A management method, device and computer storage medium related to video memory

Abstract

The embodiment of the invention discloses a video memory management method and device and a computer storage medium. The video memory management method can comprise the following steps that: a graphics processing unit GPU driver allocates a physical space to a created buffer object in a video memory based on a set allocation logic; the GPU drive program performs memory mapping MMAP operation on the created buffer area object so as to establish a mapping relationship between a physical address of the created buffer area object in the video memory and a virtual address in a system memory; the GPU drive program writes the mapping relation into a mapping table of a memory management unit MMU of a GPU; and the graphic application program writes data to be buffered into a buffer physical space allocated by the GPU drive program in the video memory through the created buffer object through the GPU drive program.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of video memory management of graphics processors, and in particular, to a video memory management method, device, and computer storage medium. Background technique [0002] Computing devices typically have at least one programmable processing element, often referred to as a central processing unit (CPU); in addition, computing devices typically have other programmable processors for various types of specialized processing, Such as a graphics processing unit (GPU, Graphic Processing Unit) for graphics processing operations. A GPU typically includes multiple cores or processing elements designed to execute instructions and data provided by the CPU on parallel data streams based on the same instructions, thereby enabling the GPU to be more efficient than the CPU in algorithms that process large blocks of data in parallel. Generally speaking, the CPU acts as the system master of the compu...

AI Technical Summary

Problems solved by technology

The data of the buffer object is usually large. If the GPU copies the buffer object from the system memory to the GPU memory before each parallel operation, the copying process will cause a large performance loss.

Images