Thread lock optimization method

Abstract

The present invention provides a thread lock optimization method. The method comprises the following steps: determining a key section in a memory; and using a synchronization primitive provided by the CPU, controlling a thread lock to take charge of protecting the key section, before a thread enters the key section, allowing to obtain the access right of the thread lock firstly, and ensuring that at the same time only one thread can enter the key section. The key section is a memory space stored with shared variables, and the determination of the key section is designated by a user. According to the thread lock optimization method disclosed by the present invention, the synchronization expense between the common lock threads in the original universal system is avoided, only a small number of CPU primitives is used, safety between threads is realized, the rate is increased by a hundred times, and performance of synchronous parallel access to the same locked area is improved by at least a hundred times, so that the performance is significantly enhanced, the CPU occupancy rate is significantly reduced, and the conflict between the threads is reduced.

Description

technical field [0001] The invention relates to the technical field of computers, in particular to a thread lock optimization method. Background technique [0002] In the prior art, relying on the threading model provided by the operating system, such as POSIX (Portable Operating System Interface) thread locks, mutual exclusion between threads can access the same memory object, so as to achieve the purpose of safely sharing data between threads. [0003] However, POSIX thread mutexes or POSXI-compatible thread mutexes are implemented through operating system callbacks. Since system callbacks need to switch between user space and operating system kernel space multiple times, POSIX thread mutexes are a kind of Heavy lock structure. In a large-scale concurrent system, implementing thread safety requires the use of a large number of mutexes, which are also an operating system resource. In a large-scale concurrent system, frequent calls to mutex locks cause frequent switching b...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an optimization method for thread locks, realize a general non-blocking lock structure, and efficiently realize safe sharing of data between threads, so as to solve the problems caused by the use of mutex locks in large-scale concurrent systems in the prior art The problem of low performance and 100% CPU occupancy of the CAS scheme, while maintaining the same generality as the POSIX mutex, lowers the theoretical and technical requirements for the correct use of the caller

Images