Mirroring has been enabled by default for a long time. All the runtime
code that inspected the flag has long been removed. This is just
cleaning up some leftover code.
PiperOrigin-RevId: 328421064
Change-Id: I193b67f1f12f96ebb8b6822a780d88f4b82e77af
When executing on a remote worker, we may have to copy the TensorHandle
for each executed op. To avoid duplicated work, we expand the
TensorHandle to keep track of mirrors which are tied to the lifetime of
the TensorHandle. If a mirror already exists on a remote worker, no
additional copy is needed.
The change consists of the following:
- Add map of remote mirrors in TensorHandle.
- Add `mirror` boolean argument to EagerCopyToDevice which indicates to try
configuring a mirror if possible.
- Add Device argument to RemoteAddress to handle mirrors.
- Expose a ContextMirroringPolicy for the EagerContext. We plan to add
additional policies in the future, such as local tensor mirroring.
- Rename ContextDevicePlacementPolicy variables to be consistent with
ContextMirroringPolicy.
PiperOrigin-RevId: 253945140
Currently, the only place we increase the refcount is in remote tensor handle
destructors.
Unfortunately, since tensorhandle/context lifetime is tied to python objects, it did not seem possible to enforce deletion of the context to be after the deletion of all tensor handles. This also means that if you remote tensors, you will not be able to delete the EagerContext. This is mostly fine, since we don't support deletion of eager contexts in any case, and we only run into this issue during python teardown.
PiperOrigin-RevId: 246178464
In order to take advantage of the type system to help enforce ownership, this
change refactors DeviceMgr to use std::unique_ptr<Device> instead of Device*'s.
It also updates all callers to use the new types.
PiperOrigin-RevId: 222645861
This change moves //tensorflow/contrib/lite to //tensorflow/lite in preparation
for TensorFlow 2.0's deprecation of contrib/. If you refer to TF Lite build
targets or headers, you will need to update them manually. If you use TF Lite
from the TensorFlow python package, "tf.contrib.lite" now points to "tf.lite".
Please update your imports as soon as possible.
For more details, see https://groups.google.com/a/tensorflow.org/forum/#!topic/tflite/iIIXOTOFvwQ
@angersson and @aselle are conducting this migration. Please contact them if
you have any further questions.
PiperOrigin-RevId: 219536476
