Ray tracing is a technique used for generating highly realistic computer graphics images. In this paper, we explore the design of a simple but extremely parallel, multi-threaded, multi-core processor architecture that performs real-time ray tracing. Our architecture, called TRaX for Threaded Ray eXecution, consists of a set of thread states that include commonly used functional units for each thread and share large functional units through a programmable interconnect to maximize utilization. The memory system takes advantage of the application's read-only access to the scene database and write-only access to the frame buffer output to provide efficient data delivery with a relatively simple structure. Preliminary results indicate that a multi-core version of the architecture running at a modest speed of 500 MHz already provides real-time ray traced images for scenes of a complexity found in video games. We also explore the architectural impact of a ray tracer that uses procedural (computed) textures rather than image-based (look-up) textures to trade computation for reduced memory bandwidth.
展开▼
