IP Library › Granted Patent US 12,536,731
Granted Patent B2
US 12,536,731 · App. 17/488,167 · Granted Jan 27, 2026

Acceleration structures with delta instances

Inventors: Konstantin I. Shkurko (Orlando, FL); Matthäus G. Chajdas (Munich, DE); Michael Mantor (Orlando, FL)
Assignee: Advanced Micro Devices, Inc.
G06T15/06G06T3/4007G06T9/001G06T15/08G06T17/10G06T2210/21
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,536,731
App. No.
17/488,167
Granted
Jan 27, 2026
Kind
B2
Abstract

Described herein is a technique for performing ray tracing operations. The technique includes encountering, at a non-leaf node, a pointer to a bottom-level acceleration structure having one or more delta instances; identifying an index associated with the pointer, wherein the index identifies an instance within the bottom-level acceleration structure; and obtaining data for the instance based on the pointer and the index.

Claims (34)

1 . A method for performing ray tracing operations, the method comprising:

encountering a delta instance of a bottom-level acceleration structure of a bounding volume hierarchy that stores data of positional differences between vertices of geometry of a leaf node of a base instance of a bottom-level acceleration structure and vertices of geometry of a leaf node of the delta instance, wherein the geometry of the base instance and geometry of the delta instance are renderable geometry, wherein the base instance and the delta instance have a shared box node sized to bound children of the shared box in the delta instance and children of the shared box node in the base instance;

obtaining data for the delta instance based on the positional differences; and

rendering the geometry of the leaf nodes of the base instance and the leaf nodes of the delta instance using the data, wherein the rendering includes traversing the delta instance by traversing to a child of the shared box node in the delta instance based on results of an intersection test between a ray and the shared box node.

2 . The method of claim 1 , wherein obtaining the data comprises decompressing the delta instance based on the base instance.

3 . The method of claim 1 , wherein the delta instance includes compressed data that is compressed using an identity compression technique.

4 . The method of claim 1 , wherein the delta instance includes compressed data that is compressed using a delta compression technique.

5 . The method of claim 1 , wherein the delta instance includes compressed data that is compressed using a shared pointer technique.

6 . The method of claim 1 , wherein the delta instance includes compressed data that is compressed using a shared box node technique.

7 . The method of claim 1 , wherein obtaining the data includes interpolating between two instances of a bottom-level acceleration structure based on an interpolation coefficient.

8 . The method of claim 1 , wherein the encountering occurs as part of a ray intersection test.

9 . A system for performing ray tracing operations, the system comprising:

a memory; and

a plurality of compute units that are communicatively coupled to the memory, wherein the plurality of compute units are configured to form

a bounding volume hierarchy including a bottom-level acceleration structure having one or more delta instances, and

a ray tracing pipeline,

wherein the ray tracing pipeline is configured to:

encounter a delta instance of a bottom-level acceleration structure of a bounding volume hierarchy that stores data of positional differences between vertices of geometry of a leaf node of a base instance of a bottom-level acceleration structure and vertices of geometry of a leaf node of the delta instance, wherein the geometry of the base instance and geometry of the delta instance are renderable geometry, wherein the base instance and the delta instance have a shared box node sized to bound children of the shared box in the delta instance and children of the shared box node in the base instance;

obtain data for the delta instance based on the positional differences; and

render the geometry of the leaf nodes of the base instance and the leaf nodes of the delta instance using the data, wherein the rendering includes traversing the delta instance by traversing to a child of the shared box node in the delta instance based on results of an intersection test between a ray and the shared box node.

10 . The system of claim 9 , wherein obtaining the data comprises decompressing the delta instance based on the base instance.

11 . The system of claim 9 , wherein the delta instance includes compressed data that is compressed using an identity compression technique.

12 . The system of claim 9 , wherein the delta instance includes compressed data that is compressed using a delta compression technique.

13 . The system of claim 9 , wherein the delta instance includes compressed data that is compressed using a shared pointer technique.

14 . The system of claim 9 , wherein the delta instance includes compressed data that is compressed using a shared box node technique.

15 . The system of claim 9 , wherein obtaining the data includes interpolating between two instances of the bottom-level acceleration structure based on an interpolation coefficient.

16 . The system of claim 9 , wherein the encountering occurs as part of a ray intersection test.

17 . A non-transitory computer-readable medium storing instructions that, when executed by a processor, causes the processor to perform ray tracing operations, the ray tracing operations including:

encountering a delta instance of a bottom-level acceleration structure of a bounding volume hierarchy that stores data of positional differences between vertices of geometry of a leaf node of a base instance of a bottom-level acceleration structure and vertices of geometry of a leaf node of the delta instance, wherein the geometry of the base instance and geometry of the delta instance are renderable geometry, wherein the base instance and the delta instance have a shared box node sized to bound children of the shared box in the delta instance and children of the shared box node in the base instance;

obtaining data for the delta instance based on the positional differences; and

rendering the geometry of the base instance and the delta instance using the data, wherein the rendering includes traversing the delta instance by traversing to a child of the shared box node in the delta instance based on results of an intersection test between a ray and the shared box node.

18 . The non-transitory computer-readable medium of claim 17 , wherein obtaining the data comprises decompressing the delta instance based on the base instance.

19 . The non-transitory computer-readable medium of claim 17 , wherein the delta instance includes compressed data that is compressed using an identity compression technique.

20 . The non-transitory computer-readable medium of claim 17 , wherein the delta instance includes compressed data that is compressed using a delta compression technique.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Apr 22, 2022
From: SHKURKO, KONSTANTIN I.; CHAJDAS, MATTHÄUS G.; MANTOR, MICHAEL
To: ADVANCED MICRO DEVICES, INC.
Reel/Frame 059687/0140 →
Continuity (1)
Related Publication 20230097562A1 · Mar 30, 2023
References Cited (9)
US 20090167763A1 · Waechter · 2009 [cited by examiner]
US 20110080403A1 · Ernst · 2011 [cited by examiner]
US 20160071312A1 · Laine · 2016 [cited by examiner]
US 20170287203A1 · Vaidyanathan · 2017 [cited by examiner]
US 20180190013A1 · Wald · 2018 [cited by examiner]
US 20230215091A1 · Drabinski · 2023 [cited by examiner]
GB 2593946A · 2021 [cited by examiner]
JP 2009193095A · 2009 [cited by examiner]
Lefrancois, M.-K. & Gautron, P., “DX12 Raytracing tutorial—Part 1”, NVIDIA Corporation, retrieved from: https://developer.nvidia.com/rtx/raytracing/dxr/dx12-raytracing-tutorial-part-1 on Sep. 28, 2021, 4 pages. [cited by applicant]