Performance driven reliable link design for networks on chips

ABSTRACT

With decreasing feature size of transistors, the interconnect wire delay is becoming a major bottleneck in current Systems on Chips (SoCs). Another effect of shrinking feature size is that the wires are becoming unreliable as they are increasingly susceptible to various noise sources such as cross-talk, coupling noise, soft errors etc. Increasing importance of wire delay and reliability has lead to a communication centric design approach, Networks on Chip (NoC), for building complex SoCs. Current NoC communication design methodologies are based on conservative design approaches and consider worst case operating conditions for link design, resulting in large latency penalty for data transmission. In order to sub-stantially decrease the link delay and thereby increase system performance an aggressive design approach is needed. In this work we present Terror, timing error tolerant communication system, for aggressively designing the links of NoCs. In our methodology, instead of avoiding timing errors by a worst-case design, we do aggressive design by tolerating timing errors. Simulation results show large latency savings (up to 35%) for the Terror based system compared to traditional design methodology.

REFERENCES

• 1: Luca Benini , Giovanni De Micheli, Networks on Chips: A New SoC Paradigm, Computer, v.35 n.1, p.70-78, January 2002 [doi>10.1109/2.976921]
• 2: R. Ho, K. Mai, M. Horowitz, "The Future of Wires", Proceedings of the IEEE, pp. 490--504, April 2001.
• 3: Kathirgamar Aingaran , Fabian Klass , Chin-Man Kim , Chaim Amir , Joydeep Mitra , Eileen You , Jamil Mohd , Sai-Keung Dong, Coupling Noise Analysis for VLIS and ULSI Circuits, Proceedings of the 1st International Symposium on Quality of Electronic Design, p.485, March 20-22, 2000
• 4: Kenneth L. Shepard , Vinod Narayanan, Noise in deep submicron digital design, Proceedings of the 1996 IEEE/ACM international conference on Computer-aided design, p.524-531, November 10-14, 1996, San Jose, California, United States
• 5: Naresh R. Shanbhag, Reliable and Efficient System-on-Chip Design, Computer, v.37 n.3, p.42-50, March 2004 [doi>10.1109/MC.2004.1274003]
• 6: William J. Dally , Brian Towles, Route packets, net wires: on-chip inteconnectoin networks, Proceedings of the 38th conference on Design automation, p.684-689, June 2001, Las Vegas, Nevada, United States [doi>10.1145/378239.379048]
• 7: L. P. Carloni, K. L. McMillan, A. L. Sangiovanni Vincentelli, "Theory of Latency-Insensitive Design", IEEE Trans. on CAD of ICs and Systems, Vol.20, No.9, pp. 1059--1076, Sep 2001.
• 8: Yi Zhao , Li Chen , Sujit Dey, On-Line Testing of Multi-Source Noise-Induced Errors on the Interconnects and Buses of System-on-Chips, Proceedings of the 2002 IEEE International Test Conference, p.491, October 07-10, 2002
• 9: Jean Walrand , Pravin Varaiya, High-performance communication networks, Morgan Kaufmann Publishers Inc., San Francisco, CA, 1996
• 10: Augustus K. Uht, Going Beyond Worst-Case Specs with TEAtime, Computer, v.37 n.3, p.51-56, March 2004 [doi>10.1109/MC.2004.1274004]
• 11: Dan Ernst , Nam Sung Kim , Shidhartha Das , Sanjay Pant , Rajeev Rao , Toan Pham , Conrad Ziesler , David Blaauw , Todd Austin , Krisztian Flautner , Trevor Mudge, Razor: A Low-Power Pipeline Based on Circuit-Level Timing Speculation, Proceedings of the 36th annual IEEE/ACM International Symposium on Microarchitecture, p.7, December 03-05, 2003
• 12: Michele Favalli , Cecilia Metra, Low-level error recovery mechanism for self-checking sequential circuits, 1997 Workshop on Defect and Fault-Tolerance in VLSI Systems, p.234-242, October 20-22, 1997
• 13: M. Singh, S. M. Nowick, "MOUSETRAP: Ultra-High-Speed Transition-Signaling Asynchronous Pipelines", Proc. ICCD 01, Sept. 2001.
• 14: Eric Dupont , Michael Nicolaidis , Peter Rohr, Embedded Robustness IPs for Transient-Error-Free ICs, IEEE Design & Test, v.19 n.3, p.56-70, May 2002
• 15: F. Klass, "Semi-dynamic and dynamic flip-flops with embedded logic", VLSI Circuits 98, pp. 108--109, June 1998.
• 16: Satoshi Matsushita, Design experience of a chip multiprocessor merlot and expectation to functional verification, Proceedings of the 15th international symposium on System Synthesis, October 02-04, 2002, Kyoto, Japan [doi>10.1145/581199.581223]
• 17: B. Ackland et al., "A single-chip, 1.6-billion, 16-b MAC/s multiprocessor DSP", IEEE Journal of Solid-State Circuits, vol. 35, Issue 3, pp. 412--424, Mar 2000.
• 18: M. B. Taylor et al., "Evaluation of the Raw Microprocessor. An Exposed-Wire-Delay Architecture for ILP and Streams", ISCA 2004.
• 19: A. Jain et al., "A 1.2GHz Alpha Microprocessor with 44.8GB/s. Chip Pin Band-width", ISSCC Digest of Technical Papers, pp.240--241, Feb. 2001.

INDEX TERMS

Keywords:
aggressive design, link, networks on chips, performance, reliability