A task migration algorithm for power management on heterogeneous multicore

DOI

10.2991/icamcs-16.2016.74How to use a DOI?

Keywords

cache contention, power consumption, heterogeneous multicore system, cache miss.

Abstract

The advantages of the increased chip integration for multicore technology exponentially improve the performance of the processor chip. However, with the increased number of cores, the power consumption of the system has become an increasingly important problem. Compared with homogeneous multicore, the main advantage of heterogeneous architecture on the power consumption is that it allows for the inclusion of processors whose design is specialized for particular types of jobs, and for jobs to be assigned to a processor best suited for that job. In most multicore platforms, cache resources is a scarce resource, different cores share on chip caches. Without effective management by the scheduler, such caches can cause thrashing that severely degrades system performance. An excellent task scheduling algorithm is an effective method to reduce Cache contention. The purpose of this paper is through the program feature classification reduce process cache contention, let the process which can better make use of the performance of processor execute upon high frequency cores, improving the processor performance reduce the execution time to achieve the purpose of reducing power consumption in a heterogeneous multicore system.

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Manman Peng
AU  - Wen Luo
PY  - 2016/06
DA  - 2016/06
TI  - A task migration algorithm for power management on heterogeneous multicore
BT  - Proceedings of the 2016 5th International Conference on Advanced Materials and Computer Science
PB  - Atlantis Press
SP  - 350
EP  - 356
SN  - 2352-5401
UR  - https://doi.org/10.2991/icamcs-16.2016.74
DO  - 10.2991/icamcs-16.2016.74
ID  - Peng2016/06
ER  -