Software Issues - Methodology of Parallel Computing

Software Issues - Methodology of Parallel Computing

• Simple, but general and extensible to many more nodes is domain decomposition

• Most successful concurrent machines with many nodes have obtained good performance from "Data Parallelism" or "Domain Decomposition"

• Problem is an algorithm applied to data set
  - Obtain parallelism by acting on data concurrently.
  • The three architectures considered here differ as follows:
    - MIMD-DM - Processing and Data Distributed - MIMD-SM - Processing Distributed but memory shared - SIMD-DM - Synchronous Processing on Distributed Data

    Previous slide

    Next slide

    Back to the first slide

    View Graphic Version