Dynamic visualization of computations on the internet
We present the design of a visualization tool which can be incorporated in parallel/distributed code to track computational parameters with respect to convergence and the distribution of the work remaining for the parallel processes. We describe its functionality, algorithms and implementation in Java, as it is applied to a package for multivariate integration. Implementation aspects include message passing and control flow in its interaction with the application code. The approach is mainly asynchronous and based on multi-threading.
Unable to display preview. Download preview PDF.
- 1.E. de Doncker, A. Gupta, J. Ball, P Early, and A. Genz, Parint: A software package for parallel integration, in 10th ACM International Conference on Supercomputing, Kluwer Academic Publishers, 1996, pp. 149–156.Google Scholar
- 2.E. de Doncker, A Gupta, and P. Early, Two methods for load balanced distributed integration, in Lecture Notes in Computer Science, vol. 1067, Springer-Verlag, 1996, pp. 562–570.Google Scholar
- 3.E. W. Dijkstra, Co-operating sequential processes, in Programming Languages, F. Genuys, ed., Academic Press, New York, 1965, pp. 43–112.Google Scholar