Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / d213aab71d07b797948dcf9b0cf69b258047c5fe / llvm / lib / CodeGen / SelectionDAG / DAGCombiner.cpp
  1. 418da22 Revert patch ofr218493 by David Xu · 11 years ago
  2. 64f661e Redundant store instructions should be removed as dead code by David Xu · 11 years ago
  3. 6a42292 Use SDValue bool operator to reduce code. No functional change. by Sanjay Patel · 11 years ago
  4. b67bd26 Refactor reciprocal square root estimate into target-independent function; NFC. by Sanjay Patel · 11 years ago
  5. 62ac736 Optionally enable more-aggressive FMA formation in DAGCombine by Hal Finkel · 11 years ago
  6. bb29221 Replace dead links to "Hacker's Delight" with general references. NFC. by Sanjay Patel · 11 years ago
  7. 8239eaa Add DAG combine for shl + add of constants. by Matt Arsenault · 11 years ago
  8. 7bd228a Combine fmul vector FP constants when unsafe math is allowed. by Sanjay Patel · 11 years ago
  9. f7aff68 Build correct vector filled with undef nodes by David Xu · 11 years ago
  10. 394c333 Group unsafe fmul math folds together for easier reading. No functional change. by Sanjay Patel · 11 years ago
  11. f4b7a6b Fix the FIXME that was just added in r217390 - remove a bunch of redundant fold permutations. by Sanjay Patel · 11 years ago
  12. 8170dea group unsafe math folds together for easier reading by Sanjay Patel · 11 years ago
  13. 75cc90e Allow vector fsub ops with constants to get the same optimizations as scalars. by Sanjay Patel · 11 years ago
  14. f4b8deb clean up; NFC by Sanjay Patel · 11 years ago
  15. c1a7121 Fix interference caused by fmul 2, x -> fadd x, x by Matt Arsenault · 11 years ago
  16. 965de30 Fix comment and unnecessary check for FP build_vectors. by Matt Arsenault · 11 years ago
  17. e19006e Enable splitting indexing from loads with TargetConstants by Hal Finkel · 11 years ago
  18. 51e6fa2 Revert "Revert '[DAGCombiner] Split up an indexed load if only the base pointer value is live'" by Hal Finkel · 11 years ago
  19. ccd2676 Move FNEG next to FABS and make them more similar, so it's easier that they can be refactored. NFC. by Sanjay Patel · 11 years ago
  20. 3eb910b Do not introduce new shuffle patterns after operation legalization if SHUFFLE_VECTOR by Owen Anderson · 11 years ago
  21. 50cbfc5 Janitorial services: "Don’t duplicate function or class name at the beginning of the comment." by Sanjay Patel · 11 years ago
  22. e28d57d Remove local TLI vars that are just duplicates of the class var. No functional change. by Sanjay Patel · 11 years ago
  23. 78614bf Use local vars to improve readability. No functional change. by Sanjay Patel · 11 years ago
  24. 159f127 Use local variable in visitFADD. No functional change. by Sanjay Patel · 11 years ago
  25. ae402a3 Group unsafe-math optimizations for fsub into one block. No functional change. by Sanjay Patel · 11 years ago
  26. a828f2b Use local variable to improve readability. by Sanjay Patel · 11 years ago
  27. e1d1294 Simplify creation of a bunch of ArrayRefs by using None, makeArrayRef or just letting them be implicitly created. by Craig Topper · 11 years ago
  28. 4627679 Use range based for loops to avoid needing to re-mention SmallPtrSet size. by Craig Topper · 11 years ago
  29. 2cdea4c name change: isPow2DivCheap -> isPow2SDivCheap by Sanjay Patel · 11 years ago
  30. ff8b883 DAGCombiner: Make concat_vector combine safe for EVTs and concat_vectors with many arguments. by Benjamin Kramer · 11 years ago
  31. 6cc0042 Fix fmul combines with constant splat vectors by Matt Arsenault · 11 years ago
  32. b23bad1 [DAGCombiner] Improve the folding of target independet shuffles to Undef. by Andrea Di Biagio · 11 years ago
  33. 35d3133 optimize vector fneg of bitcasted integer value by Sanjay Patel · 11 years ago
  34. 7cd15be [SDAG] Fix a bug in the DAG combiner where we would fail to return the by Chandler Carruth · 11 years ago
  35. ace8e1e [DAGCombiner] Improved target independent vector shuffle combine rule. by Andrea Di Biagio · 11 years ago
  36. 6b2f5b4 [x86] Fold extract_vector_elt of a load into the Load's address computation. by Michael J. Spencer · 11 years ago
  37. 8e5beb6 Optimize vector fabs of bitcasted constant integer values. by Sanjay Patel · 11 years ago
  38. 40dbd38 [SDAG] Fix a really, really terrible bug in the DAG combiner. by Chandler Carruth · 11 years ago
  39. d913448 Remove the TargetMachine forwards for TargetSubtargetInfo based by Eric Christopher · 11 years ago
  40. cde4eb5 [x86] Don't add nodes to the combined set (and prune subsequent by Chandler Carruth · 11 years ago
  41. 2ef6744 fix for PR20354 - Miscompile of fabs due to vectorization by Sanjay Patel · 11 years ago
  42. ce45be0 [AArch64] Teach DAGCombiner that converting two consecutive loads into a vector load is not a good transform when paired loads are available. by James Molloy · 11 years ago
  43. 1806697 [SDAG] Refactor the code which deletes nodes in the DAG combiner to do by Chandler Carruth · 11 years ago
  44. 9d5a8c2 Fix issues with ISD::FNEG and ISD::FMA SDNodes where they would not be constant-folded by Owen Anderson · 11 years ago
  45. 09b8cde White space fix. by Louis Gerbarg · 11 years ago
  46. 67474e3 Make sure no loads resulting from load->switch DAGCombine are marked invariant by Louis Gerbarg · 11 years ago
  47. 4fc09b3 Retain alignment requirements for load->selects modified by DAGCombine by Louis Gerbarg · 11 years ago
  48. b143274 [SDAG] Add DEBUG logging to the legalizer, fixing a "bug" found by by Chandler Carruth · 11 years ago
  49. 411fb40 [SDAG] When performing post-legalize DAG combining, run the legalizer by Chandler Carruth · 11 years ago
  50. 197a1e2 Store nodes only have 1 result. by Matt Arsenault · 11 years ago
  51. 94bd553 [SDAG] Start plumbing an assert into SDValues that we don't form one by Chandler Carruth · 11 years ago
  52. 9f4530b [SDAG] Introduce a combined set to the DAG combiner which tracks nodes by Chandler Carruth · 11 years ago
  53. cc39b67 AA metadata refactoring (introduce AAMDNodes) by Hal Finkel · 11 years ago
  54. 17020f9 [AArch64] Lower sdiv x, pow2 using add + select + shift. by Chad Rosier · 11 years ago
  55. 9a0051c [SDAG] Make the DAGCombine worklist not grow endlessly due to duplicate by Chandler Carruth · 11 years ago
  56. 3c0012b [SDAG,cleanup] Switch the DAG combiner over to use the spelling by Chandler Carruth · 11 years ago
  57. 24ceb0c [SDAG] Rather than using a narrow test against the one dummy node on the by Chandler Carruth · 11 years ago
  58. 0fb2013 [DAGCombiner] Improve the shuffle-vector folding logic. by Andrea Di Biagio · 11 years ago
  59. 1eb0230 Revert "[x86] Fold extract_vector_elt of a load into the Load's address computation." by Michael J. Spencer · 11 years ago
  60. 7f3e11e CodeGen: don't form illegail EXTLOAD operations. by Tim Northover · 11 years ago
  61. bd5555c [DAGCombiner] Add more rules to fold shuffles. by Andrea Di Biagio · 11 years ago
  62. 2152a6c [DAGCombiner] Avoid calling method 'isShuffleMaskLegal' on illegal vector types. by Andrea Di Biagio · 11 years ago
  63. 3960a95 [DAGCombiner] Add more rules to combine shuffle vector dag nodes. by Andrea Di Biagio · 11 years ago
  64. 67d8b2e [DAGCombiner] Fix a crash caused by a missing check for legal type when trying to fold shuffles. by Andrea Di Biagio · 11 years ago
  65. 3332b70 Revert "Revert r212640, "Add trunc (select c, a, b) -> select c (trunc a), (trunc b) combine."" by Matt Arsenault · 11 years ago
  66. b2921c7 [DAG] Further improve the logic in DAGCombiner that folds a pair of shuffles into a single shuffle if the resulting mask is legal. by Andrea Di Biagio · 11 years ago
  67. f862ce8 Revert r212640, "Add trunc (select c, a, b) -> select c (trunc a), (trunc b) combine." by NAKAMURA Takumi · 11 years ago
  68. cbd44c5 Make it possible for ints/floats to return different values from getBooleanContents() by Daniel Sanders · 11 years ago
  69. 71224b0 [AArch64]Fix an assertion failure in DAG Combiner about concating 2 build_vector. by Hao Liu · 11 years ago
  70. 658c557 Add trunc (select c, a, b) -> select c (trunc a), (trunc b) combine. by Matt Arsenault · 11 years ago
  71. f0a33b7 [SDAG] At the suggestion of Hal, switch to an output parameter that by Chandler Carruth · 11 years ago
  72. d261e98 [DAG] Teach how to combine a pair of shuffles into a single shuffle if the resulting mask is legal. by Andrea Di Biagio · 11 years ago
  73. b844e72 [SDAG] Build up a more rich set of APIs for querying build-vector SDAG by Chandler Carruth · 11 years ago
  74. beeacac [x86] Revert r212324 which was too aggressive w.r.t. allowing undef by Chandler Carruth · 11 years ago
  75. 5d79bb5 [x86] Generalize BuildVectorSDNode::getConstantSplatValue to work for by Chandler Carruth · 11 years ago
  76. f236bb1 Fix ppcf128 component access on little-endian systems by Ulrich Weigand · 11 years ago
  77. 7783b0a Revert "SelectionDAG: Enable (and (setcc x), (setcc y)) -> (setcc (and x, y)) for vectors" by Tom Stellard · 11 years ago
  78. 3787b12 SelectionDAG: Don't use MVT::Other to determine legality of ISD::SELECT_CC by Tom Stellard · 11 years ago
  79. b9a0233 SelectionDAG: Enable (and (setcc x), (setcc y)) -> (setcc (and x, y)) for vectors by Tom Stellard · 11 years ago
  80. f99dd64 [X86] Add target combine rules for horizontal add/sub. by Andrea Di Biagio · 11 years ago
  81. 4db1abe [DAG] Expose NoSignedWrap, NoUnsignedWrap and Exact flags to SelectionDAG. by Andrea Di Biagio · 11 years ago
  82. 446a527 [X86] Add two combine rules to simplify dag nodes introduced during type legalization when promoting nodes with illegal vector type. by Andrea Di Biagio · 12 years ago
  83. 82111f1 Convert a vselect into a concat_vector if possible by Filipe Cabecinhas · 12 years ago
  84. f375d80 [x86] Fold extract_vector_elt of a load into the Load's address computation. by Michael J. Spencer · 12 years ago
  85. 2c77fe5 Revert "[DAGCombiner] Split up an indexed load if only the base pointer value is live" by Hal Finkel · 12 years ago
  86. a0653a3 Rename ComputeMaskedBits to computeKnownBits. "Masked" has been by Jay Foad · 12 years ago
  87. 5d78558 [DAGCombiner] Split up an indexed load if only the base pointer value is live by Adam Nemet · 12 years ago
  88. 820e041 DAGCombine: prevent formation of illegal ConstantFP nodes. by Tim Northover · 12 years ago
  89. 7f6daf1 [ARM64] Prevent bit extraction to be adjusted by following shift by Weiming Zhao · 12 years ago
  90. 2eb60fd Tidy up whitespace. by Jim Grosbach · 12 years ago
  91. 633d99b Convert AddNodeIDNode and SelectionDAG::getNodeIfExiists to use ArrayRef<SDValue> by Craig Topper · 12 years ago
  92. dd5e16d Convert one last signature of getNode to take an ArrayRef of SDUse. by Craig Topper · 12 years ago
  93. da4841b DAGCombiner: Simplify code a bit, make more transforms work with vectors. by Benjamin Kramer · 12 years ago
  94. 48d114b Convert SelectionDAG::getNode methods to use ArrayRef<SDValue>. by Craig Topper · 12 years ago
  95. ad01687 Rip out X86-specific vector SDIV lowering, make the corresponding DAGCombiner transform work on vectors. by Benjamin Kramer · 12 years ago
  96. 4dae598 DAGCombiner: Turn divs of vector splats into vectorized multiplications. by Benjamin Kramer · 12 years ago
  97. c636d15 Fix an infinite loop bug in DAG Combine about keeping transfering between ANY_EXTEND and SIGN_EXTEND. by Hao Liu · 12 years ago
  98. 1b9dde0 [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE by Chandler Carruth · 12 years ago
  99. 863a789 DAGCombiner: don't optimise non-existant litpool load by Tim Northover · 12 years ago
  100. a9bf246 Revert r191049/r191059 as it can produce wrong code (see PR17975). by Robert Lougher · 12 years ago