Comprehensive Modeling Methodologies for NoC Router Estimation

Networks-on-Chip (NoCs) are increasingly used in many-core architectures. ORION2.0 [18] is a widely adopted NoC power and area estimation tool that is based on circuit-level templates, which use specific logic structures to model implementation of different router components. ORION2.0 estimation models can have large errors (up to 185%) versus actual implementation, often due to a mismatch between the actual router RTL and the templates assumed, as well as the effects of optimization tools in modern design flows. In this work, we propose comprehensive parametric and non-parametric modeling methodologies that fundamentally differ from logic template based approaches in that the estimation models are derived from actual physical implementation data. Specifically, we propose a new parametric modeling methodology as well as improvements to previous work on non-parametric modeling. Our work on parametric modeling proposes (1) ORION NEW models, developed using a new methodology that does not assume any logic templates for router components, (2) the use of parametric regression to fit the new models to post-layout power and area values, and (3) modeling extensions that enable more detailed flit-level power estimation when integrated with simulation tools such as GARNET. The ORION NEW methodology analyzes netlists of NoC routers that have been placed and routed by commercial tools, and then performs explicit modeling of control and data paths followed by regression analysis to create highly accurate NoC models. Our work on non-parametric modeling expands on our previous work [17] and demonstrates the use of four widely used non-parametric modeling techniques: Radial Basis Functions, Kriging, Multivariate Adaptive Regression Splines, and Support Vector Machine Regression. The estimation models are also derived from actual physical implementation data. We observe that non-parametric modeling techniques have low overhead and complexity of modeling, and are highly accurate. When compared with actual implementations, our modeling methodologies achieve average estimation errors of no more than 9.8% across microarchitecture, implementation, and operational parameters as well as multiple router RTL generators. These methodologies are being implemented in an ORION3.0 distribution [19], [66].

[1]  Raphael T. Haftka,et al.  Surrogate-based Analysis and Optimization , 2005 .

[2]  Andrew B. Kahng,et al.  Improved on-chip router analytical power and area modeling , 2010, 2010 15th Asia and South Pacific Design Automation Conference (ASP-DAC).

[3]  Alexander J. Smola,et al.  Support Vector Method for Function Approximation, Regression Estimation and Signal Processing , 1996, NIPS.

[4]  Bill Lin,et al.  Comprehensive Modeling Methodologies for NoC Router Estimation Andrew B. Kahng Bill Lin Siddhartha Nath , 2012 .

[5]  Sri Parameswaran,et al.  NoCEE: energy macro-model extraction methodology for network on chip routers , 2005, ICCAD-2005. IEEE/ACM International Conference on Computer-Aided Design, 2005..

[6]  Georges G. E. Gielen,et al.  Performance modeling of analog integrated circuits using least-squares support vector machines , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[7]  Jih-Sheng Shen,et al.  A low-power crossroad switch architecture and its core placement for network-on-chip , 2005, ISLPED '05. Proceedings of the 2005 International Symposium on Low Power Electronics and Design, 2005..

[8]  Vittorio Zaccaria,et al.  OSCAR: An Optimization Methodology Exploiting Spatial Correlation in Multicore Design Spaces , 2012, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[9]  Andrew B. Kahng,et al.  Explicit modeling of control and data for improved NoC router estimation , 2012, DAC Design Automation Conference 2012.

[10]  Timothy W. Simpson,et al.  Metamodels for Computer-based Engineering Design: Survey and recommendations , 2001, Engineering with Computers.

[11]  Frank Liu,et al.  A General Framework for Spatial Correlation Modeling in VLSI Design , 2007, 2007 44th ACM/IEEE Design Automation Conference.

[12]  Andrew B. Kahng,et al.  ORION 2.0: A fast and accurate NoC power and area model for early-stage design space exploration , 2009, 2009 Design, Automation & Test in Europe Conference & Exhibition.

[13]  A. A. Ilumoka.  Efficient prediction of crosstalk in VLSI interconnections using neural networks , 2000, IEEE 9th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.00TH8524).

[14]  Jia Xinzhang,et al.  Analysis on the effect of regression and correlation models on the accuracy of Kriging model for IC , 2009, 2009 IEEE International Conference of Electron Devices and Solid-State Circuits (EDSSC).

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

[16]  Orion: a power-performance simulator for interconnection networks , 2002, MICRO 35.

[17]  Andrew B. Kahng,et al.  Accurate Machine-Learning-Based On-Chip Router Modeling , 2010, IEEE Embedded Systems Letters.

[18]  Hans C. van Houwelingen,et al.  The Elements of Statistical Learning, Data Mining, Inference, and Prediction. Trevor Hastie, Robert Tibshirani and Jerome Friedman, Springer, New York, 2001. No. of pages: xvi+533. ISBN 0‐387‐95284‐5 , 2004 .

[19]  Niraj K. Jha,et al.  GARNET: A detailed on-chip network model inside a full-system simulator , 2009, 2009 IEEE International Symposium on Performance Analysis of Systems and Software.

[20]  Li-Shiuan Peh,et al.  Flow control and micro-architectural mechanisms for extending the performance of interconnection networks , 2001 .

[21]  M. B. Yelten,et al.  Demystifying Surrogate Modeling for Circuits and Systems , 2012, IEEE Circuits and Systems Magazine.

[22]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[23]  Ashutosh Kumar Singh,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction , 2010 .

[24]  E. S. Siah,et al.  Fast parameter optimization of large-scale electromagnetic objects using DIRECT with Kriging metamodeling , 2004, IEEE Transactions on Microwave Theory and Techniques.

[25]  David M. Brooks,et al.  Accurate and efficient regression modeling for microarchitectural performance and power prediction , 2006, ASPLOS XII.

[26]  Sonja Kuhnt,et al.  Design and analysis of computer experiments , 2010 .

[27]  J. Freidman,et al.  Multivariate adaptive regression splines , 1991 .

[28]  Miss A.O. Penney.  (b) , 1974, The New Yale Book of Quotations.

[29]  Søren Nymand Lophaven,et al.  Aspects of the Matlab toolbox DACE , 2002 .

[30]  Luca Benini,et al.  Area and Power Modeling for Networks-on-Chip with Layout Awareness , 2007, VLSI Design.

[31]  Axel Jantsch,et al.  A High Level Power Model for the Nostrum NoC , 2006, 9th EUROMICRO Conference on Digital System Design (DSD'06).

[32]  Sudhakar Yalamanchili,et al.  Power constrained design of multiprocessor interconnection networks , 1997, Proceedings International Conference on Computer Design VLSI in Computers and Processors.

[33]  Gianluca Palermo,et al.  PIRATE: A Framework for Power/Performance Exploration of Network-on-Chip Architectures , 2004, PATMOS.

[34]  Luca Benini,et al.  Analysis of power consumption on switch fabrics in network routers , 2002, DAC '02.

[35]  Sally A. McKee,et al.  Efficiently exploring architectural design spaces via predictive modeling , 2006, ASPLOS XII.

[36]  T. Simpson,et al.  Comparative studies of metamodeling techniques under multiple modeling criteria , 2000 .

[37]  R. Brereton,et al.  Support vector machines for classification and regression. , 2010, The Analyst.

[38]  W. Dally,et al.  Route packets, not wires: on-chip interconnection networks , 2001, Proceedings of the 38th Design Automation Conference (IEEE Cat. No.01CH37232).

[39]  Simon W. Moore,et al.  The design and implementation of a low-latency on-chip network , 2006, Asia and South Pacific Conference on Design Automation, 2006..

[40]  S. Nassif,et al.  Accurate Spatial Estimation and Decomposition Techniques for Variability Characterization , 2010, IEEE Transactions on Semiconductor Manufacturing.

[41]  Karam S. Chatha,et al.  A power and performance model for network-on-chip architectures , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[42]  Seung Eun Lee,et al.  A high level power model for Network-on-Chip (NoC) router , 2009, Comput. Electr. Eng..

[43]  Vittorio Zaccaria,et al.  A correlation-based design space exploration methodology for multi-processor systems-on-chip , 2010, Design Automation Conference.

[44]  Fernando Gehm Moraes,et al.  NoC Power Estimation at the RTL Abstraction Level , 2008, 2008 IEEE Computer Society Annual Symposium on VLSI.

[45]  Rob A. Rutenbar,et al.  Statistical Blockade: A Novel Method for Very Fast Monte Carlo Simulation of Rare Circuit Events, and its Application , 2007, 2007 Design, Automation & Test in Europe Conference & Exhibition.