Implement direct server-to-server communication#331
Merged
Conversation
borzunov
force-pushed
the
server-to-server
branch
from
June 24, 2023 02:30
4b94939 to
9aebf0a
Compare
borzunov
force-pushed
the
server-to-server
branch
from
June 24, 2023 02:32
9aebf0a to
234bafb
Compare
borzunov
commented
Jun 24, 2023
| module_backends: Dict[str, TransformerBackend], | ||
| *, | ||
| push_manager: SyncManager, | ||
| session_pipes: Dict[str, Tuple[PipeConnection, threading.Lock]], |
Collaborator
Author
There was a problem hiding this comment.
rpc_push() may be received by a connection handler different from the one holding the inference session, so we use some multiprocess communication here.
borzunov
commented
Jul 1, 2023
| assert isinstance(block_uid, str) and isinstance(metadata, dict) | ||
| return block_uid, inputs, metadata | ||
|
|
||
| async def rpc_push(self, request: runtime_pb2.ExpertRequest, context: P2PContext) -> runtime_pb2.ExpertResponse: |
Collaborator
Author
There was a problem hiding this comment.
TODO: This can be stream-to-unary handler, so that (a) the previous server doesn't have to make a new connection each time and (b) we don't have to parse metadata at this stage each time (now it's done to find session_id). I'm not sure if it affects performance much though, so I'd postpone that to a later PR.
borzunov
commented
Jul 1, 2023
| span = None | ||
| try: | ||
| if not self._chosen_spans or not self._server_sessions or attempt_no >= 1: | ||
| # If there is a failed server session, this code closes it |
Collaborator
Author
There was a problem hiding this comment.
This code was moved to a separate method InferenceSession._update_sequence() to simplify this method.
borzunov
force-pushed
the
server-to-server
branch
from
July 1, 2023 23:50
73a0563 to
49bbce4
Compare
borzunov
force-pushed
the
server-to-server
branch
2 times, most recently
from
July 2, 2023 00:36
0e5ce37 to
2445aa5
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 2, 2023 00:37
2445aa5 to
cef5662
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 2, 2023 02:21
7d73698 to
429146e
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 2, 2023 02:24
429146e to
9a23c1e
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 5, 2023 13:01
bccdc21 to
845c172
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 5, 2023 13:06
845c172 to
0635968
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 9, 2023 22:58
c5dfba2 to
a637fd2
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 9, 2023 22:58
a637fd2 to
7e204d0
Compare
borzunov
force-pushed
the
server-to-server
branch
from
July 9, 2023 23:32
9427cae to
5b3f180
Compare
borzunov
commented
Jul 10, 2023
| inputs = self.history # Pass full inputs including prefix | ||
| else: | ||
| inputs = inputs[:, -n_input_tokens:] # No need to pass prefix further | ||
|
|
Collaborator
Author
There was a problem hiding this comment.
Refactored: This code was moved from InferenceSession.step() to _ServerInferenceSession.step(), since it's actually about one server only. The overall structure is more clear this way.
borzunov
marked this pull request as ready for review
July 10, 2023 00:16
justheuristic
approved these changes
Jul 10, 2023
artek0chumak
pushed a commit
to artek0chumak/petals
that referenced
this pull request
Jul 11, 2023
fthrvi
pushed a commit
to fthrvi/nakshatra
that referenced
this pull request
Jun 1, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Implement #226.
TODO:
manager.Queue())next_servers