Adapt3D: A Framework for Supporting CFD and MHD Modeling
Professor John E. Dorband
Research Associate Professor and MCC Chief Scientist
Computer Science and Electrical Engineering
University of Maryland, Baltimore County
1:00pm Friday 21 October 2011, ITE227
Adapt3D is a software framework that supports 3-D computational fluid dynamics (CFD) and magnetohydrodynamics (MHD) modeling on an unstructured mesh. It facilitates adaptive refinement and execution in a parallel computing environment, either on shared memory or distributed memory computer architectures. The framework was designed primarily to support parallel computing of simulations on a continuum expressed by an adaptively refined unstructured mesh. It was designed to cleanly differentiate the code that manages an unstructured mesh on a parallel computer and the code that performs the numerics on the components of the mesh. The intent is to ease the burden of the engineer or scientist when it come to understanding parallel computing or adaptive refinement, allowing him to spend his time understanding the computational science rather than the computer science.
Dr. John Dorband received a BA in Math and Physics from Northwest Nazarene University in 1972 and a PhD in Computer Science from The Pennsylvania State University in 1985. He work for NASA for 21 years doing research in parallel and high performance hardware architectures and software. He has developed parallel compilers, algorithms and applications. He was a member of the group that developed the concept of developing high performance computing architectures from commodity components (the Beowulf project).
Host: Yelena Yesha