NVIDIA / NVIDIA/cuda-samples

p2pBandwidthLatencyTest

Open
#392 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C++
Stars
9.6k
Forks
2.4k
Avg merge
53m
Merged PRs (30d)
1

Description

Seeing the following on a 4xH100-80GB system

bidirectional
GPU0-GPU0 = 2959.19 GB/s
GPU0-GPU1 = 260.89 (GB/s)

The spec says GPU0-GPU1 NVlink bandwidth is 300 GB/s. Is it possible to close the gap ?

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the p2pBandwidthLatencyTest sample and reproduce the bidirectional results on the reported 4xH100-80GB system. Compare GPU0-GPU0 and GPU0-GPU1 measurements with the stated NVLink specification, then document whether the gap is expected or identify the sample or setup behavior that requires correction.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.