Document how new problems support Nsight Compute profiling - #171
Merged
Merged
Conversation
This was referenced Sep 18, 2026
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up: hosted setup and provenance corrections are in #172.
New problem authors need a real NCU profile path to produce hardware-counter reports. Add a contributor guide, linked from the README, covering the
custom_kernelNVTX range, benchmark selection, child-process capture, separate NCU/PyTorch paths, result logging, and GPU validation.--profileuses normal Popcorn registration and the hosted GPU Mode API. Users need no provider SDK, account, credentials, or endpoint configuration.--profile-brevexplicitly selects Brev; failures never switch providers. The guide explains that hosted tasks must be synced by an operator and local evaluator edits do not update the service.Validation: compiled all three Python examples in their intended scope, checked relative links and whitespace, and checked the contract against
problems/linalg/qr_v2and the shared NVIDIA evaluator at51e22db671d36c1c76091c43c36a44546ba324a1. The QR v2 benchmark-0 B200 integration run returned a report and text/CSV exports through the revised API with no Modal executable or credentials available to the CLI. This does not establish support for every problem or metric.Companion CLI: gpu-mode/popcorn-cli#80. Backend: gpu-mode/kernelbot#530. The CLI release depends on that API/runner deployment.