Forum 06/12/2003

“Data Processing for EM”

Bill Young -Research Computing, The Scripps Research Institute, La Jolla, CA

The computational demands of EM data processing can be very high. The two widely users codes SPIDER and EMAN currently use different paradigms to run in parallel and have different computational demands. To obtain optimum performance you should match the code to the hardware. I will present a brief overview of the shared computational resources available for data processing, with an emphasis on EM refinement and obtaining optimal performance.