| Commit message (Collapse) | Author | Age |
|
|
|
| |
PiperOrigin-RevId: 216479972
|
|
|
|
| |
PiperOrigin-RevId: 215808649
|
|
|
|
| |
PiperOrigin-RevId: 215710849
|
|
|
|
| |
PiperOrigin-RevId: 215073584
|
|
|
|
|
|
|
|
| |
Currently, we run tests on machines with GPUs based on the "gpu" tag, and the
tests automatically adapt to whether a GPU is available. Creating two targets,
one tagged with "gpu" and one not, will make us run the tests in both modes.
PiperOrigin-RevId: 215045035
|
|\
| |
| |
| | |
PiperOrigin-RevId: 214793113
|
| |
| |
| |
| |
| |
| |
| |
| | |
This change reduce the size of //tensorflow/tools/pip_package:simple_console_windows's zip file from 1000027677 bytes to 47690474 bytes for a CPU build. For GPU build, it will avoid going over 4GB when multiple CUDA compatibility are specified.
To fix #22390
PiperOrigin-RevId: 214764423
|
| |
| |
| |
| | |
PiperOrigin-RevId: 214557082
|
| |
| |
| |
| |
| |
| |
| |
| | |
Bazel does not allow python rules to directly depend on c++ rules.
So I have to separately manage static dependencies, unfortunately avoiding
"kernels" option for now.
PiperOrigin-RevId: 214532631
|
| |
| |
| |
| | |
PiperOrigin-RevId: 213346439
|
|\ \
| | |
| | |
| | |
| | |
| | | |
Intel-tensorflow:shaohua/fix_gcc6.3_build_link_issue
PiperOrigin-RevId: 213208519
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The commit contains following components to support TensorFlow on ROCm platform
- bazel build system
- continuous integration logic
Authors:
- Jack Chung: jack.chung@amd.com
- Jeffrey Poznanovic: Jeffrey.Poznanovic@amd.com
- Peng Sun: Peng.Sun@amd.com
|
| |
| |
| |
| | |
PiperOrigin-RevId: 210929192
|
| |
| |
| |
| |
| |
| | |
tensorflow.bzl
PiperOrigin-RevId: 210859798
|
| |
| |
| |
| |
| | |
* Fixed the dependencies with `str(Label(...))` so that third_party codes that refer to
the tensorflow as a submodule get it properly resolved.
|
|\ \
| | |
| | |
| | | |
PiperOrigin-RevId: 209623532
|
| | |\
| |_|/
|/| | |
|
| | |
| | |
| | |
| | |
| | |
| | | |
Avoids whitespace changes in a later CL.
PiperOrigin-RevId: 209168698
|
| |\ \ |
|
| |/ /
|/| |
| | |
| | | |
PiperOrigin-RevId: 208894550
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 208838993
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Previously, specifying --define=using_mkl_dnn_only=true would cause
MKL-ML-dependent code to be #ifdef'd out, but dependencies on MKL-ML itself
would still be present. This change makes all library dependencies on MKL
properly select MKL-ML, MKL-DNN, or both, depending on the selected
configuration.
PiperOrigin-RevId: 208710102
|
|\ \ \
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 208565050
|
| | |\ \
| |_|/ /
|/| | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The existing feature macros are named INTEL_MKL to indicate that any flavor of
MKL is available, INTEL_MKL_ML to indicate that *only* MKL-ML is available
(i.e. MKL-DNN is not), and DO_NOT_USE_ML to indicate that *only* MKL-DNN is
available (i.e. MKL-ML is not).
This change renames INTEL_MKL_ML to INTEL_MKL_ML_ONLY and DO_NOT_USE_ML to
INTEL_MKL_DNN_ONLY. The meanings of the macros have not changed.
This change also adds a few sanity checks to mkl_util.h that ensures that the
combination of INTEL_MKL, INTEL_MKL_ML_ONLY, and INTEL_MKL_DNN_ONLY is
logically consistent: the *_ONLY macros may not both be defined, and if either
of them is defined, bare INTEL_MKL must also be defined.
PiperOrigin-RevId: 208313735
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 208283367
|
| | |\ \
| |_|/ /
|/| | | |
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 207176147
|
| | |\ \
| |_|/ /
|/| | |
| | | | |
based on PR review comments.
|
| | | |
| | | |
| | | |
| | | |
| | | | |
RELNOTES: n/a
PiperOrigin-RevId: 206822820
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
dependencies.
PiperOrigin-RevId: 206473913
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 206197083
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
@bazel_tools//tools/cpp:current_cc_toolchain //tools/defaults will be removed soon
PiperOrigin-RevId: 206187625
|
| |_|/
|/| |
| | |
| | | |
Signed-off-by: shaohua <shaohua.zhang@intel.com>
|
| | |\
| |_|/
|/| | |
|
| | |
| | |
| | |
| | | |
* Added nGraph bridge as a third_party to be built with TensorFlow based on user selection.
* Added a limited set of C++ unit tests to verify the correctness of the computation
|
| | |
| | |
| | |
| | |
| | |
| | | |
LGPL, including fix for the Android/Apple version of code (C++ macros-es fix)
PiperOrigin-RevId: 205842327
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
These checks were necessary when we used RTLD_GLOBAL to expose TF symbols to custom ops, since :framework and :lib pulled in implementations. They're now header-only. Ideally we'd switch the checks to framework_internal_impl and lib_internal_impl, but that would require visibility for those rules (thus making it more likely they'd get included in silly places). So this change disables the check for dynamic builds, on the theory that accidentally relying on implementation rules is much more difficult than it was with a static build.
Should allow tf_custom_op_libraries to depend on GPU kernels (which depend on core:gpu_lib which depends on :framework).
PiperOrigin-RevId: 205472434
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 205427089
|
|\ \ \
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 204662666
|
|\ \ \ \
| |_|_|/
|/| | |
| | | | |
PiperOrigin-RevId: 204530874
|
| |_|/
|/| | |
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 203987173
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 203868794
|
|/ / |
|
| | |
|
|/ |
|
|
|
|
| |
PiperOrigin-RevId: 202260254
|
|
|
|
| |
PiperOrigin-RevId: 201301504
|