DocumentCode
651313
Title
Relational STE and theorem proving for formal verification of industrial circuit designs
Author
O´Leary, John ; Kaivola, Roope ; Melham, Tom
Author_Institution
Intel Corp., Hillsboro, OR, USA
fYear
2013
fDate
20-23 Oct. 2013
Firstpage
97
Lastpage
104
Abstract
Model checking by symbolic trajectory evaluation, orchestrated in a flexible functional-programming framework, is a well-established technology for correctness verification of industrial-scale circuit designs. Most verifications in this domain require decomposition into subproblems that symbolic trajectory evaluation can handle, and deductive theorem proving has long been proposed as a complement to symbolic trajectory evaluation to enable such compositional reasoning. This paper describes an approach to verification by symbolic simulation, called Relational STE, that raises verification properties to the purely logical level suitable for compositional reasoning in a theorem prover. We also introduce a new deductive theorem prover, called Goaled, that has been integrated into Intel´s Forte verification framework for this purpose. We illustrate the effectiveness of this combination of technologies by describing a general framework, accessible to non-experts, that is widely used for verification and regression validation of integer multipliers at Intel.
Keywords
circuit analysis computing; formal verification; functional programming; network synthesis; symbol manipulation; theorem proving; Goaled deductive theorem proving; Intel Forte verification framework; compositional reasoning; correctness verification; flexible functional-programming framework; formal verification; industrial-scale circuit designs; integer multipliers; logical level; model checking; regression validation; relational STE; symbolic simulation; symbolic trajectory evaluation; Boolean functions; Cognition; Computational modeling; Computer languages; Functional programming; Integrated circuit modeling; Trajectory;
fLanguage
English
Publisher
ieee
Conference_Titel
Formal Methods in Computer-Aided Design (FMCAD), 2013
Conference_Location
Portland, OR
Type
conf
DOI
10.1109/FMCAD.2013.6679397
Filename
6679397
Link To Document