[Core][Distributed] add coordinator to reduce code duplication in tp and pp #5293
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
GeauxEric
reviewed
Jun 7, 2024
zhuohan123
approved these changes
Jun 12, 2024
vllm/distributed/parallel_state.py
Outdated
| _WORLD: Optional[GroupCoordinator] = None | ||
|
|
||
|
|
||
| def get_world() -> GroupCoordinator: |
There was a problem hiding this comment.
How about the following changes? Make the function name easier to understand:
get_world->get_world_group?get_tp->get_tp_group?get_pp->get_pp_group?
|
@youkaichao Great! I love this refactoring PR 👍 |
robertgshaw2-redhat
pushed a commit
to neuralmagic/nm-vllm
that referenced
this pull request
Jun 16, 2024
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
joerunde
pushed a commit
to joerunde/vllm
that referenced
this pull request
Jun 17, 2024
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Jun 27, 2024
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Jul 8, 2024
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Jul 24, 2024
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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Outdated suggestions cannot be applied.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.
This PR adds coordinator, the last piece in #3587 . The main benefit is that all distributed stuff (e.g. complicated selection logic in communication ops, prev/next rank logic, initialization&destruction logic) can be shared for all tp/pp code.
After this PR, we explicitly require the code writer specify which group it wants to operate on:
get_world()get_tp()get_pp()After that, they can just call
all_reduce,prev_rank... without any further changes.