acm-header
Sign In

Communications of the ACM

ACM TechNews

High-Performance Computing Turns to Apps to Cut Cost and Frustration


View as: Print Mobile App Share:
IBM Blue Gene supercomputer

IBM Blue Gene supercomputer

Credit: IBM

The concept of apps could enable more companies to take advantage of high-performance computing (HPC), considering the cost of HPC hardware and software makes adoption a challenge for all but the largest manufacturers.

With the app model, the focus is on producing software that can do one task, compared to a large system that is designed to do many functions. There is now an effort to take code that contains a very small, self-contained calculation designed for one task, such as modeling and simulating the flow of a fluid though a pipe, and turn it into an app.

For example, NanoHub.org, a U.S. National Science Foundation-supported platform, offers a way for companies to make code available as apps. The platform offers access to research software, and users will be able to do things such as model nanoscale electronic devices. Apps also address the need to find expertise to run HPC systems. "The challenge with expertise in artificial intelligence-type software is there are potentially lots of decision points," says NanoHub's George Adams. "If you focus your app on a narrower application area, the number of decision points is reduced and you do a better job of incorporating the expertise into the software."

From Computerworld 
View Full Article

Abstracts Copyright © 2012 Information Inc., Bethesda, Maryland, USA 


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account