Support TensorFlow pluggable devices #2144
Open
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PluggableDevice architecture offers a plugin mechanism for registering devices with TensorFlow without the need to make changes in TensorFlow code. It relies on C APIs to communicate with the TensorFlow binary in a stable manner. #2038 also asks for support of pluggable devices.
This PR adds the support for pluggable device. It will try to load all .so files under the given directory from argument --tensorflow_plugins by TF_LoadPluggableDeviceLibrary. With current exported symbol "TF_", the package size is increased 4M and total ratio is increased to 101%.
We will need to add alwaylink=1 in TensorFlow kernels_experimental library (tensorflow/tensorflow#60786).