parallel Computer Algebra based on Maple on Java

Speaker: Wolfgang Schreiner

Time: Tuesday 2nd October 2000, at 13.15

Place: Room 2.33

Abstract:

We describe an ongoing project on the parallelization of symbolic algorithms written in the language of the computer algebra system Maple. For this purpose, we have developed Distributed Maple, a parallel programming environment that is based on a distributed coordination framework written in Java. This framework exposes a parallel programming model that is modelled after principles from functional programming, logic programming, and dataflow. We have implemened with the help of this system a variety of algorithms in algebraic geometry (based on sequential versions in the CASA library developed at our stitute) and achieved significant speedups on heterogeneous clusters of PCs and workstations, on distributed memory systems, and on shared memory multiprocessors.