Data-Flow Algorithms for Parallel Matrix Computations

In this article the authors develop some algorithms and tools for solving matrix problems on parallel processing computers. Operations are synchronized through data-flow alone, which makes global synchronization unnecessary and enables the algorithms to be implemented on machines with very simple operating systems and communication protocols. As examples, an algorithm that forms the main modules for solving Liapounov matrix equations is presented. The authors compare this approach to wave front array processors and systolic arrays, and note its advantages in handling missized problems, in evaluating variations of algorithms or architectures, in moving algorithms from system to system, and in debugging parallel algorithms on sequential machine