University of Cambridge > Talks.cam > Computer Laboratory Wednesday Seminars > Acceleration of scientific computing using graphics hardware

Acceleration of scientific computing using graphics hardware

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Timothy G. Griffin.

Developers of scientific computing codes crave performance (FLOPS) but only a tiny minority can afford to commission bespoke hardware. As a result, commodity hardware is co-opted into the HPC role. Most scientific applications are inherently data parallel and, fortunately, the stagnation in CPU clock speed has coincided with cheap high performance networking infrastructure so that PC clusters are now ubiquitous. Another way to exploit parallelism is through multi-core shared memory devices and commodity PCs with 2, 4 (8…) cores are now commonplace. However, equally prevalent, and far more powerful, are graphics processing units (GPUs). GPUs typically have over 100 cores and generate an order of magnitude greater floating point performance than CPUs. This seminar introduces the GPU as a science co-processor with fluid dynamics simulations as the example application.

This talk is part of the Computer Laboratory Wednesday Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2017 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity