[MAIN][NVFP4] Support NVFP4 MOE with Proper Padding

[zhongbozhu]

What does this PR do ?

Contributing NVFP4 MOE utils available in dev branch back to main as it has been validated with E2E training: b01ad5b

[Already merged] TE dependency already available in TE main branch: NVIDIA/TransformerEngine#2215

TE PR for fused FP4 Grouped kernels want 64 aligned size in M, so also need to update accordingly in Mcore: NVIDIA/TransformerEngine#2351

⚠️ For major changes (either in lines of code or in its impact), please make sure to first share discuss a design-doc with the team.

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

(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.

[copy-pr-bot]

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.

[yaox12]

LGTM

[deepakn94]

LGTM.

[deepakn94]

CUDA graph support will come in a separate MR?

[zhongbozhu]

CUDA graph support will come in a separate MR?

We cannot really capture the cuda graph for moe layers.

Regarding more fine grained moe captures, main branch currently don't have many of them, while there is a PR that is trying to get those features into dev first: #1917

[zhongbozhu]

@yaox12 @yanring Currently the dev branch lacks the fix of the activation recompute, will send a standalone PR to dev about it.

#2036

[deepakn94]

CUDA graph support will come in a separate MR?

We cannot really capture the cuda graph for moe layers.

Regarding more fine grained moe captures, main branch currently don't have many of them, while there is a PR that is trying to get those features into dev first: #1917

Yup, I meant everything apart from the MoE layers.

[zhongbozhu]

Rebased to latest main

[zhongbozhu]

This PR also bumps up align size of FP4 to 64 from 32 based on TE PR: NVIDIA/TransformerEngine#2351

[yaox12]

/ok to test 08141bb