Yi Xu
ABSTRACT
The emerging on-chip optical interconnection has become a promising candidate for future network design because of its advantages in high bandwidth density, low propagation delay and dynamic power consumption. However, a key challenge of on-chip optics is the high static power consumption which dominates the total network power. Hence, it is imperative to design an energy-efficient optical network architecture with high throughput while consuming low static power. In conventional optical crossbars, static channel allocation results in low channel utilization and network throughput, while full channel sharing requires a significant number of microrings, which incurs high static power.
To obtain high network throughput with low power consumption, this paper proposes a nanophotonic crossbar architecture with light-weight distributed arbitration. Network channels are allocated to an owner node, but can also be used by a few other nodes during idle time. The number of microring resonators is greatly reduced compared to the full channel sharing architecture. The arbitration is also simplified due to the small number of nodes sharing a channel. Every node can use the statically assigned channel to avoid starvation and borrow an additional idle channel to improve the utilization of the network. We intelligently select the network nodes that should share a channel to increase the chance of successful borrowing with low probability of conflict. The energy efficiency of the proposed network architecture is evaluated in terms of energy efficiency (throughput/watt) and Energy-delay2(ED2) using synthetic traces and traffic traces from PARSEC benchmarks. The simulation results show that our design can improve energy efficiency by 34% and 26% and improve ED^2 by 73% and 45% compared to Single-write-multi-read (SWMR) crossbars and Multi-write-multi-read (MWMR) crossbars respectively.
- J. Ahn et al. Devices and architectures for photonic chip-scale integration. Appl. Phy. A, pages 989--997, 2009.Google Scholar
Cross Ref
- J. Balfour and W. J. Dally. Design tradeoffs for tiled cmp onchip networks. In ICS, 2006. Google ScholarDigital Library
- C. Batten. Designing nanophotonic interconnection networks. In Workshop on the Interaction between Nanophotonic Devices and Systems, 2010.Google Scholar
- C. Batten et al. Building manycore processor-to-dram networks with monolithic silicon photonics. In Hot Interconnects, pages 21--30, 2008. Google ScholarDigital Library
- S. Beamer et al. Re-architecting dram memory systems with monolithically integrated silicon photonics. In ISCA, pages 129--140, 2010. Google ScholarDigital Library
- C. Bienia et al. The parsec benchmark suite: Characterization and architectural implications. In PACT, pages 72--81, 2008. Google ScholarDigital Library
- D. Chiarulli et al. Multiprocessor interconnection networks using partitioned passive star topologies and distributed control. In Int'l Workshop on MPPOI, 1994.Google Scholar
- Cianchetti et al. Phastlane: a rapid transit optical routing network. In ISCA, pages 441--450, 2009. Google ScholarDigital Library
- ITRS. International technology roadmap for semiconductors. Technical report, 2010.Google Scholar
- A. Joshi et al. Silicon-photonic clos networks for global on-chip communication. In 3rd NOCS, pages 124 --133, 2009. Google ScholarDigital Library
- J. Kim et al. Flattened butterfly topology for on-chip networks. In MICRO, 2007. Google ScholarDigital Library
- N. Kirman et al. Leveraging optical technology in future bus-based chip multiprocessors. In MICRO, pages 492--503, 2006. Google ScholarDigital Library
- N. Kirman and J. F. Martínez. A power-efficient all-optical on-chip interconnect using wavelength-based oblivious routing. In ASPLOS, pages 15--28, 2010. Google ScholarDigital Library
- P. Koka et al. Silicon-photonic network architectures for scalable, power-efficient multi-chip systems. SIGARCH Comput. Archit. News, 38, 2010. Google ScholarDigital Library
- G. Kurian et al. Atac: a 1000-core cache-coherent processor with on-chip optical network. In PACT, pages 477--488, 2010. Google ScholarDigital Library
- C. Nitta et al. Addressing system-level trimming issues in on-chip nanophotonic networks. In HPCA, pages 122--131, 2011. Google ScholarDigital Library
- C. Nitta et al. Resilient microring resonator based photonic networks. In MICRO, 2011. Google ScholarDigital Library
- M. Palesi et al. Noxim, an open network-on-chip simulator. http://noxim.sourceforge.net.Google Scholar
- Y. Pan et al. Firefly: Illuminating future network-on-chip with nanophotonics. In ISCA, pages 429--440, 2009. Google ScholarDigital Library
- Y. Pan et al. Flexishare: Channel sharing for an energy-efficient nanophotonic crossbar. In HPCA, pages 1--12, 2010.Google ScholarCross Ref
- A. Shacham et al. Photonic networks-on-chip for future generations of chip multiprocessors. IEEE Trans. Comput., 57:1246--1260, September 2008. Google ScholarDigital Library
- D. Vantrease et al. Corona: System implications of emerging nanophotonic technology. In ISCA, pages 153--164, 2008. Google ScholarDigital Library
- D. Vantrease et al. Light speed arbitration and flow control for nanophotonic interconnects. In MICRO, pages 304--315, 2009. Google ScholarDigital Library
- D. Vantrease et al. Atomic coherence: Leveraging nanophotonics to build race-free cache coherence protocols. In HPCA, pages 132 --143, 2011. Google ScholarDigital Library
- Y. Xu et al. A composite and scalable cache coherence protocol for large scale cmps. In ICS, pages 285--294, 2011. Google ScholarDigital Library
- Y. Xu et al. Tolerating process variations in nanophotonic on-chip networks. In ISCA, 2012. Google ScholarDigital Library
- J. Xue et al. An intra-chip free-space optical interconnect. In ISCA, pages 94--105, 2010. Google ScholarDigital Library
- M. T. Yourst. Ptlsim: A cycle accurate full system x86--64 microarchitectural simulator. In ISPASS, pages 23--24, 2007.Google ScholarCross Ref
Index Terms
- Channel borrowing: an energy-efficient nanophotonic crossbar architecture with light-weight arbitration
Recommendations
mNoC: Large Nanophotonic Network-on-Chip Crossbars with Molecular Scale Devices
Moore's law and the continuity of device scaling have led to an increasing number of cores/nodes on a chip, creating a need for new mechanisms to achieve high-performance and power-efficient Network-on-Chip (NoC). Nanophotonics based NoCs provide for ...
Enabling High-Performance Crossbars through a Floorplan-Aware Design
ICPP '12: Proceedings of the 2012 41st International Conference on Parallel ProcessingNetworks-on-Chip (NoC) with low-radix switches forming a simple and planar topology is typically accepted as the right interconnection infrastructure for current Chip Multi Processor and high-end Multi Processor System-on-Chip. This is mainly due to its ...
Silicon-photonic network architectures for scalable, power-efficient multi-chip systems
ISCA '10Scaling trends of logic, memories, and interconnect networks lead towards dense many-core chips. Unfortunately, process yields and reticle sizes limit the scalability of large single-chip systems. Multi-chip systems break free of these areal limits, but ...
Comments