In this work, the impact of physical/logical network topol- ogy on parallel matrix computation is studied on an Intel Touchstone DELTA mesh and three generations of hyper- Cube multiprocessors. These machines are representative Of the continued development of distributed-memory Message-passing multiprocessors in the current decade. As the processor architecture and the network hardware Continue to improve, it is important that the software being Developed can easily take advantage of such improve- Ment. The author shows that for a collection of mathemati- Cal software fundamental to parallel matrix computation, This objective is accomplished by basing the parallel algo- Rithms and the software development on a logical Subcube-grid network topology.
展开▼