Commits · 6f87b50be67a3be9d9b55f1102c3ecd992a5ed28 · Roger Ferrer / llvm-epi · GitLab

Apr 24, 2023
- [mlir][linalg] Add support for lowering pack with outer_dims_perm. · 6f87b50b
  Hanhan Wang authored Apr 21, 2023
  6f87b50b
Apr 22, 2023
- [mlir][linalg] Expose lowerPack and lowerUnPack utils. · ddcc5072
  Hanhan Wang authored Apr 21, 2023
  ddcc5072
Apr 21, 2023
- [MLIR][Vector] Enable masked vectorizaton of contraction ops · c8557c7c
  Diego Caballero authored Apr 21, 2023
  c8557c7c
- [mlir][linalg] Refine `tensor.extract` vectorisation · a3ae3931
  Andrzej Warzynski authored Apr 12, 2023
  a3ae3931
Apr 20, 2023
- [mlir] Prevent implicit downcasting to interfaces · 6089d612
  Rahul Kayaith authored Apr 16, 2023
  6089d612
- [mlir][arith] Add arith.constant materialization helper · 00e3566d
  Rahul Kayaith authored Apr 16, 2023
  00e3566d
- [mlir][Affine][NFC] Wrap dialect in "affine" namespace · 4c48f016
  Matthias Springer authored Apr 20, 2023
  4c48f016
Apr 18, 2023
- [mlir][affine][NFC] Split `reifyValueBound` in two functions · 912fedfb
  Matthias Springer authored Apr 18, 2023
  912fedfb
- [mlir][linalg] Promote operands for convolution vectorization · 7517e246
  Lei Zhang authored Apr 17, 2023
  7517e246
Apr 13, 2023
- [mlir][Vector] Add a masked vectorization of tensor.pad · 8c5ad0a2
  Nicolas Vasilache authored Apr 13, 2023
  8c5ad0a2
- [mlir][Linalg] Fix hoist padding through scf.for iter_arg · b3ed6e54
  Nicolas Vasilache authored Apr 04, 2023
  b3ed6e54
- [mlir][Linalg] NFC - Extract an IndexingUtils from Linalg/Utils · cf9e88ff
  Nicolas Vasilache authored Apr 13, 2023
  cf9e88ff
Apr 12, 2023
- [mlir][Linalg] Allow linalg.copy to be vectorized with masking · 1edfb4b9
  Nicolas Vasilache authored Apr 12, 2023
  1edfb4b9
- [mlir][Linalg] Add support for tiling tensor.pad to scf.forall · 7e5b10b9
  Nicolas Vasilache authored Apr 12, 2023
  7e5b10b9
Apr 11, 2023
- [mlir][linalg] Enable propagation of pack/unpack ops through non-elementwise · b4563ee1
  Quinn Dawkins authored Apr 01, 2023
  b4563ee1
- [mlir][Linalg] NFC - Improve debug messages around padding · caaa8e5e
  Nicolas Vasilache authored Apr 11, 2023
  
  caaa8e5e
- [mlir][linalg] Replace `getUpperBoundForIndex` implementation · eabb6ccd
  Matthias Springer authored Apr 11, 2023
  eabb6ccd
Apr 04, 2023
- [mlir][Linalg] NFC - Refactor hoist padding transform and separate packing... · 1445c11f
  Nicolas Vasilache authored Apr 03, 2023
  1445c11f
Apr 03, 2023
- [mlir][Vector] Enable masking for ops with index semantics · 52177820
  Diego Caballero authored Apr 03, 2023
  52177820
- [mlir][Vector][NFC] Small vector masking clean-up · 04798db4
  Diego Caballero authored Apr 03, 2023
  04798db4
Apr 02, 2023
- [mlir] [Linalg] Limit Detensorize pass to FunctionOpInterface · b6d9e30b
  Felix Schneider authored Apr 02, 2023
  b6d9e30b
Mar 31, 2023
- [mlir][Linalg] Fix Generalize transform · 46cfdfb5
  Nicolas Vasilache authored Mar 29, 2023
  46cfdfb5
Mar 29, 2023
- [mlir][Vector] Enable masked vectorization of linalg.fill · f18a8612
  Diego Caballero authored Mar 29, 2023
  f18a8612
Mar 28, 2023
- [mlir][MemRef] Move transform related functions in Transforms.h · faafd26c
  Quentin Colombet authored Mar 28, 2023
  faafd26c
- [mlir] Apply ClangTidy readability fix (NFC) · cdeaeeeb
  Adrian Kuegel authored Mar 28, 2023
  
  cdeaeeeb
Mar 24, 2023
- [mlir][linalg] Convert input type to accumulator type in im2col patterns · 4c797666
  Quinn Dawkins authored Mar 18, 2023
  4c797666
- [mlir][vector] Hoist redundant singleton vector transfer reads · 6d0de19e
  harsh-nod authored Mar 24, 2023
  6d0de19e
- [mlir][linalg] Use affine apply in im2col gather index calculations · f5150ee3
  Quinn Dawkins authored Mar 24, 2023
  f5150ee3
- [mlir][linalg] Refactor convolution to img2col conversion to use gather semantics · c8280309
  Quinn Dawkins authored Feb 23, 2023
  c8280309
Mar 20, 2023
- [mlir] Apply ClangTidy performance finding (NFC). · 499e1a54
  Adrian Kuegel authored Mar 20, 2023
  
  499e1a54
Mar 16, 2023
- Fix asan error uncovered after landing D145133. · 1da04b04
  Mahesh Ravishankar authored Mar 16, 2023
  1da04b04
- [mlir][TilingInterface] Modify `TilingInterface` methods to better return the... · 809e3d8c
  Mahesh Ravishankar authored Mar 01, 2023
  809e3d8c
- [ADT][mlir] Fix gcc build error · e95d9c92
  Jakub Kuderski authored Mar 15, 2023
  e95d9c92
- [ADT] Allow `llvm::enumerate` to enumerate over multiple ranges · a0a76804
  Jakub Kuderski authored Mar 15, 2023
  a0a76804
Mar 15, 2023
- [ADT][mlir][NFCI] Do not use non-const lvalue-refs with enumerate · 8c258fda
  Jakub Kuderski authored Mar 15, 2023
  8c258fda
Mar 14, 2023
- [mlir][Transform] NFC - Various API cleanups and use RewriterBase in lieu of PatternRewriter · 1cff4cbd
  Nicolas Vasilache authored Mar 13, 2023
  1cff4cbd
Mar 10, 2023
- [mlir][NFC] reifyResultShapes: Add extra error checking · 758329dc
  Matthias Springer authored Mar 10, 2023
  758329dc
Mar 08, 2023
- [mlir][linalg] Downscale 2D convolution with unit dimensions to 1D convolution · 991945f4
  Devajith Valaparambil Sreeramaswamy authored Mar 08, 2023
  991945f4
- [mlir][linalg] Add vectorization support for conv_1d · 5299953a
  Devajith Valaparambil Sreeramaswamy authored Mar 08, 2023
  5299953a
- [mlir][linalg] Refine how contiguous loads are identified · 7a078b65
  Andrzej Warzynski authored Mar 05, 2023
  7a078b65