Low-power and error protection coding for network-on-chip traffic

The power consumption of the network-on-chip communication backbone is explored and the effectiveness of low-power encoding and error protection techniques is analysed. For the switch under the study, a Nostrum deflective routing switch, simulations and power analysis suggest that only a minor fraction of the power is dissipated in the logic blocks, whereas the major part is due to the interconnection wires. The authors have investigated a number of low-power and data protection mechanisms and studied their impact on power consumption of the whole network. The bus-invert encoding scheme and a limited set of Hamming data protection codes have been implemented on both data link and at the network layer. However, it turned out that all low-power data encoding schemes have little potential to decrease power consumption due to the significant overhead. On the other hand, error protection mechanisms have a significant potential to decrease power consumption because they allow to operate the network at a lower voltage. The authors' experiments show a 20% decrease of power consumption for a given error rate and for a given performance.

[1]  Chi-Ying Tsui,et al.  Re-configurable bus encoding scheme for reducing power consumption of the cross coupling capacitance for deep sub-micron instruction bus , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[2]  Johnny Öberg,et al.  Reducing Power and Latency in 2-D Mesh Globally Pseudochronous Locally Synchronous Clocking , 2004 .

[3]  Axel Jantsch,et al.  A fault model notation and error-control scheme for switch-to-switch buses in a network-on-chip , 2003, First IEEE/ACM/IFIP International Conference on Hardware/ Software Codesign and Systems Synthesis (IEEE Cat. No.03TH8721).

[4]  Karam S. Chatha,et al.  Quality-of-service and error control techniques for network-on-chip architectures , 2004, GLSVLSI '04.

[5]  Mani B. Srivastava,et al.  A survey of techniques for energy efficient on-chip communication , 2003, Proceedings 2003. Design Automation Conference (IEEE Cat. No.03CH37451).

[6]  André K. Nieuwland,et al.  Why transition coding for power minimization of on-chip buses does not work , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[7]  Dake Liu,et al.  Power consumption estimation in CMOS VLSI chips , 1994, IEEE J. Solid State Circuits.

[8]  Luca Benini,et al.  Networks on Chips : A New SoC Paradigm , 2022 .

[9]  Axel Jantsch,et al.  A network on chip architecture and design methodology , 2002, Proceedings IEEE Computer Society Annual Symposium on VLSI. New Paradigms for VLSI Systems Design. ISVLSI 2002.

[10]  Luca Benini,et al.  Power optimization of core-based systems by address bus encoding , 1998, IEEE Trans. Very Large Scale Integr. Syst..

[11]  R. Siegmund,et al.  Adaptive Partial Businvert Encoding for power efficient data transfer over wide system buses , 2000, Proceedings 13th Symposium on Integrated Circuits and Systems Design (Cat. No.PR00843).

[12]  Axel Jantsch,et al.  Power analysis of link level and end-to-end data protection in networks on chip , 2005, 2005 IEEE International Symposium on Circuits and Systems.

[13]  Axel Jantsch,et al.  Networks on chip , 2003 .

[14]  Dinesh Pamunuwa,et al.  Modelling and analysis of interconnects for deep submicron systems-on-chip. , 2003 .

[15]  L. Benini,et al.  Analysis of power consumption on switch fabrics in network routers , 2002, Proceedings 2002 Design Automation Conference (IEEE Cat. No.02CH37324).

[16]  Rajesh K. Gupta,et al.  Leakage aware dynamic voltage scaling for real-time embedded systems , 2004, Proceedings. 41st Design Automation Conference, 2004..

[17]  Nikil D. Dutt,et al.  Low power address encoding using self-organizing lists , 2001, ISLPED '01.