Distributed Computing for Localized and Multilayer Visualizations
Abstract
The aim of this paper is to develop an approach to visualizations that benefits from distributed computing. Three schemes of process distribution are considered: parallel, pipeline, and expanding pipeline computations. Expanding pipeline structure synthesizes the advantages and traits of both parallel and pipeline computations. In expanding pipeline computing, a novel approach presented in this paper, a multiplicity of processes are concurrently developed in parallel and knotted processor pipelines. The theoretical foundations for expanding pipeline computing as a computational process are in the domains of alternating Turing machines, molecular computing, and E-machines. Expanding pipeline computing constitutes the development of the conventional pipeline architecture aimed at utilization of implicit parallel structures existing in algorithms. Such structures appear in various kinds of visualization. Image deriving and processing is a field that provides diverse opportunities for utilization of the advantages of distributed computing. The most relevant to the distributed architecture is stratified visualization with its two cases based on data localization and layer separation. Visualization is treated here as a special case of simulation. The conceptual approach to distributed computing developed in this paper have been applied to visualization in a computer support system, which is utilized in radiology and namely, for the noninvasive treatment of brain aneurysms.
Cite
@article{arxiv.cs/0111022,
title = {Distributed Computing for Localized and Multilayer Visualizations},
author = {Mark Burgin and Walter Karplus and Damon Liu},
journal= {arXiv preprint arXiv:cs/0111022},
year = {2007}
}