Stream compaction for rasterization

摘要

A single instruction multiple data (SIMD) processor with a given width may operate on registers of the same width completely filled with fragments. A parallel set of registers are loaded and tested. The fragments that fail are eliminated and the register set is refilled from the parallel set.

主权项

1. A method comprising:
assembling a parallel set of m fragments in n registers; testing individual fragments of said m fragments; eliminating fragments of said m fragments based on said testing to end up with p fragments where p is less than m and creating empty spaces in said n registers; and merging the p fragments after eliminating some, but not all, of the m fragments to entirely fill the n registers with fragments that were not eliminated by accessing additional fragments from another set of fragments to fill register vacancies not filled by the p fragments.