[SPMD][PoC] XLAShardedTensor & mark_sharding API #3476
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
hello, could you please tell me the commit id of your pytorch? |
4b4cbac to
109803b
Compare
JackCaoG
reviewed
May 9, 2022
fed00ac to
ff409e4
Compare
|
Landing, I will create a separate PR for PJRT integration. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This contributes to #3871 . Test GSPMD feature in PyTorch/XLA with a single host 8 core setup. This PR aims to provide example implementations of
XLAShardedTensorandmark_shardingAPI to annotate different sharding strategies (e.g., tiled, replicated)Additionally, the full PoC example requires removing a constraint where only one core can be assigned to a replica. This is not feasible with the current runtime client, and I will also investigate and see if we can remove such a restriction and compile & execute the partitioned computation in XRT.
We will also merge with the TPU PjRT sometime soon, replacing XRT with PjRT and continue the SPMD experimentation, use PjRT SPMD as a reference implementation. cc @will-cromar