Application testing method and device, computer equipment and storage medium

Abstract

The embodiment of the invention discloses an application testing method and device, computer equipment and a storage medium. The method comprises the steps of acquiring a target test case of a target application suitable for a mainstream application platform; testing an application function support result of the target application on a heterogeneous application platform according to the target test case; and under the condition of determining that the application function support result passes the application function test, testing the application performance of the target application in the heterogeneous application platform. According to the technical scheme provided by the embodiment of the invention, functional defects and performance difference problems of the heterogeneous application platform can be quickly checked, so that the test efficiency and the test quality of the application are improved, and the running effect of the application on the heterogeneous application platform is ensured.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of testing, and in particular, to an application testing method, device, computer equipment, and storage medium. Background technique [0002] With the accumulation of massive data and the rapid development of hardware equipment, people's demand for computing power is also increasing. Pure CPU (Central Processing Unit / Processor, central processing unit) resources can no longer meet people's computing needs. Therefore, Various types of heterogeneous accelerator cards such as GPU (Graphics Processing Unit, Graphics Processing Unit) accelerator cards are applied. Configuring a heterogeneous accelerator card in a mainstream platform can constitute a heterogeneous platform, which can improve the processing efficiency of the computing platform. [0003] At present, most heterogeneous platforms build corresponding application support based on the basic computing libraries (also called dep...

AI Technical Summary

Problems solved by technology

[0004] However, because the ecology of heterogeneous platforms is not yet perfect, and the dependent libraries of some open source applications are developed based on the mainstream platform environment, for example, some third-party frameworks derived from the deep learning framework PyTorch, such as apex and DeepSpeed, lack functionality on heterogeneous platforms. support

Therefore, if some open source codes applied in the mainstream environment are migrated to the heterogeneous environment, there will be problems such as abnormal functions or performance, which will cause the open source application to fail to run successfully.

For example, the current support for deep learning application examples based on heterogeneous platform ecology mainly has the following problems: (1) Some deep learning applications cannot be directly run on heterogeneous platforms, and some dependent libraries involved cannot directly implement functionality Support, if there is a dependency problem, it is necessary to manually check and determine the underlying dependency problem step by step; (2), although the application program of deep learning can run successfully on the heterogeneous platform, there is still a certain difference in performance from the mainstream platform, resulting in poor running effect

Images