[Core] Add data parallelism override support#623
Closed
smfirmin wants to merge 1 commit into
Closed
Conversation
smfirmin
requested review from
CatherineSue,
YouNeedCryDear,
beiguo218,
pallasathena92 and
slin1237
as code owners
github-actions
Bot
added
inferenceservice
smfirmin
changed the title
Collaborator
|
duplicate of #611 |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What this PR does
Adds explicit data parallelism override support for accelerator-specific runtime configuration.
When
tensorParallelismOverride.dataParallelSizeis set for a selected accelerator class, OME now updates matching runtime arguments using the supported SGLang and vLLM aliases:--dp-size--dp--data-parallel-sizeThis mirrors the existing tensor and pipeline parallelism override behavior.
Why we need it
The API already exposes
dataParallelSize, but the reconciler only applied tensor parallel and pipeline parallel sizing. This meant accelerator-specific data parallel sizing could be configured but would not be reflected in generated runtime command or args.Data parallel sizing is needed for MoE deployments where effective parallelism depends on TP x DP.
Fixes #
How to test
go test ./pkg/controller/v1beta1/inferenceservice/components git diff --checkChecklist
make testpasses locally