Google Git
Sign in
rust/EnzymeAD/Enzyme/HEAD
  1. 6a21ecc Compute AugmentWithJuliaObjectType offsets from the DataLayout (#3184) by Valentin Churavy · 3 days ago main
  2. 272285f Make extractBLAS allocation-free (#3183) by Valentin Churavy · 6 days ago v0.0.292
  3. 0c45114 Add clang enzyme_notypeanalysis and enzyme_ta_norecur attributes (#3072) by Mark Abate · 8 days ago
  4. 3e7cc67 MLIR: give the affine atomic an ordering, and lower it to the enzyme one (#3179) by William Moses · 10 days ago
  5. 4ef11d1 Use GlobalVariable::getAlign, which replaced getAlignment (#3178) by William Moses · 10 days ago
  6. 277ecb5 [mlir] Batch AD ops (#3175) by Paul Berg · 2 weeks ago
  7. e55a996 Forward terminator handler: shadow only results the terminator forwards (#3173) by William Moses · 2 weeks ago
  8. fc6bb33 Emit diagnostic instead of segfaulting when CreateReverseDiff/CreateForwardDiff returns null (#3172) by Copilot · 2 weeks ago v0.0.291
  9. 44156f3 Fix tanh derivative: use 1-tanh(x)^2 instead of 1/cosh(x)^2 for numerical stability (#2908) by Copilot · 2 weeks ago
  10. cea1c8e Drop RustDebugInfo.cpp from the bazel build (#3171) by William Moses · 2 weeks ago
  11. c47f4c6 Remove use of deprecated llvm::PointerUnion::get (#3157) by Mark Abate · 2 weeks ago
  12. c445feb MLIR: never merge identical blocks in the region-inlining sweeps (#3166) by William Moses · 2 weeks ago
  13. 16d3db7 Add Fortran OpenMP tests (#3160) by Joe Wallwork · 2 weeks ago
  14. ee0d955 Add workflow for Shell linting (#3144) by Joe Wallwork · 2 weeks ago
  15. 71f672b Add MLIR derivatives for `math.exp2` and `math.log2` (#3155) by Vimarsh Sathia · 3 weeks ago
  16. 3fdebd6 Update LLVM 19 guards for trigonometric functions (#3153) by Nicholson Koukpaizan · 4 weeks ago
  17. 2593327 mlir: Use the checkpoint period for the outer trip count (#3151) by Paul Berg · 4 weeks ago
  18. 33c495b mlir: Allocate the $b$ checkpoints at once in binomial checkpointing (#3147) by Paul Berg · 4 weeks ago
  19. 0338786 PreserveNVVM: the primal of a custom rule is pinned in two places (#3149) by William Moses · 4 weeks ago
  20. af01140 PreserveNVVM: let a pipeline without custom rules drop their linkage (#3148) by William Moses · 4 weeks ago
  21. d101bb0 MLIR: forward mode for llvm.extractvalue and llvm.insertvalue (#3141) by William Moses · 4 weeks ago
  22. 7500f31 MLIR: forward mode for affine.parallel (#3140) by William Moses · 4 weeks ago
  23. c0a377e Add a sumMIT integration test for the Enzyme AD material tests (#3122) by William Moses · 4 weeks ago
  24. cd98af9 MLIR: one seam for the shadow of an inactive stored value (#3139) by William Moses · 4 weeks ago
  25. d06faa4 MLIR: memcpy derivatives and shadow stores of pointers (#3132) by William Moses · 4 weeks ago
  26. 7d8105c MLIR: shadow stores of mutable values in reverse mode (#3136) by William Moses · 4 weeks ago
  27. f94a37e MLIR: derivatives for enzyme.affine_atomic_rmw (#3134) by William Moses · 4 weeks ago
  28. 01f39e7 MLIR: one getBaseObject for gradient utils and alias queries (#3133) by William Moses · 4 weeks ago
  29. fa6e20e Use hasTerminator() in freeCache (#3125) by Manuel Drehwald · 5 weeks ago
  30. 22c3d42 Use original function's LoopInfo in legalRecompute (#3124) by Manuel Drehwald · 5 weeks ago
  31. e0d0dee MLIR: honor enzyme_dupnoneed by eliding primal stores (#3121) by William Moses · 5 weeks ago
  32. 54e6d5c Support LLVM 24's BranchInst split via branch helpers, keeping every earlier major (#3119) by William Moses · 5 weeks ago
  33. 1b8c5bf MLIR: differentiate through debug intrinsics (#3117) by William Moses · 5 weeks ago
  34. 3587b69 ReverseRetOpt: bail on ops whose outputs do not line up with their activities (#3118) by William Moses · 5 weeks ago
  35. ab0bd5b MLIR: carry access alignment into memref/LLVM/affine load-store adjoints (#3115) by William Moses · 5 weeks ago
  36. 7f91b3b Recognize the CUDA deallocations (#3112) by William Moses · 5 weeks ago
  37. 26050c4 Declare helper entry points under the caller's naming convention (#3111) by William Moses · 5 weeks ago
  38. 6ff84be Utils: read argument-memory effects in isReadOnly/isWriteOnly (#3103) by William Moses · 5 weeks ago
  39. a33f4ea4 Differentiate an llvm.call (#3079) by William Moses · 5 weeks ago
  40. 22af85a Route getFirstNonPHI through a Utils.h wrapper (#3081) by Valentin Churavy · 5 weeks ago
  41. db618fa Size a cloned or zeroed alloca from its type (#3108) by William Moses · 5 weeks ago v0.0.290
  42. 124dc49 mlir: unify checkpointing implementations (#3092) by Paul Berg · 5 weeks ago
  43. e2ae4c0 Describe the __enzyme_* call markers in one place (#3086) by William Moses · 5 weeks ago
  44. 002eed6 MLIR: differentiating a bodyless function is an error, not a crash or a zero (#3104) by William Moses · 5 weeks ago
  45. 9a5a4fc MLIR: forward-mode tangent for arith.select (#3107) by William Moses · 5 weeks ago
  46. c1c312e MLIR dataflow activity: a summarized function still needs its own analysis (#3105) by William Moses · 5 weeks ago
  47. 5525eba Give the derivative its own comdat rather than the primal's (#3100) by William Moses · 5 weeks ago
  48. 324a167 Remove a pop nobody reads when its cache has no other pop (#3099) by William Moses · 5 weeks ago
  49. 4f44481 Give llvm.intr.memset a forward-mode rule (#3098) by William Moses · 5 weeks ago
  50. 1d575ed Keep a branch's condition when neither successor takes an argument (#3096) by William Moses · 5 weeks ago
  51. 4313651 Say an unknown summary set the way the readers of summaries read it (#3097) by William Moses · 5 weeks ago
  52. 73a5c7e Decide whether to skip an op with the analyzer everything else uses (#3095) by William Moses · 5 weeks ago
  53. 21bd8a9 Give lifetime markers a forward-mode rule (#3093) by William Moses · 5 weeks ago
  54. b88dd07 Do not walk one map while inserting into it (#3094) by William Moses · 5 weeks ago
  55. 84c6fae Give each differentiated function its own dataflow solver (#3091) by William Moses · 5 weeks ago
  56. f411cef Join a map of sets by what the other side has (#3090) by William Moses · 5 weeks ago
  57. d486232 Prefer annotations for Enzyme's registration attributes, fixing templates (#3085) by William Moses · 5 weeks ago
  58. caf58be A result nothing differentiates has no shadow to be given (#3088) by William Moses · 5 weeks ago
  59. 3174c81 Provide a fpm.toml for the Fortran interface (#3080) by Valentin Churavy · 5 weeks ago
  60. 6d05b52 Differentiate an llvm.func in forward mode (#3087) by William Moses · 5 weeks ago
  61. 9f8e67c Load adjoint: a mutable type takes a shadow, not an accumulation (#3078) by William Moses · 5 weeks ago
  62. 715e9c0 Give math.fma a derivative (#3077) by William Moses · 5 weeks ago
  63. 84b85f8 Add exp10 intrinsic for LLVM >=18 (#2998) (#3001) by Benjamin Coveler · 5 weeks ago
  64. 1361dee Report mismatched activity as a warning, not just an -Rpass remark (#3076) by William Moses · 5 weeks ago
  65. 7a2a626 Drop identity address-space casts in the Julia alloca lowering (#3074) by William Moses · 5 weeks ago
  66. d4f78fe Use Utils.h wrappers for version-sensitive LLVM APIs, and lint for it (#3071) by William Moses · 5 weeks ago
  67. 50c11dc Serialize TypeTree offsets as signed constants (#3070) by William Moses · 5 weeks ago
  68. 381ed79 Fix compilation warnings (#3068) by Paul Berg · 5 weeks ago
  69. 8ead65c Add batch mode test (#3053) by Joe Wallwork · 5 weeks ago
  70. 83ebc86 Separate out Fortran tests at different optimisation levels (#3054) by Joe Wallwork · 5 weeks ago
  71. 38a4b38 Add more Fortran tests (#3020) by Joe Wallwork · 5 weeks ago
  72. 1db6e64 build an actual flang plugin instead of symlinking (#3056) by Valentin Churavy · 5 weeks ago
  73. 5c6240c Add pass to hoist allocations (#3051) by Paul Berg · 5 weeks ago
  74. 4f4e9c4 Preserve attributes on loops built while differentiating (#3059) by William Moses · 5 weeks ago
  75. dcf44fb MLIR fast math derivatives (#3057) by William Moses · 5 weeks ago
  76. 8c8ce94 Make the min cut graph's adjacency iteration deterministic (#3058) by William Moses · 5 weeks ago
  77. a19edda mlir: Simplify enzyme ops before each remover, not only once per function (#3050) by William Moses · 6 weeks ago
  78. f9c0694 mlir: Only delegate loop-carried caches to the loop remover (#3049) by William Moses · 6 weeks ago
  79. d9bbeb0 Properly consider ptrsizehit addr (#3047) by William Moses · 6 weeks ago
  80. 4974fa0 Bind IRBuilder calls that share an argument list to locals (#3046) by William Moses · 6 weeks ago
  81. 24f30b2 Add a lint job for order-dependent IR emission (#3045) by William Moses · 6 weeks ago
  82. cdad01e Bind op-creating calls to locals so IR emission order is deterministic (#3044) by William Moses · 6 weeks ago
  83. 4c5c97a mlir: Fix Binomial checkpointing schedule and support for mutable memory + memory space (#3042) by William Moses · 6 weeks ago
  84. a9b96ed [TA] Don't recurse infinitely on self-referential constant globals (#3038) by William Moses · 6 weeks ago
  85. 6598bf0 mlir: fix overly conservative check when marking readonly (#3039) by Jacob Mai Peng · 6 weeks ago
  86. da4a9aa Don't add non-instructions to the min-cut recompute graph (#3035) by William Moses · 6 weeks ago
  87. 8335ea9 Relax dynamic-loop derivative signature check (#3021) by Christopher Albert · 6 weeks ago
  88. a90dc1d Avoid getTerminator assertion on incomplete derivative block (#3023) by Christopher Albert · 6 weeks ago
  89. 755e0cb mlir: loop-invariant cache for cached allocas (#3030) by Jacob Mai Peng · 6 weeks ago
  90. e2a7fce fix (#3033) by Aiden Grossman · 6 weeks ago
  91. 74e17f5 Remove deprecated use of make_scope_exit (#3034) by Aiden Grossman · 6 weeks ago
  92. f3955b2 Prefer OptionalPassInfoMixin (#2996) by Aiden Grossman · 6 weeks ago
  93. 0ed71b2 [mlir] Make non-looping, non-nested RegionBranchOpInterface ops movable by min-cut (#3024) by Jacob Mai Peng · 6 weeks ago
  94. e4afe54 Speed up large constant ta (#3002) by William Moses · 6 weeks ago
  95. 65d455a mlir: mark loads of readonly pointers as movable during mincut (#2734) by Jacob Mai Peng · 6 weeks ago
  96. 76e110a [MLIR][Mincut] Prefer last equal sized element in chain to cache (#3017) by William Moses · 6 weeks ago
  97. a7c8be5 mlir: mark outlined functions from enzyme.autodiff_region private (#3015) by Jacob Mai Peng · 6 weeks ago
  98. 891aa61 [MLIR][Mincut] Add test for optimality issue for mincut (#3013) by William Moses · 6 weeks ago
  99. 8551ebb Fixes for mutable memory in scf.for reverse (#3009) by Paul Berg · 6 weeks ago
  100. 1316e42 Update to later llvm (#3007) by William Moses · 6 weeks ago