A bi-objective heuristic for heterogeneous MPSoC design space exploration

Recent technology advances allow new generation reconfigurable-based embedded systems to contain a large number of cores and reconfigurable logic elements. Consequently, to take benefit of such very powerful hybrid reconfigurable MPSoC, designers need tools to explore the large design space of the possible configurations. In this paper, we develop a new hybrid bi-objective genetic and parallel variable neighborhood descent algorithm (GA-PVNS) to determine close to optimal configurations for heterogeneous FPGA-based MPSoC (Ht-MPSoC). Our exploration method aims to optimize simultaneously two objectives, namely area on the FPGA and execution time, taking advantage of the Genetic Algorithm (GA) diversification ability and the intensification provided by Variable Neighborhood Search (VNS) algorithm. Our design space includes Ht-MPSoC with private and shared HW accelerators on FPGA. Compared to exact methods, our algorithm determines very satisfactory multi-objective configurations in a very reduced execution time.

[1]  José R. Álvarez,et al.  Artificial Neural Nets Problem Solving Methods , 2003, Lecture Notes in Computer Science.

[2]  Mohamed Abid,et al.  Design Space Exploration for Customized Asymmetric Heterogeneous MPSoC , 2014, 2014 17th Euromicro Conference on Digital System Design.

[3]  Lothar Thiele,et al.  Multiobjective evolutionary algorithms: a comparative case study and the strength Pareto approach , 1999, IEEE Trans. Evol. Comput..

[4]  Vittorio Zaccaria,et al.  Multicube Explorer: An Open Source Framework for Design Space Exploration of Chip Multi-Processors , 2010, ARCS Workshops.

[5]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[6]  H. Ishibuchi,et al.  Multi-objective genetic local search for minimizing the number of fuzzy rules for pattern classification problems , 1998, 1998 IEEE International Conference on Fuzzy Systems Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98CH36228).

[7]  Pierre Hansen,et al.  Variable Neighborhood Search , 2018, Handbook of Heuristics.

[8]  Markus Rupp,et al.  Design Space Exploration with Evolutionary Multi-Objective Optimisation , 2007, 2007 International Symposium on Industrial Embedded Systems.

[9]  Liang Chen,et al.  Shared reconfigurable fabric for multi-core customization , 2011, 2011 48th ACM/EDAC/IEEE Design Automation Conference (DAC).

[10]  David Novo,et al.  Selective Flexibility: Creating Domain-Specific Reconfigurable Arrays , 2013, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[11]  DebK.,et al.  A fast and elitist multiobjective genetic algorithm , 2002 .

[12]  Giovanni Beltrame,et al.  A comparative evaluation of multi-objective exploration algorithms for high-level design , 2014, TODE.

[13]  Nigel P. Topham,et al.  Design-Space Exploration of Resource-Sharing Solutions for Custom Instruction Set Extensions , 2009, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[14]  Hisao Ishibuchi,et al.  Multi-objective genetic local search algorithm , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[15]  Majid Sarrafzadeh,et al.  Area-efficient instruction set synthesis for reconfigurable system-on-chip designs , 2004, Proceedings. 41st Design Automation Conference, 2004..

[16]  Babak Falsafi,et al.  The HiPEAC Vision , 2010 .