Task scheduling method and system based on task stealing

Abstract

The invention discloses a task scheduling method and system based on task stealing. The method comprises the steps that a task dependence graph is constructed, and a dependence point serves as a callback function to be registered into a callback container of a depended point; a lock-free double-end queue is allocated to threads in a thread pool respectively and is emptied, and root nodes are put at the bottoms of the lock-free double-end queues of the threads in a polling mode; if the lock-free double-end queues of the threads are not empty, the nodes are taken out of the bottoms of the lock-free double-end queues and are executed; if the lock-free double-end queues of the threads are empty, the nodes are stolen from the tops of the lock-free double-end queues of other threads, the stolen nodes are pressed into the bottoms of the lock-free double-end queues of the threads and are taken out for execution; after execution of all node tasks is completed, the in-degrees of the nodes in the task dependence graph are recovered to original values, and blockage to a main thread is ended. The task scheduling method and the system are oriented to large task-level parallel application programs, and the performance of traditional task-level parallel application programs can be effectively improved.

Description

technical field [0001] The invention belongs to the technical field of computer parallel programming, and more specifically relates to a task scheduling method and system based on task stealing. Background technique [0002] The transistor size and power consumption, which are constantly approaching the physical limit, severely limit the development of single-core processors in computers. It is no longer possible to improve program performance just by waiting for chip manufacturers to launch new processors as before. To further improve application performance, only rely on the method of integrating multiple cores into a single CPU and parallelizing the application. The era of single-core serial is over, and programmers are beginning to move towards the era of multi-core parallelism. [0003] Traditional parallel programming models (including MPI and earlier versions of OpenMP) are intended only for expert-level, experienced programmers, or applications that can only adapt t...

AI Technical Summary

Problems solved by technology

When the basic block where the dependent task is located begins to execute, the execution of a large number of irrelevant tasks in the basic block where the dependent task is located will cause the required data in the cache to be swapped out, resulting in poor program locality

Images