Post-doc and PhD student positions in Parallel and Scientific Computing
Our research in numerical linear algebra and parallel computing is internationally well-established and includes the development of new and smarter algorithms and library software (e.g., contributions to LAPACK, ScaLAPACK, SLICOT, ESSL). The goal is to efficiently manage deep memory hierarchies in state-of-the-art super- and parallel computers. The research, which is being conducted in collaboration with IBM T.J. Watson Research Center and several universities in Europe and USA, results in computational science and engineering building blocks that can supply the high performance necessary for researchers and engineers to carry out reliable computing with necessary resolution and high accuracy. One outcome is recursive blocked algorithms and hybrid data structures for dense matrix library software that have the potential to automatically adjust to a memory hierarchy. The research in numerical linear algebra comprises the whole spectrum of problems including perturbation theory and estimation of the conditioning of the problem, development of algorithms and computable error bounds, and their robust and efficient implementation in state-of-the-art software modules. Several of the problems considered are motivated by our collaboration within NICONET, an international association and network for numerics in control systems design and analysis.
HPC2N is a national center for scientific and parallel computing, providing the Swedish research community with first class services and world-leading expertise in high performance computing, grid computing, scientific visualization and virtual reality. HPC2N is hosting the most powerful academic computer system (Akka) in the Nordic countries (5 376 cores, an IBM blade cluster based on Intel’s 2.5 GHz Xeon quad-core processor, www.hpc2n.umu.se/resources/Akka). Akka is ranked #39 on the June 2008 TOP500 list and #16 on the associated GREEN500 list.
The Post-doc and PhD student positions are open in the following two projects:
- High Performance and Parallel Algorithms for Matrix Computations on Multicore-type Architectures
The research includes the development of methods, techniques, and parallel algorithms as well as high-quality library software that deliver close to practical peak performance on today’s evolving deep memory HPC multicore architectures. Main focus is on new methods for solving fundamental linear algebra problems that appear in most applications using, e.g., recursion, multi-level blocking, message passing and threading, and new matrix storage formats.
The positions in this project are funded by the Swedish Research Council (VR) and Umeå University.
- Matrix Pencil Computations in Computer-Aided Control System Design: Theory, Algorithms and Software Toolswww.cs.umu.se/~bokg/ssf/index.htm
The research includes the development of theory, algorithms, tools, and high-quality software for solving linear algebra problems with applications in control systems design and analysis. Examples are computing subspaces and canonical structure information, solving important (algebraic and differential) matrix equations, as well as complicated nearness problems. Other topics are large-scale eigenvalue problems (EVP), including product and periodic EVPs.
The positions in this project are funded by the Swedish Foundation for Strategic Research (SSF) and Umeå University.
For some recent outcome (publications, PhD Theses etc) of the research in these projects, see (www.cs.umu.se/~bokg/). The current positions are aimed to strengthen and broaden our current research, and include collaboration with our international academic and industrial partners. More specific information of the open positions follows below.
1-2 Post-doc positions in Parallel and Scientific Computing
(Dnr 315-2670-08)
The Department of Computing Science invites applicants for 1-2 Post-doc positions in Parallel and Scientific Computing. The positions are aimed for post doctoral research (one year with a possibility for a one year extension) jointly with researchers in our group and with our collaboration partners. Qualified applicants should have a PhD degree or equivalent in Computing Science, Scientific Computing or some other closely related field relevant for the position. The PhD degree should preferably not be more than three years old. A solid background is desirable in Computer Science and Numerical Analysis and strong programming skills including programming for parallel computers. Very good communication skills in written and spoken English are also required.
1-2 PhD student positions in Parallel and Scientific Computing (Dnr 313-2630-08)
The Department of Computing Science further invites applicants for 1-2 PhD positions (doktorandtjänst) in Parallel and Scientific Computing. A PhD position (doktorandtjänst) involves primarily doctoral research and course work (forskarutbildning). The employment extends over a period of four years, provided that expected study and research results are achieved. The position may also include part-time work (up to 20%) as a teaching assistant. If so, the total time of employment is extended accordingly.
Eligible candidates are required to hold a MSc degree in Computing Science, Scientific Computing or a closely related subject. Good communication skills in written and spoken English are also required.
The application can be submitted either electronically or in paper form and should contain:
- A cover letter including a short description of your research interests; for the post-doc position also a statement of previous research merits.
- A curriculum vitae.
- Copies of degree certificates.
- A publication list (may not be applicable if applying for a PhD student position).
- Reprints/copies of selected publications, numbered according to the publication list. (For the PhD student position this must include Bachelors/Masters thesis, for the Post-doc position the PhD thesis).
- For Post-doc applicants: a research plan (maximum 3 pages).
- Contact information for three reference persons.
For further information about the positions, please contact Professor Bo Kågström, phone +46 90 786 5419, e-mail bokg@cs.umu.se.
Union information is available from SACO, +46-(0)90-786 53 65, SEKO civil, +46-(0)90-786 52 96 and ST, +46-(0)90-786 54 31.
Your complete application, marked with the relevant reference number (315-2670-08 or 313-2670-08, see above) should be sent to jobb@umu.se (with reference number on the subject line) the Registrar, Umeå University, SE-901 87 Umeå, Sweden, to arrive September 24, 2008 at the latest.
