Single-instruction-set heterogeneous multi-core system static task scheduling method

A heterogeneous multi-core, task scheduling technology, applied in the direction of multi-programming devices, etc., can solve problems such as unrealistic, inability to guarantee the quality of the solution, complexity and multiple constraints of scheduling problems, etc., to achieve short completion time, initialization efficiency and effective The effect of individual improvement and low power consumption

Inactive Publication Date: 2013-01-09
CAPITAL NORMAL UNIVERSITY
View PDF2 Cites 28 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The scheduling problem has the characteristics of complexity and multiple constraints. It is an NP-complete problem in the combinatorial optimization problem. It is unrealistic to use the exhaustive method to search for the optimal scheduling problem.
The heuristic algorithm has good time complexity and the ability to obtain near-optimal solutions under certain conditions, but the heuristic algorithm is a local optimization and cannot guarantee the qu...

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Single-instruction-set heterogeneous multi-core system static task scheduling method
  • Single-instruction-set heterogeneous multi-core system static task scheduling method
  • Single-instruction-set heterogeneous multi-core system static task scheduling method

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0084] From figure 1 As shown in the DAG diagram, it can be seen that T 2 The height value is 1, T 5 The height value of is 2, defined according to the height value, T 2 must be at T 5 Before execution, in practice, this constraint is not necessary, from figure 1 It can be seen that T 2 and T 5 There is no constraint relationship between them, as long as T 3 (no matter on which core) the output data after execution is passed to T 5 The processor core where T 5 can be executed, T 2 available at T 5 run after that. In the same way, figure 2 In the DAG graph shown, if height values ​​are used, then T 2 The height value is 0, T 3 with T 5 The height value is 1, then T 2 must be at T 5 done before, in practice, this constraint is also unnecessary, T 2 During execution, T 3 with T 5 is waiting in vain, in order to make way for T on another processor core 4 up and running as early as possible, should allow T 3 before T 2 run. Assuming that the execution time o...

example 2

[0086] In order to simulate and test the effect of multi-objective optimization genetic algorithm when the number of tasks is large, the algorithm is implemented in C language. Randomly generate DAG graphs and task-related attributes of 30 tasks. The processor core configuration is set to two types, 4-core (1 fast core and 3 slow cores) and 8-core (2 fast cores and 6 slow cores), where the main frequency of the fast core is twice that of the slow core. When there is a program running, the power consumption of the processor core is mainly dynamic power consumption, which is related to the main frequency and voltage. The approximate power consumption of the slow core is 2 energy consumption units, and the power consumption of the fast core is 5 energy consumption units. (approximately 2.5 times that of the slow core), assuming that no program is running, the power consumption of the fast core is 1 energy consumption unit, and the power consumption of the slow core is 0.5 energy ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

Provided is a single-instruction-set heterogeneous multi-core system static task scheduling method. The method includes five steps: step 1, population initialization; step 2, fitness value calculation; step 3, selection operator operation; step 4, cross operator operation; and step 5 variation operator operation. A local sequence represents an executing sequence of two tasks without depending relations, population initialization efficiency and effective individual are greatly improved, the executing sequence of the tasks is determined through a pre-order-relation matrix, and defects of a traditional height value method are overcome. The method can widen a hunting range of optimal individuals. When the population scale is large enough, a part of optimal solutions missed by the height value method can be found so as to obtain a more optimal scheduling sequence. For the same task set, finishing time of the whole task set is short, power consumption is low, and a purpose of energy conservation and consumption reduction is achieved.

Description

technical field [0001] The invention relates to a single instruction set heterogeneous multi-core system static task scheduling method based on a multi-objective optimization genetic algorithm (multi-objective optimization genetic algorithm), which belongs to the technical field of computer system structure. Background technique [0002] The rise of cloud computing has led to a rapid increase in energy consumption in data centers, but energy consumption has not been effectively utilized. Single instruction set heterogeneous multi-core processor is a new type of architecture proposed in recent years. Compared with homogeneous multi-core processors, this processor has better performance per power ratio, but at the same time it brings great challenges to task scheduling. . [0003] The scheduling problem has the characteristics of complexity and multiple constraints, and it is an NP-complete problem in the combinatorial optimization problem. It is unrealistic to use the exhaus...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F9/46
Inventor 徐远超谭旭范东睿张浩王达宋风龙张志敏
Owner CAPITAL NORMAL UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap