Blue Waters GPU Hackathon
Call for Team Applications
Date: September 10-14, 2018
General-purpose Graphics Processing Units (GPGPUs) potentially offer exceptionally high memory bandwidth and performance for a wide range of applications. The challenge in utilizing such accelerators has been learning how to program them. These hackathons are intended to help overcome this challenge for new GPU programmers and also to help existing GPU programmers to further optimize their applications. Any and all GPU programming paradigms are welcome. We encourage teams working on machine learning and/or deep learning projects to participate.
The goal of the hackathon is for user groups of large hybrid CPU-GPU systems to send teams of at least 3 developers along with either (1) a (potentially) scalable application that could benefit from GPU accelerators, or (2) an application running on accelerators that needs optimization for current and future GPU generations.
There will be intensive mentoring during this 5-day hands-on workshop, with the goal that the teams leave with applications running on GPUs, or at least with a clear roadmap of how to get there. Each team will be assigned mentors who have extensive experience in programming GPUs, many of them are from Dr. Wen-mei Hwu's lab and from NVIDIA, and they are knowledgeable in developing GPU-capable compilers and helping define standards such as OpenACC. These mentors have extensive experience working with science and application teams.
"But we don’t even know how to program a GPU…"
Programming experience with OpenACC or CUDA is not a requirement. You will learn quickly in the intense 5-day schedule, working on your code 95% of the time with the mentors. In the weeks preceding the hackathon, you will have opportunities to attend training to prepare you for the event. Prior GPU experience is not required!
Target Audience and Format
We are looking for teams of 3-8 developers with a scalable** application to port to (or optimize on) the GPU accelerators available at the University of Illinois, including the Blue Waters system at NCSA and systems supported by the Department of Electrical and Computer Engineering. Collectively the team should know the application intimately. If application is a suite of applications, no more than two per team is allowed and a minimum of 2 people per application must attend.
(** by scalable we really mean node-to-node communication implemented, but don’t be discouraged to apply if your application is less than scalable. We are also looking for breadth of application areas.)
How to Apply to Participate
Applications are closed - decisions will be provided to all applicants by Friday, July 20.
All participants will have access to GPU systems at the University of Illinois including the Blue Waters computing system and systems supported by the IBM-Illinois Center for Cognitive Computing Systems Research.. For GPU beginners, we recommend starting with OpenACC, but we are open to other GPU programming paradigms (e.g. CUDA, OpenMP4+) for those that have applications that already have some partial GPU port. We also host a number of events on programming hybrid architectures. Please visit the OLCF training events page for more info.
Participation in the training event is free of charge. Lunches will be provided daily. Breakfast, dinner, and local accommodations are at your own expense.
Information about local accommodations and other logistical details will be provided upon acceptance to the event.
The partners in this venture include Blue Waters, the National Center for Supercomputing Applications (NCSA), and the Department of Electrical and Computer Engineering at the University of Illinois; the Oak Ridge Leadership Computing Facility (OLCF); NVIDIA; and IBM.NCSA will be the host site for the event.
Please send any questions to firstname.lastname@example.org.Updated: July 12, 2018