Taking full advantage of the tremendous power of GPU-based supercomputers for science and engineering requires the writing of efficient software. An opportunity to receive intensive mentoring and five days of hands-on training in the development of that valuable skill is on the horizon. The Computational Science Initiative (CSI) at Brookhaven National Laboratory (BNL) will host its 3rd GPU Hackathon, a free event, on September 23–27.
The Hackathon’s organizing committee is soliciting applications from current and prospective GPU users and welcomes all programming paradigms. Previous programming experience with GPUs is not a requirement.
The organizing committee stated that teams from the high-energy physics community and the US Department of Energy’s Exascale Computing Project (ECP) are especially encouraged to participate.
Accepted teams will be mentored by GPU programming experts from industry, national labs, and universities. Access to GPU resources will be provided shortly before and during the Hackathon for teams to develop and test their codes.
The event’s website provides more details and a team application. Deadline for applications is June 30. Notification of acceptance will be issued by July 15. All other logistics are available on the website.
The Hackathon is jointly organized by BNL, University of Delaware, and Oak Ridge National Laboratory, and sponsored by ECP and the HEP Center for Computational Excellence.