无缝利用异构 XSEDE 资源以加速高价值功能神经影像数据集的处理
神经元与认知
2018-05-21 v3
摘要
我们描述了在 Open Science Grid 上处理海量高价值人类神经影像数据集的技术工作，通过机会式利用空闲 HPC 资源将计算能力提升了 5 倍以上。在极少的软件开发投入且未对其他 HPC 用户造成明显竞争性干扰的情况下，该工作在 7 个月内提供了 15,000,000 核时。
引用
@article{arxiv.1805.00786,
title = {Seemless Utilization of Heterogeneous XSede Resources to Accelerate Processing of a High Value Functional Neuroimaging Dataset},
author = {Don Krieger and Paul Shepard and Ben Zusman and Anirban Jana and David O. Okonkwo},
journal= {arXiv preprint arXiv:1805.00786},
year = {2018}
}
备注
8 pages, 10 figures, supplementary materials. Proceedings of the 2018 PEARC Conference, Practice & Experience in Advanced Research Computing, Pittsburgh, PA, July, 2018. arXiv admin note: substantial text overlap with arXiv:1710.05246. author note: The overlap with arXiv:1710.05246 is primarily in "background" material which includes a description of the solver and sample scientific findings