Corporate Affiliates Program

Modeling and Simulations

PSC can provide expertise in the development, implementation and execution of fundamental and applied models and simulations.

This visualization represents simulation of a one-and-a-half stage turbine. Color corresponds to temperature. Such simulations pinpoint likely blade hot spots, which helps in the design of cooling conduits and blade coatings. Electrical power generation is a multi-billion dollar global business, with developing countries creating a growing demand for the 21st century. To gain an edge in this fiercely competitive market, the key for companies like Westinghouse is more efficient turbines. High performance computing can help design more efficient turbines, by enabling engineers to solve problems involved with real turbine configurations, not just simplified versions. After four months of work, we had code that was more accurate and much faster than its sequential predecessor. Jobs that would have taken three months before now run in less than 12 hours.

This visualization represents simulation of a one-and-a-half stage turbine. Color corresponds to temperature. Such simulations pinpoint likely blade hot spots, which helps in the design of cooling conduits and blade coatings.

This image shows what regions in a subject's brain are involved in a memory task. This kind of study leads to improved understanding of In 1996, researchers and physicians at PSC, CMU, and UPMC created and demonstrated a breakthrough capability for viewing the brain during mental activity. By linking a magnetic resonance imaging (MRI) scanner at UPMC with the center's CRAY T3E supercomputing system, data from a subject's brain was processed faster than it was acquired by the scanner, making it possible to view a realistic 3-D image of brain activity while the subject was in the scanner. Thanks to progress in the efficiency of our complex statistical image analysis software and in our ability to cluster today's affordable PC-class workstations to act as parallel supercomputers, we can now achieve such real-time performance at a small fraction of the cost of a supercomputer. This will enable hospitals to perform real-time fMRI scans in-house.

This image shows what regions in a subject's brain are involved in a memory task. This kind of study leads to improved understanding of "working memory", which affects clinical treatment of schizophrenia and amnesias.

This is a snapshot from simulations of a protein called the villin headpiece subdomain. A multi-year collaboration between PSC, Dr. Peter Kollman's group at UCSF and other members of the computational structural biology research community resulted in the longest-ever simulation of the protein folding process using parallelized software. The results, reported in SCIENCE - the leading U.S. scientific journal - offer new insight into this fundamental process of life, which addresses the question of how in the first few microseconds of its existence a protein transforms from a stretched-out chain of amino acids to a folded structure.

This is a snapshot from simulations of a protein called the villin headpiece subdomain.

This image shows the results of computing the magnetic moments of a 512-atom unit-cell of iron above its Curie temperature, when the magnetic field begins to break down. The magentic moment at each atom (arrowhead vectors) is has a corresponfing constraining field (translucent cones) as a result of the constrained DFT model. Color indicates vector magnitude. Since 1993, Dr. Yang Wang, now at PSC, has been a member of one of the world's leading computational materials science teams, led by Dr. Malcolm Stocks of Oak Ridge National Laboratory. One of their main achievements has been the development of the Locally Self-Consistent Multiple Scattering (LSMS) algorithm and software. This is a highly accurate and efficient method for the first-principles quantum-level computation of the properties of alloys, designed from the start to exploit large-scale parallel supercomputers. It is being successfully applied to the study of magnetism in realistic materials, characterized by unit cells containing hundreds and even thousands of atoms. In November 1998, the team used a CRAY T3E-1200 provided by SGI to simulate a 1,458-atom unit cell of iron. This was the world's first fully fledged scientific computation to be clocked in excess of one trillion operations per second, and won the team the prestigious Gordon Bell Prize for 1998. The prize honors the year's best achievement in high performance computing.

This image shows the results of computing the magnetic moments of a 512-atom unit-cell of iron above its Curie temperature, when the magnetic field begins to break down. The magentic moment at each atom (arrowhead vectors) has a corresponding constraining field (translucent cones) as a result of the constrained DFT model. Color indicates vector magnitude.

For further information, please contact the Corporate Relations Office, Pittsburgh Supercomputing Center, (412) 268-4960 or by e-mail, corp-relations@psc.edu.