μC-OS II-based Operating System design for cluster in NoC-based MPSoC
Jianxin Ma, Fangfa Fu, Zhaochi Liu, Zixu Wu, Jinxiang Wang
Abstract
Jianxin Ma, Fangfa Fu, Zhaochi Liu, Zixu Wu, Jinxiang Wang
Abstract
The traditional architecture of Network on Chip (NoC) in which each PE only has one core can't make the best of router bandwidth and may cause wasting of resources. To avoid these problems, the Cluster has been introduced to the MPSoC NoC-based to improve the performance. In this paper, an Embedded Multi-core Operating System based on μC-OS II has been designed for a cluster which is quad-core SMP. The Embedded Multi-core Operating System adopts a distributed multi-task scheduling strategy based on a global priority to support multi-task scheduling in the cluster. And the master-slave scheme is used in the initialize stage of the system to cut down OS memory footprint. Moreover, to realize the multi-task parallel executing on the multi-core, the paper developed a task communication function for the inner-cluster and inter-cluster based on the MMPI. Finally, we realize the system by using the MicroBlaze IP core on the Xilinx Virtex-6 ML605.
OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The traditional architecture of Network on Chip (NoC) in which each PE only has one core can't make the best of router bandwidth and may cause wasting of resources. To avoid these problems, the Cluster has been introduced to the MPSoC NoC-based to improve the performance. In this paper, an Embedded Multi-core Operating System based on μC-OS II has been designed for a cluster which is quad-core SMP. The Embedded Multi-core Operating System adopts a distributed multi-task scheduling strategy based on a global priority to support multi-task scheduling in the cluster. And the master-slave scheme is used in the initialize stage of the system to cut down OS memory footprint. Moreover, to realize the multi-task parallel executing on the multi-core, the paper developed a task communication function for the inner-cluster and inter-cluster based on the MMPI. Finally, we realize the system by using the MicroBlaze IP core on the Xilinx Virtex-6 ML605.
Key concepts: MPSoC, MicroBlaze, Computer science, Embedded system, Scheduling (production processes), Multi-core processor, Operating system, Computer architecture