Victoria E. Howle
Department of Mathematics & Statistics
Texas Tech University
Office: Mathematics 243
Phone: (806) 834-8770
My research is in applied mathematics with a focus mainly on numerical linear algebra. My main research interests have been in physics-based preconditioning for incompressible fluid flow problems, develing iterative methods and preconditions for the solution of highly ill-conditioned systems that arise in faulted electrical power networks, and fault-tolerant linear algebra.
Numerical linear algebra: linear algebra is a fundamental component of scientific simulations. It is also often a computational bottleneck. I study methods for accelerating the solution of large linear systems in problems where standard techniques tend to fail. Some of these have included physics-based methods for incompressible fluids problems and for network problems in electrical power systems.
Fault-tolerant computing methods: As computer architectures become increasingly complex, systems failures become more prevalent. These include hard and solf failures. Hard failures are where a computational node in a large parallel simulation fails during the simulation; soft errors are more subtle. Extreme-scale computers operate at such low voltages that bit-flips can occur during a simulation. The affects of soft error on the simulation can be unpredictable and sometimes catastophic. I am interested in studying the affects of soft errors as well as developing new fault-resilient algorithms for scientific simulation.
Background & Other Info
Some useful Links
Last updated: August 2013
Copyright © 2008-2013, Victoria Howle and Department of Mathematics & Statistics, Texas Tech University All rights reserved.