Abstract: In a graphics pipeline of a graphics processor, a method for a unified primitive description for rasterization. The method includes receiving a group of primitives from a graphics application, wherein the group includes different types of primitives and the types of primitives include line primitives, point primitives and triangle primitives. For each of the types of primitives, the method includes generating a corresponding parallelogram, wherein the parallelogram has four sides disposed along an x-axis and a y-axis, and computing an inside y-axis mid point and an outside y-axis mid point based on the four sides. The parallelogram is controlled to represent to each of the primitive types respectively by adjusting a location of the inside y-axis mid point or the outside y-axis mid point.

Abstract: Floating-point texture filtering units leverage existing fixed-point filter circuits. Groups of floating-point texture values are converted to products of a fixed-point mantissa and a scaling factor that is the same for each texture value in the group. The fixed-point mantissas are filtered using a fixed-point filter circuit, and the filtered mantissa is combined with the scaling factor to determine a floating-point filtered value. Multiple floating-point filter results may be combined in a floating-point accumulator circuit. The same fixed-point filter circuit may also be used to filter fixed-point texture data by providing fixed-point input path that bypasses the format conversion and a fixed-point accumulator.

Abstract: An apparatus and method for fairly arbitrating between clients with varying workloads. The clients are configured in a pipeline for processing graphics data. An arbitration unit selects requests from each of the clients to access a shared resource. Each client provides a signal to the arbitration unit for each clock cycle. The signal indicates whether the client is waiting for a response from the arbitration unit and whether the client is not blocked from outputting processed data to a downstream client. The signals from each client are integrated over several clock cycles to determine a servicing priority for each client. Arbitrating based on the servicing priorities improves performance of the pipeline by ensuring that each client is allocated access to the shared resource based on the aggregate processing load distribution.

Abstract: Floating-point texture filtering units leverage existing fixed-point filter circuits. Groups of floating-point texture values are converted to products of a fixed-point mantissa and a scaling factor that is the same for each texture value in the group. The fixed-point mantissas are filtered using a fixed-point filter circuit, and the filtered mantissa is combined with the scaling factor to determine a floating-point filtered value. Multiple floating-point filter results may be combined in a floating-point accumulator circuit. The same fixed-point filter circuit may also be used to filter fixed-point texture data by providing fixed-point input path that bypasses the format conversion and a fixed-point accumulator.

Abstract: Circuits, methods, and apparatus provide for the storage of texture descriptors in a graphics memory. Since the texture descriptors are stored in a graphics memory, they do not need to be stored in the graphics processor itself, thus reducing graphics processor circuitry and cost. This allows more textures to be associated with each graphics primitive, thereby improving image realism.

Abstract: A scalable pipelined pixel shader that processes packets of data and preserves the format of each packet at each processing stage. Each packet is an ordered array of data values, at least one of which is an instruction pointer. Each member of the ordered array can be indicative of any type of data. As a packet progresses through the pixel shader during processing, each member of the ordered array can be replaced by a sequence of data values indicative of different types of data (e.g., an address of a texel, a texel, or a partially or fully processed color value). Information required for the pixel shader to process each packet is contained in the packet, and thus the pixel shader is scalable in the sense that it can be implemented in modular fashion to include any number of identical pipelined processing stages and can execute the same program regardless of the number of stages.