A relatively new and very dynamic technology is the General Purpose Graphic Processor Unit. The significance of the GPGPU is that it enables the transformation of an ordinary PC into an HPC (Highly Parallel Computer or High Performance Computer) at a cost of only $50-$500. The presentation will explain what GPGPUs are, their importance in the evolution of computing, their capabilities and limitations and finally how to start using them. The increased throughput from GPGPUs can enable more robust solutions in many disciplines: signal and image processing, bioinformatics, CAD, and other engineering oriented, computation-bound areas. The presentation will touch on the constructs of parallel computation and how to turn algorithms from serial execution to massively parallel execution in order to unblock compute-bound problems.

Mr. Morgenstern is a systems engineer who gets stupid computers to automatically reach intelligent conclusions when faced with massive volumes of noisy, inconsistent and sometimes contradictory sensor data. He has developed many successful products and systems that integrate sensors with image and/or signal processing. He has developed algorithms for the automatic extraction of information from a diverse set of sensors including multiband video cameras, FLIRs, multispectral scanners, LIDAR and other 3D imagers, Radar, Synthetic Aperture Radar, medical EKG, medical ultrasound, UV and XRay sensors. He was a part of the core team that developed the Black Beam Interferometer for 3D inspection of high quality surfaces. He has worked in remote sensing, industrial inspection, space observation and military sensing. He participated in the design of the Thematic Mapper earth resources sensor. Lately he has developed algorithms on massively parallel computers and also on GPGPU parallel systems. Mr. Morgenstern holds the BS degree from the University of Michigan in Physics and a MS degree in Systems Engineering from the University of Michigan School of Engineering. Mr. Morgenstern is a member of IEEE and SPIE.

These are two stand-alone apps that document the installed gpu(s) and exercise them with a dozen demos. These are really handy for anyone who purchases a gpu board and wants to verify hardware capability, workings, driver versions, etc.

http://www.softsea.com/review/NVIDIA-System-Tools.html

http://downloads.guru3d.com/GPU-Caps-Viewer-1.7.0-download-1584.html