NVIDIA-NeMo / NVIDIA-NeMo/RL

MCore refit backend nvshmem does not function

Open
#3,646 1 comment 0 reactions 1 assignee Claimed by @shanmugamr1992 View on GitHub
bug Speed
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

**Describe the bug**

The nvshmem backend for MCore refit is currently broken, with a known fix in #3405, and an unknown number of possible issues hiding behind that known fix. There are reports of performance degradation, depending on which version of nvshmem is used.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.