Andreas Kloeckner

Research Interests

• High-Order Numerical Methods for Partial Differential Equations (mainly elliptic and hyperbolic)
• Integral Equation Methods and Fast Algorithms
• High-Performance Scientific Computing
• Software Infrastructure for Scientific Computing

More:

• Google Scholar
• Arxiv
• CV
• Software on Github
• ORCID

Group

Software

My group and I have created and continue to maintain a fair amount of open-source software, most of which relates to scientific computing and our research in some way:

• Pytential: Asymptotically Fast 2D/3D Layer Potential Evaluation
• Loopy: Transformation-Based Generation of High-Performance CPU/GPU Code
• Boxtree: High-performance oct/quadtrees, plus FMM traversals
• Meshmode: High-Order Meshes and Discontinuous Function Spaces
• PyOpenCL: Pythonic Access to OpenCL, with Arrays and Algorithms
• PyCUDA: Pythonic Access to CUDA, with Arrays and Algorithms
• Relate: Online Environment for courses, see for example https://relate.cs.illinois.edu
• PuDB: Debugger for Python with a Text UI
• Pymbolic: Symbolic computation in Python

Many more… (including meshpy, modepy, pyvisfile, cgen, genpy, pymetis, …)

Bio Blurb

Andreas Klöckner is an associate professor in the scientific computing area within the Department of Computer Science at the University of Illinois at Urbana-Champaign. His research focuses on high-order accurate integral equation methods and fast algorithms for elliptic boundary value problems as well as code transformation for high-performance scientific computing. He is the recipient of a 2017 National Science Foundation CAREER Award. In support of his research, Dr. Klöckner has released numerous scientific software packages. Previously, he was a Courant Instructor at the Courant Institute of Mathematical Sciences at New York University with Leslie Greengard, after obtaining his PhD degree from the Division of Applied Mathematics at Brown University in 2010 with Jan Hesthaven.

Prospective Students: Questions and Answers

Q: Why Scientific Computing at Illinois?

• UIUC has one of the top computer science programs in the US and globally, with strong groups in HPC, compilers, architecture, AI, security, software engineering, systems (among many others).
• UIUC has many top-ranked engineering programs—-and thus many possible ways to make your numerics research count in applications. We are often involved in PSAAP Centers, which are one avenue for such collaboration, also involving national laboratories, with whom we have a close relationship.
• We, the scientific computing area in CS at Illinois are one of the larger scientific computing groups in the country, with many faculty and many students. Being situated in a CS department enables us to focus on math, methods, algorithms, and implementation, depending on application need, providing breadth as well as depth across these areas. Many of us see value in and publish open-source software tools as one of the pillars of scientific computing.

Q: What are you looking for in prospective students?

I am always looking for passionate, hard-working students to join my research group. The most important quality I look for in students is self-motivation: as a PhD student you need to take charge and proactively drive your own research agenda. I also like students who are energetic and curious, which are critical qualities to effectively identify interesting research problems and find solutions. In terms of technical skills, I expect students to have both mathematical and computational training. That is, you should be comfortable around, say, both Sobolev spaces and GPUs.

Q: What should I do if I want to work with you?

The best thing to do is to read several of my recent papers, and then write me a brief email (with your CV attached) explaining: 1) which of those papers most resemble the work you would like to do and why, 2) why you want to get a PhD, and 3) what research problems you are interested in.

Keep in mind that my research focuses on the numerical solution of partial differential equations. I am also interested in high-performance computing and tools to support it. If you aren’t interested in those areas, then I am not the right advisor for you. You can mention my name in your application so that I can easily find your case. Make sure you select the “Scientific Computing” area in your graduate application.

Advice for Current Grad Students

In the form of links because (1) other people have already said it well and (2) you don’t have to believe me—believe someone more senior!

• Managing Your Advisor by Nick Feamster
• Some Modest Advice for Graduate Students by Steven C. Stearns (original)
• Why I gave your Paper a Strong Reject by Matt Welsh

(Thanks to Gang Wang, whose page I used as a template for the prior two sections.)