High-performance computing (HPC) uses supercomputers and computer clusters to solve advanced computation problems that cannot be addressed using a typical workstation or laptop. A supercomputer is by definition at the forefront of current processing capacity, particularly in regards to calculation speed. Super computers have been in existence since the 1960s, but it wasn't until the 1990s that machines with thousands of processors began to appear. By the end of the 20th century, machines with thousands of "off-the-shelf" processors became the norm.
HPC is used for tasks which are highly computation-intensive, or data-intensive, or both. Questions of science with once unattainable answers are being simulated and manipulated using HPC, unveiling statistics and approximations that classic lab experiments cannot accommodate. At William & Mary, HPC is applied to a diverse and ever-changing set of topics including: ocean modeling, quantum and particle physics, health sciences, acoustics, mathematics, operations research, and econometrics.
Jones basement holds our own HPC cluster dubbed "SciClone" at William & Mary. The SciClone project was conceived in the late 1990's to fulfill a dual role as a scientific computing resource and as a test bed for heterogeneous cluster computing. Initial funding was secured in 1999 from the National Science Foundation, and the first phase of the system went into production in August 2000. Since then, both the concept and the hardware have evolved. The focus is now squarely on research and instructional computing.
Now in its fifth generation of hardware, SciClone has a theoretical peak performance of 15 teraflops, or fifteen trillion floating point operations per second. This measure of compute capacity describes how many basic arithmetic functions (such as additions and multiplications) can be performed within one second under ideal circumstances.
SciClone is available free of charge to any member of the William & Mary faculty, staff, or student body with a legitimate need for its services in any discipline. For more information contact Eric Walter, manager of HPC, or email firstname.lastname@example.org.