DocumentCode :
3377169
Title :
Efficient trace-driven metaheuristics for optimization of networks-on-chip configurations
Author :
Kahng, Andrew B. ; Lin, Bill ; Samadi, Kambiz ; Ramanujam, Rohit Sunkam
Author_Institution :
ECE Dept., Univ. of California San Diego, La Jolla, CA, USA
fYear :
2010
fDate :
7-11 Nov. 2010
Firstpage :
256
Lastpage :
263
Abstract :
As industry moves towards many-core chips, networks-on-chip (NoCs) are emerging as a scalable communication fabric for interconnecting the cores. With increasing core counts, there is a corresponding increase in communication demands in multi-core designs to facilitate high core utilization, and a consequent critical need for high-performance NoCs. Another megatrend in advanced technologies is that power has become the most critical design constraint. In this paper, we focus on trace-driven virtual channel (VC) allocation in application-specific NoCs. We propose a new significant VC failure metric to capture the impact of VCs on network performance and efficiently drive NoC optimization. Our proposed metaheuristics achieve up to 38% reduction in the number of VCs under a given average packet latency constraint. In addition, compared to a recently proposed trace-driven VC allocation approach, we obtain up to an O(|L|) speedup, where |L| is total number of links in the network, with no degradation in the quality of results.
Keywords :
integrated circuit design; logic design; multiprocessor interconnection networks; network-on-chip; NoC optimization; VC failure metric; application-specific NoC; average packet latency constraint; communication demands; critical design constraint; high core utilization; high-performance NoC; many-core chips; multicore designs; network performance; networks-on-chip configurations; scalable communication fabric; trace-driven metaheuristics; trace-driven virtual channel allocation; Benchmark testing; Delay; Network topology; Resource management; Routing; Runtime;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer-Aided Design (ICCAD), 2010 IEEE/ACM International Conference on
Conference_Location :
San Jose, CA
ISSN :
1092-3152
Print_ISBN :
978-1-4244-8193-4
Type :
conf
DOI :
10.1109/ICCAD.2010.5654164
Filename :
5654164
Link To Document :
بازگشت