Decision-Theoretic Design Space Exploration of Multiprocessor Platforms

This paper presents an efficient technique to perform design space exploration of a multiprocessor platform that minimizes the number of simulations needed to identify a Pareto curve with metrics like energy and delay. Instead of using semi-random search algorithms (like simulated annealing, tabu search, genetic algorithms, etc.), we use the domain knowledge derived from the platform architecture to set-up the exploration as a discrete-space Markov decision process. The system walks the design space changing its parameters, performing simulations only when probabilistic information becomes insufficient for a decision. A learning algorithm updates the probabilities of decision outcomes as simulations are performed. The proposed technique has been tested with two multimedia industrial applications, namely the ffmpeg transcoder and the parallel pigz compression algorithm. Results show that the exploration can be performed with 5% of the simulations necessary for the most used algorithms (Pareto simulated annealing, nondominated sorting genetic algorithm, etc.), increasing the exploration speed by more than one order of magnitude.

[1]  Karen A. F. Copeland.  Design and Analysis of Experiments, 5th Ed. , 2001 .

[2]  Cagkan Erbas,et al.  System-level modelling and design space exploration for multiprocessor embedded system-on-chip architectures , 2006 .

[3]  Vittorio Zaccaria,et al.  A Sensitivity-Based Design Space Exploration Methodology for Embedded Systems , 2002, Des. Autom. Embed. Syst..

[4]  Lothar Thiele,et al.  Comparison of Multiobjective Evolutionary Algorithms: Empirical Results , 2000, Evolutionary Computation.

[5]  Andrzej Jaszkiewicz,et al.  Multiple objective metaheuristic algorithms for combinatorial optimization , 2001 .

[6]  Vincenzo Catania,et al.  Efficient design space exploration for application specific systems-on-a-chip , 2007, J. Syst. Archit..

[7]  Niraj K. Jha,et al.  MOGAC: a multiobjective genetic algorithm for the co-synthesis of hardware-software embedded systems , 1997, 1997 Proceedings of IEEE International Conference on Computer Aided Design (ICCAD).

[8]  Erik Brockmeyer,et al.  Storage Management Programmable Process , 2002 .

[9]  Andy D. Pimentel,et al.  A systematic approach to exploring embedded system architectures at multiple abstraction levels , 2006, IEEE Transactions on Computers.

[10]  Francky Catthoor,et al.  Systematic dynamic memory management design methodology for reduced memory footprint , 2006, TODE.

[11]  Vittorio Zaccaria,et al.  An efficient design space exploration methodology for multiprocessor SoC architectures based on response surface methods , 2008, 2008 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation.

[12]  Erik Brockmeyer,et al.  Data Access and Storage Management for Embedded Programmable Processors , 2002, Springer US.

[13]  Marco Laumanns,et al.  Performance assessment of multiobjective optimizers: an analysis and review , 2003, IEEE Trans. Evol. Comput..

[14]  SilvanoCristina,et al.  Multi-objective design space exploration of embedded systems , 2005 .

[15]  Viktor K. Prasanna,et al.  Rapid design space exploration of heterogeneous embedded systems using symbolic search and multi-granular simulation , 2002, LCTES/SCOPES '02.

[16]  Grant Martin,et al.  Overview of the MPSoC design challenge , 2006, 2006 43rd ACM/IEEE Design Automation Conference.

[17]  Richard S. Sutton,et al.  Introduction to Reinforcement Learning , 1998 .

[18]  Lee W. Schruben,et al.  An experimental procedure for simulation response surface model identification , 1987, CACM.

[19]  David J. Lilja,et al.  Simulation of computer architectures: simulators, benchmarks, methodologies, and recommendations , 2006, IEEE Transactions on Computers.

[20]  Maurizio Palesi,et al.  Multi-objective design space exploration using genetic algorithms , 2002, Proceedings of the Tenth International Symposium on Hardware/Software Codesign. CODES 2002 (IEEE Cat. No.02TH8627).

[21]  Cristina Silvano,et al.  Decision-theoretic exploration of multiProcessor platforms , 2006, Proceedings of the 4th International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS '06).

[22]  Frank Vahid,et al.  Soft-core Processor Customization using the Design of Experiments Paradigm , 2007, 2007 Design, Automation & Test in Europe Conference & Exhibition.

[23]  Yaochu Jin,et al.  A Critical Survey of Performance Indices for Multi-Objective Optimisation , 2003 .

[24]  Martin Lukasiewycz,et al.  Efficient symbolic multi-objective design space exploration , 2008, 2008 Asia and South Pacific Design Automation Conference.

[25]  Peter Norvig,et al.  Artificial Intelligence: A Modern Approach , 1995 .

[26]  Marco Laumanns,et al.  An efficient, adaptive parameter variation scheme for metaheuristics based on the epsilon-constraint method , 2006, Eur. J. Oper. Res..

[27]  Jason Cong,et al.  Platform-Based Behavior-Level and System-Level Synthesis , 2006, 2006 IEEE International SOC Conference.

[28]  H. Ishibuchi,et al.  Multi-objective genetic algorithm and its applications to flowshop scheduling , 1996 .

[29]  Bernhard Sendhoff,et al.  A critical survey of performance indices for multi-objective optimisation , 2003, The 2003 Congress on Evolutionary Computation, 2003. CEC '03..

[30]  Margaret J. Robertson,et al.  Design and Analysis of Experiments , 2006, Handbook of statistics.

[31]  Luca Fossati,et al.  ReSP: A Nonintrusive Transaction-Level Reflective MPSoC Simulation Platform for Design Space Exploration , 2009, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[32]  Pier Luca Lanzi,et al.  Ant colony optimization for mapping and scheduling in heterogeneous multiprocessor systems , 2008, 2008 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation.

[33]  Nikil Dutt,et al.  An Enhanced Power Estimation Model for On-Chip Caches , 2004 .

[34]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[35]  Vittorio Zaccaria,et al.  Multi-objective design space exploration of embedded systems , 2003, J. Embed. Comput..

[36]  Alberto L. Sangiovanni-Vincentelli,et al.  Benefits and challenges for platform-based design , 2004, Proceedings. 41st Design Automation Conference, 2004..

[37]  Vittorio Zaccaria,et al.  Discrete Particle Swarm Optimization for Multi-objective Design Space Exploration , 2008, 2008 11th EUROMICRO Conference on Digital System Design Architectures, Methods and Tools.

[38]  Jörg Henkel,et al.  System-level exploration for Pareto-optimal configurations in parameterized systems-on-a-chip , 2001, IEEE/ACM International Conference on Computer Aided Design. ICCAD 2001. IEEE/ACM Digest of Technical Papers (Cat. No.01CH37281).

[39]  Luca Fossati,et al.  ReSP: A non-intrusive Transaction-Level Reflective MPSoC Simulation Platform for design space exploration , 2008, 2008 Asia and South Pacific Design Automation Conference.