Skip to content

Yjeong/cmk nccltest add b300 noht #110

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
youngjeong46 merged 4 commits into from
Sep 16, 2026
Merged

Yjeong/cmk nccltest add b300 noht #110

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
1fa873d
B300 noht cmk nccl allreduce test
youngjeong46 Sep 16, 2026
6bbebfa
readme edit
youngjeong46 Sep 16, 2026
8ee01db
solved merge conflict
youngjeong46 Sep 16, 2026
8be71ce
small edit
youngjeong46 Sep 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion cmk-nccltests/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,5 +43,6 @@ This directory contains Kubernetes manifests for running [NCCL tests](https://gi

## Configuration notes

- **Topology file**: B200, H100, and H200 jobs mount `/etc/crusoe/nccl_topo` from the host and set `NCCL_TOPO_FILE` to the SKU-specific XML. This path is pre-populated on Crusoe GPU nodes.
- **Topology file**: B300, B200, H100, and H200 jobs mount `/etc/crusoe/nccl_topo` from the host and set `NCCL_TOPO_FILE` to the SKU-specific XML. This path is pre-populated on Crusoe GPU nodes.
- **CUDA image**: B200 and H200 use `nccl-tests:13.0.1-ubuntu24.04-nccl-2.29.2-1`; H100 uses `nccl-tests:12.8.1-ubuntu24.04-nccl-2.26.5-1`. Both images are hosted at `ghcr.io/crusoecloud/nccl-tests`.
- **B300 No Hyperthreading**: Certain B300 nodes provided by Crusoe disables Hyperthreading for certain CPU performance profile. Those VMs have a dedicated topology file, and therefore should use its own yaml file as indicated by `nccl-b300-noht.yaml`.
Loading