Skip to content

Improve Nest GPU Utilization 5/N: Async H2D in Dataloader#15920

Open
MahmoudAshraf97 wants to merge 2 commits into
NVIDIA-NeMo:mainfrom
MahmoudAshraf97:ssl-pin-memory
Open

Improve Nest GPU Utilization 5/N: Async H2D in Dataloader#15920
MahmoudAshraf97 wants to merge 2 commits into
NVIDIA-NeMo:mainfrom
MahmoudAshraf97:ssl-pin-memory

Conversation

@MahmoudAshraf97

@MahmoudAshraf97 MahmoudAshraf97 commented Jul 17, 2026

Copy link
Copy Markdown
Contributor

Important

The Update branch button must only be pressed in very rare occassions.
An outdated branch is never blocking the merge of a PR.
Please reach out to the automation team before pressing that button.

What does this PR do ?

This PR closes the gap between ASR and NEST SSL gpu utilization during training
since AudioNoiseBatch uses a custom return type, torch dataloader pin_memory is a no-op and all transfers are pageable, implementing this raised the utilization on 8xH100 machine from 80% to almost 100%

This is a 1k test run using the same config before and after

gpu_util_comparison

This PR doesn't overlap with the changes in #15589 and can be merged cleanly
Collection: [Note which collection this PR will affect]

Changelog

  • Add specific line by line info of high level changes in this PR.

Usage

  • You can potentially add a usage example below
# Add a code snippet demonstrating how to use this 

GitHub Actions CI

The Jenkins CI system has been replaced by GitHub Actions self-hosted runners.

The GitHub Actions CI will run automatically when the "Run CICD" label is added to the PR.
To re-run CI remove and add the label again.
To run CI on an untrusted fork, a NeMo user with write access must first click "Approve and run".

Before your PR is "Ready for review"

Pre checks:

  • Make sure you read and followed Contributor guidelines
  • Did you write any new necessary tests?
  • Did you add or update any necessary documentation?
  • Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
    • Reviewer: Does the PR have correct import guards for all optional libraries?

PR Type:

  • New Feature
  • Bugfix
  • Documentation

If you haven't finished some of the above items you can still open "Draft" PR.

Who can review?

@stevehuang52

Additional Information

  • Related to # (issue)

Signed-off-by: Mahmoud Ashraf <hassouna97.ma@gmail.com>
Signed-off-by: Mahmoud Ashraf <hassouna97.ma@gmail.com>
@copy-pr-bot

copy-pr-bot Bot commented Jul 17, 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.

@svcnvidia-nemo-ci svcnvidia-nemo-ci added the waiting-on-maintainers Waiting on maintainers to respond label Jul 19, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ASR community-request waiting-on-maintainers Waiting on maintainers to respond

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants