Science Grid This Week
Computational Chemistry Workshop Features Collaborations Large and Small
Computational chemists may work individually or in small groups, but they need access to the same networks of computing resources developed for large-scale science and engineering projects. Researchers developing the Computational Chemistry Grid are working to provide chemists with such resources using the GridChem application and supercomputing resources from five collaborating institutions.

A GridChem tutorial kicked off the fifth annual Computational Chemistry Workshop, held April 17-19 at the University of Illinois’ National Center for Supercomputing Applications in Urbana, Illinois. The workshop featured keynote speakers from the fields of chemistry and computing, including Sangtae Kim from the National Science Foundation, as well as presentations on the CCG and other grid projects from physics, molecular biology, nanotechnology and engineering.

"The CCG and GridChem are leveraging existing pieces of software and technology and using bits of glue to create an environment that will provide a collection of grid-based resources for chemical physics applications," explained the NCSA’s John Towns. "We’re trying to provide experimental chemists with an easier interface to simulation technology."

Computational chemists require large amounts of computing power to calculate molecular properties such as dipole moments and vibrational frequencies and to simulate molecular behavior. GridChem will launch and monitor those and other calculations and simulations on CCG computing resources, which already include over 400 computer processors and 3.5 million CPU hours annually available.

In his address to the workshop, Kim, the director of the NSF’s Shared Cyberinfrastructure Division, spoke about the need for collaboration among scientists in different disciplines in the development of middleware, the name given to the category of software that connects individual users to grid computing resources.

"Cyberinfrastructure is a system of interconnected computing resources, and as different communities develop middleware and applications to access the infrastructure, they face many of the same challenges," noted Kim. "Middleware is creating linkages across scientific disciplines as well as across space and time in computing environments."