llnl / llnl/benchpark

Laghos: in container

Open
#720 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
78
Forks
52
Avg merge
4d 19m
Merged PRs (30d)
10

Description

We need a reproducer for building a container on one machine (e.g., an LC machine), and running it on another machine (e.g., AWS).

- [ ] What type of a container? Docker?
- [ ] Is it possible to build a `container` version of a package? If yes, we could use Spack to build the container.
- [ ] Running is difficult, particularly if file systems are in play - at the very least, we have to store the caliper files etc.

Study performance difference on communication cost of how you build the container wrt MPI.
- [ ] Hybrid method: install MPI in the container. Compile the code with MPI X, ideally the same implementation as on AWS.
- [ ] Host MPI: bind in MPI when you build it, bind another MPI when run on AWS.

Contributor guide

No contributing guide indexed for this repository

Research direction

No files or tests are named. Start by resolving the container format, Spack packaging approach, and MPI strategy for building on an LC machine and running on AWS. Done means a reproducible container workflow that preserves required Caliper files and measures communication-cost differences between hybrid and host-MPI approaches.

Written by the indexing model from the issue text.

Assessment

Tech stack
aws, docker
Domain
cloud, hpc, infrastructure, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.