Hewlett-Packard announced this week it has won a contract with the federal energy department to build a $24.5 million supercomputer for its Pacific Northwest National Laboratory (PNNL) that will run on the popular open source operating system.
The new supercomputer is expected to have total peak performance of more than 8.3 teraflops, or roughly 8,300 times faster than a current PC.
Further, the Linux-based supercomputer is expected to be more than 30 times faster, have 50 times more disk space and have 10 times more memory than PNNL's current IBM computer, one of the world's most powerful when installed five years ago. Calculations that currently take a month to complete could be done in one day on the planned system.
The Linux supercomputer primarily will be used to study complex chemical problems that form the basis for advances in life sciences, subsurface transport, material design, atmospheric chemistry and combustion.
Though supercomputers traditionally have been used for compute-intensive problems, the evolution of Grid and distributed technologies are offering cost-effective ways for scientists, researchers and engineers to harness the computing power of large networks, including the Internet, by enabling problems to be broken down and parceled out to hundreds or even thousands of individual computers.
Editor's note: For more information about Grid and distributed computing, visit Grid Computing Planet.