Abstract: GPU fragment programs can be used to render images in a computer system. These fragment programs are generated from render trees, which specify one or more filters or functions to be applied to an input image to render an output image. It is not uncommon for successive frames to require application of substantially the same filters. Therefore, rather than regenerate and recompile new fragment programs for successive corresponding render trees, the render trees are substantially uniquely identified and cached. Thus, when a render tree is received, it can be identified, and this identifier (such as a hash) can be used to determine whether a corresponding fragment program has already been generated, compiled and cached. If so, the corresponding cached fragment program is retrieved and executed. If not, a fragment program for the newly received render tree is generated and cached.

Abstract: A device receives input that includes definitions of components of a computational pipeline, where the components include one or more buffers, one or more kernels, and one or more stages within a control graph. The device generates, based on the input, kernel signatures for a graphics processor, where the kernel signatures compile into an executable streaming program for the computational pipeline. The device also generates, based on the input, host-side runtime code to execute the streaming program.

Abstract: GPU fragment programs can be used to render images in a computer system. These fragment programs are generated from render trees, which specify one or more filters or functions to be applied to an input image to render an output image. It is not uncommon for successive frames to require application of substantially the same filters. Therefore, rather than regenerate and recompile new fragment programs for successive corresponding render trees, the render trees are substantially uniquely identified and cached. Thus, when a render tree is received, it can be identified, and this identifier (such as a hash) can be used to determine whether a corresponding fragment program has already been generated, compiled and cached. If so, the corresponding cached fragment program is retrieved and executed. If not, a fragment program for the newly received render tree is generated and cached.

Abstract: A device receives input that includes definitions of components of a computational pipeline, where the components include one or more buffers, one or more kernels, and one or more stages within a control graph. The device generates, based on the input, kernel signatures for a graphics processor, where the kernel signatures compile into an executable streaming program for the computational pipeline. The device also generates, based on the input, host-side runtime code to execute the streaming program.

Abstract: A development application leverages the programmability of shader execution units in the graphics processing subsystem to make graphics processing subsystem state data accessible to applications executed outside the graphics processing subsystem. The development application modifies shaders to include state output instructions adapted to direct a shader execution unit to copy graphics processing subsystem state data to a location in the computer system that is accessible to applications executed outside of the graphics processing subsystem. Following the execution of the state output instructions, the shader execution unit can be halted or can continue executing the shader. The development application can modify the shader to include state restoration instructions adapted to direct the shader execution unit to set state data of the graphics processing subsystem to previous or new values.

Abstract: The present invention is directed to a method for processing, in a computer system, a computer program having a plurality of operations. The method features calling a dynamic programming routine to generate a schedule for executing a subgroup of the plurality of operations by modeling operations of a computational processor associated with the computer system to minimize a computational cost of placing the computer system in a final machine state (finMS).

Abstract: The disclosed embodiments relate to a rendering cluster that renders an image of a scene object. The rendering cluster may comprise an illumination node that produces illumination output based on lighting properties of the scene object and a material node that produces material output based on material properties of the scene object. The illumination output is combined with the material output to form the image.

Abstract: The disclosed embodiments relate to a rendering cluster that renders an image of a scene object. The rendering cluster may comprise an illumination node that produces illumination output based on lighting properties of the scene object and a material node that produces material output based on material properties of the scene object. The illumination output is combined with the material output to form the image.