Empirical results are presented on the performance of parallel computations, with respect to various performance criteria, under different assumptions of the underlying computer architecture. The performance criteria used are the Parallel Index, the Speedup, the Utilization, the Efficiency, the Redundancy, the Compression and a definition of the Quality of the resultant computation. The underlying architectures assumed are parallel processor organizations of both the SIMD and MIMD varieties, with limited and unlimited degrees of physical parallelism.
|Original language||English (US)|
|Number of pages||10|
|State||Published - Jan 1 1980|
|Event||Proc Int Conf Parallel Process 9th - Harbor Springs, MI, USA|
Duration: Aug 26 1980 → Aug 29 1980
All Science Journal Classification (ASJC) codes
- Hardware and Architecture