A parallel algorithm based on CPU/GPU dynamic coordination control and scheduling1

被引：0

作者：

Jiang, Hanyang ^{[1
,2
]}

Wang, Danhua ^{[1
]}

机构：

[1] Hengyang Normal University, Hengyang,421002, China

[2] Hunan Provincial Key Laboratory of Intelligent Information Processing and Application, Hengyang,421002, China

来源：

Acta Technica CSAV (Ceskoslovensk Akademie Ved) | 2017年 / 62卷 / 01期

关键词：

Energy efficiency - Scheduling - Computing power - Parallel algorithms - Electric power utilization - Synchronization - Green computing;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

CPU/GPU dynamic coordination technology is gradually applied to the development of parallel algorithms, but the current algorithm has some shortcomings in terms of efficiency and power consumption. Therefore, it is necessary to further study the parallel algorithm based on dynamic coordination control and scheduling. Based on the power cap technology, this paper sets up the device frequency and the single computing node of the synchronization system by studying the synchronization system of the computer, controlling DVFS and mapping mode. In order to avoid power failure and load imbalance, this paper has developed a new empirical model of performance and synchronization system, these models make the most of the power of the system to reduce the power consumption to a predetermined level, Good settings can be done. © 2017 Institute of Thermomechanics CAS, v.v.i.

引用

页码：361 / 369