Data Parallel Simulation of Self-organizing Maps on Hypercube Architectures

Erich Schikuta and Claus Weidmann
Inst. of Applied Computer Science, Dept. of Data Engineering\\
University of Vienna,
Rathausstr. 19/4, A-1010, Vienna, Austria

Email: schiki@ifs.univie.ac.at


Abstract:

In this paper a data parallel model for the parallel simulation of self-organizing maps is proposed. This approach is based on the SPMD (single program multiple data) model and utilizes the highly specialized programming environment of the Vienna FORTRAN Compilation System (VFCS) for the parallelization process. It allows an easy development of the simulation system as a sequential program attributed with data distribution information shifting the difficult parallelization process to the compiler. An analysis of the runtime behavior of the developed parallel programs for the simulation of self-organizing maps for different data distribution schemes is given, which justifies the efficiency of the proposed approach.


WSOM'97