计算机应用与软件2024，Vol.41Issue(12)：15-22,28,9.DOI:10.3969/j.issn.1000-386x.2024.12.003
GPU异构集群的协同计算引擎设计研究
COLLABORATIVE COMPUTING ENGINE DESIGN FOR GPU HETEROGENEOUS CLUSTER
摘要
Abstract
The cooperation of GPU and multi-core CPU can improve the efficiency of big data computing.However,users need to consider the parallel algorithm logic and collaborative computing process in the application field at the same time,which increases the programming difficulty for GPU heterogeneous collaborative computing.Therefore,based on the analysis of collaborative computing between and within GPU heterogeneous cluster nodes,the process logic is extracted,and a collaborative computing engine combining coarse and fine granularity is proposed,which automatically generates collaborative execution plans,and users only need to pay attention to the design and implementation of the algorithm in the application field.Experimental results show that the performance loss is less than 4.2%compared with the manual programming scheme.Therefore,the collaborative computing engine can be used in the development and application of GPU general computing,which can effectively improve the development efficiency of ordinary users.关键词
GPU协同计算/有限自动机/计算引擎/模板技术Key words
GPU collaborative computing/Finite state machine/Computing engine/Template technology分类
信息技术与安全科学引用本文复制引用
李清清,于欣宁,王海峰..GPU异构集群的协同计算引擎设计研究[J].计算机应用与软件,2024,41(12):15-22,28,9.基金项目
山东省重点研发计划项目(2019GGX101003) （2019GGX101003）
山东省重点研发计划(重大科技创新工程)项目(2019JZZY010134). （重大科技创新工程）
