News & Announcements

April 30, 2013

"Lois Curfman McInnes, Senior Computational Scientist"

What role do you play at Argonne, and what kind of work you do?

I do research on computational science; I work on developing algorithms and software for the solution of large-scale nonlinear systems arising in simulations based on partial differential equations, which are used to model many phenomena in science and engineering. One aspect of my work that I really enjoy is that it's interdisciplinary, involving collaboration among applied mathematicians, computer scientists and researchers in various scientific and engineering domains.

Many of our projects are team-based, and that’s been the foundation for much of my research, including work on the Portable, Extensible Toolkit for Scientific Computing (PETSc). My research has been in close collaboration with incredibly talented people. That's one of the things I like about computational science, especially computational science here at Argonne — partnerships working together to tackle challenges in science.

What attracted you to work at Argonne?

Argonne's Mathematics and Computer Science division is well known for global leadership and innovation in numerical software and high-performance computing. When I began working at Argonne in 1993, I was excited and grateful to be able to be a part of this unique work environment, and I certainly feel the same way now.

Can you explain the focus of your research on numerical software?

My focus is solving systems of nonlinear equations that arise in models based on partial differential equations — in particular, developing algorithms that solve such systems efficiently on high-performance computers. An area of current emphasis is the development of composable solvers for extreme-scale architectures, where we design efficient and robust algorithms by combining several pieces that leverage understanding of particular parts of a problem. Developing scalable solvers for multiphysics problems is another focus. These are challenging because multiphysics models couple together two or more models that traditionally have been solved separately. These new combinations raise many new challenges in modeling, algorithms and software.

What does a typical work day look like for you?

A ...

 

Continue reading: http://www.anl.gov/articles/lois-curfman-mcinnes