A Pattern-Based SpGEMM Library for Multi-Core and Many-Core Architectures
文献类型:期刊论文
作者 | Xie, Zhen2,3; Tan, Guangming2,3; Liu, Weifeng1; Sun, Ninghui2,3 |
刊名 | IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS |
出版日期 | 2022 |
卷号 | 33期号:1页码:159-175 |
ISSN号 | 1045-9219 |
关键词 | Libraries Sparse matrices Prediction algorithms Neural networks Predictive models Memory management Tuners SpGEMM spare BLAS sparse format auto-tuning neural network |
DOI | 10.1109/TPDS.2021.3090328 |
英文摘要 | General sparse matrix-matrix multiplication (SpGEMM) is one of the most important mathematical library routines in a number of applications. In recent years, several efficient SpGEMM algorithms have been proposed, however, most of them are based on the compressed sparse row (CSR) format, and the possible performance gain from exploiting other formats has not been well studied. And some specific algorithms are restricted to parameter tuning that has a significant impact on performance. So the particular format, algorithm, and parameter that yield the best performance for SpGEMM remain undetermined. In this article, we conduct a prospective study on format-specific parallel SpGEMM algorithms and analyze their pros and cons. We then propose a pattern-based SpGEMM library, that provides a unified programming interface in the CSR format, analyses the pattern of two input matrices, and automatically determines the best format, algorithm, and parameter for arbitrary matrix pairs. For this purpose, we build an algorithm set that integrates three new designed algorithms with existing popular libraries, and design a hybrid deep learning model called MatNet to quickly identify patterns of input matrices and accurately predict the best solution by using sparse features and density representations. The evaluation shows that this library consistently outperforms the state-of-the-art library. We also demonstrate its adaptability in an AMG solver and a BFS algorithm with 30 percent performance improvement. |
资助项目 | National Key Research and Development Program of China[2017YFB0202105] ; National Key Research and Development Program of China[2016YFB0201305] ; National Key Research and Development Program of China[2016YFB0200803] ; National Key Research and Development Program of China[2016YFB0200300] ; National Natural Science Foundation of China[61521092] ; National Natural Science Foundation of China[91430218] ; National Natural Science Foundation of China[31327901] ; National Natural Science Foundation of China[61472395] ; National Natural Science Foundation of China[61432018] ; National Natural Science Foundation of China[61671151] |
WOS研究方向 | Computer Science ; Engineering |
语种 | 英语 |
出版者 | IEEE COMPUTER SOC |
WOS记录号 | WOS:000673452600001 |
源URL | [http://119.78.100.204/handle/2XEOYT63/17499] |
专题 | 中国科学院计算技术研究所期刊论文_英文 |
通讯作者 | Xie, Zhen |
作者单位 | 1.China Univ Petr, Dept Comp Sci & Technol, Beijing 102249, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100864, Peoples R China 3.Chinese Acad Sci, Inst Comp Technol, State Key Lab Comp Architecture, Beijing 100864, Peoples R China |
推荐引用方式 GB/T 7714 | Xie, Zhen,Tan, Guangming,Liu, Weifeng,et al. A Pattern-Based SpGEMM Library for Multi-Core and Many-Core Architectures[J]. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2022,33(1):159-175. |
APA | Xie, Zhen,Tan, Guangming,Liu, Weifeng,&Sun, Ninghui.(2022).A Pattern-Based SpGEMM Library for Multi-Core and Many-Core Architectures.IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,33(1),159-175. |
MLA | Xie, Zhen,et al."A Pattern-Based SpGEMM Library for Multi-Core and Many-Core Architectures".IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 33.1(2022):159-175. |
入库方式: OAI收割
来源:计算技术研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。