Participants: NCSA staff and relevant affiliates
Location: An NCSA conference room, in person
Date: TBD (postponed again)
Participants
Please indicate your interest in participating in this workshop by
- adding your name to the list below
T. Andrew Manning
Jeff Terstriep
- Todd Nicholson
- Bruno Abreu
- Joshua Allen
Amy Schuele- Jeremy Enos
- Jake Rundall
- Galen Arnold
- Gregory Bauer
- Jim Phillips
- Darren Adams
Overview
This workshop's purpose is to have a meeting where NCSA staff working on projects involving batch processing, asynchronous jobs, and/or HPC can meet one another and exchange ideas about these topics. The goal is to facilitate personal introductions and cross-pollination between organizationally disparate project teams that do similar technical work in this area.
Links
Agenda
- [11:00-11:15] Introduction (Andrew Manning)
- [11:15-11:35] Presentation A
- [11:35-11:55] Presentation B
- [12:00-13:00] Lunch and Topical Discussions
- [13:00-13:20] Presentation C
- [13:25-13:45] Presentation D
Presentations
Format: Presentations will be 20 min talks with 10 min for questions and discussion, typically during the talk in context instead of at the end.
- Presentation A
- Presenter: Jake Rundall
- Title: Overview of NCSA's HPC clusters
- Presentation B
- Presenter: Jake Rundall
- Title: Fair-share scheduling in Slurm — how can we make it make sense and seem fair?
- Introduce the issues and current state of affairs. Follow up with a "topical discussion".
- Presentation C
- Presenter: Andrew Manning
- Title: Job management in a Kubernetes-based science gateway
- Presenter: Andrew Manning
- Presentation D
- Presenter: Bruno Abreu
- TItle: Overview of batch processing in the C3 AI Suite
Topical discussions
Topical discussions allow us to share information on topics in a less formal way than presentations yet with some focus provided by the discussion leaders. The duration of these discussions is flexible.
Topics
- Traditional HPC Environments
- Batch Scheduling
- Current Environments
- Delta, HOLL-I, HAL, Campus Cluster, Nightingale, vForge
- Cloud Computing
- Openstack/Radiant
- Kubernetes
- C3 AI Suite
- Hybrid Environments
- Batch Oriented - HPCng, AWS ParallelCluster, StackHPC, etc
- Cloud Oriented -
- Asynchronous job management
Identity and Access Management (IAM)
- Use cases
1 Comment
Timothy Andrew Manning
It is not obvious where this wiki page should reside. If there is a more appropriate wiki space and parent page please comment your recommendation