Research on Parallel Algorithm Library

在这项研究中，我们已经开发了有效的并行算法和算法库，我们的第一个贡献是开发一个可重新配置网格的可视化模拟器，并实现了一个在可重新配置网格上运行的多个并行算法。我们已经开发了以下特点：接受使用C类语言的算法描述，可视化并行算法的执行，以及提供各种统计信息，包括运行时间和延迟时间。实现并行算法包括总和和前缀-总和问题、排序和排序以及图形算法，如可达性和最低权重生成树。Fur\， we\ed an efficient geometric algorithm that computes the convex hull of n points in the plane in O（log log n） on the reconfigurable mesh of sizeイイD8nイエD8×イD8nイエD8。这个算法是对已知最好的算法的改进，它在O（log n）时间执行相同的任务。第二，我们在无线电通信环境下开发了一系列并行算法。我们开发的并行算法包括领导电子、初始化、路由、排序和排序。例如，我们开发了一个初始化算法，该算法在O（n）时间内为每个处理器在O（log n）时间内执行广播/接收操作。

In this research, we have been developed efficient parallel algorithms and algorithm library.Our first contribution is to develop a visualized simulator for reconfigurable meshes, and implemented a number of parallel algorithms running on the reconfigurable mesh. The simulator we have developed has the following feature : accept algorithm description using C-like language, visualize the execution of the parallel algorithm, and supply various statistic information including running time, and delay time. The parallel algorithms implemented include sum and prefix-sums problem, ranking and sorting, and graph algorithms such as reachability and minimum weight spanning tree. Further, we developed an efficient geometric algorithm that computes the convex hull of n points in the plane in O(log log n) on the reconfigurable mesh of size ィイD8nィエD8×ィイD8nィエD8. This algorithm is an improvement of the best known algorithm that performs the same task in O(log n) time.Second, we developed a number of parallel algorithms under the radio communication environment. The parallel algorithms we developed include leader electron, initializing, routing, ranking, and sorting. for example, we developed an initializing algorithm that gives n processors unique ID in [1, n] in O(n) time with each processor performing broadcast/receive operations in O(log n) time.