Intersection test for ray tracing

By accelerating the ray intersection process in a layered and serial mode, the intersection test of the ray tracing system is optimized, solving the rendering efficiency problem on resource-constrained devices and achieving more efficient ray tracing rendering.

CN114331802BActive Publication Date: 2026-11-17IMAGINATION TECH LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202111130835.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-09-30
Filing Date
2021-09-26
Publication Date
2026-11-17
Estimated Expiration
2041-09-26

AI Technical Summary

Technical Problem

Existing ray tracing systems struggle to efficiently perform intersection tests between rays and scene geometry when rendering images in real time, especially on resource-constrained mobile devices, resulting in excessive processing power and storage requirements.

Method used

By employing a hierarchical acceleration structure and a serial mode ray intersection process, a limited number of new ray requests are generated, intersection information is stored, and ray traversal is optimized, thereby reducing storage and computational requirements.

Benefits of technology

It improves the rendering efficiency of ray tracing systems on resource-constrained devices, reduces the risk of memory overflow, and decreases processing time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114331802B_ABST
    Figure CN114331802B_ABST
Patent Text Reader

Abstract

This invention discloses an intersection test for ray tracing. A system and method for performing ray intersection tests in a ray tracing system are provided. The ray tracing system uses a hierarchical acceleration structure comprising multiple nodes, each node identifying one or more elements that can be intersected by rays. The system utilizes a serial mode ray intersection process, wherein a finite number of new ray requests are generated when a ray intersects with an enclosing volume.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to intersection testing for ray tracing. Background Technology

[0002] Ray tracing systems simulate how light rays interact with a scene. For example, ray tracing techniques can be used in graphics rendering systems configured to generate images from a 3D scene description. These images can be photorealistic or achieve other goals. For instance, animated films can be created using 3D rendering techniques. A 3D scene description typically includes data defining the geometry within the scene. This geometric data is usually defined by primitives, typically triangles, but can sometimes be other shapes such as polygons, lines, or points.

[0003] Ray tracing mimics the natural interaction of light with objects in a scene, and complex rendering features can be naturally generated from ray-traced 3D scenes. Ray tracing can be relatively easily parallelized at the pixel-by-pixel level because pixels are typically independent of each other. However, in cases such as ambient occlusion, reflections, and caustics, the distribution and varying positions and directions of light rays in a 3D scene make it difficult to pipeline the processing involved in ray tracing. Ray tracing allows for the rendering of realistic images, but it typically requires high levels of processing power and large working memory, making it potentially difficult to implement for real-time rendering (e.g., for gaming applications), especially on devices with strict limitations on silicon area, cost, and power consumption, such as mobile devices (e.g., smartphones, tablets, laptops, etc.).

[0004] At a very broad level, ray tracing involves: (i) intersection testing to identify intersections between rays and geometry (e.g., primitives) in the scene, and (ii) shading, including performing certain processes (e.g., by executing a shader procedure) in response to identifying the intersections to determine how the intersections contribute to the image being rendered. The execution of the shader procedure may cause additional rays to be emitted into the scene. These additional rays may be referred to as “secondary rays.”

[0005] Identifying intersections between rays and geometry in a scene involves a great deal of processing. In a very simple approach, every ray can be tested for every primitive in the scene, and then the closest intersection can be identified once all intersection hits have been determined. This approach is impractical for scenes that may have millions or billions of primitives, where the number of rays to process could also be millions. Therefore, ray tracing systems typically use acceleration structures that characterize the geometry in a scene in a way that reduces the work required for intersection testing. However, even with existing acceleration structures, it is difficult to perform intersection testing at rates suitable for real-time rendering (e.g., for gaming applications), especially on devices with strict limitations on silicon area, cost, and power consumption, such as mobile devices (e.g., smartphones, tablets, laptops, etc.).

[0006] Modern ray tracing architectures typically use an accelerated architecture based on bounding volume hierarchies—specifically, bounding box hierarchies. Primitives are grouped together into bounding volumes that surround the primitives. These bounding volumes are then grouped together to form larger bounding volumes that surround them. Intersection testing thus becomes easier because if a ray misses a bounding volume, it doesn't need to be tested against any of the bounding volume's child nodes.

[0007] Intersection testing is performed by traversing the hierarchical structure. If a given ray "hits" an enclosing volume (node), the ray needs to be tested against every child node of that enclosing volume (node). This continues downwards through the hierarchical structure until the ray either doesn't hit any of the node's child nodes or hits at least one primitive. Testing a ray against a node requires retrieving from memory (i) a description of the ray (typically defined by ray information including at least the origin and direction) and (ii) a description of the node's geometry (enclosing volume coordinates or primitive coordinates). This information is indicated by a ray request that identifies the ray and the node. Summary of the Invention

[0008] This summary is provided to introduce, in a simplified form, a series of concepts further described below in the detailed description. This summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter.

[0009] A system and method for performing ray intersection testing in a ray tracing system. The ray tracing system uses a hierarchical acceleration structure comprising multiple nodes, each node identifying one or more elements that can be intersected by rays. The system utilizes a serial mode ray intersection process, wherein when a ray intersects with an enclosing volume, a finite number of new ray requests are generated, and intersection information that can be used to identify enclosing volumes for which no ray requests were generated is stored.

[0010] A computer-implemented method is provided for performing an intersection test between one or more rays and elements identified by nodes of a hierarchical acceleration structure, wherein: the rays are defined by ray information; and the nodes identify one or more elements for the intersection test, wherein at least some elements identified by the overall hierarchical acceleration structure are represented by another node of the hierarchical acceleration structure.

[0011] The computer-implemented method includes iteratively performing the following serial mode ray intersection process: obtaining one or more ray requests, each ray request identifying a ray and a node in a hierarchical structure that identifies the element to which the ray of the ray request will be tested for intersection; processing the one or more ray requests and the hierarchical acceleration structure to identify any intersections between the ray of the ray request and the element identified by the node of the ray request for each ray request; and for each ray request: determining whether the ray intersects with more than a first predetermined number of elements represented by other nodes of the hierarchical acceleration structure; and in response to the ray intersecting with more than a first predetermined number of elements represented by other nodes, generating up to a first predetermined number of new ray requests, each new ray request identifying the ray of the ray request and another node representing the element identified as intersecting with the ray of the ray request.

[0012] The first predetermined number can be, for example, 1. Preferably, the first predetermined number is no greater than 12, for example, no greater than 8, for example, no greater than 4. In some examples, the first predetermined number is 1, and the step of generating at most the first predetermined number of new ray requests includes generating a single new ray request.

[0013] The step of generating a maximum of a first predetermined number of new ray requests may include generating a second predetermined number of new ray requests for a ray, each new ray request identifying the ray of the ray request and a corresponding other node in the hierarchy, wherein the other node represents an element identified as intersecting with the ray of the ray request. The second predetermined number is equal to or less than the first predetermined number. In one example, the second predetermined number is 1.

[0014] The computer-implemented method may further include generating and storing intersection information for the ray in response to the ray intersecting with more than a first predetermined number of elements represented by other nodes. The intersection information can be used to identify any other nodes that have not generated a new ray request for them and represent elements identified as intersecting with the ray request.

[0015] In some examples, the intersection information identifies: the node represented by the ray request; and information identifying which other nodes associated with the element identified by the node have generated new ray requests.

[0016] The computer-implemented method may further include: for each ray request, in response to the number of intersections between the ray and elements represented by other nodes being zero: determining whether the ray of the ray request is associated with any stored intersection information; and in response to determining that the ray is associated with stored intersection information, processing the stored intersection information to generate a third predetermined number of new ray requests, each new ray request identifying the ray and a node in the hierarchy for which no ray request has yet been generated and which represents the element with which the ray intersects.

[0017] Intersection information may include one or more sets of one or more data blocks, each set of data blocks representing a different ray, and each data block may identify: a node that identifies more than a first predetermined number of elements with which the corresponding ray intersects, each element being represented by another node; and information that identifies which elements represented by other nodes for the corresponding node have generated ray requests for the corresponding ray.

[0018] In some examples, the steps of processing the stored intersection information include: identifying the most recently stored data block for the ray; processing the most recently stored data block to generate a third predetermined number of new ray requests, each new ray request identifying the ray and nodes in the hierarchy for which no ray request has yet been generated and which represent elements intersecting with the ray; and updating the stored intersection information.

[0019] One or more data blocks in each set may include a linked list of data blocks.

[0020] Optionally, the intersection information is stored in a memory pool shared by all rays.

[0021] The computer-implemented method may further include, for each ray request: in response to the number of intersections between the ray and an element represented by another node being between one and a first predetermined number, including an end value, generating a new ray request for each intersecting element represented by another node, each new ray request identifying the ray of the ray request and the corresponding other node representing the intersecting element.

[0022] Specifically, when the ray requested by the ray has no intersection with any element represented by another node in the hierarchical structure, it is not necessary to generate and / or store intersection information. Therefore, the method may include not generating and / or storing intersection information if the ray requested by the ray fails to intersect with more than a first predetermined number of elements represented by other nodes.

[0023] The step of obtaining one or more ray requests may include, if available, selecting and obtaining at least one of the new ray requests generated in a previous iteration of the serial mode ray intersection process.

[0024] The selection of a new ray request can depend on its age. For example, the oldest ray request may have a higher probability of being selected than the youngest ray request. In one example, the oldest ray request is selected before the youngest. The age of a new ray request refers to the length of time since the ray request was initially generated. This method can allow, for example, the oldest new ray request to be selected first, to reduce the average time spent by rays traversing the hierarchical structure, i.e., undergoing intersection tests.

[0025] The selection of a new ray request can depend (if available) on the age of the stored intersection information associated with the ray identified by the new ray request. In some examples, the number of new ray requests selected and obtained depends on the total size of the stored intersection information.

[0026] The number of new ray requests selected and obtained may be reduced in response to the total size of the stored intersection information exceeding a first predetermined threshold size.

[0027] In other examples, the number of selected and obtained ray requests is further reduced in response to the total size of the stored intersection information exceeding a second predetermined threshold size, which is greater than a first predetermined threshold size.

[0028] This provides a "braking" method for acquiring light requests, such that the number of light requests acquired decreases as free memory space diminishes. This reduces the likelihood of memory overflow. By using a gradual braking method, such as at least two predetermined thresholds, the process of acquiring light requests is less likely to encounter bottlenecks.

[0029] When the total size of the stored intersection information exceeds a first predetermined threshold size, the selection of at least one new ray request may depend on the age of the stored intersection information associated with the ray identified by the ray request.

[0030] In some examples, when the total size of the stored intersection information exceeds a first predetermined threshold size, the selection of at least one new ray request includes at least one ray request that identifies the ray associated with the oldest stored intersection information.

[0031] An intersection testing system is also proposed for performing intersection tests between one or more rays and elements identified by nodes of a hierarchical acceleration structure, wherein: the rays are defined by ray information; and the nodes identify one or more elements for the intersection test, wherein at least some elements identified by the overall hierarchical acceleration structure are represented by another node of the hierarchical acceleration structure.

[0032] The intersection testing system includes: a ray request processor configured to iteratively obtain one or more ray requests, each ray request identifying a ray and a node in a hierarchical structure that identifies the element to which the ray of the ray request will be intersected; and an intersection testing processor configured, for each ray request obtained in the iteration, to process one or more ray requests and the hierarchical acceleration structure to identify any intersections between the ray of the ray request and the element identified by the node of the ray request for each ray request; and for each ray request, to determine whether the ray of the ray request intersects with more than a first predetermined number of elements represented by other nodes of the hierarchical acceleration structure; and in response to the ray intersecting with more than a first predetermined number of elements represented by other nodes, to generate up to a first predetermined number of new ray requests, each new ray request identifying the ray of the ray request and another node representing the element identified as intersecting with the ray of the ray request.

[0033] A ray tracing system or graphics processing system is also proposed, configured to perform any of the methods described herein. A ray tracing system or graphics processing system including the intersection test system described herein is also proposed. The intersection test system, ray tracing system, or graphics processing system can be contained in hardware on an integrated circuit.

[0034] A method for manufacturing an intersection testing system, ray tracing system, or graphics processing system as described herein is also provided using an integrated circuit manufacturing system.

[0035] A method is also provided for manufacturing the intersection test system, ray tracing system, or graphics processing system as briefly described above using an integrated circuit manufacturing system. The method includes: processing a computer-readable description of the intersection test system, ray tracing system, or graphics processing system using a layout processing system to generate a circuit layout description of an integrated circuit including the intersection test system, ray tracing system, or graphics processing system; and manufacturing the intersection test system, ray tracing system, or graphics processing system based on the circuit layout description using an integrated circuit generation system.

[0036] A computer-readable code is also provided, configured to cause any of the methods described herein to be performed when the code is executed. A computer-readable storage medium is also provided on which the computer-readable code described herein is encoded. The storage medium may be a non-transitory computer-readable storage medium. When the computer-readable code is executed on a computer system, it causes the computer system to perform any of the methods described herein.

[0037] An integrated circuit definition dataset is also provided, which, when processed in an integrated circuit manufacturing system, configures the integrated circuit manufacturing system to manufacture an intersection test system, a ray tracing system, or a graphics processing system as described herein.

[0038] A non-transitory computer-readable storage medium is also provided, on which a computer-readable description of an intersection testing system, a ray tracing system, or a graphics processing system as briefly described above is stored, which, when processed in an integrated circuit manufacturing system, causes the integrated circuit manufacturing system to manufacture an integrated circuit including the intersection testing system, the ray tracing system, or the graphics processing system.

[0039] A non-transitory computer-readable storage medium is also provided, on which a computer-readable description of an intersection testing system, a ray tracing system, or a graphics processing system as briefly described above is stored. When the computer-readable description is processed in an integrated circuit manufacturing system, it causes the integrated circuit manufacturing system to: process the computer-readable description of the intersection testing system, the ray tracing system, or the graphics processing system using a layout processing system to generate a circuit layout description of an integrated circuit including the intersection testing system, the ray tracing system, or the graphics processing system; and manufacture the intersection testing system, the ray tracing system, or the graphics processing system according to the circuit layout description using an integrated circuit generation system.

[0040] An integrated circuit manufacturing system is also provided, which is configured to manufacture any intersection testing system, ray tracing system or graphics processing system as described herein.

[0041] The integrated circuit manufacturing system may include: a non-transitory computer-readable storage medium storing a computer-readable description of an intersection testing system, a ray tracing system, or a graphics processing system as briefly described above; a layout processing system configured to process the computer-readable description to generate a circuit layout description of an integrated circuit including the intersection testing system, the ray tracing system, or the graphics processing system; and an integrated circuit generation system configured to manufacture the intersection testing system, the ray tracing system, or the graphics processing system according to the circuit layout description.

[0042] The layout processing system can be configured to determine the location information of the logic components of a circuit derived from an integrated circuit description in order to generate a circuit layout description of the integrated circuit that includes the intersection test system, ray tracing system, or graphics processing system.

[0043] As will be apparent to those skilled in the art, the above features can be appropriately combined, and can be combined with any aspect of the examples described herein. Attached Figure Description

[0044] The example will now be described in detail with reference to the accompanying drawings, in which:

[0045] Figure 1 The scene is shown as divided according to the surrounding volume structure;

[0046] Figure 2 Indicates used for Figure 1 The layered acceleration structure surrounding the volume structure shown;

[0047] Figure 3 This is a simplified block diagram of a system for testing the intersection of rays, based on an example.

[0048] Figure 4 This is a flowchart illustrating one method;

[0049] Figure 5 This is a more detailed block diagram of the system for performing intersection tests on light rays, based on the example.

[0050] Figure 6 The data structure for storing data stored by the storage module of the serial mode light controller is shown;

[0051] Figure 7 A computer system in which a graphics processing system is implemented is shown; and

[0052] Figure 8 An integrated circuit manufacturing system for generating integrated circuits that include a graphics processing system is shown.

[0053] The accompanying drawings illustrate various examples. Those skilled in the art will understand that the element boundaries (e.g., boxes, groups of boxes, or other shapes) shown in the drawings represent one example of a boundary. In some examples, it may be that one element can be designed as multiple elements, or multiple elements can be designed as one element. Where appropriate, common reference numerals are used throughout the drawings to indicate similar features. Detailed Implementation

[0054] The following description is given by way of example to enable those skilled in the art to make and use the invention. The invention is not limited to the embodiments described herein, and various modifications to the disclosed embodiments will be readily apparent to those skilled in the art.

[0055] The embodiments will now be described by way of example only.

[0056] Before explaining the examples of the disclosed systems in detail, it will be helpful to explain the examples of the acceleration architectures used.

[0057] Figure 1Scene 100 is shown, comprising a number of different objects or primitives (shown as circles, squares, and triangles). The scene can conceptually be divided into bounding volumes 101-112, or axis-aligned bounding boxes (AABBs). Thus, each bounding volume represents a portion of the scene. Conceptually, each bounding volume defines or contains one or more elements, where an element can be a bounding volume, a primitive, or another suitable element structure used in a ray tracing architecture (e.g., triggering an instance using another hierarchical structure if intersected by rays). For example, the first set of bounding volumes 101-104 can each define one or more other bounding volumes. The second set of bounding volumes 105-112 can each define one or more primitives in the scene. For example, bounding volume 105 can define the first set of primitives.

[0058] In the example shown, the first enclosing volume 101 defines two other enclosing volumes 104 and 105; the second enclosing volume 102 defines two other enclosing volumes 106 and 107; and the third enclosing volume defines three other enclosing volumes 108, 109, and 110. One of the other enclosing volumes 104 of the first enclosing volume, the “first other enclosing volume,” subsequently defines two additional enclosing volumes 111 and 112.

[0059] It is obvious that if a ray fails to intersect with the enclosing volume, it can be inferred that the ray will not intersect with any element defined or contained by that enclosing volume. Therefore, using an enclosing volume provides a mechanism for reducing the number of intersection tests that need to be performed.

[0060] Scene 100 can be represented by a hierarchical acceleration structure 20, such as Figure 2 As shown, the hierarchical acceleration structure 20 is formed by multiple nodes 201-212 branching from the root node 200 (representing scene 100). Each node 201-212 represents a corresponding bounding volume 101-112. Specifically, each node may include a list (sometimes labeled "test list") or another set identifying any elements contained in the bounding volume represented by the node. In this way, each node identifier may be one or more elements (e.g., other bounding volumes or primitives) intersected by rays. In the illustrated hierarchical acceleration structure 20, such identifiers are shown as branches from nodes.

[0061] Methods for determining the bounding volume of nodes used to construct a hierarchical structure are known in the art and can be performed in a top-down manner (e.g., starting at the root node and working down the hierarchy) or in a bottom-up manner (e.g., starting at the leaf node and working up the hierarchy).

[0062] Rays can be manipulated through a hierarchical acceleration structure, starting from the root node, to perform intersection tests on the rays. Any intersection between a ray and an element identified by a node (e.g., an element branching from a node) is identified. If a ray is determined to intersect with another enclosing volume (which itself is represented by another node in the hierarchical acceleration structure), the ray is tested against the elements of the node represented by that enclosing volume, thus manipulating the ray through the hierarchical structure.

[0063] Methods for determining the intersection point between a ray and an element are well-known and typically involve processing the geometry of the element and data on the ray to identify whether the ray intersects with the element.

[0064] Leaf nodes in a hierarchical structure identify at least one primitive or single instance (examples of specific types of bounding boxes, which will be briefly explained below). Some hierarchical structures may include only leaf nodes identifying instances. Other hierarchical structures may include only leaf nodes identifying primitives.

[0065] The primitives (circles, triangles, and squares) in this example are simple geometric shapes; therefore, each of them can be described using a single primitive. More complex objects can be described by multiple primitives. As is known to those skilled in the art, triangular primitives are common in graphics applications. However, the scope of this disclosure is not limited to triangular primitives. Specifically, in ray tracing, primitives can be "programmatic" in the sense that they can be defined programmatically.

[0066] Figure 1 and Figure 2 The example shown depicts a simple scenario where all objects in the scene are represented by primitives directly identified in a hierarchical acceleration structure.

[0067] However, in some examples, objects can be represented by "instances," which are enclosing volumes of a particular type, and provide another example of an element that can be identified by the leaf nodes of the hierarchical acceleration structure 20.

[0068] Effectively, instances locate and orient predefined objects within a scene. Instances are associated with a low-level acceleration structure (BLAS) that describes the predefined object within its "local space," for example, identifying the relative position of the primitives forming the object with respect to some predefined coordinate space (which may differ from the scene's "world space," depending on the orientation of a given instance of an object within the scene). The BLAS itself can be formed from nodes that identify the bounding volumes of primitives that define / enclose predefined objects in a manner similar to the previously described structure 20. Thus, instances of objects in a scene can be specified by a so-called instance transformation matrix, which maps the object (and its associated bounding volume hierarchy) as defined in the object's "local space" to the desired position / orientation in "world space." The instance transformation matrix can also perform other operations on the underlying objects, such as stretching or reflection. Therefore, when creating acceleration structures for wider scenes, instantiation allows for the multiple use of a single BLAS to represent multiple versions or "instances" of the same object in different positions and orientations within the scene.

[0069] If an intersection between a ray and an instance is detected, this can trigger a ray traversal of the BLAS associated with the instance, thereby determining any intersections between the ray and the primitives of the predefined object.

[0070] In this way, the instance is treated as a type of enclosing volume, which, when intersected by rays, triggers further intersection tests of the rays with one or more nodes of different hierarchical acceleration structures, rather than intersection tests with other nodes of the original hierarchical acceleration structure.

[0071] Intersection with instances triggers ray traversal of secondary hierarchical acceleration structures, known as low-level acceleration structures (BLAS). When one or more BLAS are available, a hierarchical acceleration structure with nodes identifying instances used to trigger traversals of BLAS (e.g., structure 20) is typically referred to as a top-level acceleration structure (TLAS). Thus, a TLAS describes the scene at a high level, starting from the top-level root node and terminating in the leaf nodes that identify instances (identifying BLAS) and / or the lowest-level primitives. An example of a TLAS terminates only in the lowest-level instances (i.e., it does not contain any primitives itself). The precise nature of a TLAS will depend on implementation details.

[0072] The necessary operations for traversing BLAS are well known in the art, such as transforming the parameters of the "world space" ray or the geometry of the "local space" node using an instance transformation matrix to place the ray and (the "local space" node's) geometry in the same coordinate system to perform an intersection test, and are not described in detail for the sake of brevity.

[0073] Figure 3A block diagram of an intersection test system 300 for performing intersection tests using rays from a ray tracing system, according to an example, is shown. It will be understood that this block diagram is part of a larger ray tracing system, and the other components of the system are outside the scope of this disclosure.

[0074] System 300 includes a light request processor (RRP) 310, an intersection test processor (ITP) 320, and a light storage library (RS) 330.

[0075] Ray storage 330 stores information about all rays that need to be intersected for a specific scene. The stored information may include at least the unique identifier of the ray and the geometric information of the ray (e.g., the source and direction of the ray relative to the scene, and the possible ray range of the boundary of a segment of the ray to be intersected). For example, the unique identifier of the ray may simply be the memory address of ray storage 330.

[0076] The Intersection Test Processor (ITP) 320 may include an Intersection Test Scheduler (ITS) 321 that coordinates intersection tests between rays and scene elements (e.g., enclosing volumes, including instances or primitives). The Intersection Test Processor receives information identifying the rays to be tested, and for each ray, receives information identifying the elements to be tested against (e.g., information identifying a list of elements to be tested against). The Intersection Test Scheduler may communicate with one or more Intersection Test Units (ITU) 322, 323 to perform intersection tests. Specifically, the test unit may perform the actual intersection test after known operations by comparing the geometry of the elements with the geometry of the rays. The Intersection Test Scheduler 321 may schedule the execution of the test units 322, 323.

[0077] The ray request processor 310 submits information about the intersection test to be performed to the intersection test processor 320. Specifically, the ray request processor 310 can obtain and / or use information about rays stored in the ray storage 330 to generate a ray request and submit the information to the intersection test processor 320, which enables the intersection test processor to coordinate the intersection test. The ray request can identify the ray and the nodes of the elements for which the ray is to be tested.

[0078] In some examples, the ray request processor 310 is configured to provide the intersection test processor 320 with information about scene elements involved in the intersection test of the processed ray request. This process may include requesting data from memory 340 containing the geometry of each element to be involved in the intersection test.

[0079] The ray request processor 310 may include or communicate with a ray request memory / buffer (RRM) 315. The ray request memory stores ray requests to be processed.

[0080] In a specific example, ray request processor 310 receives ray requests (each identifying a ray and a node in a hierarchical acceleration structure, the node identifying an element for which to test the ray) and bundles the rays into ray stacks. A ray stack identifies rays to be tested against elements of the same node. Thus, a ray stack is actually a group of organized ray requests; that is, it can be viewed as ray requests identifying more than one ray (rather than a single ray). These ray stacks can then be passed to intersection test processor 320 for processing. Therefore, ray request processor 310 can be a coherence acquisition unit.

[0081] Each ray stack may be accompanied by geometric information about the elements of a node (which identifies the element to be tested against its ray stack), which can be obtained from memory 340 by RRP 310. This facilitates intersection testing between these elements and the rays of the ray stack.

[0082] Information about rays stored in ray storage 330 can be updated based on processing performed by the ray request processor and / or intersection test processor. For example, this information may include data (e.g., a counter) indicating whether the intersection test process with the ray has been completed.

[0083] The intersection testing system 300 can process rays in two different modes described in this disclosure: parallel mode or serial mode. In the context of this document, a "mode" is a processing path or procedure performed on rays according to their type. Therefore, the intersection testing system 300 can operate simultaneously in parallel and serial modes, processing some rays in parallel mode and others in serial mode. The advantages of processing in serial mode will become apparent from the following disclosure.

[0084] Rays processed in parallel mode can be labeled "parallel mode ray," and rays processed in serial mode can be labeled "serial mode ray." Therefore, two types of rays are defined: parallel mode rays processed in parallel mode and serial mode rays processed in serial mode. A ray request for a serial mode ray can be labeled "serial mode ray request," and a ray request for a parallel mode ray can be labeled "parallel mode ray request." Each ray can be associated with a mode indicator; for example, the mode indicator is stored as part of the ray information stored in ray storage 330, indicating which of the two modes the ray should be processed in. Other methods for identifying the current mode of a ray will be apparent, such as by appropriately labeling the ray's identity or associating the ray's identity with the mode via a separate lookup table.

[0085] Figure 4 This is a flowchart illustrating an iterative execution process 400 used to perform ray intersection tests using a hierarchical acceleration structure. All rays, both serial and parallel, can undergo this process. Process 400 can be... Figure 3 The intersection test system 300 described in the document is executed.

[0086] Process 400 can be executed after the ray and the hierarchical acceleration structure (for which the test ray is to be executed) have been obtained. In other words, process 400 can only be executed when the ray and the hierarchical structure are defined. Therefore, process 400 provides an intersection testing procedure between existing and defined ray and node representations in the hierarchical structure.

[0087] Process 400 includes step 410 of obtaining a light request. This step is performed by light request processor 310.

[0088] Each ray affected by process 400 is associated with a ray request. The ray request contains ray information, which identifies the ray (e.g., the ray's origin and direction, and possibly the ray range indicating a portion of the ray to be tested) and identifies the nodes (of the hierarchical acceleration structure) of the elements for which the ray of the ray request is to be intersected.

[0089] When a ray is first submitted to process 400 (i.e., not associated with a previous ray request), the node identified by the ray is the root node of the hierarchical acceleration structure. The generation of the first ray request for the node can be performed by the ray request processor (e.g., by receiving and processing a new ray from the ray store 330).

[0090] When there is no request for ray generation or a new ray is available, the ray exits or the process completes (400).

[0091] Process 400 includes a subprocess 420 that processes ray requests to identify the element (identified by the node of the ray request) to which each ray of the ray interacts. In other words, subprocess 420 includes determining the intersection point between the (ray of the ray) and the corresponding element (identified by the node of each ray request).

[0092] Subprocess 420 may include a coherence collection step 421, which includes gathering or bundling multiple rays into ray sets (“ray stacks”). Each ray stack identifies rays to be tested against elements of the same node. This is performed by processing acquired ray requests to identify groups of ray requests that have / identify the same node, thereby determining which rays form part of the same ray stack.

[0093] As explained earlier, a ray stack is actually a set of organized ray requests, that is, it can actually be viewed as a ray request that identifies more than one ray (rather than a single ray).

[0094] If executed, step 421 is performed by the light request processor 310.

[0095] Subprocess 420 may also include a test step 422, in which each ray stack (or ray request, if step 421 is not performed) is tested for each node (an element identified by it) associated with the ray stack. Specifically, each ray of the ray stack (or ray request) is tested for each element of a node in the ray stack to identify any intersection between any ray and any element. In this way, for each ray of the ray stack and each element identified by a node, it is determined whether the ray intersects with that element. This effectively identifies any intersection between each ray of the ray stack and any element contained within the enclosing volume represented by the nodes of the ray stack.

[0096] Methods for performing intersection tests between elements (e.g., enclosing volumes or primitives) and rays are well known in the art and typically involve processing the geometry of the element and the ray in order to determine whether the ray intersects with the element.

[0097] Step 422 is performed using the intersection test processor 320. The intersection test processor may, for example, use ITS 321 to coordinate the intersection tests using test units 322-323.

[0098] As previously explained, in some examples, each node is or includes a list that identifies elements conceptually branching off from that node (conceptually, elements defined or contained within the enclosing volume represented by the node). For each element (of geometry) in the list for a particular node, each ray of the ray stack (associated with that particular node) is tested to identify any intersections between the ray and the element.

[0099] In this way, subprocess 420 can identify the intersection between the ray identified by the obtained ray request and the element of the corresponding node identified by the obtained ray request.

[0100] Using the described coherence collection step 421 means that each time step 421 is performed, the list and (geometric) information about the list elements need to be retrieved from memory only once, for example, rather than for each individual ray request.

[0101] Steps 410 and sub-process 420 are common to both serial and parallel mode rays. Therefore, ray requests for both serial and parallel mode rays can be processed in the manner previously described.

[0102] Process 400 further includes subprocess 430, which performs one or more actions based on the intersection of any identified points. The actions performed in subprocess 430 depend on whether the ray (request) is a serial mode ray (request) or a parallel mode ray (request).

[0103] Therefore, subprocess 430 may include determining whether a ray is a serial or parallel ray. This can be performed, for example, by checking the identity of the ray in step 435 (e.g., consulting a ray repository).

[0104] For parallel mode ray requests, the parallel mode ray request generation step is performed in process 440. For serial mode ray requests, the serial mode ray request generation step is performed in process 450.

[0105] Process 440 includes generating a new ray request for each intersection between the requested ray and the bounding volume identified by the node of the ray request. Each new ray request identifies the (same) ray of the ray request and a subsequent node representing the bounding volume. The ray identified in each new ray request is the same ray as the (original) ray request that caused the generation of the new ray request (i.e., a ray with the same properties). Therefore, in process 440, each identified intersection between the requested ray and the bounding volume (contained within the bounding volume represented by the node identified by the ray request) triggers the generation of a new (parallel mode) ray request. The new ray request identifies the parallel mode ray and the node represented by the bounding volume that intersects with the parallel mode ray. Therefore, if multiple bounding volume intersections are detected for a given parallel mode ray in subprocess 420, multiple new ray requests can be generated in process 440.

[0106] In this way, the number of new ray requests generated each time a parallel mode ray is executed in process 440 is unlimited.

[0107] The generation of a new ray request can be performed by the ray request processor 310 (e.g., based on information about the identified intersections received from the intersection test processor 320) or by the intersection test processor 320 itself. Therefore, step 440 can be performed by the ray request processor or (less preferably) the intersection test processor 320.

[0108] The newly generated parallel mode ray requests can be stored in the ray request memory 315. Each newly generated parallel mode ray request is obtained in step 410 and processed in step 420 in a subsequent iteration of process 400.

[0109] For parallel mode rays, if there is no identified intersection between the ray and any further enclosing volume identified by any ray-requested node for the ray, then the ray has successfully traversed the hierarchical acceleration structure (i.e., the ray has missed all objects in the scene, or it has intersected with primitives).

[0110] This can be checked using a ray storage counter for each parallel mode ray. This counter can be stored in a ray storage library 330 (which stores information about each ray). The counter is initialized to 1 (when a ray is first submitted to the process). Each time a new ray request is generated (i.e., an intersection is identified), the counter for that ray can be incremented by 1 for each enclosing volume of the intersection (via communication from the intersection test processor to the ray storage library 330). Each time a ray request is fully processed (e.g., when the ray of the ray request undergoes an intersection iteration test), the counter for that ray is decremented by 1. The counter thus effectively tracks the number of ray requests associated with a particular ray. When the counter reaches 0, the ray is "complete" and can then be deleted from the ray storage library.

[0111] Based on the foregoing, it is evident that a single parallel mode ray can be associated with a potentially infinite number of different ray requests. This is because a new ray request is generated each time the interaction between the ray and the surrounding volume is determined.

[0112] This disclosure recognizes that, in some cases, parallel-mode raying may result in the generation of an extremely large number of new ray requests (e.g., if the ray intersects with more than a threshold number of elements in the scene). This may, for example, fill the ray request memory or otherwise (e.g., in the ray request processor) cause an overflow.

[0113] This disclosure proposes a novel processing mode, designated "serial mode," which overcomes these problems by effectively limiting the number of new ray requests (for a given serial mode ray request) that can be generated during each iteration of process 400. In some examples, rays may initially be processed in parallel mode and, if an overflow condition occurs, are either discarded (e.g., from process 400) or marked and reprocessed in serial mode.

[0114] The following article describes the operations performed on a serial mode ray request according to an embodiment.

[0115] The serial mode ray request generation process 450, performed only on serial mode rays, includes generating a maximum of a first predetermined number of new ray requests if at least one intersection point between the ray identified to the ray request and the enclosing volume is found. Thus, the number of new ray requests for serial mode rays is limited to no more than the first predetermined number. Each new ray request again identifies the (same) ray that triggered or caused the generation of the new ray request and another node representing the element identified as intersecting with the ray request. In this way, each new ray request differs from the original ray request, i.e., the ray request that triggered or caused the generation of the new ray request, because it identifies different nodes, while the ray identified by the new ray request remains unchanged from the original ray request.

[0116] Preferably, the process 450 further includes a step for generating and storing intersection information (if necessary). Intersection information is information that can be used to identify other nodes for which no new ray request has been generated and to indicate elements intersecting with rays for which ray requests have already been identified. Intersection information is only needed to be generated and stored when it is determined that a serial pattern ray (for a specific serial pattern ray request) intersects with more than a first predetermined number of enclosing volumes.

[0117] A specific example of process 450 will be described below.

[0118] For serial mode ray (requests): If more than a first predetermined number N1 (e.g., 1) intersections with the enclosing volume (identified by the node requesting the ray) are identified in step 422, then subprocess 450 executes step 451 to generate a second predetermined number N2 (e.g., 1) of new ray requests for the ray, and step 452 to generate and store intersection information. The second predetermined number is less than or equal to the first predetermined number, such that no more than the first predetermined number of new ray requests are generated for the serial mode ray.

[0119] Each of the second predetermined number N2 new ray requests identifies a serial pattern ray with a hierarchical structure and another node (e.g., a list). The other node represents the bounding volume (i.e., geometric information) that determines its intersection with the ray. Each new ray request represents a different bounding volume.

[0120] The intersection information identifies the node associated with the (original) ray request, that is, the node associated with the original ray request, and the node identified in turn identifies the enclosing volume that intersects with the ray more than a first predetermined number. The intersection information also identifies information suitable for identifying other nodes (branched from the node of the original ray request) for which a new ray request has been generated.

[0121] Therefore, in some cases, not every intersection of a serial-mode ray with an identifier results in the generation of a new ray request. Instead, only the maximum number of new ray requests are generated, and data that can be used to identify which enclosing volumes (represented by nodes) have not yet generated ray requests for is stored for future use.

[0122] This effectively limits the number of new ray requests that can be generated based on a single ray request for a serial-mode ray. This avoids or reduces the possibility of memory or processor overflow in the ray request processor.

[0123] For serial mode ray (requests): If the number of intersections between the identified ray and the enclosing volume (represented by the node associated with the ray request) exceeds 0 and does not exceed a first predetermined number N1 (i.e., between 1 and N1, including the endpoints), then process 450 performs step 453 of generating a new ray request for each identified intersection. Each new ray request identifies the ray of the serial mode ray request and a node representing the enclosing volume with which the ray intersects. Therefore, serial mode ray (requests) can be processed in a similar manner to parallel mode ray requests.

[0124] For serial mode ray (requests): If no intersection is identified, process 450 performs step 454 to determine or check whether the ray is associated with any stored intersection information. If the ray is associated with stored intersection information, it is processed in step 455 to generate a maximum of a third predetermined number N3 (e.g., 1) of new ray requests. The third predetermined number N3 may be equal to the first predetermined number.

[0125] Step 455 may also include requesting an update to the stored intersection information based on the generated new rays.

[0126] In some examples, step 453 further includes (if fewer than a first predetermined number N1 of intersections with the enclosing volume is identified) determining or checking whether a ray is associated with any stored intersection information. If a ray is associated with stored intersection information, it is processed to generate additional new ray requests (e.g., such that the total number of ray requests generated by step 453 is up to a fourth predetermined number N4, which may be the same as the first predetermined number and / or the third predetermined number).

[0127] In some examples, each ray is associated with a maximum number of active ray requests, for example, a maximum of a fifth predetermined number of active ray requests N5. In these examples, the intersection result of all ray requests for a particular ray is processed such that the next iteration of process 400 generates no more than the fifth predetermined number N5 of ray requests. This process may include, for example, modifying one or more of the first to fourth predetermined values.

[0128] Therefore, for serial mode ray (requests), process 450 does not generate ray requests for each intersection with an identifier of the enclosing volume (an element represented by another node in the hierarchical structure), but rather generates a maximum number of new ray requests for each iteration of the process. Information about all other intersections is stored until the ray does not intersect with any tested enclosing volume (or optionally, with fewer than a predetermined number of intersections). In this case, the stored information is used to generate the next set of one or more ray requests.

[0129] A ray can exit the recursive process when no intersection requested by any ray is identified and the stored intersection information is not identified or no longer identifies any information identifying any untested elements. Therefore, it is no longer necessary to use a counter with serial mode rays.

[0130] At least one of the newly generated serial mode ray requests can be obtained in step 410 of a subsequent iteration of process 400 (and processed in step 420). The exact number of serial mode ray requests obtained and processed may depend on available storage space, such as for intersection information to be stored. In other examples, all newly generated serial mode ray requests are obtained in step 410.

[0131] As previously explained, the proposed method for processing serial mode light prevents or reduces the possibility of memory buffer overflows for the light request processor.

[0132] The intersection information of a specific ray can be stored in a LIFO or similar data storage module. Specifically, this data storage module can be designed to process the information of the most recently identified intersection for a ray before the less recently identified intersection.

[0133] For each ray, the intersection information may include at least one data block, each data block including: information about nodes identifying elements where more than a first predetermined number of rays initially intersect; and information about which of these elements ray requests have been generated. Therefore, the intersection information may be formed by one or more sets of each data block, each set representing a different ray.

[0134] If more than one data block is stored for a given serial mode ray, then preferably, (when processing intersection information to generate the next(s) ray(s) request) the most recently stored data block is used to generate the next(s) ray(s) request.

[0135] For ease of processing, each data block can be linked to another; for example, each data block can point to another data block on the same ray. This can be done using a linked list (described below).

[0136] Intersection information can be stored in a common memory, that is, memory that is not pre-allocated for each ray, so that intersection information for all rays can be stored in the same memory. The common memory is preferably on-chip memory.

[0137] Figure 5 A more detailed description of an intersection test system 500 according to an embodiment of the present invention is shown. The intersection test system 500 is configured to perform both serial and parallel mode processing of light rays.

[0138] As previously described, the intersection testing system includes a ray request processor 310, an intersection testing processor 320, and a ray storage library 330. Here, the ray request processor 310 is a coherence acquisition unit.

[0139] The ray storage library 330 may include a ray information storage library (RIS) 532. The ray information storage library 532 may store information about all rays being processed (parallel mode and serial mode). The stored information may include a ray ID (providing a unique identifier for the ray), a mode indicator (indicating the mode the ray is operating in), and a counter (the operation of which has been previously described for parallel mode rays).

[0140] The intersection test system 500 may also include a serial mode ray controller (SMRC) 535. This may be configured as a separate processing module, or in some examples, may be incorporated into the ray storage library 330 (e.g., as an additional function of the ray storage library).

[0141] The serial mode ray controller 535 is configured to store information about serial mode ray requests to be processed by the ray request processor. Specifically, the serial mode ray controller can receive information about the intersections between serial mode rays and (scene) elements, and determine how to generate further ray requests for the serial mode rays.

[0142] The serial mode optical controller 535 can be configured to handle optical request processor 310 (i.e., in...). Figure 4 The serial mode light request is obtained in step 410 of the process 400 described in the document.

[0143] The serial mode optical controller 535 may include two storage modules (both preferably on-chip storage modules). The serial mode optical controller 535 may include a serial ID (SID) library (SIDS) 536 (which may also be labeled "serial mode optical library") and a shared intersection information library (SIIS) 537.

[0144] Figure 6 An example of a data structure 600 for storing data stored by a storage module of a serial mode light controller 535 is shown to improve understanding of the context.

[0145] SID storage 536 may include a fixed number of entries 611-61N, each associated with a different serial mode ray. Each entry in the SID storage (“SID entry”) may include a unique serial identifier (serial ID or SID) 61A. In some examples, this is simply a unique memory address for the SID entry.

[0146] The ray information storage library 532 can be adapted to store data indicating the serial ID associated with the ray for rays operating in serial mode (i.e., identifying which entry in the SID storage library is associated with the ray). This data can be stored in the same memory space previously allocated for use as a counter (if the ray operates in parallel mode). This reveals that the counter is not used in the serial mode method, and it is proposed to change the use of the memory space to reduce data storage waste.

[0147] Each entry in the SID store 536 may also contain information 61B of the next node (or optionally, multiple nodes), which identifies the element for which an intersection test is to be performed on the serial mode ray, i.e., information that can be used to generate the next(s) ray(s) request for said ray. This is in Figure 6 The information shown is "Next Node Information".

[0148] The shared intersection information store 537 stores intersection information for all serial pattern rays stored in the serial ID store. The memory space for the shared intersection information is preferably not pre-allocated per serial ray; that is, each entry in the SID store has no dedicated space. In other words, the shared intersection information store is a pool of shared memory resources for intersection information. In some examples, different data blocks of intersection information forming the same serial pattern ray can, for example, use linked lists pointing to each other to associate different (unallocated) portions of the memory with the same serial pattern ray.

[0149] The intersection information of the rays can be stored in a shared intersection information store 537 as one or more “SIIS entries” or “data blocks” 621-62N (e.g., each is a different entry or memory address in the shared intersection information store). Each data block 621-62N includes information 62A about nodes that identify more than a first predetermined number of enclosing volumes where the rays initially intersect (i.e., nodes with more than a first predetermined number of enclosing volumes intersected by the serial mode rays). Each data block also includes information 62B about which of these enclosing volumes have generated corresponding ray requests (i.e., which intersecting enclosing volumes have been processed). This can be in the form of an “offset” value indicating which elements associated with a node have been processed, for example, indicating how far down the list of elements (associated with that node) a ray request has been generated.

[0150] Each data block can also point to another data block associated with the serial ray via pointer 62C. This effectively forms a linked list of serial rays.

[0151] Each entry in SID store 536 (“SID entry”) may further include (if relevant, e.g., if the intersection information of the serial mode ray is stored in shared intersection information store 537) a pointer 61C pointing to an entry in the shared intersection information store. Specifically, the pointer may point to the latest data block of the serial mode ray.

[0152] The serial ID storage 536 can store up to a maximum predetermined number of serial mode ray entries. If more than this number of rays wish to enter serial mode, they can be temporarily cached, for example, in the serial mode buffer 560 until space becomes available. A more detailed embodiment of this method will be described later.

[0153] Return to reference Figure 5 When the ray request processor 310 receives a ray request, the SMRC provides the serial mode ray request. For each ray request, the SMRC uses next-node information to identify the serial mode ray with a SID and a list of nodes.

[0154] When a serial mode ray request is made to the RRP, entries (data blocks) in the SIID can be retained (when the expected ray request intersects with more than N1 enclosing volumes).

[0155] If a ray request provided by the SMRC results in the identification of more than N1 intersections (within the bounding volumes identified by the ray requesting node), the next node information (for the SIIS entry associated with the serial mode ray) is updated to identify the nodes represented by N2 of these bounding volumes. A new data block is generated and stored in the shared intersection information store 537, for example at a reserved SIIS entry, where the data block uses pointer 62C (if present) to point to the previously most recently stored data block of that serial mode ray (i.e., the most recently stored data block excluding new data blocks stored in this process). The pointer to the SID entry of the serial mode ray is also updated to point to the most recently stored shared intersection information (i.e., to the newly stored SIIS entry).

[0156] If a ray request provided by SMRC results in the identification of 1 to N1 intersections (with the bounding volumes identified by the ray request node), then the next node information (for entries associated with serial mode rays) is updated to identify the node representing each intersecting bounding volume (for that ray). If the SIIS entry is reserved, this reservation for the ray request can be removed.

[0157] In some examples, if fewer than N1 intersections are identified, the intersection information of the ray is processed (if any) to identify one or more nodes for testing.

[0158] As previously explained, intersection information can identify one or more nodes, each node identifying more than N1 bounding volumes to which the associated rays intersect. Intersection information also indicates which of those bounding volumes a ray request has been generated for.

[0159] However, this information alone may not be sufficient to determine which bounding volumes intersected by rays did not generate ray requests for them. This is because it may not be known which bounding volumes identified by the node intersect with rays. In other words, the node may have identified bounding volumes with which rays do not intersect, but which bounding volumes resulted in or did not result in an intersection are not stored as part of the intersection information 62N.

[0160] To overcome this, the next ray request generated for a serial mode ray can identify a node that identifies the bounding volumes of more than N1 related rays. The result of this intersection test (which identifies the intersecting bounding volumes) can then be processed along with information identifying which of those bounding volumes a ray request has been generated for to determine subsequent ray requests for that ray (avoiding bounding volumes that have already been tested).

[0161] Therefore, elements of nodes that have already been examined can be retested. While this may increase the number of intersection tests performed to process a single ray, the proposed method avoids the need to store data that specifically identifies nodes representing each intersection box (instead, identifies the “parent node” that identifies the enclosing volume of the intersection).

[0162] Preferably, as the amount of available or unused memory space in the shared intersection information storage 537 decreases, the number of light requests submitted by the serial mode light controller 535 to the light request processor 310 also decreases. In other words, referencing Figure 4 The number of ray requests obtained in step 410 can be reduced in response to the increased memory usage due to the increased intersection information. This effectively reduces the possibility of the shared intersection information storage 537 overflowing.

[0163] This can be accomplished using a gradual braking method. In this method, the amount of storage capacity used can have multiple thresholds, and each time a threshold is exceeded, fewer light requests are submitted to the light request processor (from the serial mode light controller 535). Therefore, there can be a first and / or a second predetermined threshold / percentage that, when exceeded, causes (…) Figure 4 (In step 410, less light is requested.)

[0164] Controlled motion avoids a sharp drop in performance and reduces the likelihood that a serial mode ray will be processed one ray at a time (e.g., reducing the likelihood that the shared intersection information store will become permanently full).

[0165] Preferably, the age of the serial mode ray is used to allow the ray request to proceed. Specifically, the older the serial mode ray (in the serial mode ray controller 535), the greater the likelihood that it will be submitted to the ray request processor in the ray request.

[0166] Therefore, the serial mode ray controller 535 can also store (for each serial mode ray) time information indicating the time since the serial mode ray was first assigned a SID relative to other assigned SIDs. For example, each entry in the SID store 536 can point to another older entry in the SID store, thus indicating the age of the entry in the SID store 536.

[0167] To improve performance, in some embodiments, ray requests can be generated and obtained for serial-mode raying, where any identified intersection of elements (e.g., elements branching from the node associated with the ray request) known to not result in the generation of a new ray request. For example, no new ray request is generated when the element identified by the node consists only of primitives. In other words, for nodes known not to generate new ray requests, the mechanism for limiting the number of generated ray requests can be bypassed. Therefore, the limitation imposed by the braking method can apply only to ray requests where the element identified by the node of the ray request is a bounding volume (non-leaf node).

[0168] One implementation of this braking mechanism could be achieved by using two programmable thresholds and a final fixed-level threshold. Each programmable threshold specifies the number of oldest serial-mode rays (or serial-mode ray requests) that can proceed when the number of unused SIIS entries falls below the corresponding threshold. The final threshold specifies that only the oldest serial-mode ray can proceed, i.e., the ray request has been passed to and processed by the RRP.

[0169] The final threshold is based on the maximum depth constant supported by the hardware (i.e., the maximum depth of the hierarchical acceleration structure). The final threshold can be set such that when only this number of linked list stack entries are available / free, ray requests for only the earliest serial mode ray in the SID store are allowed to be passed to the RRP.

[0170] If only a single threshold is used, a hardware-only approach to braking will work, where the threshold is based on the maximum depth of the hierarchical acceleration structure. However, in this case, SIIS can fill the data of many partially complete serial-mode rays, meaning a "bottleneck" is formed where only one serial-mode ray proceeds at a time.

[0171] This disclosure recognizes that by using appropriate braking (e.g., gradually reducing the number of ongoing serial mode rays), the final threshold can be completely avoided, and thus the average number of active serial mode rays throughout the intersection test procedure can be increased.

[0172] In an exemplary scenario: it can be assumed that each serial mode ray requires an average of 16 SIIS entries. With storage capacity for 2048 SIIS entries and a maximum of 256 serial mode rays storing SIDS, the embodiment can implement two programmable braking thresholds for which: when 1024 SIIS entries are free / available, only 128 serial mode rays can proceed; and when 512 SIIS entries are available, only 32 serial mode rays are allowed to proceed. A hardware limit (“hard limit”) can also exist, allowing only one serial mode ray (preferably the oldest serial mode ray) to proceed when 64 list entries are available.

[0173] SIIS does not have the storage capacity to store 16 entries for all 256 serial mode rays, as this would require 4096 SIIS entries. Instead: the first 1024 SIIS entries will be used to store 4 SIIS entries for each serial mode ray; the next 512 SIIS entries will be used to store 4 entries for the oldest 128 serial mode rays; the next 448 SIIS entries will be used to store 14 entries for the oldest 32 serial mode rays; and the last 64 SIIS entries will be used to store 64 entries for the oldest serial mode.

[0174] In this way, at least 22 entries are provided to the oldest 32 serial mode rays, which means that the (average) processing rate is at least one 32 serial mode rays.

[0175] Therefore, the gradual braking method avoids continuous occupation of SIIS resources and allows a consistent processing rate (no less than a predetermined number of serial mode beams at a time) to continue.

[0176] The proposed numbers and thresholds are merely exemplary, and those skilled in the art will be able to readily adjust the proposed thresholds to suit specific implementations. In some examples, the first threshold is reached when a first predetermined percentage of the available memory for the (intersection) information, such as 50% to 80%, is occupied. In some examples, the second threshold is reached when a second predetermined percentage of the available memory for the (intersection) information, such as 70% to 90%, is occupied.

[0177] The proposed braking method can also be used to reduce the number of light requests as the amount of available memory / storage device for other forms of information about light ("information related to one or more light rays") decreases, for example, reaching a predetermined threshold. For example, the proposed braking method for reducing the number of light requests can be employed when the amount of available memory for storing light requests decreases, for example, exceeding different thresholds.

[0178] For example, a first threshold may be reached when a first predetermined percentage of the available memory containing information about light, such as 50% to 80%, is occupied. In some examples, a second threshold may be reached when a second predetermined percentage of the available memory containing information about light, such as 70% to 90%, is occupied.

[0179] If executed, the proposed braking method is performed during the iterative intersection test process, which is only performed after the defined rays and hierarchical acceleration structure (reducing the number of ray requests obtained in response to available memory). In other words, the process of identifying the intersection points between predefined rays and predefined hierarchical structures (e.g., Figure 4 The proposed braking method is utilized during process 400. In this way, the use of memory does not directly depend on the number of rays to be tested or the format of the hierarchical structure. Instead, the number of ray requests obtained in each iteration of the ray intersection process is controlled.

[0180] In other words, the iterative intersection test process (or ray intersection process) executed by the system can only be performed after the rays and / or hierarchical acceleration structure have been defined and / or obtained by the system. In this way, the number of ray requests obtained is independent of the number of rays and / or the format of the hierarchical structure. This method provides improved flexibility in performing ray intersection processes (e.g., allowing an unlimited number of rays and an unlimited format for the hierarchical acceleration structure) while taking memory usage into account.

[0181] Back Figure 5 The intersection test system 500 may also include a serial mode buffer 560. As shown, this can be a separate component or integrated into the light storage library 330. The serial mode buffer 560 can receive information about light that has been discarded by the light request processor (e.g., if the light request processor's memory buffer overflows), i.e., it stores light that is desired to enter serial mode (i.e., to be processed using serial mode). When a serial mode light entry becomes available (e.g., when the previous serial mode light is completed), the serial mode buffer 560 can then pass the stored light to the serial mode light controller 535.

[0182] A polling method can be used to pass stored rays from the serial mode buffer 560 to the serial mode ray controller 535. This ensures that over time, all rays desired to enter serial mode will be passed to the serial mode ray controller 535. An alternative method could be to use a first-in, first-out (FIFO) buffer.

[0183] The aforementioned disclosure proposes a novel method for handling intersections between rays and bounding volumes, specifically how to generate one or more new ray requests based on these intersections. Intersections between rays and other types of elements (e.g., primitives or instances) can be handled in a conventional manner. For example, an intersection between a ray and a primitive can trigger a shader procedure, while an intersection with an instance can trigger ray processing through a lower-level hierarchy.

[0184] In a particular example, all predetermined values ​​N1, N2, N3, and N4 (and N5, if applicable) are equal to each other, and in another example, they are all equal to one.

[0185] As previously described, all rays can initially be processed in parallel mode, and if a ray requests the processor to overflow or runs out of memory, the ray currently being processed can be discarded or marked. Any discarded / marked rays are then reprocessed in serial mode, for example, using the serial mode method described previously.

[0186] In some examples, marked rays can continue to be processed in parallel mode (e.g., but a certain number of ray requests generated when the ray request processor overflows are discarded). Allowing rays to continue may be valuable because if an intersection with a primitive is found, this could lead to a reduction in the ray's range (which in some ray processing systems could cause the ray to exit the intersection test process).

[0187] The general format for data entries in SID repository 536 and shared intersection information repository 537 has been previously described. More precise examples of suitable formats for data entries are described below with reference to Tables 1 and 2. Table 1 shows exemplary parameters (e.g., fields) for a single data entry in SID repository 536. Table 2 shows exemplary parameters (e.g., fields) for a single data entry in shared intersection information repository 537.

[0188] In the following description, it is assumed that the predetermined values ​​N1, N2, N3, and N4 are all equal to 1. However, those skilled in the art will be able to modify the method for different predetermined values.

[0189] When initializing a new serial mode ray, the following parameters are set for the SID entry of that serial mode ray: STATE (set to a value indicating that the SID is in use); NODE ADDRESS (identifies the first node with elements for which the serial mode ray is to be tested, i.e., the root node); TYPE (identifies the type of node associated with NODE ADDRESS); OFFSET (initialized to 0); RID (identifies the ray ID of the serial mode ray, stored in the ray information store). PRIORITY can also be updated (e.g., to point to the most recently stored SID entry for another ray, to identify the order of the rays).

[0190] When a serial mode ray undergoes an intersection test, the SMRC passes a ray request to the ray request processor (from the NODE ADDRESS), which identifies the serial mode ray and the node with the elements to be tested against it.

[0191] If the intersection test identifier serial mode ray does not intersect any enclosing volume (or no enclosing volume is at or beyond the offset of the OFFSET identifier), and the POINTER VALID parameter indicates that SIID does indeed store an entry associated with the SID entry, then update the values ​​of the following parameters for the SID entry: NODE ADDRESS (instead read the NODE ADDRESS of the entry pointed to by POINTER in SIIS); OFFSET (update to the OFFSET value of the entry pointed to by POINTER in SIIS); TYPE (update to the TYPE value of the entry pointed to by POINTER of the SID entry in SIIS); POINTER (update to the value of STACK POINTER); and POINTER VALID (update to the value of STACK POINTER VALID). The associated SIIS entry (i.e., the SIIS entry pointed to by POINTER before the update) can then be deleted or marked for deletion / overwrite.

[0192] If the intersection test identifier serial mode ray intersects with one enclosing volume (at or beyond the offset of the OFFSET identifier), then update the following parameters of the SID entry: NODE ADDRESS (identifies the node representing the intersecting enclosing volume), TYPE (identifies the type of element identified by the identified node), and OFFSET (reinitializes to 0).

[0193]

[0194] Table 1

[0195] If the intersection test indicates that a serial mode ray intersects with more than one enclosing volume (at or beyond the offset of the OFFSET identifier), a new SIIS entry is created for that serial mode ray. The new SIIS entry is set with the following parameter groups: NODE ADDRESS (identifies a node with more than one enclosing volume that defines the intersection of the serial mode ray), TYPE (identifies the element type of the node), STACK POINTER (identifies any other recently stored SIIS entry for this serial mode ray, if it exists), and STACK POINTER VALID (identifies whether STACK POINTER points to another SIIS entry). The SID entry for the serial mode ray is updated as follows: NODE ADDRESS (identifies the node representing the first identified enclosing volume beyond the OFFSET), TYPE (identifies the element type of the node newly identified by NODE ADDRESS), OFFSET (initialized to 0 for the new node), POINTER (points to the newly stored SIIS entry), and POINTER VALID (indicates pointing to a valid SIIS entry). The OFFSET parameter of the newly stored SIIS entry is also updated.

[0196]

[0197] Table 2

[0198] The intersection testing system according to this disclosure can be provided as part of a ray tracing system. The ray tracing system may include one or more intersection testing systems and may implement one or more shader programs. The ray tracing system may be provided as part of a graphics processing system.

[0199] Figure 7 A computer system in which such a graphics processing system can be implemented is shown. The computer system includes a CPU 702, a GPU 704, memory 706, and other devices 714, such as a display 716, a speaker 718, and a camera 719. A processing block 710 (corresponding to intersection test systems 300, 500) is implemented on the GPU 704. In other examples, the processing block 710 may be implemented on the CPU 702. Components of the computer system can communicate with each other via a communication bus 720. A storage library 712 (corresponding to memory 112) is implemented as part of memory 706.

[0200] Figure 3 and Figure 5The intersection test system is shown as comprising multiple functional blocks. This is merely illustrative and not intended to define a strict division between different logical elements of such an entity. Each functional block can be provided in any suitable manner. It should be understood that the intermediate values ​​formed by the intersection test system described herein do not need to be physically generated by the intersection test system at any point in time, and may merely represent logical values ​​that conveniently describe the processing performed by the intersection test system between its inputs and outputs.

[0201] The intersection testing systems (and ray tracing and / or graphics processing systems incorporating them) described herein can be contained in hardware on an integrated circuit. The intersection testing systems described herein can be configured to perform any of the methods described herein. Generally, any of the functions, methods, techniques, or components described above can be implemented in software, firmware, hardware (e.g., a fixed logic circuit system), or any combination thereof. The terms “module,” “function,” “component,” “element,” “cell,” “block,” and “logic” are used herein to generally denote software, firmware, hardware, or any combination thereof. In the case of a software implementation, a module, function, component, element, cell, block, or logic represents program code that, when executed on a processor, performs a specified task. The algorithms and methods described herein can be executed by one or more processors executing code that causes the processor to execute the algorithm / method. Examples of computer-readable storage media include random access memory (RAM), read-only memory (ROM), optical disk, flash memory, hard disk storage, and other memory devices that can use magnetic, optical, and other techniques to store instructions or other data and can be accessed by a machine.

[0202] As used herein, the terms computer program code and computer-readable instructions refer to any kind of executable code for a processor, comprising code expressed in machine language, interpreted language, or scripting language. Executable code includes binary code, machine code, bytecode, code defining integrated circuits (e.g., hardware description languages ​​or netlists), and code expressed in programming languages ​​such as C, Java, or OpenCL. Executable code can be, for example, any kind of software, firmware, script, module, or library that, when properly executed, processed, interpreted, compiled, or run in a virtual machine or other software environment, causes the processor of a computer system that supports the executable code to perform the tasks specified by that code.

[0203] A processor, computer, or computer system can be any kind of device, machine, or special-purpose circuit, or a collection or part thereof, that has the processing power to execute instructions. A processor can be any kind of general-purpose or special-purpose processor, such as a CPU, GPU, neural network accelerator (NNA), system-on-a-chip, state machine, media processor, application-specific integrated circuit (ASIC), programmable logic array, field-programmable gate array (FPGA), etc. A computer or computer system may include one or more processors.

[0204] This invention also intends to cover software, such as hardware description language (HDL) software, that defines the configuration of hardware as described herein for designing integrated circuits or for configuring programmable chips to perform desired functions. That is, a computer-readable storage medium may be provided on which computer-readable program code in the form of an integrated circuit definition dataset is encoded, which, when processed (i.e., run) in an integrated circuit manufacturing system, configures the system to manufacture an intersection test system (or ray tracing system or graphics processing system) configured to perform any of the methods described herein, or to manufacture an intersection test system (or ray tracing system or graphics processing system) including any of the devices described herein. The integrated circuit definition dataset may be, for example, an integrated circuit description.

[0205] Therefore, a method for manufacturing a system (or ray tracing system or graphics processing system) as described herein can be provided in an integrated circuit manufacturing system. Furthermore, an integrated circuit definition dataset can be provided, which, when processed in the integrated circuit manufacturing system, causes the method for manufacturing the system (or ray tracing system or graphics processing system) to be executed.

[0206] Integrated circuit definition datasets can be in the form of computer code, such as netlists, code for configuring programmable chips, or hardware description languages ​​suitable for manufacturing at any level in integrated circuits, including register-transfer level (RTL) code, high-level circuit representations (such as Verilog or VHDL), and low-level circuit representations (such as OASIS(RTM) and GDSII). Higher-level representations (such as RTL) that logically define hardware suitable for manufacturing in integrated circuits can be processed on a computer system configured to generate manufacturing definitions of integrated circuits within the context of a software environment that includes definitions of circuit elements and rules for combining these elements to generate the manufacturing definitions of the integrated circuits defined by that representation. As is typically the case where software executes at a computer system to define a machine, one or more intermediate user steps (e.g., providing commands, variables, etc.) may be required to configure the computer system to generate the manufacturing definitions of the integrated circuits, executing code that defines the integrated circuits to generate the manufacturing definitions of said integrated circuits.

[0207] Now refer to Figure 8 Describe an example of processing integrated circuit definition datasets at an integrated circuit manufacturing system in order to configure the system for manufacturing intersecting test systems (or ray tracing systems or graphics processing systems).

[0208] Figure 8 An example of an integrated circuit (IC) manufacturing system 802 is shown, configured to manufacture an intersection test system (or ray tracing system or graphics processing system) as described in any of the examples herein. Specifically, the IC manufacturing system 802 includes a layout processing system 804 and an integrated circuit generation system 806. The IC manufacturing system 802 is configured to receive an IC definition dataset (e.g., defining an intersection test system as described in any of the examples herein), process the IC definition dataset, and generate an IC (e.g., containing the intersection test system as described in any of the examples herein) based on the IC definition dataset. The processing of the IC definition dataset configures the IC manufacturing system 802 to manufacture an integrated circuit containing an intersection test system (or ray tracing system or graphics processing system) as described in any of the examples herein.

[0209] The layout processing system 804 is configured to receive and process an IC definition dataset to determine a circuit layout. Methods for determining a circuit layout based on an IC definition dataset are known in the art and may involve, for example, synthesizing RTL code to determine a gate-level representation of the circuit to be generated, for example, in relation to logic components (e.g., NAND, NOR, AND, OR, MUX, and FLIP-FLOP components). By determining the location information of the logic components, the circuit layout can be determined based on the gate-level representation of the circuit. This can be done automatically or with user intervention to optimize the circuit layout. When the layout processing system 804 has determined the circuit layout, it can output the circuit layout definition to the IC generation system 806. The circuit layout definition may be, for example, a circuit layout description.

[0210] IC generation system 806 generates ICs according to circuit layout definitions as known in the art. For example, IC generation system 806 can implement a semiconductor device manufacturing process for generating ICs, which may include a multi-step sequence of photolithography and chemical processing steps, during which electronic circuits are progressively formed on a wafer made of semiconductor material. The circuit layout definition may be in the form of a mask, which can be used in the photolithography process to generate ICs according to the circuit definition. Alternatively, the circuit layout definition provided to IC generation system 806 may be in the form of computer-readable code, which IC generation system 806 can use to form an appropriate mask for generating ICs.

[0211] The different processes performed by the IC manufacturing system 802 can all be implemented in one location, for example, by one party. Alternatively, the IC manufacturing system 802 can be a distributed system, such that some processes can be performed at different locations and by different parties. For example, some of the following stages can be performed at different locations and / or by different parties: (i) synthesizing RTL codes representing the IC definition dataset to form a gate-level representation of the circuit to be generated; (ii) generating a circuit layout based on the gate-level representation; (iii) forming a mask according to the circuit layout; and (iv) using the mask to manufacture the integrated circuit.

[0212] In other examples, by processing an integrated circuit definition dataset at the integrated circuit manufacturing system, the system can be configured to manufacture an intersection test system (or a ray tracing system or a graphics processing system) without processing the IC definition dataset to determine the circuit layout. For example, an integrated circuit definition dataset can define the configuration of a reconfigurable processor such as an FPGA, and processing that dataset can configure the IC manufacturing system (e.g., by loading the configuration data into the FPGA) to generate a reconfigurable processor with that defined configuration.

[0213] In some embodiments, when processed in an integrated circuit manufacturing system, the integrated circuit manufacturing definition dataset can enable the integrated circuit manufacturing system to generate devices as described herein. For example, the integrated circuit manufacturing definition dataset referenced above... Figure 8 The described method configures an integrated circuit manufacturing system to produce devices as described herein.

[0214] In some examples, an integrated circuit definition dataset may include software running on hardware defined at the dataset, or software running in combination with hardware defined at the dataset. Figure 8 In the example shown, the IC generation system can be further configured by the integrated circuit definition dataset to load firmware onto the integrated circuit according to the program code defined in the integrated circuit definition dataset during the manufacturing of the integrated circuit, or otherwise provide the integrated circuit with program code for use with the integrated circuit.

[0215] Compared to known implementations, the implementation of the concepts set forth in this application in devices, apparatuses, modules, and / or systems (and in the methods implemented herein) can lead to performance improvements. Performance improvements may include one or more of increased computational performance, reduced latency, increased throughput, and / or reduced power consumption. During the manufacture of such devices, apparatuses, modules, and systems (e.g., in integrated circuits), trade-offs can be made between performance improvements and physical implementation methods, thereby improving the manufacturing process. For example, a trade-off can be made between performance improvements and layout area, thereby matching the performance of known implementations but using less silicon. This can be accomplished, for example, by reusing functional blocks in a serialized manner or sharing functional blocks among elements of the device, apparatus, module, and / or system. Conversely, the concepts set forth in this application that lead to improvements in the physical implementation of devices, apparatuses, modules, and systems (such as reduced silicon area) can be traded off for performance improvements. This can be accomplished, for example, by manufacturing multiple instances of the module within a predefined area budget.

[0216] The applicant has independently disclosed each individual feature described herein, as well as any combination of two or more such features, to the extent that such features or combinations can be implemented based on the specification as a whole, in accordance with the common knowledge of those skilled in the art, regardless of whether such features or combinations of features solve any problem disclosed herein. In view of the foregoing description, those skilled in the art will understand that various modifications can be made within the scope of this invention.

Claims

1. A computer-implemented method for performing an intersection test between one or more rays and elements (101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112) identified by nodes (200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212) of a hierarchical acceleration structure (20), wherein: Light is defined by light information; A node identifies one or more elements used for intersection testing, wherein at least some of the elements identified by the overall hierarchical acceleration structure are represented by another node of the hierarchical acceleration structure; The computer-implemented method includes iteratively performing the following serial-mode ray intersecting process: Obtain (410) one or more ray requests, each ray request identifying a ray and a node of the hierarchical acceleration structure, the node identifying the element for which the ray of the ray request will undergo an intersection test; Processing (420) the one or more ray requests and the hierarchical acceleration structure to identify, for each ray request, any intersection between the ray of the ray request and the element identified by the node of the ray request; and For each ray request: Determine whether the ray intersects with more than a first predetermined number of elements represented by other nodes of the hierarchical acceleration structure; and In response to the ray intersecting with more than the first predetermined number of elements represented by other nodes, (451) a maximum of the first predetermined number of new ray requests are generated, each new ray request identifying the ray of the ray request and another node representing the element identified as intersecting with the ray of the ray request.

2. The computer-implemented method of claim 1, wherein, The first predetermined number is 1, and generating (451) a maximum of the first predetermined number of new ray requests includes generating a single new ray request.

3. The computer-implemented method of claim 1 or 2, wherein, Generating a maximum of the first predetermined number of new ray requests includes: Generate (451) a second predetermined number of new ray requests for the ray, each new ray request identifying the ray of the ray request and a corresponding other node of the hierarchical acceleration structure, wherein the other node represents an element identified as intersecting with the ray of the ray request.

4. The computer-implemented method according to claim 1 or 2, further comprising, in response to the ray intersecting with more than the first predetermined number of elements represented by other nodes, generating and storing (452) intersection information for the ray, the intersection information being usable to identify any other nodes that have not generated a new ray request for them and represent elements identified as intersecting with the ray requested by the ray.

5. The computer-implemented method of claim 4, wherein, The intersection information identifies: the node represented by the ray request (62A); and information identifying which other nodes associated with the element identified by the node have generated new ray requests (62B).

6. The computer-implemented method of claim 4, further comprising, for each ray request, responding to the fact that the number of intersections between the ray and elements represented by other nodes is zero: Determine (454) whether the ray requested by the ray is associated with any stored intersection information; and in response to determining that the ray is associated with stored intersection information: The stored intersection information (455) is processed to generate a third predetermined number of new ray requests, each new ray request identifying the ray and a node in the hierarchical acceleration structure for which no ray request has yet been generated for the ray and which represents the element to which the ray intersects.

7. The computer-implemented method of claim 6, wherein, The intersection information includes one or more sets of one or more data blocks (621, 622, 623, 624, 62N), each set of data blocks representing a different ray, and each data block is identified as follows: Node (62A), the node identifying more than the first predetermined number of elements, each element being represented by another node with which the corresponding ray is determined to intersect; as well as Information (62B) identifies which elements represented by other nodes have generated ray requests for the corresponding ray for the corresponding node.

8. The computer-implemented method according to claim 7, wherein, Processing the stored intersection information includes: Identify the most recently stored data block for the light ray; Process the latest stored data block to generate the third predetermined number of new ray requests, each new ray request identifying the ray and a node in the hierarchical acceleration structure for which no ray request has yet been generated and which represents the element intersecting with the ray; and Update the stored intersection information.

9. The computer-implemented method according to claim 4, wherein, The intersection information is stored in a memory pool (537) shared by all rays.

10. The computer-implemented method of claim 1, further comprising, for each light request: In response to the number of intersections between the ray and the element represented by other nodes being between one and the first predetermined number, including the end value, a new ray request (453) is generated for each intersecting element represented by another node, each new ray request identifying the ray of the ray request and the corresponding other node representing the intersecting element.

11. The computer-implemented method according to claim 1, wherein, Obtaining (410) one or more ray requests includes, if available, selecting and obtaining at least one of the new ray requests generated in a previous iteration of the serial mode ray intersection process.

12. The computer-implemented method according to claim 11, wherein, The selection of a new ray request depends on the age of the new ray request.

13. The computer-implemented method according to claim 11, wherein, The selection of the new ray request depends on the age of the stored intersection information associated with the ray identified by the new ray request.

14. The computer-implemented method according to claim 11, wherein, The number of new ray requests selected and obtained depends on the total size of the stored intersection information.

15. The computer-implemented method according to claim 14, wherein, The number of selected and acquired new ray requests decreases in response to the total size of the stored intersection information exceeding a first predetermined threshold size.

16. The computer-implemented method according to claim 15, wherein, The number of selected and obtained ray requests is further reduced in response to the total size of the stored intersection information exceeding a second predetermined threshold size, which is greater than the first predetermined threshold size.

17. The computer-implemented method according to claim 15, wherein, When the total size of the stored intersection information exceeds the first predetermined threshold size, the selection of the at least one new ray request depends on the age of the stored intersection information associated with the ray identified by the ray request.

18. The computer-implemented method according to claim 17, wherein, When the total size of the stored intersection information exceeds the first predetermined threshold size, the selection of the at least one new ray request includes at least one ray request that identifies the ray associated with the oldest stored intersection information.

19. An intersection testing system (300, 500) for performing an intersection test between one or more rays and elements (101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112) identified by nodes (200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212) of a hierarchical acceleration structure (20), wherein: Light is defined by light information; A node identifies one or more elements used for intersection testing, wherein at least some of the elements identified by the overall hierarchical acceleration structure are represented by another node of the hierarchical acceleration structure; The intersection testing system includes: A ray request processor (310) is configured to iteratively obtain (410) one or more ray requests, each ray request identifying a ray and a node of the hierarchical acceleration structure, the node identifying the element for which the ray of the ray request will undergo an intersection test; and Intersection test processor (320), the intersection test processor being configured to handle ray requests obtained in each iteration: Processing (420) the one or more ray requests and the hierarchical acceleration structure to identify, for each ray request, any intersection between the ray of the ray request and the element identified by the node of the ray request; and For each ray request: Determine whether the requested ray intersects with more than a first predetermined number of elements represented by other nodes of the hierarchical acceleration structure; and In response to the ray intersecting with more than the first predetermined number of elements represented by other nodes, (453) a maximum of the first predetermined number of new ray requests are generated, each new ray request identifying the ray of the ray request and another node representing the element identified as intersecting with the ray of the ray request.

20. A non-transitory computer-readable storage medium having computer-readable code encoded thereon, the computer-readable code being configured to, when executed, cause a method to perform an intersection test between one or more rays and elements identified by nodes of a hierarchical acceleration structure, wherein: Light is defined by light information; A node identifies one or more elements used for intersection testing, wherein at least some of the elements identified by the overall hierarchical acceleration structure are represented by another node of the hierarchical acceleration structure; The computer-implemented method includes iteratively performing the following serial-mode ray intersecting process: Obtain one or more ray requests, each ray request identifying a ray and a node of the hierarchical acceleration structure, the node identifying the element for which the ray of the ray request will undergo an intersection test; Process the one or more ray requests and the hierarchical acceleration structure to identify, for each ray request, any intersection between the ray of the ray request and the element identified by the node of the ray request; and For each ray request: Determine whether the ray intersects with more than a first predetermined number of elements represented by other nodes of the hierarchical acceleration structure; and In response to the ray intersecting with more than the first predetermined number of elements represented by other nodes, a maximum of the first predetermined number of new ray requests are generated, each new ray request identifying the ray of the ray request and another node representing the element identified as intersecting with the ray of the ray request.

Citation Information

Patent Citations

  • Architectures for parallelized intersection testing and shading for ray-tracing rendering

    CN102037497A

  • Ray tracing system architecture and methods

    CN102282591A