Dynamic Shortest Paths using JavaScript on GPUs

Dynamic Shortest Paths using JavaScript on GPUs
复制标题

在 GPU 上使用 JavaScript 的动态最短路径

DOI:
--
复制
发表时间:
2015
期刊:
影响因子:
--
通讯作者:
A. Ingole
A. Ingole
中科院分区:
--
文献类型:
--
作者:
A. Ingole

文献摘要

被引文献

相似文献

互联网上的信息正在迅速增长,其处理需要高速的硬件和软件基础设施。JavaScript现在是Web应用程序的一个组成部分,这些应用程序执行从在线表单中的错误检查到处理Google地图的任务。由于其交互性,JavaScript应用程序的性能至关重要,特别是在处理大量不断变化的数据时。因此,最近一直在追求JavaScript代码的并行化。在这项工作中,我们的目标是GPU上的动态图算法的GPU并行化。我们目前的实现,并实现了有效的并行化的动态单源最短路径计算。我们比较了增量,减量和完全动态的版本对他们的静态对应,并显示高达约10%的更新,动态处理的GPU是有益的。
Information on the internet is growing rapidly and its processing needs high-speed infrastructure, both in hardware and software. JavaScript is now an integral ingredient of web applications which perform tasks ranging from error checking in online forms to processing Google maps. Due to their interactive nature, performance of JavaScript applications is critical, especially while handling huge volumes of evolving data. Therefore, parallelization of JavaScript code has been pursued in the recent past. In this work, we target GPU parallelization of dynamic graph algorithms on GPUs. We present implementation and achieve effective parallelization of dynamic single source shortest path computation. We compare the incremental, decremental and fully dynamic versions against their static counterpart and show that upto about 10% of updates, dynamic processing on GPUs is beneficial.