NSF-supported centers provide 800 million normalized units

released September 27, 2005

In the September 2005 round of allocating supercomputing time, 800 million normalized units (NUs) were meted out on supercomputing systems around the country that are supported by the National Science Foundation. This is more than the total NUs allocated in 2004's four allocation rounds combined.

Of that total, more than 410 million NUs were allocated on systems supported by NSF CORE funding. The balance was on systems supported by the NSF's Extended Terascale Facility or TeraGrid.

More than 377 million units were made available on systems at NCSA. About 162 million units were allocated on NCSA's massive Tungsten cluster. User requests for Tungsten were almost double that number, far exceeding the number available. This made Tungsten the most requested and the most allocated system in NSF's arsenal in September.

The Large and Medium Resource Allocation Committees (LRAC and MRAC) peer review proposals for time on these systems.

Scientists and engineers at research institutions throughout the U.S. apply for allocations on these high-performance systems to conduct research that would not be possible on smaller local computing systems or desktop computers.

A normalized unit is the equivalent of one hour of computing time on a single processor of a Cray X-MP supercomputer. Based on the LINPACK benchmark, it is used by the NSF programs as the standard measure for comparing time allocated on multiple systems.

 

Briefs Archive