Method for building multi-core operating environment

Abstract

The invention provides a method for a building multi-core operating environment, which is characterized in that: a plurality of pieces of cpu are divided into two groups in advance, wherein one group is on-line cpu which operates a standard Linux operating system for checking message content, and the other group is an off-line cpu for operating a specified target code; when all cpu relevant resource finishes initializing, parts of on-line cpu become the off-line cpu; the off-line cpu operates one endless loop function cpu_idle so as to embed the code inlet function of a specified target into the endless loop function cpu_idle; the off-line cpu operates the specified target code to classify and distribute received network message to the corresponding on-line cpu according to data flow. Theinvention can solve the problem that cpu cost sharp increases along with the increase of cpu amount under the Linux environment and provides a new efficient path for multi-core framework development.

Description

technical field [0001] The invention relates to computer technology, in particular to a method for establishing a multi-core operating environment. Background technique [0002] Since the rise of the dual-core concept in 2006 on the X86 platform, multi-core has been deeply rooted in the hearts of the people in less than three years. Now 4-core processors have been mass-produced, and 6-core processors will soon be available. The trend of multi-core not only prompts chip manufacturers to bring newer system architecture, higher performance, lower power consumption and more diversified functions, but also brings many new values ​​and challenges to software design. An operating system generally has two architectures on a multi-core processor hardware platform, one is a Symmetric Multi-Processor (SMP) architecture, and the other is an AMP (Asynchronous Multi-Processor) architecture. like figure 1 Shown is a schematic diagram of the SMP architecture taking 4 cores as an example. ...

AI Technical Summary

Problems solved by technology

[0004] 1) This fully concurrent architecture has high requirements for software. At present, many systems are non-concurrent and safe. If they are all converted to concurrent, the workload will be heavy, and it cannot be completed in a short time.

[0005] 2) In order to solve the problem of contention for shared resources (memory data, I / O, etc.), it is necessary to use a large number of lock mechanisms, and the operation of these locks affects the performance, and the communication overhead between CPUs will also increase with time. As the number of CPUs increases, it also rises sharply. When the number of cores increases to a certain level, the performance will no longer increase or even begin to decline.

[0010] 2) Running different tasks on the same operating system is the same mechanism, but it is easier to implement, and the corresponding performance improvement effect is also less

[0014] In addition, on the X86 platform, each network card will be assigned an irq number when it is initialized, and the interrupt affinity technology can be used to direct the interrupt number irq number to the specified CPU, so that all network traffic can be processed by one CPU. Then this cpu distributes traffic to another cpu according to the characteristics of the traffic. Although the purpose of this solution can also be achieved, because this cpu is still working in a Linux environment, the occurrence of process scheduling will cause the cache to fail, and the relationship with other cpus Some overheads between them cannot be avoided, such as IPI interrupts; and in terms of scalability, it is not as good as this solution. If there are more available cpu resources, it is very complicated to implement using interrupt affinity technology.

Images