Title :
Symbolic model-based diagnosis with auto-correction framework for arithmetic circuits
Author_Institution :
Electr. & Comput. Eng. (ECE) Dept., Univ. of Tehran, Tehran, Iran
Abstract :
Design error detection, localization and correction are considered as important and difficult tasks in the arithmetic circuit debugging process. In this paper we present a symbolic model-based diagnosis framework for arithmetic circuit debugging which can deal with modulo equivalence based on a canonical decision diagram called Modular Horner Expansion Diagram (M-HED) in order to verify the equivalence of an and-inverter-graph (AIG) representation as the gate level implementation against a polynomial expression over Z2n as the specification. If two models are not equivalent, our approach is able to automatically correct the AIG according to the specification. We evaluate our approach on several large arithmetic circuits thereby showing performance benefits of many orders of magnitude than widely accepted industrial techniques.
Keywords :
arithmetic; decision diagrams; error detection; graph theory; logic design; logic gates; and-inverter-graph representation; arithmetic circuit debugging process; autocorrection framework; canonical decision diagram; error detection design; modular Horner expansion diagram; modulo equivalence; polynomial expression; symbolic model-based diagnosis framework; Adders; Bismuth; Debugging; Integrated circuit modeling; Logic gates; Optimization; Polynomials; Arithmetic Circuit; Debugging; Diagnosis;
Conference_Titel :
Quality Electronic Design (ASQED), 2011 3rd Asia Symposium on
Conference_Location :
Kuala Lumpur
Print_ISBN :
978-1-4577-0145-0
DOI :
10.1109/ASQED.2011.6111744