Prior GPU experience is not required, as those selected will be paired with experienced mentors who will teach them how to leverage accelerated computing in their own applications or further optimise their codes.
General-purpose Graphics Processing Units (GPGPUs) potentially offer exceptionally high memory bandwidth and performance for a wide range of applications. A challenge in utilising 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 - a great opportunity for graduate students and postdocs. Any and all GPU programming paradigms are welcome.
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 road-map of how to get there. Each team will be assigned mentors from universities, national laboratories, super-computing centres, industry partners, and NVIDIA who have extensive experience in programming GPUs. Programming experience with CUDA is not a requirement although training can be offered in advance of the event.
We are looking for:
- Teams of 3-5 developers, ideally postdocs, research software engineers (RSEs) or graduate students with a scalable application to port to (or optimise on) GPU accelerators. Collectively, the team should know the application intimately.
- Mentors from universities, national laboratories, super-computing centres and industry partners to work as team mentors. Mentors will work in pairs within teams to support optimisation of the the teams code during the week.
Team applications can be made on the Sheffield GPU hackathon website: http://gpuhack.shef.ac.uk/
If you are interested in volunteering to mentor then please contact gpuhack-organisers@sheffield.ac.uk
Given the current COVID-19 situation and the current ambition to accelerate infectious disease and infrastructure modelling it would be great to attract applications from teams who are keen to work with GPU experts to improve performance and scalability of models in this area.