Skip to content
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

add tips to force NCCL comm to go through EFA #531

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

add tips to force NCCL comm to go through EFA #531

Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1573a8f
add tips to force NCCL comm to go through EFA
KeitaW Jan 23, 2025
ebb0711
Update micro-benchmarks/nccl-tests/README.md
mhuguesaws Jan 23, 2025
041677c
Update micro-benchmarks/nccl-tests/README.md
KeitaW Jan 23, 2025
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
22 changes: 22 additions & 0 deletions micro-benchmarks/nccl-tests/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -316,3 +316,25 @@ The formula defines the maximum theoretical bandwidth that can be achieved on di
* `t` : time to complete the operation. (similar to sec for Algbw and Busbw)
* `S` : number of elements being communicated (similar to count for Algbw and Busbw)
* `B` : theoretical peak bandwidth.

## 4. Tips and Tricks
Copy link
Contributor

Choose a reason for hiding this comment

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

Please find another title.


This section demonstrates NCCL tests tips and tricks useful to diagnose cluster nodes.
Copy link
Contributor

Choose a reason for hiding this comment

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

How do we know if this is bad or good?


#### Test EFA

You can force inter-GPU communications to go through EFA with the following environment variables:
KeitaW marked this conversation as resolved.
Show resolved Hide resolved


```bash
#Single node EFA
# libfabric flags
export FI_PROVIDER=efa
export FI_EFA_USE_DEVICE_RDMA=1

# NCCL Environment force disable P2P through NVlink, PCI and SHM.
mhuguesaws marked this conversation as resolved.
Show resolved Hide resolved
export NCCL_P2P_DISABLE=1
export NCCL_SHM_DISABLE=1
export NCCL_NVLS_ENABLE=0
export NCCL_NET='AWS Libfabric'
```