There are lots of different approaches developed for global static program error analysis. Most of them are concentrated on sequential programs. Analysis of parallel programs is considered complicated. The overview contains brief information on obsolete and modern approaches to global static error analysis of parallel programs.

R.N. Taylor has proposed Call History Graph to analyse anomalies in parallel programs. R. Rugina and M. Rinard have developed context-sensitive flow-sensitive pointer aliasing analysis for multithread programs to detect regular errors. In LASER laboratory, the approach of computing MHP (May Happen in Parallel) information for a Java program has been developed. M. Dwyer has proposed the extension of the classical semi-lattice model of the data flow analysis to the complete lattice model.

perchine.pdf112.18 KB