Costas Bekas

 
 

C. Bekas is a Research Staff member with the IBM Zurich Research Lab. He received his B. Eng., Msc and PhD, all from the Computer Engineering & Informatics Department, University of Patras, Greece, in 1998, 2001 and 2003 respectively. Between 2003-2005, he worked as a postdoctoral associate with prof. Yousef Saad at the Computer Science & Engineering Department, University of Minnesota, USA. He has been with IBM since September 2005.

 

News

Oct2009 Supercomputing 09. Uncertainty Quantification: Quadratic cost inverse covariance matrix estimation. How to analyze TBytes of data in less than 20min, achieving 730 TFLOPS on the BG/P Supercomputer. (link)

Aug2009 Supercomputing 09: Can Multicore Protect your Bones?

Aug2009 SIAM 09: Invited minisymp. on non-linear eigenvalue problems

May2009 SIAM Act. Group on Supercomputing Nomination Committee member

Jan2009 Professional Development Evening @ SIAM CS&E 2009

Dec2008 SIAM CS&E 2009: minisymposium on electronic structure calculations (with C. Yang)

July2008 Osteoporosis simulations on Blue Gene hit the news (google search)

June2008 Guest Editor for Elsevier Parallel Computing PMAA 08 special issue (call for papers)


In the past

May2008 Analysing human bone strength on the supercomputer, ISC 08

Jan2008 Minisymposium in Parallel Combinatorial Scientific Computing PMAA 08 (with P. Arbenz )

Feb2008 Human bon e structure analysis, accepted in ECCOMAS 08

Jan2008 Electronic Structures minisymp. accepted in SIAM OPT. 08 (with C. Yang )


Contact Info

Computational Sciences Group

IBM Research, Zurich Research Laboratory

Saumerstrasse 4,

CH-8803, Rueschlikon, Switzerland


Email: bek AT zurich DOT ibm DOT com

P: +41 44 724 8969

F: +41 44 724 8958



Research Interests

  1. -Computational Science & Engineering

  2. -Computational Materials Science

  3. -Supercomputing / High Performance & Productivity Architectures

  4. -Numerical Linear Algebra



Education

University of Patras, CEID, Greece, 1993-2003

  1. -PhD in Computer Science, June 2003. Thesis title: Efficient Computation of Matrix Pseudospectra: Methods and Software Tools. Supervisor: Prof. E. Gallopoulos.

  2. -MsC in Computer Engineering & Informatics, May 2001. Thesis title: Towards the Efficient Computation of Matrix Pseudospectra. Supervisor: Prof. E. Gallopoulos.

  3. -Dimploma in Computer Engineering & Informatics, July 1998.



Awards & Distinctions

  1. -Bodossaki Foundation Doctoral Scholarship

  2. -Technical Chamber of Greece, Award of Excellence, 2000

  3. -CEID, Doctoral Fellowship



Major Software Projects

  1. -Car-Parinello Molecular Dynamics (CPMD)

  2. -Quantum Espresso Molecular Dynamics




Publications


Journals

10. C. Bekas, A. Curioni and W. Andreoni, Atomic Wavefunction Initialization in ab initio Molecular Dynamics using Distributed Lanczos, Parallel Computing, Elsevier, 34 (6), p.441-450, Jul 2008 (link to paper)

9. C. Bekas, E. Kokiopoulou and Y. Saad, Polynomial Filtered Lanczos Iterations with Applications in Density Functional Theory, To appear, SIAM J. Mat. Annal. Appl. (Dec 2007) (link to article)

8. Y. Zhou, Y. Saad, C. Bekas, M. Tiago and J. Chelikowsky, Diagonalization methods in PARSEC, Phys. St. Sol., 243 (9), pp. 2188-2197, 2006.

7. C. Bekas, E. Kokiopoulou, and Y. Saad. An estimator for the diagonal of a matrix, 2005. To appear in Appl. Num. Math. (2007).

6. C. Bekas, Y. Saad, M. Tiago and J. Chelikowsky. Computing Charge Densities with Partially Reorthogonalized Lanczos, March 2005. Comp. Phys. Comm., Volume 171, Issue 3, Pages 175-186.

5. C. Bekas and Y. Saad: Computation of Smallest Eigenvalues using Spectral Schur Complements, SIAM J. Sci. Comp., 27 (458), 2005.

4. C. Bekas, E. Kokiopoulou and E. Gallopoulos, The Design of a Distributed PSE for Computing Pseudospectra, Future Generation Computer Systems, 21(6), pages 930-941, Elsevier, 2005.

3. E. Kokiopoulou, C. Bekas and E. Gallopoulos, Computing Smallest Singular Triplets with Implicitly Restarted Lanczos Bidiagonalization, J. Appl. Num. Mathematics, 49(1), 2004, pages 39-61.

2. C. Bekas and E. Gallopoulos, Parallel Computation of Pseudospectra by Fast Descent, Parallel Computing, v. 28, pp. 223-242, 2002.

1. C. Bekas and E. Gallopoulos, Cobra: Parallel Path Following for Computing the Matrix Pseudospectrum, Parallel Computing, v. 27, pp. 1879-1896, 2001


Conferences

21. C. Bekas, A. Curioni, D. Christen and R. Mueller, "Multicore Computers Can Protect Your Bones!", Supercomputing 09 (poster), Portland, USA, Nov. 2009.

20. C. Bekas, A. Curioni and I. Feduova, "Low Cost High Performance Uncertainty Quantification", Worskhop on High Performance Computational Finance, Supercomputing 09, Portland, USA, Nov. 2009.

19. C. Bekas and A. Curioni, "Massively Parallel Ab Initio Electronic Structure and Molecular Dynamics Calculations", SIAM CSE 09, Miami, USA, March 2009.

18. C. Bekas, A. Curioni, P. Arbenz, C. Flaig, G. H. van Lenthe, A. J. Wirth and R. Mueller, "Extreme Scalability Challenges in Analyses of Human Bone Structures", ISC 08, Dresden, Germany, June 2008 (link to paper).

17. C. Bekas, A. Curioni, P. Arbenz, C. Flaig, G. H. van Lenthe, A. J. Wirth and R. Mueller, "Extreme Scalability Challenges in Analyses of Human Bone Structures", ECCOMAS 2008, Venice Italy, July 2008.

16. C. Bekas, Y. Saad, "Polynomial Filtered Lanczos Iterations", NUMAN 2007, Sept. 2007, Kalamata, Greece.

15. C. Bekas, A. Curioni and W. Andreoni, New scalability frontiers in ab initio electronic structure calculations using the BG/L supercomputer, Springer, Lecture Notes in Computer Science, 4699/2007, pages 1026-1035, 2007.

14. C. Bekas, E. Kokiopoulou, E. Gallopoulos and V. Simoncini, Parallel Computation of Pseudospectra Using Transfer Functions on a MATLAB-MPI Cluster Platform, 9th EuroPVM/MPI, September 29th - October 2nd, 2002, Linz, Austria.

13. C. Bekas, E. Kokiopoulou, I. Koutis and E. Gallopoulos, Towards the effective parallel computation of matrix pseudospectra, ACM, ICS 2001, Sorrento, Italy, pp 261-270.

12. C. Bekas A. Curioni and W. Andreoni. "ab initio Electronic Structure Calculations on the BG/L Supercomputer ". PMAA 06, Rennes, France, Sept. 2006.

11. C. Bekas A. Curioni and W. Andreoni. "New scalability frontiers in ab initio electronic structure calculations using the BG/L supercomputer ". PARA 06, Umea, Sweden, June 2006.

10. C. Bekas K. Petrakos and E. Gallopoulos. "Towards Computing Pseudospectra of Parameter Dependent Matrices ". In Proc. Copper Mountain Conference on Iterative Methods, March, 2004.

9. C. Bekas and Y. Saad. "AMLS and Spectral Schur Complements ". In Proc. Copper Mountain Conference on Iterative Methods, March, 2004.

8. C. Bekas, E. Gallopoulos and V. Simoncini, "Transfer Functions and Path Following for Computing Pseudospectra". Extended abstract in the (electronic) Proc. SIAM Conf. in Applied Linear Algebra, July 2003.

7. E. Kokiopoulou, C. Bekas and E. Gallopoulos, "Implicitly Restarted Harmonic Lanczos Bidiagonalization for the Computation of the Smallest Singular Values of Large Sparse Matrices", 2nd International Workshop on Parallel Matrix Algorithms and Applications (PMAA'02), 9-10 November 2002, Neuchatel, Switzerland.

6. I. Koutis , C. Bekas, E. Kokiopoulou and E. Gallopoulos. "Advances in the Theory and Computation of Pseudospectra", SIAM Annual Meeting, July 8-12, 2002, Philadephia, USA. 

5. C. Bekas, I. Koutis, E. Kokiopoulou, A. Sidiropoulos and E. Gallopoulos. "Parallel Methods for the Computation of Matrix Pseudospectra", 5th IMACS Conference on Iterative Methods in Scientific Computing, May 28-31, 2001

4. C. Bekas, E. Gallopoulos. "Computing Pseudospectra by Fast Descent". PMAA 2000, August 18-21, 2000, Neuchatel, Switzerland.

3. C. Bekas, E. Gallopoulos and V. Simoncini. "On the Computational Effectiveness of Transfer Function Approximations to the Matrix Pseudospectrum". In Proc. Copper Mountain Conference on Iterative Methods, Vol. II, April 2000

  1. 2.C. Bekas, O. Bertrand, E. Gallopoulos and B. Philippe. "New Continuation Methods for the Computation of the matrix Pseudospectrum" HERCMA 98, Athens University of Economics and Business.

  2. 1.C. Bekas and E. Gallopoulos. Cobra: A hybrid method for computing the matrix pseudospectrum. In Proc. Copper Mountain Conference on Iterative Methods, 1998.




Invited Talks


- AMLS, Spectral Schur Complements and Iterative Computation of Eigenvalues, Scientific Computing Seminar, Lawrence Berkeley National Laboratory, Jan. 18, 2005

- Large Scale Computational Challenges in Materials Science, Computer Science Department Colloquium, The College of William and Mary, Jan 28, 2005

- Polynomial filtered Lanczos iterations with applications in Density Functional Theory, Seminar for Applied Mathematics, ETHZ, Zurich, Dec. 14, 2005

- Polynomial filtered Lanczos iterations, Minisymposium on eigenvalue methods, ICIAM 07, July 2007



DISCLAIMER

The opinions on this web page are not necessarily shared by the IBM Corporation. Copyrights of papers held in this page are with the respective societies or publishers, please respect them.  The preprints offered here are only for the promotion of scientific research.


Last Revision: June 2009

Laconic Bio