Skip to content

Forced load imbalance#2917

Merged
Victarry merged 5 commits into
NVIDIA:devfrom
nanz-nv:nanz/load_balance
Jan 19, 2026
Merged

Forced load imbalance#2917
Victarry merged 5 commits into
NVIDIA:devfrom
nanz-nv:nanz/load_balance

Conversation

@nanz-nv

@nanz-nv nanz-nv commented Jan 13, 2026

Copy link
Copy Markdown
Contributor

What does this PR do ?

This PR tries to add a utility to force load imbalance among experts.

⚠️ For major changes (either in lines of code or in its impact), please make sure to first share a design doc with the team. If you're unsure what's the best way to do so, contact the @mcore-oncall.

Contribution process

flowchart LR
    A[Pre-checks] --> B[PR Tests]
    subgraph Code Review/Approval
        C1[Expert Review] --> C2[Final Review]
    end
    B --> C1
    C2 --> D[Merge]
Loading

Pre-checks

  • I want this PR in a versioned release and have added the appropriate Milestone (e.g., Core 0.8)
  • I have added relevant unit tests
  • I have added relevant functional tests
  • I have added proper typing to my code Typing guidelines
  • I have added relevant documentation
  • I have run the autoformatter.sh on my PR

Code review

The following process is enforced via the CODEOWNERS file for changes into megatron/core. For changes outside of megatron/core, it is up to the PR author whether or not to tag the Final Reviewer team.

For MRs into `main` branch

Feel free to message or comment the @mcore-oncall to help accelerate your merge into main. The less complex your PR is, the faster it will be approved and merged!

(Step 1): Add PR label Expert Review

(Step 2): Collect the expert reviewers reviews

  1. Attach the Expert Review label when your PR is ready for review.
  2. GitHub auto-assigns expert reviewers based on your changes. They will get notified and pick up your PR soon.

⚠️ Only proceed to the next step once all reviewers have approved, merge-conflict are resolved and the CI is passing.
Final Review might get declined if these requirements are not fulfilled.

(Step 3): Final Review

  1. Add Final Review label
  2. GitHub auto-assigns final reviewers based on your changes. They will get notified and pick up your PR soon.

(Optional Step 4): Cherry-pick into release branch

If this PR also needs to be merged into core_r* release branches, after this PR has been merged, select Cherry-pick to open a new PR into the release branch.

For MRs into `dev` branch The proposed review process for `dev` branch is under active discussion.

MRs are mergable after one approval by either eharper@nvidia.com or zijiey@nvidia.com.

Merging your PR

Any member of core-adlr and core-nemo will be able to merge your PR.

@nanz-nv nanz-nv requested review from a team as code owners January 13, 2026 03:55
@copy-pr-bot

copy-pr-bot Bot commented Jan 13, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

'The default value 1e-3 is same as that used in DeepSeekV3.')
group.add_argument('--moe-router-force-load-balancing', action='store_true',
help='[Experimental] Force override routing to balance token distribution using random logits for MoE routers, supporting naive top-k and group-limited top-k. This experimental feature is for benchmarking purposes only!')
group.add_argument('--moe-router-force-biased', type=float, default=None,

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall LGTM. I think the argument name and help message could be refined to make it more clear to users.

Suggested changes:

    group.add_argument('--moe-router-force-expert-bias-std', type=float, default=None,
                       help='[Experimental] Apply random expert bias in normal distribution with specified std to router logits. The random seeds are shared  across all ranks.'
                       'If the specified value is positive, generates new random bias each forward pass. '
                       'If the specified value is negative, generates bias once per layer and reuses it (abs value is std).'
                       'This experimental feature is for benchmarking purposes only!')

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks. The help message has been updated.

@Victarry

Copy link
Copy Markdown
Contributor

/ok to test 9c46233

@Skylion007

Copy link
Copy Markdown
Contributor

Thank you, this is very useful for MLSys debugging.

@yaox12

yaox12 commented Jan 15, 2026

Copy link
Copy Markdown
Member

The lint issue needs to be fixed.

auto-merge was automatically disabled January 15, 2026 13:49

Head branch was pushed to by a user without write access

@Victarry

Copy link
Copy Markdown
Contributor

/ok to test 1689973

@Victarry

Copy link
Copy Markdown
Contributor

/ok to test 38a11bb

@github-actions

Copy link
Copy Markdown
Contributor

Thank you for your contribution!

NVIDIA Megatron-LM is currently transitioning to development on Github. We will aim to review your PR after we complete our transition and stabilize our Github development process.

Thank you for your understanding.

"""[Experimental] Force load balancing with random logits for MoE router, supports naive topk
and group-limited topk. This is an experimental feature and only for benchmark."""

moe_router_force_biased: Optional[float] = None

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

One API issue is that can the be fused together as an enum string parsing with the current Boolean value for force load balancing? Or at least have some though going into other force routing strategies for benchmarking, for instance, you may want to try the worst case of all experts being biased to the same GPU group to test worst case memory scenarios.

@Victarry Victarry added this pull request to the merge queue Jan 19, 2026
Merged via the queue into NVIDIA:dev with commit bd8411c Jan 19, 2026
47 of 54 checks passed
@chtruong814 chtruong814 removed the needs-follow-up Issue needs follow-up label Jan 19, 2026
@yanring

yanring commented Feb 11, 2026

Copy link
Copy Markdown
Contributor

Communicated with Nan. Nan will find some time to file the main PR in the near future.

nanz-nv added a commit to nanz-nv/Megatron-LM that referenced this pull request Feb 12, 2026
Co-authored-by: Dennis(Zhenhuan) Liu <denliu@nvidia.com>
@nanz-nv nanz-nv mentioned this pull request Feb 12, 2026
6 tasks
Victarry added a commit to nanz-nv/Megatron-LM that referenced this pull request Mar 3, 2026
Co-authored-by: Dennis(Zhenhuan) Liu <denliu@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants