llnl / llnl/benchmarks

Create four problem configurations for Branson

Open
#25 4 comments 0 reactions 3 assignees View on GitHub

@rfhaque is already working on this.

Since Jan 16, 2026.

Dominant language
No language data
Stars
5
Forks
1
Avg merge
1d 13m
Merged PRs (30d)
3

Description

The following problem configurations need to be made for Branson.
These problems should have scripts / input decks.
Scripts / input decks must be version controlled, either in the llnl/benchmarks or in lanl/branson
We need to determine if we will use a branch on lanl/branson or a tag or ..

In order of priority:

- [ ] GPU, decomposed, history, SoA
- [ ] GPU, decomposed, event, SoA
- [ ] GPU, decomposed, event, AoS
- [ ] CPU, decomposed, history, SoA

We need to do a throughput study on the GPUs to understand the GPU saturation.

- [x] Single node, plot the computation rate (segments/second or particles/second) as a function of number of particles, and find the point where the speed saturates. Possibly starting with 2M particles per GPU, and increasing the number of particles until the problem OOMs.
- [ ] Need an initial problem defined, single node and hopefully multi-node
- [ ] Build and run on AMD and NVIDIA GPUs

Contributor guide

No contributing guide indexed for this repository

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.