Commits

Breadcrumbs

History for

onquantize-graph-docs

User selector

All users

All time

Commit History

Commits on Jan 13, 2017

• Fix the Windows cmake build.

  Show description for 2cf39e0
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  2cf39e0

  Copy full SHA for 2cf39e0

Commits on Jan 12, 2017

• Switch open-source to use jemalloc for CPU Tensor memory allocation, gRPC, and other places where we call malloc/free.

  Show description for 83c6e0c
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  83c6e0c

  Copy full SHA for 83c6e0c

Commits on Dec 13, 2016

• Automated rollback of change 141555939

  Show description for 77d641c
  

  ringw

  authored and

  tensorflower-gardener

  committed
  77d641c

  Copy full SHA for 77d641c

Commits on Dec 9, 2016

• Automated rollback of change 141209775

  Show description for 63dc340
  

  ringw

  authored and

  tensorflower-gardener

  committed
  63dc340

  Copy full SHA for 63dc340

• Merge changes from github.

  Show description for 1cb9689
  

  andrewharp

  authored and

  tensorflower-gardener

  committed
  1cb9689

  Copy full SHA for 1cb9689

Commits on Dec 6, 2016

• Add a ParallelForWithWorkerId function.

  Show description for d0f8bda
  

  ringw

  authored and

  tensorflower-gardener

  committed
  d0f8bda

  Copy full SHA for d0f8bda

Commits on Nov 17, 2016

• Merge changes from github.

  Show description for 54e5000
  tensorflower-gardener

  committed
  54e5000

  Copy full SHA for 54e5000

Commits on Nov 2, 2016

• Delete old Eigen threadpool and associated code.

  Show description for bdd7615
  tensorflower-gardener

  committed
  bdd7615

  Copy full SHA for bdd7615

Commits on Nov 1, 2016

• Added the ability to wait for queues to start running before returning from

  Show description for 5de9d3c
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  5de9d3c

  Copy full SHA for 5de9d3c

Commits on Oct 25, 2016

• Fix reversed sense of #ifndef NDEBUG in definition of TF_DCHECK_OK.

  Show description for e3c67f5
  tensorflower-gardener

  committed
  e3c67f5

  Copy full SHA for e3c67f5

Commits on Oct 14, 2016

• Add TF_DCHECK_OK to status.h, a debug-only version of

  Show description for a176836
  tensorflower-gardener

  committed
  a176836

  Copy full SHA for a176836

Commits on Oct 11, 2016

• Reduce lock contention in BlockingCounter. This matters mostly when the shard count is large, so this change does not have a significant effect on existing use cases, where typically count == numbe…

  Show description for fc3b501
  tensorflower-gardener

  committed
  fc3b501

  Copy full SHA for fc3b501

• Small header changes and a Doxygen tweak to shape the C++ API a bit better.

  Show description for 120c3f1
  tensorflower-gardener

  committed
  120c3f1

  Copy full SHA for 120c3f1

Commits on Sep 23, 2016

• Automated rollback of change 133885298

  Show description for 527e5ca
  tensorflower-gardener

  committed
  527e5ca

  Copy full SHA for 527e5ca

Commits on Sep 21, 2016

• Automated rollback of change 133839671

  Show description for 0ae4e2e
  tensorflower-gardener

  committed
  0ae4e2e

  Copy full SHA for 0ae4e2e

• Add bitmap implementation to tensorflow/core/lib/core

  Show description for 9717135
  tensorflower-gardener

  committed
  9717135

  Copy full SHA for 9717135

Commits on Sep 9, 2016

• Remove (unused) POSIX-specific include from platform-independent code.

  Show description for b3e3089
  

  mrry

  authored and

  tensorflower-gardener

  committed
  b3e3089

  Copy full SHA for b3e3089

Commits on Sep 8, 2016

• Adjust the cost_per_unit for Shard() in SparseFeatureCrossOp to reduce Context

  Show description for 3b784e5
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  3b784e5

  Copy full SHA for 3b784e5

Commits on Aug 26, 2016

• io::InputBuffer: add ReadVarint32().

  Show description for 2f3328b
  

  Zongheng Yang

  authored and

  tensorflower-gardener

  committed
  2f3328b

  Copy full SHA for 2f3328b

Commits on Aug 25, 2016

• Fix a lost wakeup bug in `Notification::WaitForNotificationWithTimeout()`.

  Show description for b04c9f4
  

  mrry

  authored and

  tensorflower-gardener

  committed
  b04c9f4

  Copy full SHA for b04c9f4

Commits on Aug 24, 2016

• Move host_info.h internal, move public functions into cpu_info.h

  Show description for 46231cf
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  46231cf

  Copy full SHA for 46231cf

Commits on Aug 2, 2016

• Allocate Task explicitly to make sizeof(Task) smaller and to speed

  Show description for bec135b
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  bec135b

  Copy full SHA for bec135b

Commits on Aug 1, 2016

• Make default construction of Tasks cheaper.

  Show description for 8ca0466
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  8ca0466

  Copy full SHA for 8ca0466

Commits on Jul 23, 2016

• Performance improvements for gRPC-based TensorFlow system. Rather than

  Show description for ceff248
  tensorflower-gardener

  committed
  ceff248

  Copy full SHA for ceff248

Commits on Jul 19, 2016

• Add Context and WithContext classes to platform. The ThreadPool will construct a Context upon a request to schedule a callback and set the context when calling the callback in a different thread.

  Show description for c30ad8d
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  c30ad8d

  Copy full SHA for c30ad8d

Commits on Jul 6, 2016

• Parse tf.Example using arena allocation. Dense string benchmark shows 10-20% speedup.

  Show description for c4c19f1
  tensorflower-gardener

  committed
  c4c19f1

  Copy full SHA for c4c19f1

Commits on Jun 27, 2016

• Adds a "currentThreadIndex" method to Eigen's ThreadPoolDevice. Use it to handle per-thread buffer allocation for the tileable executor without resorting to thread_local that is not fully supported…

  Show description for 7a62f1e
  tensorflower-gardener

  committed
  7a62f1e

  Copy full SHA for 7a62f1e

Commits on Jun 25, 2016

• Merge changes from github.

  Show description for 67324b1
  tensorflower-gardener

  committed
  67324b1

  Copy full SHA for 67324b1

Commits on Jun 13, 2016

• Replace internal google Status class with tensorflow::Status in suggested usage.

  Show description for 4a92a23
  tensorflower-gardener

  committed
  4a92a23

  Copy full SHA for 4a92a23

Commits on Jun 3, 2016

• Cleanup: Get rid of unused copy of simple threadpool implementation. The implementation is identical to the simple threadpool in Eigen (can be selected by not defining EIGEN_USE_NONBLOCKING_THREAD_…

  Show description for 89fb09e
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  89fb09e

  Copy full SHA for 89fb09e

Commits on Jun 2, 2016

• Update copyright for 3p/tf/core.

  Show description for c8b59c0
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  c8b59c0

  Copy full SHA for c8b59c0

Commits on May 24, 2016

• Enforce max_parallelism in work sharder by reverting to old sharding code if max_parallelism is less than number of workers in the pool.

  Show description for 67ddfa5
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  67ddfa5

  Copy full SHA for 67ddfa5

Commits on May 19, 2016

• Add optional max_parallelism parameter to tensorflow::ThreadPool::ParallelFor, such that limits on number of tasks provided to work sharder can be respected.

  Show description for c4119be
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  c4119be

  Copy full SHA for c4119be

Commits on May 10, 2016

• tensorflow: finer-grained Shard parallelization

  Show description for 8e37ef5
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  8e37ef5

  Copy full SHA for 8e37ef5

Commits on May 5, 2016

• Merge changes from github.

  Show description for 8bf6ef1
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  8bf6ef1

  Copy full SHA for 8bf6ef1