The Design And Realization Of Light Graphics Engine Based On Domestic GPU

The early computers did not include the GPU.Many of the processing functions and related calculations of the graphics were undertaken by the CPU.As the software requirements and computer graphics processing operations increased,the graphics operations supported by the CPU alone no longer satisfied the computer.The birth of GPU has broken this situation,GPU can enhance the graphics rendering effect,by directly programming the GPU to achieve rendering technology.However,programming the GPU directly on the computer is difficult in scene design and cannot be combined with modeling tools,lacking some key technologies.Therefore,research on graphics engine development has become necessary.Currently,the graphics engines on the market are based on high-performance GPU development,and the performance on domestic GPU is not good.A graphics engine that can target domestic GPU is necessary.The light graphics engine supports Windows and Linux systems,uses OpenGL as the underlying graphics library,provides basic mathematical support,implements graphics algorithms,and builds a programmable rendering pipeline that transforms the 3D model into screen display pixels through a series of transformations.According to requirements of the light graphics engine,the system architecture of the light graphics engine is designed.Design and implement data management module,file module,space scene management module,rendering module and GUI management module.The file module focuses on model import and xml format file processing.The space scene management module focuses on the management of scene nodes through octree technology.The rendering module implements the rendering pipeline,including vertex buffers,textures,materials,light,LOD terrain.Finally,through a series of basic functional benchmarks,the light graphics engine is optimized for a specific domestic GPU.The light graphics engine can render satisfactory results on a domestic GPU,with complete functionality and scalability for subsequent functions.Finally,I look forward to the future improvement work.