System for fast intersection of secondary rays with geometric objects in ray tracing
The present disclosure describes a fast intersection between secondary rays and geometric objects for a global illumination ray tracing. The Acceleration Structures of prior art are replaced by a new and novel device—a Dynamically Aligned Structure (DAS), a means for carrying out the intersection between secondary rays and scene geometry in large groups of rays, gaining high speed and lowering computational complexity. Its reduced power consumption is suitable to consumer level computing devices.
1. A computer based system for fast intersections between secondary rays and objects in ray tracing of a three dimensional scene, utilizing graphics pipeline, comprising:
at least one graphics processor with memory,
at least one general purpose processors with memory,
geometric data base of the scene,
render target memory;
wherein in runtime,
(I) graphics processor shoots first projection of rays at a cluster of existing hit points for acquiring depth map of the existing hit points; and
(II) graphics processor shoots second projection of rays through said cluster of existing hit points, utilizing the acquired depth map, for creating secondary rays; and
(III) graphics processor renders with the secondary rays the three dimensional scene, or part thereof, storing the intersections of secondary rays with geometrical objects to a render target memory; and
(IV) existing hit points are matched with points of intersections of related secondary rays and the render target; and
(V) the identity of objects represented at the points of intersection of secondary rays with the render target are located.
2. The system of claim 1 , wherein the graphics processor is a discrete GPU having a hardware graphics pipeline.
3. The system of claim 1 , wherein the graphics processor is an integrated GPU having a hardware graphics pipeline.
4. The system of claim 1 , wherein the graphics processor is a hybrid graphics system of at least one integrated GPU and at least one discrete GPU, all having a hardware graphics pipeline.
5. The system of claim 1 , wherein the general purpose processor is a multicore CPU having multiple processing cores.
6. The system of claim 1 , wherein the existing hit points are primary or secondary hit points.
7. The system of claim 1 , wherein the first projection renders existing hit points only, disregarding the geometric data of the three dimensional scene.
8. The system of claim 1 , wherein the acquiring depth map of the existing hit points is done by disregarding the geometric data of the scene.
9. The system of claim 1 , wherein the acquiring depth map is implemented by Z-buffering mechanism.
10. The system of claim 1 , wherein the rays of projection that encounter existing hit points are divided into two segments.
11. The system of claim 1 , wherein the division point between two segments is where a ray encounters an existing hit point.
12. The system of claim 1 , wherein secondary ray comprises the second segment.
13. The system of claim 1 , wherein all geometric objects within the range of the first segment are discarded.
14. The system of claim 1 , wherein a ray of the first projection that does not encounter an existing hit point does not become a secondary ray.
15. The system of claim 1 , wherein secondary rays render geometric data of the three dimensional scene.
16. The system of claim 1 , wherein all the intersection points between the secondary rays and the geometric data of the scene are stored in the render target memory.
17. The system of claim 1 , wherein all the intersection points between the secondary rays and the geometric data are located in furtherance of the existing hit points.
18. The system of claim 1 , wherein the tasks of matching existing hit points with related intersections and identifying the data objects at the intersections, can be performed either by graphics processing unit or by general purpose processing units of the system.
19. The system of claim 1 , wherein matching an existing hit point with a corresponding intersection point in a render target is done by matching coordinates.
20. The system of claim 1 , wherein locating the identity of objects at the points of intersection is done by inspecting the render target.