I know, I know... parallel processing has been around for years. Until recently, it has mainly been used in high-performance computing applications such as supercomputers and PC-based grid and cluster computing.  So what's the news? The technology is now finding its way into servers and sub-systems. What's facilitating this shift? Give a shout out to multicore processor technology!

 

Developers of high-end medical imaging systems like computed tomography (CT) imaging devices are definitely on-board. Systems like this require incredible compute power and are ideal candidates for parallel processing. Intel® Quad-Core processors are making parallel processing possible at the system level. By delivering four execution cores within a single processor, images that took several hours to process ten years ago can be processed in seconds. Wow!

 

While the benefits of parallel processing are great, there is a cost. There is a tremendous amount of knowledge needed to take maximum advantage of the technology. Surprisingly, this skill-set doesn't always reside within design teams. Don't get me wrong – these folks are incredibly talented, but understanding data flow, interdependencies and micro-architecture is critical. Here's the good news...  To get the benefit of parallel processing, while limiting the overhead required, companies can use a growing number of application-specific sub-systems that offer the compute power to handle complex image processing.

 

Today there are many factors to consider when deciding on a parallel processing compute solution for medical imaging. To help navigate the decision making process, visit:
http://www.radisys.com/news_events/white_papers/white_papers_form.cfm?id=70130000000DNZY&pid=29

 

Michael Reunert

Sr. Product Line  Manager, Embedded Servers and Medical  Imaging

http://www.radisys.com

 

Message Edited by serenajoy on 03-11-2009 08:28 PM
Message Edited by pmahler_intel on 03-12-2009 08:59 AM