Parallel texture caching

The creation of high-quality images requires new functionality and higher performance in real-time graphics architectures. In terms of functionality, texture mapping has become an integral component of graphics systems, and in terms of performance, parallel techniques are used at all stages of the graphics pipeline. In rasterization, texture caching has become prevalent for reducing texture bandwidth requirements. However, parallel rasterization architectures divide work across multiple functional units, thus potentially decreasing the locality of texture references. For such architectures to scale well, it is necessary to develop efficient parallel texture caching subsystems. We quantify the effects of parallel rasterization on texture locality for a number of rasterization architectures, representing both current commercial products and proposed future architectures. A cycle-accurate simulation of the rasterization system demonstrates the parallel speedup obtained by these systems and quantities inefficiencies due to redundant work, inherent parallel load imbalance, insufftcient memory bandwidth, and resource contention. We find that parallel texture caching works well, and is general enough to work with a wide variety of rasterization architectures.