Flexible collective communication tuning architecture applied to Open MPI
Flexible collective communication tuning architecture applied to Open MPI
复制标题
应用于 Open MPI 的灵活集体通信调优架构
DOI：
--
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
E. Jeannot
中科院分区：
文献类型：
--
作者：
G. Fagg;Jelena Pjesivac;G. Bosilca;T. Angskun;J. Dongarra;E. Jeannot
集体通信对于现代高性能应用程序来说是无价的，尽管这些通信模式的大多数用户并不总是想知道它们的内部工作原理。集合体的实现通常留给中间件开发人员，例如那些提供 MPI 库的开发人员。由于许多此类库被设计为通用且可移植，因此 MPI 开发人员通常提供仅适合知识丰富的用户的内部调整选项，并允许一定程度的自定义。本文提出的工作不仅旨在提供一组非常有效的集体操作以与 Open MPI 实现一起使用，而且还使它们的控制和调整变得简单和灵活。此外，本文通过在运行时动态调整 MPI Alltoallv 算法，展示了所提出的框架灵活性的一个新颖示例。
Collective communications are invaluable to modern high performance applications, although most users of these communication patterns do not always want to know their inner most working. The implementation of the collectives are often left to the middle-ware developer such as those providing an MPI library. As many of these libraries are designed to be both generic and portable the MPI developers commonly offer internal tuning options suitable only for knowledgeable users that allow some level of customization. The work presented in this paper aims not only to provide a very efficient set of collective operations for use with the Open MPI implementation but also to make the control and tuning of them straightforward and flexible. Additionally this paper demonstrates a novel example of the proposed frameworks flexibility, by dynamically tuning a MPI Alltoallv algorithm during runtime.