Fault tolerance in memristive crossbar-based neuromorphic computing systems

Xu，Qi1; Chen，Song2; Geng，Hao3; Yuan，Bo4; Yu，Bei3; Wu，Feng2; Huang，Zhengfeng1

2019

10.1016/j.vlsi.2019.09.008

INTEGRATION-THE VLSI JOURNAL   影响因子和分区

0167-9260

1872-7522

In recent years, neuromorphic computing systems (NCS) based on memristive crossbar have provided a promising solution to enable acceleration of neural networks. However, Stuck-at faults in the memristor devices significantly degrade the computing accuracy of NCS. In this paper, we propose an effective fault tolerant framework for memristive crossbar-based neuromorphic computing systems. First, a fault tolerance-aware hierarchical clustering method is proposed to partition weight connections of a sparse neural network into clusters. Then, for each cluster, memristive crossbar configuration is proposed to determine a suitable size of the crossbar with consideration of both hardware cost and successful mapping rate. Next, an integer linear programming formulation is developed to derive a connection-memristor mapping for fault tolerance. Finally, an efficient matching-based heuristic algorithm is further proposed to speed-up the fault-tolerant mapping process. Experimental results show that the proposed fault tolerant framework can improve the successful mapping rate and simultaneously reduce the hardware cost.

Fault tolerance Hierarchical clustering Memristive crossbar Neuromorphic computing system

SCI ; EI

英语

其他

Fundamental Research Funds for the Central Universities[JZ2019HGBZ0159] ; National Natural Science Foundation of China[61874156] ; National Natural Science Foundation of China[]

Computer Science ; Engineering

Computer Science, Hardware & Architecture ; Engineering, Electrical & Electronic

WOS:000503087800008

Elsevier B.V.

20193907464982

Cluster analysis ; Fault tolerance ; Heuristic algorithms ; Integer programming ; Mapping ; Memristors

Surveying:405.3 ; Semiconductor Devices and Integrated Circuits:714.2 ; Digital Computers and Systems:722.4 ; Computer Software, Data Handling and Applications:723 ; Computer Programming:723.1 ; Optimization Techniques:921.5

COMPUTER SCIENCE

2-s2.0-85072389380

Scopus

1.Department of Electronic Science and TechnologyHefei University of Technology,China
2.School of MicroelectronicsUniversity of Science and Technology of China,China
3.Department of Computer Science and EngineeringThe Chinese University of Hong Kong,China
4.Department of Computer Science and EngineeringSouthern University of Science and Technology,China

Xu，Qi,Chen，Song,Geng，Hao,et al. Fault tolerance in memristive crossbar-based neuromorphic computing systems[J]. INTEGRATION-THE VLSI JOURNAL,2019,70:70-79.

Xu，Qi.,Chen，Song.,Geng，Hao.,Yuan，Bo.,Yu，Bei.,...&Huang，Zhengfeng.(2019).Fault tolerance in memristive crossbar-based neuromorphic computing systems.INTEGRATION-THE VLSI JOURNAL,70,70-79.

Xu，Qi,et al."Fault tolerance in memristive crossbar-based neuromorphic computing systems".INTEGRATION-THE VLSI JOURNAL 70(2019):70-79.