Achieving performance beyond the CPU: GPUs, co-processors, accelerators, and other technologies

Attendance via H.323 Videoconference (hosted by Barcelona), or online via Scopia (http://video-conference.bsc.es/).

Dr Adrian Jackson, EPCC, The University of Edinburgh

Over the past five to ten years a number of different technologies have been employed to build the fastest HPC systems. GPGPUs are the commonest such technology used today but they are only one of a series of processing technologies used to improve the performance of a program beyond what a standard processor could deliver. This talk will discuss the details of GPGPUs, FPGAs, and the Cell processor, describing their basic features and architecture, how they are used to achieve improved performance for scientific programs, and some of the issues/challenges to utilising them. The material will provide an introduction to all the technologies at a basic level.

Connection details

The presentation is available to all HPC-Europa2 visitors and researchers in Irish Third Level Institutes.  Please contact info@tchpc.tcd.ie if you wish to connect and listen to this presentation