Computer hardware architecture and data structures for triangle binning to support incoherent ray traversal

摘要

A new hardware architecture defines an indexing and encoding method for accelerating incoherent ray traversal. Accelerating multiple ray traversal may be accomplished by organizing the rays for minimal movement of data, hiding latency due to external memory access, and performing adaptive binning Rays may be binned into coarse grain and fine grain spatial bins, independent of direction.

主权项

1. Digital circuitry for binning objects in a scene, the digital circuitry comprising:
inputs to receive data representative of a plurality of partitioning planes that define a grid in the scene and cells comprising the grid and data representative of a plurality of input objects; first digital circuits configured to:
bin one or more of the input objects into one or more cells (first dirty cells) of the grid based on vertices of the input objects; andgenerate a first bitmap representative of the grid and comprising a plurality of bits which correspond to the cells of the grid, a bit being set when the corresponding cell is one of the first dirty cells; second digital circuits configured to:
generate rays that originate on edges of the input objects and are coplanar with the input objects;identify second dirty cells that contain at least a portion of the input objects using intersections between the rays that originate on the edges of the input objects and are coplanar with the input objects and the partitioning planes that define the cells comprising the grid; andgenerate a second bitmap representative of the grid and comprising a plurality of bits which correspond to the cells of the grid, a bit being set when the corresponding cell is one of the second dirty cells; and a combining circuit operable to combine the first and second bitmaps to generate an object grid block bitmap representative of dirty cells in the grid.