"Smart" design space sampling to predict Pareto-optimal solutions

Many high-level synthesis tools offer degrees of freedom in mapping high-level specifications to Register-Transfer Level descriptions. These choices do not affect the functional behavior but span a design space of different cost-performance tradeoffs. In this paper we present a novel machine learning-based approach that efficiently determines the Pareto-optimal designs while only sampling and synthesizing a fraction of the design space. The approach combines three key components: (1) A regression model based on Gaussian processes to predict area and throughput based on synthesis training data. (2) A "smart" sampling strategy, GP-PUCB, to iteratively refine the model by carefully selecting the next design to synthesize to maximize progress. (3) A stopping criterion based on assessing the accuracy of the model without access to complete synthesis data. We demonstrate the effectiveness of our approach using IP generators for discrete Fourier transforms and sorting networks. However, our algorithm is not specific to this application and can be applied to a wide range of Pareto front prediction problems.

[1]  Michael T. M. Emmerich,et al.  Single- and multiobjective evolutionary optimization assisted by Gaussian random field metamodels , 2006, IEEE Transactions on Evolutionary Computation.

[2]  Xavier Gandibleux,et al.  A survey and annotated bibliography of multiobjective combinatorial optimization , 2000, OR Spectr..

[3]  James C. Hoe,et al.  Fast and accurate resource estimation of automatically generated custom DFT IP cores , 2006, FPGA '06.

[4]  Franz Franchetti,et al.  Formal datapath representation and manipulation for implementing DSP transforms , 2008, 2008 45th ACM/IEEE Design Automation Conference.

[5]  Markus Püschel,et al.  Computer generation of streaming sorting networks , 2012, DAC Design Automation Conference 2012.

[6]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[7]  Vittorio Zaccaria,et al.  ReSPIR: A Response Surface-Based Pareto Iterative Refinement for Application-Specific Design Space Exploration , 2009, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[8]  Qingfu Zhang,et al.  Expensive Multiobjective Optimization by MOEA/D With Gaussian Process Model , 2010, IEEE Transactions on Evolutionary Computation.

[9]  Pedro C. Diniz,et al.  A compiler approach to fast hardware design space exploration in FPGA-based systems , 2002, PLDI '02.

[10]  Joshua D. Knowles,et al.  ParEGO: a hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems , 2006, IEEE Transactions on Evolutionary Computation.

[11]  Lothar Thiele,et al.  Modular design space exploration framework for embedded systems , 2005 .

[12]  Marco Laumanns,et al.  SPEA2: Improving the Strength Pareto Evolutionary Algorithm For Multiobjective Optimization , 2002 .

[13]  Andreas Krause,et al.  Information-Theoretic Regret Bounds for Gaussian Process Optimization in the Bandit Setting , 2009, IEEE Transactions on Information Theory.

[14]  Gary B. Lamont,et al.  Evolutionary Algorithms for Solving Multi-Objective Problems (Genetic and Evolutionary Computation) , 2006 .

[15]  Chaitali Chakrabarti,et al.  Accurate models for estimating area and power of FPGA implementations , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[16]  Lothar Thiele,et al.  The Hypervolume Indicator Revisited: On the Design of Pareto-compliant Indicators Via Weighted Integration , 2007, EMO.

[17]  Eckart Zitzler,et al.  Indicator-Based Selection in Multiobjective Search , 2004, PPSN.

[18]  Niu Gang,et al.  Area and delay estimation for FPGA implementation of coarse-grained reconfigurable architectures , 2006 .