Skip to content
Draft
Changes from all commits
Commits
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
45 changes: 45 additions & 0 deletions working-groups/benchmarking/resource-request/NERSC_proposal.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
# NERSC Proposal Information Needed:

## Project Title
HPSF Benchmarking

## Senior Investigators
* Chris Siefert, SNL
* Olga Pearce, LLNL
* Adam Lavely, LBNL

## Primary Funding Source (ideally this is DOE Office of Science. This will need funding agency, PM and grant numbers)
* We don't have one?

## Project summary and goals: Provide a brief summary of your project easily understood by people outside your field.
## What will this project accomplish? What is the significance of this work?

High Performance Computing (HPSF) Benchmarking Working Group aims to bring together projects and organizations
to create an automated way to continuously test the performance of evolving HPC software stacks,
ensuring that the components work and perform well.
In 2026, we will stand up software stack testing by means of several approaches in order to evaluate these approaches,
identify challenges and design solutions.

## Detailed Description for DOE Program Managers: Provide a more technically detailed explanation of your research for proposal reviewers.


## Enter a URL for a relevant web page describing your project.
* https://github.com/hpsfoundation/tac/blob/main/working-groups/benchmarking/README.md
* Do we want to add Trilinos and Benchpark here?

## Computational Resources - Enter the number of NERSC computational hours requested CPU and GPU

## If you are applying for GPU time, please tell us about your code's readiness to use GPUs.

## Briefly describe how you estimated the compute hours (CPU and GPU) requested above as a justification for the Reviewers to determine the appropriateness of your request.

## Please list the 3rd party or community software packages (eg. math libraries) that you need NERSC to provide.
* MPI
* Blas/Lapack

## Please list the application codes or frameworks you will run at NERSC. This includes Code Name, URL and Description (and whether or not the code is GPU enabled).
* Trilinos
* Benchpark: https://github.com/LLNL/benchpark. Benchpark is an open collaborative repository for reproducible specifications of HPC benchmarks. Available benchmarks include AMG2023 (LLNL), Babelstream (UCL), Branson (LANL), Genesis (RIKEN), GPCNet (LBL,ANL,HPE), Gromacs (KTH), HPCC (UTK), HPCG (SNL,UTK), HPL (UTK), IOR, Kripke (LLNL), Laghos (CEED), LAMMPS (SNL), md-test, OSU microbenchmarks (OSU), Phloem (LLNL), Quicksilver (LLNL), QWS (RIKEN), RAJAPerf (LLNL), Remhos (SEED), Salmon (RIKEN), Sandia Microbenchmarks (SNL). Most are GPU enabled.

## List any additional HPC resources that have been awarded or are expected to be used to support this project.