Authors

Using Multithreading modeling and Static Program Code Analysis for Race Conditions Detection

An approach to static code analysis that allows race conditions detecting is described. The method focuses on real tasks with complex code constructs, including loops and branches. The method proposed is based on a simulation of multi-threaded application performance and analysis of the model. The model is suitable for real-world problems, which are defined as those where the logic and the notion of correctness attached to the concrete values of variables and history of operations on shared variables rather than to their pair-wise comparisons.

Static detection of races in code that contains loops and branching

Today analyzing data-races is a relevant problem because of complex architecture of the modern programs. The article introduces the model of concurrent threads execution that alows analyzing data-races statically. Target algorithms for the model are ones containing custom synchronization machinery and atomic operations. As opposed to known models of a concurrent execution the suggested model analyze the code at a lower level of instructions.