Sandia National Laboratories is a multi-program U.S. Department of Energy laboratory with major sites in Albuquerque, New Mexico and Livermore, California. Sandia's mission areas include national security, industrial competitiveness, energy resources and environmental quality.
High performance computing is already a key element of almost every major Sandia program, as well as programs at the other DOE laboratories. However, tremendous increases in computational capability, both hardware and software, are still needed for many applications. In the future, computer simulations must replace large-scale experiments that are no longer acceptable due to environmental consequences or costs. Simulations will also lessen dependence on expensive prototyping and retooling; entire systems will be computer simulated, redesigned, and optimized prior to actual assembly.
Some current application areas in high performance computing at Sandia include shock physics, computer design of materials, chemical reacting flows, low density flows, seismic processing, and medical imaging. Our recent highlights include: a series of Jupiter impact simulations alerted astrophysicists to the fact that fireballs would be visible when a series of comet fragments struck Jupiter last July; optoelectronics and catalysts with application-specific properties have been designed and optimized on high performance computers; low density flow models have been used to analyze and optimize semiconductor manufacturing processes; and a Sandia-led industrial consortium is drastically reducing the time required to generate computational grids for unstructured finite-element calculations.
The Department of Energy is the Nation's largest consumer of computational scientists and engineers and, consequently, has a stake in insuring that there is an adequate supply of this human resource. The Department supports computational science activities and programs at the K-12, undergraduate, and graduate levels, as well as post-doctoral research opportunities in its laboratories. Several universities have now heeded the call, and "programs" in scientific computation are becoming more prevalent. The successful programs are interdisciplinary in nature, are oriented towards applications and problem solving, and use computation as an essential component.
In general, we would expect a computational science graduate to possess knowledge of applied mathematics and numerical methods; knowledge of and exposure to high-performance computing; general knowledge of computer science; knowledge in a science discipline; interest or experience in solving real problems; and the ability to interact in a team environment. Successful computational science activities are best accomplished by people with these qualifications working cooperatively in a team environment with applied mathematicians, and computer scientists. It is this environment that has contributed to the success of many of Sandia's programs in high performance computing.