Systematic debugging of parallel programs based on collective breakpoints

Abstract

The paper introduces the concept of collective breakpoints and classifies the possible parallel breakpoints comparing their mechanisms. Based on the collective breakpoints the macrostep-by-macrostep execution mode has been defined. After introducing the concept of the execution tree and meta-breakpoints the systematic debugging of message passing parallel programs is explained. The main features and distributed structure of DIWIDE, a macrostep debugger is described. The integration of DIWIDE into the GRADE and WINPAR parallel programming environments is outlined. An algorithm is shown how to generate automatically the collective breakpoints in the GRADE environment.