SCREMS: A Parallel Computer Cluster For Multiphysics & Multiscale Modeling of Subsurface & Surface Flows

University Of Texas At Austin, Austin TX

Investigators

Abstract

The Center for Subsurface Modeling (CSM), part of the Texas Institute for Computational and Applied Mathematics at The University of Texas at Austin, will purchase 64 PCs, a control workstation, as well as a storage device, in order to constuct a parallel computing platform capable of 100 Gflops. The network for the cluster (high speed Myrinet network (1Gbit) and a 100Mb ethernet network) is already in place. This platform will be dedicated to research in the mathematical sciences being led by CSM. Projects that will use this platform include large-scale multiphysics modeling of subsurface processes, multiscale modeling of flow in extremely heterogeneous rocks, finite element coastal and ocean modeling, parallel terascale iterative solvers, and Discontinuous Galerkin discretization methods. A substantial component of each of these projects is the development of new scalable parallel algorithms as well as implementation and testing of both proof-of-concept as well as of production-quality codes for simulation of cases driven by energy and environment applications. A dedicated limited-access resource greatly expedites such development. The Center for Subsurface Modeling, part of the Texas Institute for Computational and Applied Mathematics at The University of Texas at Austin, will purchase equipment needed to construct a medium size parallel computer capable of 100 billion arithmetic operations per second. This computer will be dedicated to research in the mathematical and computational sciences being led by CSM. Researchers involved in the project have decades of experience in development of robust, accurate and fast computational techniques. The new cluster will allow for large-scale computational modeling and testing of such techniques which promise to improve ability to understand, predict, monitor, and control natural and engineered processes occuring in surface and subsurface. Significant examples include optimization of oil and gas recovery, analysis of contamination and remediation scenarios in subsurface and surface waters like aquifer and coastal environments. The access to a dedicated computing resource will greatly expedite such development.