Single-machine multi-core parallel model checking method with high performance

Abstract

The invention discloses a single-machine multi-core parallel model checking method with high performance. The method comprises the following steps of: (1) establishing an initial node in a composite state space of a to-be-detected system model and a to-be-detected attribute; (2) mapping the initial node into a vertex of a Ligra (Lightweight Graph Processing Framework for Shared Memory) calculation chart by a state space generator, then gradually generating successor nodes of the initial node, wherein the initial node and the generated successor nodes form a state space, the initial node is a node in a first level, the node in an ith level includes successor nodes in an (i+1)th level of N layers, the successor nodes of each layer include a plurality of nodes, a current state space is mapped into the Ligra calculation chart after all successor nodes belonging to nodes in the same level are generated each time, then whether a receivable cycle is in the Ligra calculation chart is checked by calling a receivable cycle detector, and if the receivable cycle is in the Ligra calculation chart, model checking is stopped. The method has the characteristics of high performance, good expandability and the like.

Description

technical field [0001] The invention belongs to the fields of software engineering, formalized methods and parallel computing, and specifically relates to a high-performance parallel linear sequential logic model detection method in a single-machine multi-core environment. Background technique [0002] Model checking (Model Checking) is a very important automatic verification technology. It was first proposed by Clarke and Emerson in 1981 respectively, and it mainly verifies the modal / propositional properties of finite-state concurrent systems through explicit state search or implicit fixed point calculation. Model checking has been used in the analysis and verification of many safety-critical systems, communication protocols, control systems, and safety authentication protocols in recent years. The model detector is used to conduct a comprehensive inspection of the security and activity of the hardware and software system to ensure that the system meets the security charac...

AI Technical Summary

Problems solved by technology

Although single-machine multi-core computing based on shared memory can improve computing efficiency, using existing single-machine multi-core parallel linear sequential logic model detectors, such as Divine, etc., requires multiple threads to process the entire state space. In order to reduce the competition between threads, usually need Divide the entire state space to each thread, which reduces competition but introduces a large segmentation overhead, resulting in poor performance

Images