DPD: a distributed program debugger for the REM environment

The development of a distributed program debugger (DPD) for the remote execution manager (REM) environment is discussed. The model for DPD and REM consists of multiple processors, each with its own private memory and each capable of running several processes. DPD debugs at a higher level of abstraction than sequential debuggers, specifically at the process communication and control level. Synchronization and communication faults are more easily detected when events are the sending and receiving of messages or the creation and termination of processes. When a fault is detected in a distributed program, DPD has the ability to stop the program, roll back, and replay. During the replay, the user can control and change the behavior of the program in order to detect the synchronization faults. A graphical interface is used to show the ordering of events and the communication between processes.<>