Commits · 0f8b6686ac288cda8d14d2ec5b8ca98d188b0684 · Roger Ferrer / llvm-epi · GitLab

Apr 29, 2021
- [mlir] fix shared-lib build · 28ab7ff2
  Alex Zinenko authored Apr 29, 2021
  
  28ab7ff2
- [mlir] support max/min lower/upper bounds in affine.parallel · 6841e6af
  Alex Zinenko authored Apr 29, 2021
  6841e6af
- [mlir] Affine: parallelize affine loops with reductions · 545fa378
  Alex Zinenko authored Apr 29, 2021
  545fa378
- [mlir] Fix top-level comments (NFC) · de94b185
  Lorenzo Chelini authored Apr 29, 2021
  
  de94b185
- [mlir] Add LinalgTransforms dependency on Complex · b863af5a
  Tres Popp authored Apr 29, 2021
  
  b863af5a
- [mlir] Support complex numbers in Linalg promotion · 42e5f422
  Tres Popp authored Apr 06, 2021
  42e5f422
- [mlir] Split out Python bindings entry point into a separate file · ac0a70f3
  Alex Zinenko authored Apr 22, 2021
  ac0a70f3
- [MLIR][Shape] Fix `shape.broadcast` to standard lowering · eb56fa97
  Frederik Gossen authored Apr 29, 2021
  eb56fa97
- [mlir][Linalg] Generalize linalg vectorization · b6113db9
  Nicolas Vasilache authored Apr 28, 2021
  b6113db9
- [mlir][Python][Linalg] Adding const, capture, and index support to the OpDSL. · c2be2cda
  Tobias Gysi authored Apr 29, 2021
  c2be2cda
- [mlir][python] Add `destroy` method to PyOperation. · 49745f87
  Mike Urbach authored Apr 23, 2021
  49745f87
- [mlir] Move PyConcreteType to header. NFC. · 32e2fec7
  John Demme authored Apr 28, 2021
  32e2fec7
Apr 28, 2021
- [mlir][python] Update `PyOpResult.owner` to get the parent object. · 6ff74f96
  Mike Urbach authored Apr 23, 2021
  6ff74f96
- [mlir][Linalg] Avoid changing the rank of the result in canonicalizations of subtensor. · 41849a91
  MaheshRavishankar authored Apr 28, 2021
  41849a91
- [mlir] Fix canonicalization of tiled_loop if not all opresults fold. · fa0d044c
  Alexander Belyaev authored Apr 28, 2021
  fa0d044c
- Revert "[MLIR][Shape] Concretize broadcast result type if possible" · 511ffe17
  Frederik Gossen authored Apr 28, 2021
  511ffe17
- [mlir][python] Add basic python support for GPU dialect and passes · b87219f7
  Nicolas Vasilache authored Apr 28, 2021
  b87219f7
- [mlir][python] Add python support for async dialect and passes. · e7db8408
  Nicolas Vasilache authored Apr 27, 2021
  e7db8408
- [mlir] Fix the postsubmit comments in https://reviews.llvm.org/D101445 · 9a66d334
  Alexander Belyaev authored Apr 28, 2021
  
  9a66d334
- [mlir] Add folding for tensor inputs and memref.cast in linalg.tiled_loop. · 29dbac0a
  Alexander Belyaev authored Apr 28, 2021
  29dbac0a
- [MLIR] Add ComplexToStandard conversion pass. · 2ea7fb7b
  Adrian Kuegel authored Apr 28, 2021
  2ea7fb7b
- [mlir] Fix typos (NFC) · 41b86d8a
  Lorenzo Chelini authored Apr 28, 2021
  
  41b86d8a
- [MLIR][Shape] Concretize broadcast result type if possible · dca53610
  Frederik Gossen authored Apr 28, 2021
  dca53610
- [MLIR][Shape] Canonicalize casted extent tensor operands · cb393f4c
  Frederik Gossen authored Apr 28, 2021
  cb393f4c
- [MLIR][Shape] Derive more concrete type for `shape.shape_of` · 3e037f8f
  Frederik Gossen authored Apr 28, 2021
  3e037f8f
- [mlir][Python][Linalg] Fixing typos (NFC). · 3071107c
  Tobias Gysi authored Apr 28, 2021
  
  3071107c
- [MLIR] Add and propagate section attribute for LLVM_GlobalOp · b65472d6
  Ranjith Kumar H authored Apr 28, 2021
  b65472d6
- [mlir] Support setting operand values in C and Python APIs. · 63d16d06
  Mike Urbach authored Apr 23, 2021
  63d16d06
- [MLIR][Python] Add capsule methods for pybind11 to PyValue. · 3f3d1c90
  Mike Urbach authored Apr 22, 2021
  3f3d1c90
- Handle the case of tile and pad a subset of the dimensions · 7fe20634
  Ahmed Taei authored Apr 26, 2021
  7fe20634
- [tosa][mlir] Fix FullyConnected to correctly order dimensions · cc1ae54e
  Rob Suderman authored Apr 23, 2021
  cc1ae54e
- [mlir][tosa] Add tosa.negate lowerings for quantized cases · 8f190b13
  Rob Suderman authored Apr 23, 2021
  8f190b13
Apr 27, 2021
- [mlir] Fix bug in ForwardDataFlowAnalysis solver · 82bcd985
  River Riddle authored Apr 27, 2021
  82bcd985
- [MLIR] Debug log IR after pattern applications · 6176a8f9
  Frederik Gossen authored Apr 27, 2021
  6176a8f9
- [MLIR][Shape] Remove empty extent tensor operands · f8d7bd99
  Frederik Gossen authored Apr 27, 2021
  f8d7bd99
- [MLIR][Shape] Replace single operand broadcasts with appropriate cast · 2b9b999d
  Frederik Gossen authored Apr 27, 2021
  2b9b999d
- [mlir] Add a pass to tile Linalg ops using `linalg.tiled_loop`. · 4b13b758
  Alexander Belyaev authored Apr 27, 2021
  4b13b758
- [MLIR][Linalg] Generalize splat constant folding · b003ebd6
  Frederik Gossen authored Apr 26, 2021
  b003ebd6
- [mlir] Refactor the forward dataflow propagation in SCCP into a generic framework · d07c90e3
  River Riddle authored Apr 26, 2021
  d07c90e3
- [mlir][tensors] Introduce attribute interface/attribute for tensor encoding · 23c9e8bc
  Aart Bik authored Apr 26, 2021
  23c9e8bc