NCSA officials say the system, to be called Lincoln, will be operational by next month. It will consist of 192 Dell PowerEdge servers, each one running two quad-core Intel 2.33 GHz processors. Each server will also have 16G of memory.
For additional computational power, the system will use 96 Nvidia Tesla S1070 accelerator cards. Each card is essentially a graphics processing unit repurposed for floating-point computational duties and offers an additional 500 gigaflops of power, along with 16G of memory.
"Achieving performance at the petascale and exascale and beyond may well depend on a heterogeneous mix of processors," said NCSA Director Thom Dunning, in a written statement. "The use of novel architectures for scientific computing is part of ongoing work at NCSA."
Technicians will network Lincoln with another NCSA supercomputer, Abe, which boasts a top speed of 89 teraflops. Together the two supercomputers will allow single applications to use up to 152 teraflops. The combined power of those computers would put them in eighth place among the most powerful computers in the world, as identified on the Top500's most recent list of the fastest supercomputers.
Lincoln will be used for weather modeling and to help build programming tools for researchers to make better use of multicore environments.
"There is a whole new constellation of parallel processing architectures now entering the mainstream," said Wen-mei Hwu, a University of Illinois professor who is leading a project to develop application algorithms, programming tools and software for computational accelerators. "It is crucial that we begin making use of them to drive scientific discovery and that we prepare the next generation of researchers to harness them."
Intel Set To Introduce Mobile Quad-core Processors
Microsoft Changes Virtualization Licensing Rules
