Commits

Branch selector

 master

User selector

bixia1

Datepicker

All time

Commit History

Commits on Nov 25, 2018

• [XLA:GPU] Convert the reduction implementation to use tiling scheme.

  Show description for 09a7d52
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  09a7d52

Commits on Nov 19, 2018

• [XLA:GPU] Set the requested block size for the tiling scheme to 1.

  Show description for 73f15c6
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  73f15c6

Commits on Nov 15, 2018

• [XLA:GPU] Miscellaneous clean up for 0-2-1 transpose related code.

  Show description for 387a062
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  387a062

Commits on Nov 14, 2018

• [XLA:GPU] Refactor the 0-2-1 tiling implementation to support more use cases.

  Show description for 619c775
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  619c775

Commits on Nov 12, 2018

• [XLA:CPU/GPU] Change the return type for EmitCallToIntrinsic.

  Show description for c398b00
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  c398b00

• [XLA:GPU] Limit the use of the 0-2-1 shared memory transpose implementation.

  Show description for df41194
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  df41194

Commits on Oct 31, 2018

• [XLA] Make PropagateBufferConstraint and PropagateOperandConstraint in the

  Show description for bef87d7
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  bef87d7

Commits on Oct 30, 2018

• [XLA:GPU] Remove the implicit reference of the last kernel in the thunk sequence

  Show description for f7e4143
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  f7e4143

Commits on Oct 26, 2018

• [XLA:GPU] Remove a map data structure used to support the code generation of

  Show description for 5ef3c95
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  5ef3c95

Commits on Oct 25, 2018

• [XLA:GPU] Temporarily disable the check for non-layout-changing instructions

  Show description for 822337e
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  822337e

Commits on Oct 19, 2018

• [XLA:GPU] Code clean up related to the manipulation of thunk sequence in the

  Show description for 121a06d
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  121a06d

Commits on Oct 5, 2018

• [XLA] Extend the HLO verifier to check that non-layout-changing instructions

  Show description for 03b4161
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  03b4161

Commits on Oct 3, 2018

• [XLA] Revise the way to express a CPU specific test.

  Show description for 207bea0
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  207bea0

• [XLA] Disable a test for layout changing elementwise operations.

  Show description for 2e11deb
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  2e11deb

• [XLA] Modify the function that determines whether an instruction can change

  Show description for bbe15ee
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  bbe15ee

Commits on Oct 2, 2018

• [XLA] A test that disables layout assignment should only contain layout

  Show description for 44f273e
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  44f273e

Commits on Sep 6, 2018

• [XLA:GPU] Refactor some code for fusion output handling.

  Show description for 1e9cbcc
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  1e9cbcc

Commits on Sep 4, 2018

• [XLA] Add a test case for propagating the result layout of a non-elementwise

  Show description for 5bb543d
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  5bb543d

Commits on Aug 28, 2018

• [XLA] Avoid implicit layout change for non-layout-changing instructions.

  Show description for bf7b20e
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  bf7b20e

Commits on Aug 10, 2018

• [XLA] Enhance the HLO verifier to report errors for illegal Rng instructions.

  Show description for 539b48b
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  539b48b

Commits on Aug 9, 2018

• [XLA] Add parameter allow_mixed_precision used for creating the HloVerifier

  Show description for e396384
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  e396384

Commits on Aug 7, 2018

• [XLA:CPU/GPU] Turn on random_ops_test for the CPU and GPU backends.

  Show description for f53c888
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  f53c888

• [XLA:CPU/GPU] Fix a problem in the random number generator for F16.

  Show description for bd3f56b
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  bd3f56b

Commits on Aug 6, 2018

• [XLA:CPU/GPU] Fix a problem in the random number generator.

  Show description for 7cc2dae
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  7cc2dae

• [TF:XLA] Fix testRandomUniformIsNotConstant for F16.

  Show description for 2c6ae17
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  2c6ae17

Commits on Jul 31, 2018

• [TF] Add test cases to check that random ops are stateful.

  Show description for 217dd20
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  217dd20

Commits on Jul 26, 2018

• [XLA:GPU] Enable TF RNG ops for the GPU backend.

  Show description for 8786b41
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  8786b41

Commits on Jul 25, 2018

• [XLA:CPU/GPU] Implement the parallel Philox random number generation algorithm.

  Show description for 7ff06d7
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  7ff06d7

Commits on Jul 24, 2018

• [XLA:GPU] Add an operator<< to Thunk::Kind.

  Show description for fada40a
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  fada40a

Commits on Jul 20, 2018

• [XLA] Increase the sample size for PrngTest.Uniformity256.

  Show description for 9f8256a
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  9f8256a

Commits on Jul 4, 2018

• [XLA:GPU] Enhance the tiled 0-2-1 transpose algorithm to handle fusion.

  Show description for eb8a110
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  eb8a110

Commits on Jul 2, 2018

• [XLA] Remove a bogus invalid argument message printed out when --v>=1.

  Show description for 97a040a
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  97a040a

Commits on Jun 30, 2018

• [XLA] Remove a bogus invalid argument message printed out when --v>=1.

  Show description for 26a6060
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  26a6060

Commits on Jun 22, 2018

• [XLA] Teach algebraic simplifier to convert a copy instruction to a bitcast

  Show description for 40ff081
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  40ff081

Commits on Jun 20, 2018

• Allow the use of 32 bit integer type for loop index and tensor element index.

  Show description for 4283949
  

  bixia1

  authored and

  tensorflower-gardener

  committed
  4283949

Pagination

PreviousNext