The design of parallel GIScience applications is bound to specific conventional parallel computer architectures, such as SIMD(single instruction stream and multiple data streams) and MIMD(multiple instruction streams and multiple data streams) systems (Flynn 1966). This tight-coupleing approach is problemtic in three important respects

1. Logic: the design of domain decomposition and task scheduling methods is focused on the characteritics of spatial data and the analysis operations that 
2. Generality: Any change to the characteristics of spatial data and operations requires a corresponding change to be made in the design and implementation of domain decomposition and task scheduling methods. This makes it difficult to develop generic parallel processing solutions for geographical analysis.
3. Compatibility: Solutions to domain decomposition and task scheduling are dependent upong specific parallel computer architectures. A change in architechture requires a corresponding change to the solution process even if the method of analysis were held constant

转载本文请联系原作者获取授权，同时请注明本文来自郭旦怀科学网博客。
链接地址：https://m.sciencenet.cn/blog-204718-304198.html

上一篇：如何提高记忆力
下一篇：外甥女的鼓励