Commits

Branch selector

 master

User selector

chsigg

Datepicker

All time

Commit History

Commits on Oct 19, 2019

• Get active source lane predicate from shuffle instruction.

  Show description for 6c9d67c
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  6c9d67c

  Copy full SHA for 6c9d67c

Commits on Oct 18, 2019

• Add gpu.barrier op to synchronize invocations of a local workgroup.

  Show description for a744226
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  a744226

  Copy full SHA for a744226

Commits on Oct 16, 2019

• Support custom accumulator provided as region to gpu.all_reduce.

  Show description for 671808b
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  671808b

  Copy full SHA for 671808b

Commits on Oct 10, 2019

• Mark GPU dialect as illegal when lowering to NVVM.

  Show description for 95a9f68
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  95a9f68

  Copy full SHA for 95a9f68

Commits on Oct 9, 2019

• Guard rewriter insertion point during signature conversion.

  Show description for 2ccb836
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  2ccb836

  Copy full SHA for 2ccb836

• Change to doxygen comments. NFC.

  Show description for 99dda8e
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  99dda8e

  Copy full SHA for 99dda8e

• Assert that region is not cloned into itself.

  Show description for f72910a
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  f72910a

  Copy full SHA for f72910a

Commits on Oct 7, 2019

• Change Block::getParent() to be a const function. This is only necessary because ilist_node_with_parent specifically requires a 'getParent() const' method. If/When ilist_node removes this constrain…

  Show description for 9d862a4
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  9d862a4

  Copy full SHA for 9d862a4

• Support reduction of partial warps.

  Show description for 7b968e5
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  7b968e5

  Copy full SHA for 7b968e5

Commits on Oct 4, 2019

• Fix typos, NFC.

  Show description for 140bd33
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  140bd33

  Copy full SHA for 140bd33

Commits on Oct 3, 2019

• Generalize parse/printBinaryOp to parse/printOneResultOp.

  Show description for 55993bf
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  55993bf

  Copy full SHA for 55993bf

• Automated rollback of commit b960daa3e8b8f1c2d0c279508d85e6c668d60b3e

  Show description for 1142923
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  1142923

  Copy full SHA for 1142923

Commits on Oct 1, 2019

• Change all_reduce lowering to support 2D and 3D blocks.

  Show description for b5c488b
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  b5c488b

  Copy full SHA for b5c488b

• Add verification error message for ops that require at least one operand or result.

  Show description for a40bd3f
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  a40bd3f

  Copy full SHA for a40bd3f

• Add integer shift ops to LLVM dialect.

  Show description for 3f18e53
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  3f18e53

  Copy full SHA for 3f18e53

Commits on Sep 30, 2019

• Switch comments from GPU dialect terms to CUDA terms (NFC).

  Show description for 97d0213
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  97d0213

  Copy full SHA for 97d0213

Commits on Sep 26, 2019

• Add AllReduceOp to GPU dialect with lowering to NVVM.

  Show description for 8ad20bc
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  8ad20bc

  Copy full SHA for 8ad20bc

Commits on Sep 24, 2019

• Patch CUB with http://github.com/NVlabs/cub/pull/170 to build with clang.

  Show description for b226c6a
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  b226c6a

  Copy full SHA for b226c6a

• Clone called functions into nested GPU module.

  Show description for 1058df0
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  1058df0

  Copy full SHA for 1058df0

• Clone called functions into nested GPU module.

  Show description for 23c4487
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  23c4487

  Copy full SHA for 23c4487

• Allow null Attribute for value when building GlobalOp.

  Show description for 756c06f
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  756c06f

  Copy full SHA for 756c06f

• Allow null Attribute for value when building GlobalOp.

  Show description for 1fb5535
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  1fb5535

  Copy full SHA for 1fb5535

Commits on Sep 23, 2019

• Outline GPU kernel function into a nested module.

  Show description for 44e8897
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  44e8897

  Copy full SHA for 44e8897

• Outline GPU kernel function into a nested module.

  Show description for 467e9e8
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  467e9e8

  Copy full SHA for 467e9e8

• Fix a number of Clang-Tidy warnings.

  Show description for 1a6f5b7
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  1a6f5b7

  Copy full SHA for 1a6f5b7

• Fix a number of Clang-Tidy warnings.

  Show description for ddef427
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  ddef427

  Copy full SHA for ddef427

Commits on Sep 21, 2019

• Make GlobalOp's value attribute optional.

  Show description for 2ff3122
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  2ff3122

  Copy full SHA for 2ff3122

• Make GlobalOp's value attribute optional.

  Show description for 11b44e0
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  11b44e0

  Copy full SHA for 11b44e0

Commits on Sep 19, 2019

• Add address space attribute to LLVMIR's GlobalOp.

  Show description for 9d1bb32
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  9d1bb32

  Copy full SHA for 9d1bb32

• Fix --config=cuda_clang so that it really uses clang.

  Show description for 866724d
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  866724d

  Copy full SHA for 866724d

• Outline GPU kernel function into a nested module.

  Show description for 40360ed
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  40360ed

  Copy full SHA for 40360ed

Commits on Sep 18, 2019

• Unify error messages to start with lower-case.

  Show description for 63e1945
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  63e1945

  Copy full SHA for 63e1945

Commits on Sep 16, 2019

• Add missing absl dependency.

  Show description for 1e87336
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  1e87336

  Copy full SHA for 1e87336

• Unify how errors are emitted in LaunchFuncOp verification.

  Show description for 27640a5
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  27640a5

  Copy full SHA for 27640a5

• Error out when kernel function is not found while translating GPU calls.

  Show description for b2b3f47
  

  chsigg

  authored and

  tensorflower-gardener

  committed
  b2b3f47

  Copy full SHA for b2b3f47

Pagination

Previous

Next