Method for handling of out-of-order opaque and alpha ray/primitive intersections
A hardware-based traversal coprocessor provides acceleration of tree traversal operations searching for intersections between primitives represented in a tree data structure and a ray. The primitives may include opaque and alpha triangles used in generating a virtual scene. The hardware-based traversal coprocessor is configured to determine primitives intersected by the ray, and return intersection information to a streaming multiprocessor for further processing. The hardware-based traversal coprocessor is configured to provide a deterministic result of intersected triangles regardless of the order that the memory subsystem returns triangle range blocks for processing, while opportunistically eliminating alpha intersections that lie further along the length of the ray than closer opaque intersections.
1. A system including:
memory configured to store data that defines vertices of a plurality of primitives; and
hardware circuitry comprising a result queue and operatively coupled to the memory, wherein the hardware circuitry is configured to:
determine, in an order based on the presentation of the primitives to the hardware circuitry, primitives which are intersected by a ray;
store intersection results for a plurality of primitives the ray is determined to intersect in the result queue while omitting, from the result queue, intersection results for one or more primitives which have been determined to be intersected by the ray; and
report the intersection results stored in the result queue.
2. The system of claim 1 , wherein the omitted intersection results from the result queue are omitted to provide a deterministic visualization result regardless of an order the plurality of primitives are presented to the hardware circuitry.
3. The system of claim 1 , wherein omitting the intersection results from the result queue includes replacing intersection result stored in the result queue for a primitive intersection with another intersection result for a primitive intersection.
4. The system of claim 1 , wherein the primitives which are intersected by the ray are determined in an order the primitives in a primitive range are received from the memory that is different from an order the primitives are stored in the memory.
5. The system of claim 1 , wherein the data that defines vertices of the plurality of primitives is stored in cache line sized blocks, each of the cache line sized blocks including a different set of primitives associated with at least one node of an acceleration data structure, and the cache line sized blocks are returned from the memory in an order that is different from the order the cache line sized blocks are stored in the memory.
6. The system of claim 1 , wherein each of the one or more of the primitives the ray is determined to intersect is an alpha primitive.
7. The system of claim 1 , wherein the omitted intersection results for the one or more primitives from the result queue include primitives which are provably capable of being omitted without a functional impact on visualizing a virtual scene.
8. The system of claim 7 , wherein the primitives intersected by the ray which are provably capable of being omitted without a functional impact on visualizing the virtual scene are parametrically farther away from an origin of the ray than another primitive determined to be intersected by the ray.
9. The system of claim 1 , wherein the omitted intersection results for the one or more primitives from the result queue include at least one primitive that is parametrically closer to an origin of the ray than another primitive determined to be intersected by the ray.
10. A method implemented by a hardware circuitry including a result queue, the method comprising:
receiving, from a memory external to the hardware circuitry, data that defines vertices of a plurality of primitives;
determining, in an order based on the presentation of the primitives to the hardware circuitry, primitives which are intersected by a ray;
storing intersection results for a plurality of primitives the ray is determined to intersect in the result queue while omitting, from the result queue, intersection results for one or more primitives which have been determined to be intersected by the ray; and
reporting, to a processor, the intersection results stored in the result queue.
11. The method of claim 10 , wherein the omitted intersection results from the result queue are omitted to provide a deterministic visualization result regardless of an order the plurality of primitives are returned from the memory.
12. The method of claim 10 , wherein omitting the intersection results from the result queue includes replacing intersection result stored in the result queue for a primitive intersection with another intersection result for a primitive intersection.
13. The method of claim 10 , wherein the primitives which are intersected by the ray are determined in an order the primitives are received from the memory that is different from an order the primitives are stored in the memory.
14. The method of claim 10 , wherein each of the one or more of the primitives the ray is determined to intersect is an alpha primitive.
15. The method of claim 10 , wherein the omitted intersection results for the one or more primitives from the result queue include primitives which are provably capable of being omitted without a functional impact on visualizing a virtual scene.
16. The method of claim 15 , wherein the primitives intersected by the ray which are provably capable of being omitted without a functional impact on visualizing the virtual scene are parametrically farther away from an origin of the ray than another primitive determined to be intersected by the ray.
17. The method of claim 10 , wherein the data that defines vertices of the plurality of primitives is stored in cache line sized blocks, each of the cache line sized blocks including a different set of primitives associated with at least one node of an acceleration data structure, and the cache line sized blocks are returned from the memory in an order that is different from the order the cache line sized blocks are stored in the memory.
18. The method of claim 10 , wherein the omitted intersection results for the one or more primitives from the result queue include at least one primitive that is parametrically closer to an origin of the ray than another primitive determined to be intersected by the ray and reported.
19. The method of claim 10 , wherein the plurality of primitives include opaque primitives and alpha primitives, and reporting the primitives the ray is determined to intersect includes:
identifying, to the processor, an opaque primitive intersected by the ray which is parametrically closest intersected opaque primitive to an origin of the ray; and
identifying a plurality of alpha primitives intersected by the ray and spatially located between the origin of the ray and the parametrically closest opaque primitive to the processor in an order the intersected alpha primitives are stored in the memory.
20. The method of claim 10 , wherein the plurality of primitives include opaque primitives and alpha primitives, and the result queue includes a single location for storing an opaque primitive determined to be intersected by the ray and a plurality of locations for storing a plurality of alpha primitive determined to be intersected by the ray.