Name: Erik Boman
Pronouns: he/him/his
Biography:
Erik Boman is a scientist in the Center for Computing Research at Sandia National Labs, where he has worked for over 20 years. He holds a PhD in Scientific Computing from Stanford University. His research interests are in scientific computing, parallel computing, sparse matrix algorithms, graph algorithms, and combinatorial scientific computing.
Institution/Lab: Sandia National Laboratories
Website: https://egboman.github.io/
SRP Collaboration Topic/Title: Preconditioners to speed up linear solvers
Field or research area: Applied mathematics, computer science
Please select all the topical areas that apply to your project:
Computational Science Applications (i.e., bioscience, cosmology, chemistry, environmental science, nanotechnology, climate, etc.); Computer Science (i.e., architectures, compilers/languages, networks, workflow/edge, experiment automation, containers, neuromorphic computing, programming models, operating systems, sustainable software); High-Performance Computing
Brief Abstract:
The most expensive part of many simulations is the linear solves. Preconditioners are critical to accelerate iterative solvers for large, sparse problems. We will develop and implement algebraic preconditioners that can be used as a black-box for a wide variety of matrices. We propose a new variation of incomplete factorizations. An initial implementation may be done in Matlab or Python, but the eventual goal is to do a parallel implementation and study its performance. If good results are obtained, a research publication is a strong possibility. The software may potentially become part of the Trilinos framework, and thus accessible to thousands of researchers world-wide.
Desired relevant skills, background, or interests:
Numerical linear algebra, computer programming, familiar with version control (e.g., git/github), some experience with GPU and/or parallel programming desired.
Other comments:
Do any special requirements apply? Minimum GPA (specify what GPA in comments below); In-Person Only; Permanent Resident OK
Other, specify:
Keywords:
high-performance computing; linear algebra; sparse matrices; preconditioners
Lightning Talk Title: Preconditioning for Solving Large Sparse Linear Systems