Method and device for running high-performance GPU application and computer storage medium

Abstract

The embodiment of the invention belongs to the technical field of GPUs, and discloses a method and device for operating a high-performance GPU application and a computer storage medium. The method cancomprise the steps of dividing at least one macro block for a to-be-rendered frame based on rendering complexity of a high-end GPU application; establishing a mapping relationship between the macro block and each GPU in the GPU array; dividing a rendering task of the to-be-rendered frame into rendering sub-tasks corresponding to the macro blocks; and distributing the rendering subtasks corresponding to the macro blocks to the GPUs mapped by the macro blocks according to the mapping relationship so as to execute the rendering subtasks. Through the method, the high-end GPU application can be deployed to the low-end GPU for execution, so that the effect of running the high-end GPU application through the low-end GPU is achieved.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of Graphics Processing Unit (GPU), and in particular, to a method, device, and computer storage medium for running high-performance GPU applications. Background technique [0002] With the development of GPU general-purpose technology, the application scenarios that GPUs can apply to are becoming more and more abundant. Generally, GPU applications can be divided into low-end GPU applications suitable for low-end GPUs, and mid-range GPU applications according to the performance of the GPUs they are applicable to. , High-end GPU running mid-range and high-end GPU applications. [0003] Currently, GPUs are divided into low-end GPUs (GPUs below 192G FLOPS) and mid-range GPUs (GPUs from 192G to 1024GFLOPS) according to the FLOPS (Floating-point Operation PerSecond) performance index of GPUs. As well as high-end GPUs (GPUs higher than 1024G FLOPS), it is understandable that high-end, medium-...

AI Technical Summary

Problems solved by technology

[0005] In the current various graphics applications, if you want to implement or run high-end GPU applications, you need to use high-performance high-end GPUs to meet the corresponding application requirements. Therefore, low-end GPUs cannot meet the performance requirements of high-end GPU applications.

However, in some data application platforms, the GPUs they carry are not high-end GPUs, but mid-to-low-end GPUs as the core. In this case, it is difficult for high-end GPU applications to run on these application platforms. The application scenarios of the data application platform are limited to the low-end GPU application field

Inability to implement or run high-end GPU applications

Images