Rasterization of compute workloads

Patent metadata

Field	Value
Publication number	US-12444119-B2
Application number	US-202018035507-A
Country	US
Kind code	B2
Filing date	Nov 9, 2020
Priority date	Nov 9, 2020
Publication date	Oct 14, 2025
Grant date	Oct 14, 2025

Abstract

The present disclosure relates to methods and apparatus for graphics processing, e.g., a GPU. The apparatus may receive an image including a plurality of pixels associated with one or more workgroups and one or more pixel tiles, each of the workgroups and the pixel tiles including one or more pixels of the plurality of pixels. The apparatus may determine whether the one or more workgroups are misaligned with the one or more pixel tiles. The apparatus may determine a conversion order of the one or more workgroups when the one or more workgroups are misaligned with the one or more pixel tiles, the conversion order corresponding to a common multiple of one of the one or more workgroups and one of the one or more pixel tiles. The apparatus may convert each of the one or more workgroups based on the conversion order of the one or more workgroups.

First claim

What is claimed is:

1. A method of graphics processing, comprising:
receiving an image including a plurality of pixels associated with one or more workgroups and one or more pixel tiles, each of the one or more workgroups including one or more pixels of the plurality of pixels and each of the one or more pixel tiles including one or more pixels of the plurality of pixels;
determining whether the one or more workgroups are misaligned with the one or more pixel tiles;
determining, based on a rasterization algorithm, a conversion order of the one or more workgroups when the one or more workgroups are misaligned with the one or more pixel tiles, the conversion order corresponding to a common multiple of one of the one or more workgroups and one of the one or more pixel tiles, wherein the conversion order corresponds to a rasterization order associated with adjacent fibers that access consecutive addresses; and
converting each of the one or more workgroups based on the conversion order of the one or more workgroups.

2. The method of claim 1 , wherein the conversion order corresponds to a snake pattern of the one or more workgroups and the one or more pixel tiles.

3. The method of claim 1 , wherein each of the one or more pixels in each of the one or more workgroups are converted based on a workgroup conversion order.

4. The method of claim 3 , wherein the workgroup conversion order corresponds to a zigzag pattern of the one or more pixels in each of the one or more workgroups.

5. The method of claim 1 , further comprising:
executing at least one computing program for one or more work items of each of the one or more workgroups, each of the one or more work items including one or more pixels, wherein the at least one computing program is executed based on the conversion order.

6. The method of claim 1 , wherein converting each of the one or more workgroups based on the conversion order comprises rasterizing each of the one or more workgroups based on the conversion order.

7. An apparatus for graphics processing, comprising:
memory; and
at least one processor coupled to the memory and configured to:
receive an image including a plurality of pixels associated with one or more workgroups and one or more pixel tiles, each of the one or more workgroups including one or more pixels of the plurality of pixels and each of the one or more pixel tiles including one or more pixels of the plurality of pixels;
determine whether the one or more workgroups are misaligned with the one or more pixel tiles;
determine, based on a rasterization algorithm, a conversion order of the one or more workgroups when the one or more workgroups are misaligned with the one or more pixel tiles, the conversion order corresponding to a common multiple of one of the one or more workgroups and one of the one or more pixel tiles, wherein the conversion order corresponds to a rasterization order associated with adjacent fibers that access consecutive addresses; and
convert each of the one or more workgroups based on the conversion order of the one or more workgroups.

8. The apparatus of claim 7 , wherein the conversion order corresponds to a snake pattern of the one or more workgroups and the one or more pixel tiles.

9. The apparatus of claim 7 , wherein to convert each of the one or more workgroups, the at least one processor is configured to convert each of the one or more pixels in each of the one or more workgroups based on a workgroup conversion order.

10. The apparatus of claim 9 , wherein the workgroup conversion order corresponds to a zigzag pattern of the one or more pixels in each of the one or more workgroups.

11. The apparatus of claim 7 , wherein the at least one processor is configured to:
execute, based on the conversion order, at least one computing program for one or more work items of each of the one or more workgroups, each of the one or more work items including one or more pixels.

12. The apparatus of claim 11 , wherein to execute the at least one computing program, the at least one processor is configured to execute the at least one computing program with a shader processor of a graphic processing unit (GPU).

13. The apparatus of claim 11 , wherein the at least one processor is configured to:
store each of the work items of each of the one or more workgroups in a cache.

14. The apparatus of claim 13 , wherein the at least one processor is configured to:
remove at least one pixel tile from the cache when the cache reaches a storage threshold; and
store the removed at least one pixel tile in a system memory.

15. The apparatus of claim 13 , wherein the cache is a level two (L2) cache.

16. The apparatus of claim 7 , wherein the one or more pixel tiles are based on a compression granularity of the plurality of pixels.

17. The apparatus of claim 7 , wherein to convert each of the one or more workgroups based on the conversion order, the at least one processor is configured to rasterize each of the one or more workgroups based on the conversion order.

18. A non-transitory computer-readable medium storing computer executable code for graphics processing, the code when executed by a processor causes the processor to:
receive an image including a plurality of pixels associated with one or more workgroups and one or more pixel tiles, each of the one or more workgroups including one or more pixels of the plurality of pixels and each of the one or more pixel tiles including one or more pixels of the plurality of pixels;
determine whether the one or more workgroups are misaligned with the one or more pixel tiles;
determine, based on a rasterization algorithm, a conversion order of the one or more workgroups when the one or more workgroups are misaligned with the one or more pixel tiles, the conversion order corresponding to a common multiple of one of the one or more workgroups and one of the one or more pixel tiles, wherein the conversion order corresponds to a rasterization order associated with adjacent fibers that access consecutive addresses; and
convert each of the one or more workgroups based on the conversion order of the one or more workgroups.

Assignees

• Qualcomm Inc

Inventors

• Zhang Yibin
• Ying Zilin
• Du Yun
• Qi Heng
• Yu Jiexia
• Yu Yang
• Gruber Andrew Evan
• Liang Jian
• Wang Tao
• Bourd Alexei Vladimirovich
• Zhong Gang
• Huang Minjie

Classifications

• G06T2200/28

  involving image processing hardware · CPC title

• G06T1/60

  Memory management · CPC title

• G06T15/005 Primary

  General purpose rendering architectures · CPC title

• G06T1/20 Primary

  Processor architectures; Processor configuration, e.g. pipelining · CPC title