Hierarchically Blocked Algorithms and Optimized Kernels for Dense Matrix Computations on Memory-Tiered High-Performance Computing Systems

 

 

Some recent publications and theses

The listings of publications (with support from the SSF research program) are in reversed chronological order, i.e., the most recent publications are listed first under each subtitle.

Submitted papers

  • Parallel and Cache-Efficient In-Place Matrix Storage Format Conversion. Fred Gustavson, Lars Karlsson, and Bo Kågström, ACM Trans. on Math. Software (submitted) Feb. 2010.
     
  • A Novel Parallel QR Algorithm for Hybrid Distributed Memory HPC Systems. Robert Granat, Bo Kågström, and Daniel Kressner. SIAM J. Scientific Computing (submitted), Also as Lapack Working Note #216, 2009. (pdf)
     

Refereed journal publications and invited book chapters

  • Parallel Solvers for Sylvester-type Matrix Equations with Applications in Condition Estimation, Part I: Theory and Algorithms, Robert Granat and Bo Kågström, ACM Trans. Math. Software (accepted), Feb. 2010. (pdf)
     
  • Parallel Solvers for Sylvester-type Matrix Equations with Applications in Condition Estimation, Part II: The SCASY Software Library, Robert Granat and Bo Kågström, ACM Trans. Math. Software (accepted), Feb. 2010. (pdf)
     
  • Distributed SBP Cholesky Factorization Algorithms with Near-Optimal Scheduling, Fred Gustavson, Lars Karlsson, and Bo Kågström, ACM Trans. on Math. Software, Vol. 36, No. 2, 2009 (Also as Report UMINF-07.19 and IBM Research Report RC24342). (pdf)
     
  • Parallel Eigenvalue Reordering in Real Schur Forms, Robert Granat, Bo Kågström, and Daniel Kressner. Concurrency and Computation: Practice and Experience, 21(9):1225-1250, 2009. (Also as LAPACK Working Note 192). (pdf)
     
  • RECSY and SCASY Library Software: Recursive Blocked and Parallel Algorithms for Sylvester-Type Matrix Equations with Some Applications, Robert Granat, Isak Jonsson, and Bo Kågström. In R. Ciegis et al., editor, Parallel Scientific Computing - Advances and Applications, Vol.27, pp. 3-24, Springe Optimization and Its Applications, 2009. (pdf)
     
  • Blocked Algorithms for the Reduction to Hessenberg-Triangular Form Revisited, Bo Kågström, Daniel Kressner, Enrique Quintana-Orti, and Gregorio Quintana-Orti, BIT Numerical Mathematics, 48(1):563-584, 2008, (Also as LAPACK Working Note 198). (pdf)
     
  • Block variants of Hammarling's method for solving Lyapunov equations, Daniel Kressner, ACM Trans. Math. Software, 34(1):1-15, 2008. (pdf)
     
  • Multishift Variants of the QZ Algorithm with Aggressive Early Deflation, Bo Kågström and Daniel Kressner, SIAM J. Matrix Anal. Appl., 29(1):199–227, 2006. (pdf)
     
  • Block algorithms for reordering standard and generalized Schur forms, Daniel Kressner, ACM Trans. Math. Software, 32(4):521-532, 2006.
     
  • Recursive Blocked Algorithms and Hybrid Data Structures for Dense Matrix Library Software, Erik Elmroth, Fred Gustavson, Isak Jonsson, and Bo Kågström, SIAM Review, Vol. 46, No. 1, 2004, pp. 3-45. (pdf)
     
  • Structured backward error for the linear system $A^TAx=b$, Ji-guang Sun, Journal of Natural Sciences of Heilongjiang University, 21, No.4 (2004), 4-10. (pdf)
     
  • Recursive Blocked Algorithms for Solving Triangular Systems - Part I: One-Sided and Coupled Sylvester-Type Matrix Equations, Isak Jonsson and Bo Kågström, ACM Trans. Math Software, Vol. 28, No. 4, Dec 2002, pp. 392-415. (pdf)
     
  • Recursive Blocked Algorithms for Solving Triangular Systems - Part II: Two-Sided and Generalized Sylvester and Lyapunov Matrix Equations, Isak Jonsson and Bo Kågström, ACM Trans. Math Software, Vol. 28, No. 4, Dec 2002, pp. 416-435. (pdf)

Refereed conference proceedings

  • A Framework for Dynamic Node-Scheduling of Two-Sided Blocked Matrix Computations, Lars Karlsson and Bo Kågström, In Proc. of PARA'08, accepted for publication, 2009. (pdf)
     
  • Parallel Algorithms for Triangular Periodic Sylvester-type Matrix Equations, Per Andersson, Robert Granat, Isak Jonsson and Bo Kågström, In E.Luque et al, Euro-Par 2008 Parallel Processing - 14th International Euro-Par Conference, LNCS 5168 of Lecture Notes of Computer Science, pp. 780-789, Springer , 2008. (pdf)
     
  • A Parallel Schur Method for Solving Continuous-time Algebraic Riccati Equations, Robert Granat, Bo Kågström and Daniel Kressner, In Proc. 2008 IEEE Conference on Computer Aided Control Systems Design (CACSD'08), 2008. (pdf)
     
  • Parallel Variants of the Multishift QZ Algorithm with Advanced Deflation Techniques, Björn Adlerborn, Bo Kågström, and Daniel Kressner. In Bo Kågström et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2006, Lecture Notes in Computer Science, LNCS 4699, pages 117–126, Springer, 2007. (pdf)
     
  • Recursive Blocked Algorithms for Solving Periodic Triangular Sylvester-Type Matrix Equations, Robert Granat, Isak Jonsson, and Bo Kågström, In B.Kågström et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2006, Lecture Notes in Computer Science, LNCS 4699, pages 531–539. Springer, 2007. (pdf)
     
  • Parallel Algorithms and Condition Estimators for Standard and Generalized Triangular Sylvester-Type Matrix Equations, Robert Granat and Bo Kågström, In B. Kågström et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2006, Lecture Notes in Computer Science, LNCS 4699, pages 127–136. Springer, 2007. (pdf)
     
  • Three Algorithms for Cholesky Factorization on Distributed Memory Using Packed Storage, Fred Gustavson, Lars Karlsson, and Bo Kågström, In B. Kågström et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2006, Lecture Notes in Computer Science, LNCS 4699, pages 550–559. Springer, 2007. (pdf)
     
  • A Parallel Block Iterative Method for Interactive Contacting Rigid Multibody Simulations on Multicore PCs, Claude Lacoursière, In B. Kågström et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2006, Lecture Notes in Computer Science, LNCS 4699, pages 956–965. Springer, 2007. (pdf)
     
  • Semi-automatic generation of grid computing interfaces for numerical software libraries,  E. Elmroth and R. Skelander, In J. Dongarra et al (eds.), Applied Parallel Computing. State-of-the-art in Scientific Computing.  Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2004, Lecture Notes in Computer Science, LNCS 3732, pages 404–412. Springer, 2006. (pdf).
     
  • Evaluating Parallel Algorithms for Solving Sylvester-Type Matrix Equations: Direct Transformation-Base versus Iterative Matrix-Sign-Functon-Based Methods,  Robert Granat and Bo Kågström, In J. Dongarra et al (eds.), Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2004, Lecture Notes in Computer Science, LNCS 3732, pages 719–729. Springer, 2006. (pdf)
     
  • Management of Deep Memory Hierarchies—Recursive Blocked Algorithms and Hybrid Data Structures for Dense Matrix Computations, Bo Kågström, In J. Dongarra et al., editor, Applied Parallel Computing: State of the Art in Scientific Computing, PARA 2004, Lecture Notes in Computer Science, LNCS 3732, pages 21–32. Springer, 2006. (pdf)
     
  • Combining Explicit and Recursive Blocking for Solving Triangular Sylvester-Type Matrix Equations on Distributed Memory Platforms, Robert Granat, Isak Jonsson and Bo Kågström, Euro-Par 2004 Parallel Processing, M. Danelutto et al (editors), LNCS Vol. 3149, 2004, pp. 742-750. (ps)(pdf)
     
  • Parallel ScaLAPACK-style Algorithms for Solving Continuous-Time Sylvester Matrix Equations, Robert Granat, Bo Kågström, and Peter Poromaa, Euro-Par 2003 Parallel Processing, H. Kosch et al (editors), LNCS Vol. 2790, 2003, pp. 800-809. (ps) (pdf)
     
  • RECSY - A High Performance Library for Sylvester-Type Matrix Equations, Isak Jonsson and Bo Kågström, Euro-Par 2003 Parallel Processing, H. Kosch et al (editors), LNCS Vol. 2790, 2003, pp. 810-819. (ps) (pdf) (library)

Technical reports and other publications
 

  • Blocked In-Place Transposition with Application to Storage Format Conversion, Lars Karlsson, Report UMINF 09.01, Dept. of Computing Science, Umeå University, Sweden, 2009. (pdf)
     

Theses

  • Blocked and Scalable Matrix Computations - Packed Cholesky, In-Place Transposition, and Two-Sided Transformations, Lars Karlsson, Ph Licentiate Thesis, UMINF 09.11, Dept. of Computing Science, Umeå University, Sweden, ISBN 978-91-7264-733-6, April 2009. (pdf)
     
  • Algorithms and Library Software for Periodic and Parallel Eigenvalue Reordering and Sylvester-Type Matrix Equations with Condition Estimation, Robert Granat, PhD Thesis, UMINF 07.21, Dept. of Computing Science, Umeå University, Sweden, ISBN 978-91-7264-410-6, November 2007. (pdf)
     
  • Software Tools for Matrix Canonical Computations and Web-Based Software Library Environments, Pedher Johansson, PhD Thesis, UMINF 06.30, Dept. of Computing Science, Umeå University, Sweden, ISBN 91-7264-144-X, November 2006. (pdf)
     
  • Recursive Blocked Algorithms, Data Structures, and High-Performance Software for Solving Linear Systems and Matrix Equations, Isak Jonsson, PhD Thesis, UMINF 03.17, Dept. of Computing Science, Umeå University, Sweden, ISBN 91-7305-568-9, December 2003. (pdf)

 

Swedish Foundation for Strategic Research Department of Computing Science High Performance Computing Center North