Session on High-Performance Computer Algebra
Session Organizers: Jeremy
Johnson and Marc Moreno Maza
Abstract:
Improved algorithms, better implementations, and faster computers
have enabled many previously time-consuming computer algebra computations to
be performed routinely and have extended the range of what is practically
possible to compute. However, there remains many computations that still require
excessive computing time, and while existing computer algebra systems have had some
practical success in applications, their widespread use in computational science
and engineering remains limited. Part of this is due to inherent difficulties
of exact computation; nevertheless, there are many cases where the performance achieved
by an implementation could be dramatically improved through optimized implementations
and parallel computation and the use of special purpose hardware.
While the computer algebra community has begun to incorporate high performance computing
techniques, tuning algorithms to perform
well on modern computer architectures and adapting algorithms and systems to parallel
computers can be a difficult and time consuming process; much work remains in these directions. Moreover,
there are many challenges in achieving high-performance in computer algebra algorithm
implementations due to their irregular structure and higher level data types.
Also the complexity of computer algebra systems make the incorporation of parallel computation
challenging.
This session is devoted to exploring the application of
high-performance computing to computer algebra algorithms, applications and
systems, and the research and implementation challenges this poses.
Potential Topics:
- Practical implementation and performance analysis of computer algebra algorithms
- Implementation and optimization techniques for computer algebra algorithms
- Parallel implementations of computer algebra algorithms and parallel computer algebra systems
- The application and extension of optimizing compiler techniques to
the implementation of computer algebra algorithms.
- Adapting the implementation of computer algebra algorithms to improve
performance by better utilizing the underlying hardware
- Techniques for automating the optimization and platform adaptation
of computer algebra algorithms
- Cache complexity and cache-oblivious computer algebra algorithms
- Hardware accelaration technologies (multi-cores, GPUs, FPGAs)
Conference Proceedings:
Papers presented at the session will be considered for publication in a special issue of the Journal of
Symbolic Computation dedicated to high-performance computer algebra
Call for Submissions:
If you are interested in presenting a talk in this session, please upload
an abstract by May 15, 2009 at the following URL.
http://www.easychair.org/conferences/?conf=hpcaaca2009.
Accepted contributions:
- A computer algebraist meets a computer centre director
(James Davenport)
- Towards an efficient implementation for the resolution of structured linear system
(Benoit Lacelle and Eric Schost)
- Parallel Disk-Based Computation and Computational Group Theory
(Eric Robinson, Gene Cooperman, Daniel Kunkle and Jürgen Müller)
- Multitasking Polynomial Homotopy Continuation in PHCpack
(Jan Verschelde)
- Memory Efficiency in Polynomial Multiplication
(Daniel Roche)
- Fast multiplication and its variants in Newton iteration
(Ling Ding and Eric Schost)
- Balanced Dense Polynomial Multiplication on Multi-cores
(Yuzhen Xie and Marc Moreno Maza)
- SPIRAL-Generated Modular FFTs
(Jeremy Johnson and Lingchuan Meng)
- A Note on the Performance of Sparse Matrix-vector Multiplication with Column Reordering
(Sardar Haque and Shahadat Hossain)
- Basic Polynomial Algebra Subroutines
(Yuzhen Xie and Marc Moreno Maza)
- Code Generation and Autotuning in Computer Algebra
(Jeremy Johnson, Werner Krandick, David Richardson and Anatole Ruslanov)
- Linear Algebra Modulo Tiny Primes
(Jean-Guillaume Dumas, David Saunders and Bryan Youse)
)
|