There are a wide variety of intelligence accelerators with promising performance and energy efficiency,deployed in a broad range of applications such as computer vision and speech recognition.However,programming productivity hinders the deployment of deep learning accelerators.The low-level library invoked in the high-level deep learning framework which supports the end-to-end execution with a given model,is designed to reduce the programming burden on the intelligence accelerators.Unfortunately,it is inflexible for developers to build a network model for every deep learning application,which probably brings unnecessary repetitive implementation.In this paper,a flexible and efficient programming framework for deep learning accelerators,FlexPDA,is proposed,which provides more optimization opportunities than the low-level library and realizes quick transplantation of applications to intelligence accelerators for fast upgrades.We evaluate FlexPDA by using 10 representative operators selected from deep learning algorithms and an end-to-end network.The experimental results validate the effectiveness of FlexPDA,which achieves an end-to-end performance improvement of 1.620x over the low-level library.

Lei Liu;Xiu Ma;Hua-Xiao Liu;Guang-Li Li

College of Computer Science and Technology,Jilin University,Changchun 130012,China;Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education,Jilin University Changchun 130012,China;State Key Laboratory of Computer Architecture,Institute of Computing Technology,Chinese Academy of Sciences Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China

计算机科学技术学报（英文版）

[1]Lei Liu;Xiu Ma;Hua-Xiao Liu;Guang-Li Li-.FlexPDA:A Flexible Programming Framework for Deep Learning Accelerators)[J].计算机科学技术学报（英文版）,2022(05):1200-1220

FlexPDA,Accelerators,620x

Flexible,Programming,Framework,Deep,Learning,There,are,wide,variety,intelligence,accelerators,promising,performance,energy,efficiency,deployed,broad,range,applications,such,computer,vision,speech,recognition,However,programming,productivity,hinders,deployment,deep,learning,low,level,library,invoked,high,framework,which,supports,end,execution,given,model,designed,reduce,burden,Unfortunately,inflexible,developers,build,network,every,probably,brings,unnecessary,repetitive,implementation,In,this,paper,efficient,proposed,provides,more,optimization,opportunities,than,realizes,quick,transplantation,fast,upgrades,We,evaluate,by,using,representative,operators,selected,from,algorithms,experimental,results,validate,effectiveness,achieves,improvement,over

0.561397