Multi-mapping Meshes: A New Communicating Fabric for Networks-on-Chip

Traditional mesh is a popular interconnection architecture for Networks-on-Chip (NoCs). However, with the trend towards larger number of cores in chip multiprocessors, it is unable to stand the fast-growing diameter and average distance in meshes. Recently, concentration and express channels are two countermeasures for that. In this paper, a new scheme called multi-mapping is proposed, which allows one processing element (PE) to be connected to multiple routers, and vice versa. By properly establishing the mapping relationships between PE and router, both diameter and average distance of the network are lowered while the interconnections between routers are not altered. To provide efficient and in-order communication in NoCs, we develop the X¡Y routing scheme with wormhole-switching technique used in multi-mapping meshes. The simulation results are presented to show the effectiveness of our method by comparing with traditional meshes and mesh-based express cubes under different traffic patterns.

[1]  William J. Dally,et al.  Express Cubes: Improving the Performance of k-Ary n-Cube Interconnection Networks , 1989, IEEE Trans. Computers.

[2]  Onur Mutlu,et al.  Express Cube Topologies for on-Chip Interconnects , 2009, 2009 IEEE 15th International Symposium on High Performance Computer Architecture.

[3]  William J. Dally,et al.  Performance Analysis of k-Ary n-Cube Interconnection Networks , 1987, IEEE Trans. Computers.

[4]  William J. Dally,et al.  Principles and Practices of Interconnection Networks , 2004 .

[5]  Anant Agarwal,et al.  Scalar operand networks: on-chip interconnect for ILP in partitioned architectures , 2003, The Ninth International Symposium on High-Performance Computer Architecture, 2003. HPCA-9 2003. Proceedings..

[6]  Christopher Batten,et al.  Building Manycore Processor-to-DRAM Networks with Monolithic Silicon Photonics , 2008, 2008 16th IEEE Symposium on High Performance Interconnects.

[7]  Doug Burger,et al.  Implementation and Evaluation of On-Chip Network Architectures , 2006, 2006 International Conference on Computer Design.

[8]  William J. Dally,et al.  Design tradeoffs for tiled CMP on-chip networks , 2006, ICS '06.

[9]  Anna T. Lawniczak,et al.  Performance of data networks with random links , 1999, ArXiv.

[10]  Axel Jantsch,et al.  International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2020, Singapore, September 20-25, 2020 , 2020, CODES+ISSS.

[11]  Radu Marculescu,et al.  Energy- and performance-aware mapping for regular NoC architectures , 2005, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[12]  William J. Dally,et al.  Flattened Butterfly Topology for On-Chip Networks , 2007, IEEE Comput. Archit. Lett..

[13]  Radu Marculescu,et al.  Key research problems in NoC design: a holistic perspective , 2005, 2005 Third IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS'05).