| 关于二分法的一点注记 |
| |
| 引用本文： | 王能超. 关于二分法的一点注记[J]. 华中科技大学学报(自然科学版), 1988, 0(5) |
| |
| 作者姓名： | 王能超 |
| |
| 作者单位： | 华中理工大学数学系 |
| |
| 基金项目： | 国家自然科学基金资助项目 |
| |
| 摘 要： | 1.文献[1]、[2]提出了设计并行算法的一种基本技术——所谓二分技术。在设计二分算法时,我们曾假设“处理机台数不受限制,即可按算法并行化需要提供足够多台处理机”。这似乎是一项理想化的假设,因为实际并行机系统所提供的处理机台数总是确定的,不一定能达到算法所要求的“处理机台数界”。因之可能认为:上述关于处理机台数界不受限制的假设限制了二分法的应用范围。其实这是一种误解。 我们知道,传统的串行算法之所以不一定适应新一代的计算机,是因为其并行性差。譬如递推计算只能由单处理机承担。为了充分发挥多台处理机的作用,就必须大力开发算
|
A Note on the Bisection Method |
| |
| Wang Nengchao. A Note on the Bisection Method[J]. JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE, 1988, 0(5) |
| |
| Authors: | Wang Nengchao |
| |
| Affiliation: | Wang Nengchao |
| |
| Abstract: | The practicality and effectiveness of the algorithms, designed with the bisection technique for an actual system of parallel processor, are discussed. The summation and solving of the first-order linear recurrence problem are also treated. When the number of the practical processors is smaller than that of the processor bound as given by the algorithm originally designed with the bisection technique, the computation formulae for the algorithm can be divided into groups and executed in each processor respectively. Once the size of the problem is reduced and well matched to the number of processors available, the problem can be solved in accordance with the original algorithm. |
| |
| Keywords: | Bisection method Parallel algorithm. |
| 本文献已被 CNKI 等数据库收录！ |
|
