Kokkos Kernels: Performance Portable Sparse/Dense Linear Algebra and Graph Kernels
Kokkos Kernels: Performance Portable Sparse/Dense Linear Algebra and Graph Kernels
复制标题
Kokkos Kernels：性能便携式稀疏/密集线性代数和图内核
DOI：
--
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
I. Yamazaki
中科院分区：
文献类型：
--
作者：
S. Rajamanickam;Seher Acer;Luc Berger;V. Dang;Nathan D. Ellingwood;E. Harvey;Brian Kelley;C. Trott;Jeremiah J. Wilke;I. Yamazaki
随着硬件体系结构向亿亿级发展，开发计算科学与工程（CSE）应用程序依赖于可持续软件开发的性能可移植方法。本文介绍了一个方面的性能可移植性与开发一个可移植的内核库，服务于几个CSE应用程序和软件框架的需求。我们描述Kokkos Kernels，一个用于稀疏线性代数，稠密线性代数和图形内核的内核库。我们描述了这样一个库的设计原则，并展示了一些选定的内核库的便携式性能。具体来说，我们展示了四个稀疏内核，三个密集批处理内核，两个图形内核和一个团队级算法的性能。
As hardware architectures are evolving in the push towards exascale, developing Computational Science and Engineering (CSE) applications depend on performance portable approaches for sustainable software development. This paper describes one aspect of performance portability with respect to developing a portable library of kernels that serve the needs of several CSE applications and software frameworks. We describe Kokkos Kernels, a library of kernels for sparse linear algebra, dense linear algebra and graph kernels. We describe the design principles of such a library and demonstrate portable performance of the library using some selected kernels. Specifically, we demonstrate the performance of four sparse kernels, three dense batched kernels, two graph kernels and one team level algorithm.
DOI：
10.1109/icpp.2017.59
发表时间：
2017-05
期刊：
2017 46th International Conference on Parallel Processing (ICPP)
影响因子：
--
作者：
Mustafa Kemal Tas;K. Kaya;Erik Saule
通讯作者：
Mustafa Kemal Tas;K. Kaya;Erik Saule