Commits · c27d815249638df29957e117c9e351a3e8aabc3b · Roger Ferrer / llvm-epi · GitLab

Jul 14, 2022
- [mlir] Use value instead of getValue (NFC) · c27d8152
  Kazu Hirata authored Jul 14, 2022
  
  c27d8152
Jul 13, 2022
- [mlir][NVGPU] Verifiers for nvgpu.mma.sync Op · f7d42d51
  Manish Gupta authored Jul 13, 2022
  f7d42d51
- lowerParallel is also called on unit-size, one-sided reduction dims · 6870a50f
  Benoit Jacob authored Jul 04, 2022
  6870a50f
- [mlir] Plumb through default attribute populate for extensible dialect. · 69b6454f
  Jacques Pienaar authored Jul 13, 2022
  
  69b6454f
- [mlir] Use has_value instead of hasValue (NFC) · 491d2701
  Kazu Hirata authored Jul 13, 2022
  
  491d2701
- [mlir] Allow empty lists for DenseArrayAttr. · aabfaf90
  Adrian Kuegel authored Jul 13, 2022
  aabfaf90
- [mlir][vector] Fix crash in vector.reduction canonicalization · 5f8cefeb
  Thomas Raoux authored Jul 12, 2022
  5f8cefeb
- [mlir][VectorToLLVM] Fix bug in lowering of vector.reduce fmax/fmin · 8fe076ff
  Thomas Raoux authored Jul 12, 2022
  8fe076ff
Jul 12, 2022
- [mlir] Switch create to use NamedAttrList&& · 0db084d4
  Jacques Pienaar authored Jul 12, 2022
  0db084d4
- [mlir] Remove VectorToROCDL · d6ef3d20
  Krzysztof Drewniak authored Jul 07, 2022
  d6ef3d20
- [mlir] fold more eagerly in structured op splitting · a5c802a4
  Alex Zinenko authored Jul 08, 2022
  a5c802a4
- [mlir][vector] Add accumulator operand to MultiDimReduce op · 051b36ba
  Thomas Raoux authored Jul 12, 2022
  051b36ba
- [mlir] Handle linalg.index correctly in TilingInterface · 81b62f7f
  Alex Zinenko authored Jul 08, 2022
  81b62f7f
- [mlir] Use semantically readable functions for transform op effects · e15b855e
  Alex Zinenko authored Jul 08, 2022
  e15b855e
- [mlir] Transform op for multitile size generation · 3963b4d0
  Alex Zinenko authored Jul 07, 2022
  3963b4d0
- [mlir] Allow Tile transform op to take dynamic sizes · 4e4a4c05
  Alex Zinenko authored Jul 07, 2022
  4e4a4c05
- [mlir] Add ReplicateOp to the Transform dialect · 00d1a1a2
  Alex Zinenko authored Jul 07, 2022
  00d1a1a2
- [mlir][Math] Support fold SqrtOp with constant dense. · 4d7d5c5f
  Jianjian GUAN authored Jul 11, 2022
  4d7d5c5f
- Read/write external resource alignment tag in little-endian · de9a7260
  Ulrich Weigand authored Jul 12, 2022
  de9a7260
- Fix linalg.dot over boolean tensors. · ad3a0787
  Johannes Reifferscheid authored Jul 07, 2022
  ad3a0787
- [mlir] Add support for regex within `expected-*` diagnostics · 7306dc91
  River Riddle authored Jul 07, 2022
  7306dc91
Jul 11, 2022
- [mlir][sparse] implement sparse2sparse reshaping (expand/collapse) · faa00c13
  Aart Bik authored Jul 08, 2022
  faa00c13
- [mlir][complex] Lower complex.log to libm log call · 7769505a
  Kai Sasaki authored Jul 11, 2022
  7769505a
- [OMPIRBuilder] Add support for simdlen clause · ac892c70
  Prabhdeep Singh Soni authored Jul 11, 2022
  ac892c70
- [MLIR][TOSA] Fix converting tosa.clamp and tosa.relu to linalg · 6e8e91a7
  jungpark-mlir authored Jul 11, 2022
  6e8e91a7
- [mlir][vector] Add pattern to distribute splat constant · 0af26805
  Thomas Raoux authored Jul 11, 2022
  0af26805
- [mlir][vector] Avoid creating duplicate output in warpOp · d7d6443d
  Thomas Raoux authored Jul 11, 2022
  d7d6443d
- [MLIR][Presburger] introduce MPInt to support fast arbitrary precision in Presburger · cdbc5f1e
  Arjun P authored Jul 11, 2022
  cdbc5f1e
- Revert "[MLIR][Presburger] introduce MPInt to support fast arbitrary precision in Presburger" · 1e10d35e
  Arjun P authored Jul 11, 2022
  1e10d35e
- [MLIR][Presburger] introduce MPInt to support fast arbitrary precision in Presburger · c9035df2
  Arjun P authored Jul 11, 2022
  c9035df2
- [mlir] Flip accessors to prefixed form (NFC) · 136d746e
  Jacques Pienaar authored Jul 10, 2022
  136d746e
- [mlir][Math] Support fold Log2Op with constant dense. · ad4b7fb3
  Jianjian GUAN authored Jul 07, 2022
  ad4b7fb3
Jul 09, 2022
- [mlir][vector] Relax reduction distribution pattern · 0660f3c5
  Thomas Raoux authored Jul 09, 2022
  0660f3c5
- [mlir][bufferization] Do not canonicalize to_tensor(to_memref(x)) · fc9b37dd
  Matthias Springer authored Jul 09, 2022
  fc9b37dd
- [mlir:LSP] Drop potentialy annoying completion commit characters · 9bbc0d4e
  River Riddle authored Jul 08, 2022
  9bbc0d4e
- [mlir:LSP] Add code completions for builtin signed/unsigned integers · 34b3f066
  River Riddle authored Jul 08, 2022
  
  34b3f066
- [mlir:LSP] Add support for code completing attributes and types · fe4f512b
  River Riddle authored Jul 06, 2022
  fe4f512b
- [mlir:LSP] Add support for keyword code completions · 2e41ea32
  River Riddle authored Jul 06, 2022
  2e41ea32
Jul 08, 2022
- [mlir] Add method to populate default attributes · 82140ad7
  Jacques Pienaar authored Jul 08, 2022
  82140ad7
- [mlir][tosa] Enable decomposing Conv2D also where 1 input dim is dynamic · e08a991f
  Jacques Pienaar authored Jul 08, 2022
  e08a991f