The method of constructing a schedule for parallel algorithm execution is considered in the article. This algorithm takes into account the execution time of each operation of the algorithm and the relationship of operations on the data. The method is based on an information graph in which the nodes are the operations of the algorithm, and the edges are the directions of the data transfer. As a result of the interchange of operations between computing nodes, it is possible to achieve a reduction in the execution time of the algorithm by reducing the time spent on data transfer between computing nodes and reducing the downtime of computational nodes. The algorithm can be applied both in parallel programming and in adjacent areas, for example, when scheduling tasks in distributed systems.
|Название||Lecture Notes in Computer Science |
|ISSN (печатное издание)||0302-9743|
|ISSN (электронное издание)||1611-3349|
|конференция||18th International Conference on Computational Science and Its Applications, ICCSA 2018|
|Период||2/07/18 → 5/07/18|
- Теоретические компьютерные науки
- Компьютерные науки (все)