Heterogeneous platform based multi-parallel error detection system framework

Abstract

The invention belongs to the technical field of parallel processors and particularly relates to a heterogeneous platform based multi-parallel error detection system framework. According to the framework, powerful parallel computing capability and programmability of a universal graphic processor in a heterogeneous platform are mainly utilized to simultaneously detect mainstream multiple parallel errors, including data competition, atomic contravention and sequential contravention. In the aspect of design complexity, only relatively smooth hardware complexity is required, the logic of an on-chip key path (such as high-speed cache or cache consistency) does not need to be changed, only a memory access collection module and a memory access preprocessing module are added for collecting memory access instructions possibly causing the parallel errors and providing related information of error detection separately, and an error detection algorithm realizes high parallelism by utilizing the universal graphic processor. The hardware framework provided by the invention can discover the parallel errors in the program running process and is very low in runtime overhead.

Description

technical field [0001] The invention belongs to the technical field of parallel processors, and in particular relates to a variety of parallel error detection architectures based on the processing capabilities of heterogeneous platforms (combined architectures of general-purpose processors and graphics processors). Background technique [0002] With the development of computer technology, multi-core processors are becoming more and more common. In daily life, desktops, laptops, tablets, and even mobile phones are already equipped with multi-core processors. The increase in the number of processor cores on a single chip brings potential computing power. In order to take full advantage of computing resources on multi-core devices, parallel programs are commonly used. However, writing correct parallel programs is not easy due to parallel program bugs. On the one hand, programmers are accustomed to the serialized way of thinking, which leads to errors when writing parallel pr...

AI Technical Summary

Problems solved by technology

Also, these computing units are often idle while programmers are debugging

On the other hand, although parallel errors have different manifestations, they are essentially caused by unreasonable use of shared resources.

Therefore, the focus of the program state required for detection is on the memory access records (especially the shared memory access records), which makes the detection process somewhat similar in terms of front-end collection

Images