Towards comment generation for MPI programs

Abstract

As parallel architectures are becoming more commonplace, the development of parallel software that exploits the potential parallelism is increasingly important. Like sequential programs, these programs need to be maintained, which requires considerable time understanding the maintenance tasks and related software or documentation. Studies have shown that good comments can help programmers quickly understand a program’s functions. Unfortunately, few software projects adequately comment the code. This thesis work investigates the problem of automatically generating descriptive summary comments for functions in MPI parallel programs and presents an approach to generate such descriptive summary comments for parallel MPI Java methods. Given the signature and body of a method, the automatic comment generator identifies content for the summary and generates natural language text that summarizes the method’s overall actions. A base comment generator for sequential Java is modified for a first prototype for parallel MPI programs.