

Project Goals 



SPICA ProjectThis project aims for a Skeletonbased Parallel Implementation of selected Computer Algebra Algorithms in Eden. Project GoalThe goal of this project is an abstract, mathematically profound specification and a parallel implementation of major algorithms of computer algebra. Such algorithms describe algebraic, exact, symbolic computations. Although an implementation in a functional programming language promises deep connections to the mathematical background, most implementations are imperative for the efficiency. However such approach blocks a flexible development and hence  an easy parallelization for multicores or grids. This projects selects few algorithms of computer algebra:
for a parallel implementation in Eden. The latter is a parallel extension of parallel functional language Haskell. A special focus is set on abstracting the fitting computation schemes to algorithmic skeletons. Such parallel generalized computation schemes are expressed as higher order functions and can be utilized for parallelizing whole classes of algorithms. Usage of highlyoptimized skeletons present a prospect of precision, flexibility, efficiency and versatility. As a whole, this project leads to the foundations of a flexible open parallel computer algebra system. Project TimelineThe project funding by DFG began in April 2009. As of now, skeletons and their instantiations are written and available for the following algorithms.
More information on skeletons, including the source code, is available on the designated page. TriviaSpica is the 15th brightest star, 260 light years away from the Earth. Hipparchus is believed to have observed it to discover precession of the equinoxes. To locate Spica at nighttime follow the arc of the handle of the Big Dipper to Arcturus and follow it forth for the same distance once more. The star name comes from Latin "ear of wheat (in the hand of the goddess)". 


Oleg Lobachev. EMail This page: http://www.mathematik.unimarburg.de/~lobachev 
Pages maintained by Oleg Lobachev Last change (content): 11/10/10 
