Research Opportunities for Graduate Students

If your transcript bears evidence of achievement in engineering and mathematics, and you wish to study computer science in an excellent environment with close contact to applied problems in high performance computing...

If you are able, when handed a paper you barely understand, to do self-motivated bibliographic and web searches, and to give a comprehensible short talk on it within a week...

If you cannot stop at a proof of concept, but enjoy seeing your ideas carried through to the implementation and transfer stages...

... then you probably have what it takes to be supported by me under existing or pending grants as a research assistant at ODU. Look first at the VILaP-HPCC and GAANN pages and feel free to e-mail me for further details.

Current projects of interest include:

• development of Newton-Krylov-Schwarz (NKS) methods and nonlinear Schwarz methods more generally
• application of NKS methods to systems with multiple models (different governing equations in different regions)
• tuning the parameters in NKS methods through the systematic application of optimization (e.g., parallel direct search)
• partitioning methods for unstructured grids that take into account underlying physics
• multiobjective partitioning methods that simultaneously balance computational work and communication work
• matrix-free methods in optimization
• optimal subdomain preconditioners for NKS methods, including local marching methods and multigrid methods
• reuse of Krylov subspaces in multiple, related systems
• automatic differentiation tools in computational fluid dynamics and radiation transport
• cache optimization for large-scale parallel engineering computations

(Not all of the interests above are presently funded, but some of them could be, given suitable candidates. Financial support during studies is available for doctoral candidates only, but masters candidates may find good CS 697 projects. Students seeking grant-based Research Assistantships apart from the established programs linked above should first earn an honors grade in at least one of my graduate courses in architecture or HPCC. Teaching assistantships are usually available from the department for an initial year of support, while your interests crystallize and while you find an advisor.)

Useful prerequisites in the CS department at ODU include:

• CS 417 Computational Methods & Software
• CS 486 Introduction to Parallel Computing
• CS 635 Parallel Computer Architecture
• CS 687 High Performance Scientific Computing
• CS 745 Topics in High Performance Computing