VGPU: a real time GPU emulator

A method for caching resources associated with a graphics application during emulation of a target graphics processing unit (GPU) on a host computing system, the method comprising:

• establishing a content based resource cache in memory of the host computing system;
• establishing an address based resource cache for the target GPU in the memory of the host computing system, the address based resource cache including a frame buffer;
• upon execution of the graphics application by a CPU of the host computing system, writing a graphics resource generated by a host GPU of the host computing system to the content based cache in response to a CPU instruction issued by the CPU of the host computing system, the writing the graphics resource generated by the host GPU being a part of a multi-pass rendering process;
• and
• upon rendering of graphics by the host GPU of the host computing system based at least in part on execution of the graphics application, reading a graphics resource from the address based resource cache that includes the frame buffer in response to a GPU instruction issued by the host GPU of the host computing system.

A system for emulating execution of a graphics application on a target graphics processing unit (GPU), the system comprising:

• one or more host central processing units (CPUs) to execute control logics;
• memory;
• a host graphics processing unit (GPU);
• control logic that converts commands generated by execution of the graphics application to intermediate commands that instruct the host GPU to emulate a target GPU during rendering of graphics of the graphics application, the intermediate command providing calls to a shader management control logic and a resource management control logic of the target GPU;
• control logic that converts a shader associated with the graphics application to a converted shader compatible with the host GPU under direction of the shader management control logic;
• control logic that caches one or more converted shaders in a shader cache established in the memory and accessible to the host GPU under direction of the resource management control logic;
• control logic that converts a resource associated with the graphics application to a converted resource compatible with the host GPU under direction of the resource management control logic;
• and
• control logic that caches one or more converted resources in a resource cache established in the memory and accessible to the host GPU.