GPU acceleration method and GPU acceleration device

Abstract

The invention provides a GPU (Graphics Processing Unit) acceleration method and a GPU acceleration device. The method comprises the following steps: calling a GPU driver in the kernel through a continuous physical memory allocation function in an EXA driver to make a GPU allocate a contiguous physical address buffer area for the EXA driver; filling out the contiguous physical address buffer area with to-be-accelerated data; and submitting the physical addresses in the contiguous physical address buffer area and corresponding acceleration instructions to the GPU driver to make the GPU accelerate data. According to the technical scheme provided by the invention, one buffer area allocation operation and a large number of memcpy operations are saved, the time for submitting acceleration instructions to the GPU is shortened effectively, and the efficiency of GPU acceleration is improved.

Description

technical field [0001] The present invention relates to hardware acceleration technology, in particular to a GPU acceleration method and device. Background technique [0002] Graphics acceleration structure (Ex-kaa aXeleration Architecture, EXA) driver is the interface between the graphics card in the Linux graphics system and the X window system. It is the interface layer for the Linux graphics system to achieve GPU hardware acceleration. It works on the graphics interface server Xserver in the X window system Underneath and above the graphics card's kernel graphics processing unit (Graphics Processing Unit, GPU) driver, the X window system realizes the 2D graphics card acceleration of all Linux graphics systems through the EXA driver. [0003] In the current graphics acceleration solution, before the Xserver calls the GPU through the EXA driver to perform an acceleration operation, it first allocates a discontinuous memory buffer for the EXA driver in the user space to sto...

AI Technical Summary

Problems solved by technology

[0004] The existing acceleration scheme requires two buffer allocation operations and a large number of memcpy operations before calling the GPU for acceleration operations, which delays the execution of GPU acceleration operations and makes GPU acceleration inefficient.

Images