AdapNoC: A Fast and Flexible FPGA-based NoC Simulator

被引:22
作者
Kamali, Hadi Mardani [1 ]
Hessabi, Shahin [1 ]
机构
[1] Sharif Univ Technol, Dept Comp Engn, Tehran, Iran
来源
2016 26TH INTERNATIONAL CONFERENCE ON FIELD PROGRAMMABLE LOGIC AND APPLICATIONS (FPL) | 2016年
关键词
NoC; FPGA; adaptive routing algorithm; virtualization; Dual-Clock; DESIGN; CHIP;
D O I
10.1109/FPL.2016.7577377
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Network on Chip (NoC) is the most common interconnection platform for multiprocessor systems-on-chips (MPSoCs). In order to explore the design space of this platform, we need a high-speed, cycle-accurate, and flexible simulation tool. In this paper, we present AdapNoC, a configurable cycle-accurate FPGA-based NoC simulator, which can be configured via software. A wide range of parameters are configurable in FPGA side of the proposed simulator, and the software side is implemented on an embedded soft-core processor. We transfer some parts of simulator, such as Traffic Generators (TGs) and Traffic Receptors (TRs), to software side without any degradation in simulation speed. Moreover, we implement a dual-clock architecture as an innovation in virtualization methodology, which is also capable to share idle time-slots, which helps not only simulate bigger NoCs, but also reduce simulation time drastically. Also, by employing a traffic aggregator architecture, AdapNoC provides table-based adaptive routing algorithm as a configurable parameter in router microarchitecture. We evaluate simulation time of AdapNoC by using Xilinx Virtex-6 XC6VLX240T, and demonstrate 53x-180x speed-up against BOOKSIM. Also, due to our proposed virtualization, and TGs and TRs migration to software side, we can implement a 64-node non-virtualized or a 1024-node virtualized mesh network in only %72 of Xilinx Virtex-6 XC6VLX240T resources.
引用
收藏
页数:8
相关论文
共 36 条
[1]  
Abba S., MICROPROCESSORS MICR, V38, P375
[2]  
[Anonymous], 2005, SIGARCH Comput. Archit. News
[3]  
Binkert Nathan, 2011, Computer Architecture News, V39, P1, DOI 10.1145/2024716.2024718
[4]   Thousand core chips-a technology perspective [J].
Borkar, Shekhar .
2007 44TH ACM/IEEE DESIGN AUTOMATION CONFERENCE, VOLS 1 AND 2, 2007, :746-749
[5]  
Chung E. S., 2007, PAR DISTR PROC S IEE, V1, P26
[6]   OCCN: a NoC modeling framework for design exploration [J].
Coppola, M ;
Curaba, S ;
Grammatikakis, MD ;
Locatelli, R ;
Maruccia, G ;
Papariello, F .
JOURNAL OF SYSTEMS ARCHITECTURE, 2004, 50 (2-3) :129-163
[7]  
Dally B. P., 2004, Principles and Practices of InterconnectionNetworks
[8]  
Dally WJ, 2001, DES AUT CON, P684, DOI 10.1109/DAC.2001.935594
[9]   A design flow for application-specific networks on chip with guaranteed performance to accelerate SOC design and verification [J].
Goossens, K ;
Dielissen, J ;
Gangwal, OP ;
Pestana, SG ;
Radulescu, A ;
Rijpkema, E .
DESIGN, AUTOMATION AND TEST IN EUROPE CONFERENCE AND EXHIBITION, VOLS 1 AND 2, PROCEEDINGS, 2005, :1182-1187
[10]  
Haoyuan Y., REC COMM CENTR SYST, P1