ago
  94. 521c16a Simplify the loop a bit by Chris Lattner · 21 years ago
  95. 92581c2 There is no need for separate WriteSets and PhiNodeBlocks lists. It is just a by Chris Lattner · 21 years ago
  96. 036f1e7 The PhiNodes 2D vector is only used during PHI node placement. It doesn't by Chris Lattner · 21 years ago
  97. 9e38fbf * Document instance vars better by Chris Lattner · 21 years ago
  98. 9157f04 Two small cleanups/speedups: by Chris Lattner · 21 years ago
  99. 0fa1571 * Minor cleanups by Chris Lattner · 21 years ago
  100. 5560c9d Spell `necessary' correctly. by Misha Brukman · 21 years ago