Skip to content

Conversation

@chhwang
Copy link
Contributor

@chhwang chhwang commented Jan 15, 2026

No description provided.

Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

This PR extends the MemoryChannel tutorial to support multi-node execution in addition to the existing single-node multi-process mode. The changes enable users to run the tutorial example across different physical nodes using command-line arguments.

Changes:

  • Modified the worker function to accept rank, GPU ID, and bootstrap IP:port as parameters instead of deriving rank from GPU ID
  • Updated main function to support both single-node (no arguments) and multi-node (3 arguments) execution modes
  • Enhanced documentation to explain cross-node execution with MNNVL hardware support

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File Description
examples/tutorials/03-memory-channel/bidir_memory_channel.cu Refactored to support command-line arguments for multi-node execution while maintaining backward compatibility for single-node mode
docs/tutorials/03-memory-channel.md Added new section explaining cross-node execution with usage examples

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants