Automatic online-offline method and system of CPU (central processing unit) and memory block

Abstract

The invention discloses an automatic online-offline method and system of a CPU (central processing unit) and a memory block. The method includes: according to a selected online-offline strategy, monitoring average load of the CPU of the system and memory using rate in real time through a daemon process; when online or offline conditions are met, inquiring online state bitmap of the CPU and the memory block of the system, selecting out the CPU or the memory block capable of being online and offline for online and offline operation; after the online and offline operation is performed, updating online state of the corresponding CPU or the memory block to the online state bitmap. After the strategy formulated in advance is guided in, for business requests changing in real time, a server can automatically decide whether online-offline of the CPU and the memory block is needed or not and automatically perform the online-offline operation of the CPU and the memory block, so that reaction speed is high, and decision making is better, flexibility in using the high-end server is improved, manpower management cost is saved, energy consumption of the whole system is lowered properly, and cooling of the whole system is reduced.

Description

technical field [0001] The invention relates to the technical field of operating system design for high-end servers, in particular to a method and system for automatically logging on and off the CPU and memory blocks. Background technique [0002] With the development of chip interconnection technology, more and more processors and memory are installed inside a high-end server. Correspondingly, the computing and processing capabilities of high-end servers are becoming more and more powerful, and the energy consumption of the whole machine is increasing. Inside the server, the management granularity of the processor is a core or CPU, and the management granularity of the memory is a memory block of a uniform size. During the use of high-end servers, according to different businesses and different time periods, sometimes it is not necessary for all CPUs and memory blocks to operate at the same time to meet business needs. For the sake of energy saving, some CPU and memory blo...

AI Technical Summary

Problems solved by technology

Manually carry out online and offline operations of CPU and memory blocks, which has high labor costs and slow response speed, and administrators often cannot make optimal operations, such as offline several CPUs and several memory blocks

[0004] During off-peak hours when certain services are running, administrators can manually offline process the CPU and memory blocks of the server to save energy consumption, but administrators are often unable to determine how many CPU and memory blocks are offline without affecting business operations. Down to achieve the best energy saving effect

At the same time, if there is a sudden increase in the number of business requests and the administrator is not on the job, the offline CPU and memory blocks cannot be brought online in time, resulting in the inability to fully utilize the performance of the server, and affecting business operations and user experience.

Images