A study on the implementation of 2-D mesh-based networks-on-chip in the nanometre regime
Introduction
In the deep submicrometer (DSM) regime, electrical level issues that affect signalling, timing, power and noise are challenging established design procedures. Expected trends in the future evolution of VLSI systems can be codified into the following points:
- (i)
Moore's law will continue to hold for another 10 years [1].
- (ii)
Single processors will not be able to utilize the transistors of an entire chip, and a single synchronous clock region will span only a small fraction of the chip area [2], [3].
- (iii)
Applications will be modelled as a large number of communicating tasks, where the tasks may have very different characteristics (such as control or data flow dominated) and origins (IP re-use from earlier products or external sources) [4], [5].
An architecture that enforces modularity and is suitable for this kind of heterogeneous implementation is the Network-on-Chip (NoC) architecture [6], [7], [8]. It eases the expected bottlenecks of complexity and wire delay in nanometre technologies, and promotes extensive re-use of design cores through standardization of on-chip communication. This paper considers the physical layout of likely mesh-based NoC architectures, and carries out a feasibility study in a future 65 nm technology. Some of the issues covered are layout trade-offs and wiring schemes for the network, and area, performance and power metrics for the different architectures. The cost of the network in relation to the resources with regard to power consumption is investigated when standard rail-to-rail CMOS signalling is used.
The parameters of the 65 nm technology are obtained by following guidelines outlined in the International Technology Roadmap for Semiconductors (ITRS), and by scaling from an existing technology. Some assumptions about unknowns are made in order to facilitate analysis. This may make for a somewhat rough and ready approach on occasion, but it is adequate to provide representative figures for likely future CMOS technologies.
The rest of the paper is structured in the following manner. First the background of the NoC backbone including architecture and protocol, are reviewed in Section 2. Next the modelling details, including the physical mapping of the geometry to electrical properties and calculation of area, delay and power, are covered. Section 4 details the analysis and results for the specific architectures considered in the study. We end with a discussion.
Section snippets
Architecture
A perusal of the literature shows many works that have elucidated the NoC concept and a layered protocol [6], [8]. Of these and others, the most attention to the physical level is paid in [6]. It describes a folded torus topology that fits well to VLSI implementations with a two-dimensional layout and limited wires. The proposed routing layout places the network wires on top of the resources in dedicated metal layers.
This is the most intuitive layout, but there are in general a myriad of ways
Technology scaling
A feasibility study for NoC implementations in the DSM regime requires models that accurately capture the behaviour of active and passive devices in the given technology node. This is the science of technology extrapolation, which has received a great deal of attention over the past few decades. Its importance is due to the fact that not only does predicting future trends give us an idea of what is achievable, but also has a strong influence on the evolvement of VLSI systems. Influential
Square-switch architecture
The layout, and switch and resource arrangement of the square-switch architecture is sketched out in Fig. 1. As mentioned, the area overhead for this architecture is 20%. To reduce this, there are two possibilities:
- (a)
let the resource area extend under the wire channels, creating a compromise between architectures 1 and 2;
- (b)
let the switch extend into the wire channels (i.e. shape it like a ‘+’ sign instead of a square).
Discussion and conclusions
This paper considered the physical issues related to the implementation of a packet-switched NoC. By reviewing proposals in the literature from several research groups, two simple but representative architectures were identified. Based on the ITRS, parameters for a technology expected in 2007 were derived, and used to refine the architectural details (size of a tile and switch) by considering the integration density of devices. Finally, cost and performance metrics for two likely architectures
Acknowledgements
This work was supported in part by the Swedish governmental funding agencies Sida and Vinnowa.
References (18)
- International Technology Semiconductor Roadmap (ITRS) 2001, [Online], available:...
- D. Sylvester, K. Keutzer, Getting to the bottom of deep submicron, in: Proceedings of ICCAD, 1998, pp....
- A. Hemani et al., Lowering power consumption in clock by using globally asynchronous locally synchronous design style,...
Component SoftwareBeyond Object Oriented Software
(1998)- D. Gajski, R. Dömer, J. Zhu, IP-Centric methodology and design with the SpecC language, in: A.A. Jerraya, J. Mermet...
- W.J. Dally, B. Towles, Route packets, not wires: on-chip interconnection networks, in: Proceedings of DAC, 2001, pp....
- M. Sgroi, M. Sheets, A. Mihal, K. Keutzer, S. Malik, J. Rabaey, A. Sangiovanni-Vincentelli, Addressing the...
- L. Benini, G. DeMicheli, powering networks on chip, in: Proceedings of ISSS, 2001, pp....
- M. Millberg, E. Nilsson, R. Thid, S. Kumar, A. Jantsch, The Nostrum backbone—a communication protocol stack for...
Cited by (12)
Engineering of Germanium Tunnel Junctions for Optical Applications
2018, IEEE Photonics JournalA 3D multilayer optical network on chip based on mesh topology
2016, Photonic Network CommunicationsHierarchical supporting structure for dynamic organization in many-core computing systems
2013, PECCS 2013 - Proceedings of the 3rd International Conference on Pervasive Embedded Computing and Communication SystemsCascaded microresonator-based matrix switch for silicon on-chip optical interconnection
2009, Proceedings of the IEEEFrom 2D to 3D NoCs: A case study on worst-case communication performance
2009, IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCADPhysical mapping and performance study of a multi-clock 3-dimensional network-on-chip mesh
2009, 2009 IEEE International Conference on 3D System Integration, 3DIC 2009