Commits

Branch selector

 master

User selector

nicolasvasilache

Datepicker

All time

Commit History

Commits on Oct 18, 2019

• Lower vector transfer ops to loop.for operations.

  Show description for 26416a3
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  26416a3

  Copy full SHA for 26416a3

• Implement lowering of VectorTypeCastOp to LLVM

  Show description for ce68983
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  ce68983

  Copy full SHA for ce68983

• Add EDSC support for loop.for operations

  Show description for ca21a79
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  ca21a79

  Copy full SHA for ca21a79

• Fix OSS target name GPUtoNVVMTransforms -> MLIRGPUtoNVVMTransforms

  Show description for 97c4089
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  97c4089

  Copy full SHA for 97c4089

Commits on Oct 17, 2019

• Add EDSC support for loop.for operations

  Show description for 7460e88
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  7460e88

  Copy full SHA for 7460e88

• Decouple Linalg promotion from Linalg tiling - NFC

  Show description for f07d189
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  f07d189

  Copy full SHA for f07d189

• Rename LoopNestBuilder to AffineLoopNestBuilder - NFC

  Show description for 9db8e61
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  9db8e61

  Copy full SHA for 9db8e61

Commits on Oct 16, 2019

• Fix typo in VectorOps.td

  Show description for 77b175e
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  77b175e

  Copy full SHA for 77b175e

Commits on Oct 15, 2019

• Add conversion for splat of vectors of 2+D

  Show description for 83d8620
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  83d8620

  Copy full SHA for 83d8620

Commits on Oct 14, 2019

• Fix linalg.subview behavior in (partially) static cases.

  Show description for 500a4d6
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  500a4d6

  Copy full SHA for 500a4d6

• Add lowering of VectorOps dialect to LLVM to the Linalg LLVM lowering pass

  Show description for 462bc81
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  462bc81

  Copy full SHA for 462bc81

Commits on Oct 7, 2019

• Start a minimal mlir_utils runtime library for testing debugging purposes

  Show description for ba3ebea
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  ba3ebea

  Copy full SHA for ba3ebea

• Support AllocOp terminal in Linalg::AliasAnalysis.

  Show description for 1ae2369
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  1ae2369

  Copy full SHA for 1ae2369

• Fix CMake build after adding TestOpaqueLoc.cpp

  Show description for 656e7a5
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  656e7a5

  Copy full SHA for 656e7a5

Commits on Oct 4, 2019

• Replace constexpr MemRefType::kDynamicStrideOrOffset by a MemRefType:;getDynamicStrideOrOffset() method - NFC

  Show description for bed4130
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  bed4130

  Copy full SHA for bed4130

• Add missing Linalg lowerings to allow roundtrip.mlir to lower to LLVM

  Show description for d7d83ef
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  d7d83ef

  Copy full SHA for d7d83ef

Commits on Oct 3, 2019

• Add syntactic sugar for strided memref parsing.

  Show description for 8ce19db
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  8ce19db

  Copy full SHA for 8ce19db

Commits on Oct 2, 2019

• Extract MemRefType::getStridesAndOffset as a free function and fix dynamic offset determination.

  Show description for e9ca4fa
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  e9ca4fa

  Copy full SHA for e9ca4fa

• Replace spurious `long` stride type by int64_t - NFC

  Show description for 06e6f2f
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  06e6f2f

  Copy full SHA for 06e6f2f

Commits on Oct 1, 2019

• Unify Linalg types by using strided memrefs

  Show description for 0de37a2
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  0de37a2

  Copy full SHA for 0de37a2

Commits on Sep 30, 2019

• Normalize MemRefType lowering to LLVM as strided MemRef descriptor

  Show description for 23a8d43
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  23a8d43

  Copy full SHA for 23a8d43

• Fix MemRefType::getStrides corner case

  Show description for 2a01da0
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  2a01da0

  Copy full SHA for 2a01da0

Commits on Sep 27, 2019

• Add TODO to revisit coupling of CallOp to MemRefType lowering

  Show description for 3b68ab2
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  3b68ab2

  Copy full SHA for 3b68ab2

• Promote MemRefDescriptor to a pointer to struct when passing function boundaries in LLVMLowering.

  Show description for 4c3ddfc
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  4c3ddfc

  Copy full SHA for 4c3ddfc

• Fix JitRunner.cpp Error creation pattern and reactivate tests.

  Show description for b394f9c
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  b394f9c

  Copy full SHA for b394f9c

Commits on Sep 26, 2019

• Decouple tiling from fusion in Linalg.

  Show description for 48a197f
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  48a197f

  Copy full SHA for 48a197f

Commits on Sep 24, 2019

• Normalize lowering of MemRef types

  Show description for 94d0fea
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  94d0fea

  Copy full SHA for 94d0fea

• Normalize lowering of MemRef types

  Show description for 092838f
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  092838f

  Copy full SHA for 092838f

Commits on Sep 20, 2019

• Use SmallVectorImpl in getStrides

  Show description for 8817ec1
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  8817ec1

  Copy full SHA for 8817ec1

• Fix public build

  Show description for 514475e
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  514475e

  Copy full SHA for 514475e

• Use SmallVectorImpl in getStrides

  Show description for 8de9ee5
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  8de9ee5

  Copy full SHA for 8de9ee5

• Fix public build

  Show description for 588c9e2
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  588c9e2

  Copy full SHA for 588c9e2

• Add utility to extract strides from layout map in MemRefType.

  Show description for b1e271c
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  b1e271c

  Copy full SHA for b1e271c

• Add utility to extract strides from layout map in MemRefType.

  Show description for e557e03
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  e557e03

  Copy full SHA for e557e03

Commits on Sep 6, 2019

• Simplify Linalg ABI integration with external function calls.

  Show description for 395214a
  

  nicolasvasilache

  authored and

  tensorflower-gardener

  committed
  395214a

  Copy full SHA for 395214a

Pagination

Previous

Next