Distributed processing has been a subject of recent interest due to the availability of computer networks. Over the past few years it has led to the identification of several challenging problems. One of these is the ...
详细信息
Distributed processing has been a subject of recent interest due to the availability of computer networks. Over the past few years it has led to the identification of several challenging problems. One of these is the problem of optimally distributing program modules over a distributed processing system. This paper presents an LC (Least Cost) branch-and-bound algorithm to find an optimal assignment that minimizes the sum of execution costs and communication costs. Experimental results show that, for over half of the randomly generated instances, the saving rates exceed 99%.
The importance of parallel programming increases as multiple processor systems become accepted more widely. Recent research investigates some approaches to the design and analysis of parallel algorithms, and several e...
详细信息
The importance of parallel programming increases as multiple processor systems become accepted more widely. Recent research investigates some approaches to the design and analysis of parallel algorithms, and several examples serve to illustrate the importance of interprocessor communication in parallel processing. Various techniques that are applicable in the design and analysis of parallel algorithms are reviewed, the emphasis being on those techniques that incorporate communication aspects. Included in the research are several models of synchronous and asynchronous parallel computation and their use in the analysis of algorithms. Relatively primitive methodologies for designing parallel algorithms are examined, with the conclusion that a need exists for more general and practical methodologies. Models reviewed include: 1. a shared memory parallel method, 2. a communication-oriented model, and 3. automata models.
暂无评论