arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 1101.0605cs.DCastro-ph.CO
收藏

基于全球分布式超级计算机的高性能引力N体模拟

High Performance Gravitational N-body Simulations on a Planet-wide Distributed Supercomputer

Derek Groen, Simon Portegies Zwart, Tomoaki Ishiyama, Junichiro Makino

摘要:

我们报告的性能,我们的冷暗物质宇宙学N体模拟,同时进行使用全球各地的超级计算机。我们在阿姆斯特丹(荷兰,欧洲),东京(日本,亚洲),爱丁堡(英国,欧洲)和埃斯波(芬兰,欧洲)的各种超级计算机上分布的60到750个核心上进行了模拟。无论0.32秒的网络延迟和超过30.000公里的光纤网络电缆的通信,与单个超级计算机上相同数量的核心相比,我们都能够实现约87%的性能。我们认为,使用广泛分布的超级计算机,以获得更多的计算能力在技术上是可行的,最大的障碍是由本地调度和预订政策。

英文摘要:

We report on the performance of our cold-dark matter cosmological N-body simulation which was carried out concurrently using supercomputers across the globe. We ran simulations on 60 to 750 cores distributed over a variety of supercomputers in Amsterdam (the Netherlands, Europe), in Tokyo (Japan, Asia), Edinburgh (UK, Europe) and Espoo (Finland, Europe). Regardless the network latency of 0.32 seconds and the communication over 30.000 km of optical network cable we are able to achieve about 87% of the performance compared to an equal number of cores on a single supercomputer. We argue that using widely distributed supercomputers in order to acquire more compute power is technically feasible, and that the largest obstacle is introduced by local scheduling and reservation policies.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑