Compacting results vectors between stages of graphics processing

摘要

Ray tracing, and more generally, graphics operations taking place in a 3-D scene, involve a plurality of constituent graphics operations. Responsibility for executing these operations can be distributed among different sets of computation units. The sets of computation units each can execute a set of instructions on a parallelized set of input data elements and produce results. These results can be that the data elements can be categorized into different subsets, where each subset requires different processing as a next step. The data elements of these different subsets can be coalesced so that they are contiguous in a results set. The results set can be used to schedule additional computation, and if there are empty locations of a scheduling vector (after accounting for the members of a given subset), then those empty locations can be filled with other data elements that require the same further processing as that subset.

主权项

1. A method of concurrently performing graphics computations, comprising:
determining, by a controller, using status information stored in a memory, a subset of rays from a set of rays; submitting, from the controller, the rays of the subset concurrently to a plurality of computation units, each computation unit comprising one or more processor cores, and configured to process the rays according to determined common program instructions; performing the common program instructions on the submitted rays, by the plurality of computation units, to produce results for the of rays, wherein the results comprise that some of the rays can be processed concurrently using further common program instructions, but remaining rays do not require such processing; and coalescing data for the rays that can be processed concurrently using the further common program instructions and updating the status information using the coalesced data for the rays, wherein the acts of determining and submitting are performed asynchronously from the performance of program instructions by the plurality of computation units.