A novel abstraction-guided simulation approach using posterior probabilities for verification

This paper presents a novel abstraction-guided simulation approach for multiple target states which uses posterior probabilities of the states from the abstract model, instead of abstract distances used by former abstraction-guided approaches, as the guidance of simulation. The posterior probabilities carry more precise information of the abstract model, being able to offer more effective guidance as well as allow the simulation to deal with multiple target states at a time. Experimental results show that the simulation using posterior probabilities as guidance is much more efficient than that using the abstract distances, and the multiple target states simulation framework reduces the simulation cycles effectively.

[1]  Jacob A. Abraham,et al.  Abstraction Techniques for Validation Coverage Analysis and Test Generation , 1998, IEEE Trans. Computers.

[2]  Tao Zhang,et al.  An abstraction-guided simulation approach using Markov models for microprocessor verification , 2010, 2010 Design, Automation & Test in Europe Conference & Exhibition (DATE 2010).

[3]  Avner Landver,et al.  Coverage-Directed Test Generation Using Symbolic Techniques , 1996, FMCAD.

[4]  Rajeev Motwani,et al.  Randomized Algorithms , 1995, SIGA.

[5]  Mark Horowitz,et al.  Architecture validation for processors , 1995, Proceedings 22nd Annual International Symposium on Computer Architecture.

[6]  Michael S. Hsiao,et al.  An ant colony optimization technique for abstraction-guided state justification , 2009, 2009 International Test Conference.

[7]  Michael S. Hsiao,et al.  Efficient Design Validation Based on Cultural Algorithms , 2008, 2008 Design, Automation and Test in Europe.

[8]  Alan J. Hu,et al.  An Effective Guidance Strategy for Abstraction-Guided Simulation , 2007, 2007 44th ACM/IEEE Design Automation Conference.