A number of so-called massively parallel computers exist today. While some of these support an easier-to-program shared memory programming model, the greatest potential for highly efficient use of non-shared memory parallel computers will require use of some form of non-shared memory programming model. Distributed or decentralized control applications also require use of a non-shared memory programming model. This research continues previously supported work towards developing and refining Petri-Net based models of distributed computation, and carries the models into automated and semi-automated software tools to aid programmers. The tool environment, called eXplicitly Parallel Algorithm analysis Tool (XPAT), supports graphical/visual programming as much as possible and also supports visualization of programs once written.
