[3/n tensor engine] hello tensor engine #187
Open
+798
−25
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
Add the initialize controller class `mesh_controller` that can implement the tensor engine on top of the ProcMesh/ProcActor API. The only example is currently a "hello world" that allocates and fetches a tensor. Follow up PRs will integrate creating meshes this way into the testing code more deeply and fix the issues that come up with it. This design assumes that supervision of stuff is going to be handled by the actor system and that tensor compute can just rely on that for monitoring and stuckness detection stuff. This design has no ClientActor, and the ControllerActor only exists as a Instance handle for reading messages from the workers (which send some controller messages). This does not attempt to clean up the existing RustController system yet, since it isn't feature equivalent or tested with it. Differential Revision: [D75909313](https://our.internmc.facebook.com/intern/diff/D75909313/) [ghstack-poisoned]
zdevito
added a commit
that referenced
this pull request
Jun 6, 2025
Add the initialize controller class `mesh_controller` that can implement the tensor engine on top of the ProcMesh/ProcActor API. The only example is currently a "hello world" that allocates and fetches a tensor. Follow up PRs will integrate creating meshes this way into the testing code more deeply and fix the issues that come up with it. This design assumes that supervision of stuff is going to be handled by the actor system and that tensor compute can just rely on that for monitoring and stuckness detection stuff. This design has no ClientActor, and the ControllerActor only exists as a Instance handle for reading messages from the workers (which send some controller messages). This does not attempt to clean up the existing RustController system yet, since it isn't feature equivalent or tested with it. Differential Revision: [D75909313](https://our.internmc.facebook.com/intern/diff/D75909313/) ghstack-source-id: 288027712 Pull Request resolved: #187
facebook-github-bot
added
the
CLA Signed
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.
Stack from ghstack (oldest at bottom):
Add the initialize controller class
mesh_controllerthat can implement the tensor engine on top of the ProcMesh/ProcActor API.The only example is currently a "hello world" that allocates and fetches a tensor.
Follow up PRs will integrate creating meshes this way into the testing code more deeply and fix the issues that come up with it.
This design assumes that supervision of stuff is going to be handled by the actor system and that tensor compute can just rely on that for monitoring and stuckness detection stuff.
This design has no ClientActor, and the ControllerActor only exists as a Instance handle for reading messages from the workers (which send some controller messages).
This does not attempt to clean up the existing RustController system yet, since it isn't feature equivalent or tested with it.
Differential Revision: D75909313