Fully automatic verification and error detection for parameterized iterative sequential circuits

Abstract

The paper shows how iterative parametric sequential circuits, which are most relevant in practice, can be verified fully automatically. Key observation is that monadic second-order logic on strings provides an adequate level for hardware specification and implementation. This allows us to apply the corresponding decision procedure and counter-model generator implemented in the Mona verification tool, which, for the first time, yields 'push-button' verification, and error detection and diagnosis for the considered class of circuits. As illustrated by means of various versions of counters, this approach captures hierarchical and mixed mode verification, as well as the treatment of varying connectivity in iterative designs. MotivationA clear trend towards reuse of existing designs is the emergence of parametric designs in standard libraries [13]. While such "families of circuits" have been already popular in the hardware verification community for years, where they are the best examples for induction-based reasoning in the hardware application domain, industrial practice did not feature parametric designs in standard libraries because of lack of consolidated methods for their lull-automatic treatment in the design lifecycle. Thus the pressure towards fully automated methods for the analysis, verification, and fault detection of parametric circuits and their specifications grows with the increasing demand for design reuse. Unfortunately, the standard automata theoretic techniques of the hardware community fail here, because the treatment of automata of unbounded size is required. At this stage, Basin and Klarlund discovered that monadic second-order logic on strings, although 'hopeless' from the complexity point of view, is well-behaved in many practical applications, and, in particular, well-suited for the fully automatic verification of *Fakult~t f'dr Mathematik und Informatik, Universitiit Passau, Innstr. 33, 94032 Passau (D), tel: +49 851 509.3096, fax: +49 851 509.3092, ~iziana@fmi.tmi-paasau.dl

Synthesizing Semantic Web Service Compositions with jMosel and Golog

“…A key characteristic of M2L(Str) is that it is compositional : atomic formulae are transformed into simple basic automata; for a compound formula, first the automata for its sub formulae are calculated, and then an automata synthesis operation is applied. For details refer to [17].…”

Mosel: A flexible toolset for monadic second-order logic

“…elementary gates and memory elements (say D-type flip-flops) have been defined as predicate macros on top of Mona. An example of a fairly complete library can be found in [18]. Building on these predicates, higher abstraction levels can be captured too, so that more complex predicates represent entire circuits or families of circuits.…”

“…This is due to the interpretation of the logic over strings, which implies that the parameterization allowed to express generalized behaviours is limited to the generic "length" of strings. Since strings may be taken to assume different meanings (in [16,17] sampled waveforms for control circuits, in [18] the bitwidth of a datapath), a degree of freedom in the use of the logic is still left to the application designer.…”

“…The M2L(Str) logic and its fully automatic decision procedure have allowed us to capture, in a common framework, a wide spectrum of abstraction levels, ranging from generic architecture or protocol levels [17] to the hardwareoriented register transfer and gate levels [16,18]. In particular, both behavioural and structural description styles are supported, and from both it is possible to carry out model-based analysis, verification, and error detection.…”

“…Fields of application have been the specification, verification, and synthesis, in a fully automatic manner, of relevant classes of parametric systems. In particular, the logic can be used profitably as a description language for model-based analysis of software [17] as well as hardware systems [2,16,18,19] and is therefore a good candidate formalism for hardware/software codesign. Some examples of distributed systems have been addressed too [14,15].…”