Handling post-Z coverage data in raster operations

Patent metadata

Field	Value
Publication number	US-9953455-B2
Application number	US-201313802182-A
Country	US
Kind code	B2
Filing date	Mar 13, 2013
Priority date	Mar 13, 2013
Publication date	Apr 24, 2018
Grant date	Apr 24, 2018

Abstract

Techniques are disclosed for storing post-z coverage data in a render target. A color raster operations (CROP) unit receives a coverage mask associated with a portion of a graphics primitive, where the graphics primitive intersects a pixel that includes a multiple samples, and the portion covers at least one sample. The CROP unit stores the coverage mask in a data field in the render target at a location associated with the pixel. One advantage of the disclosed techniques is that the GPU computes color and other pixel information only for visible fragments as determined by post-z coverage data. The GPU does not compute color and other pixel information for obscured fragments, thereby reducing overall power consumption and improving overall render performance.

First claim

What is claimed is:

1. A method for storing coverage information in a first render target, the method comprising:
retrieving a first coverage mask associated with a first portion of a first graphics primitive, wherein the first graphics primitive intersects a pixel that includes a plurality of samples that have passed a visibility test, and the first portion of the first graphics primitive covers at least one sample included in the plurality of samples;
retrieving, from the first render target, a second coverage mask associated with a second portion of the first graphics primitive;
performing a logical operation based on the first coverage mask and the second coverage mask to generate a modified second coverage mask;
storing, via a color raster operations unit, the modified second coverage mask in a data field in the first render target at a first location associated with the pixel; and
rendering at least one of the first portion of the first graphics primitive and the second portion of the first graphics primitive into a second rendering target based on the modified second coverage mask.

2. The method of claim 1 , wherein the visibility test comprises at least one of a depth test, a stencil test, an alpha test, and a depth bounds test.

3. The method of claim 1 , wherein the visibility test is performed after the plurality of samples have been processed by a fragment shader.

4. The method of claim 1 , wherein the plurality of samples is not processed by a fragment shader.

5. The method of claim 4 , wherein the fragment shader is in a low power state.

6. The method of claim 1 , further comprising storing the first coverage mask in a data field in the first render target at a second location associated with the pixel, wherein the first location corresponds to a first sample and the second location corresponds to a second sample.

7. The method of claim 1 , further comprising detecting that the first render target is eligible for storing coverage information.

8. The method of claim 1 , wherein the data field is selected from a plurality of data fields associated with the first render target.

9. The method of claim 1 , wherein the first coverage mask comprises a plurality of bits, and wherein each bit corresponds to a different sample in the plurality of samples.

10. The method of claim 1 , wherein the first coverage mask indicates which samples are covered by the first portion of the first graphics primitive.

11. The method of claim 10 , further comprising computing a percentage of the pixel covered by the first portion of the first graphics primitive based the first coverage mask.

12. The method of claim 1 , further comprising:
receiving a third coverage mask associated with a second portion of a second graphics primitive, wherein the second primitive intersects the pixel, and the second portion of the second graphics primitive covers at least one sample included in the plurality of samples; and
storing the third coverage mask in a data field in the first render target at a second location associated with the pixel;
wherein the third coverage mask indicates which samples are covered by the second portion of the second graphics primitive and are not obscured by the first portion of the first graphics primitive.

13. The method of claim 12 , further comprising computing a centroid of the second portion of the second graphics primitive based on the third coverage mask.

14. The method of claim 12 , further comprising:
storing a first attribute associated with the first portion of the first graphics primitive in a data field in the first render target at a third location associated with the pixel;
storing a second attribute associated with the second portion of the second graphics primitive in a data field in the first render target at a fourth location associated with the pixel;
performing a blend operation on the first attribute and the second attribute based on at least one of the first coverage mask and the third coverage mask; and
storing a result of the blend operation in a data field in the first render target at a fifth location associated with the pixel.

15. The method of claim 1 , wherein the first render target is included in a plurality of render targets that further includes a second render target that stores color information, a third render target that stores depth information, and a fourth render target that stores normal vector information.

16. The method of claim 1 , further comprising:
retrieving a third coverage mask associated with a first portion of a second graphics primitive, wherein the first graphics primitive intersects a pixel that includes a plurality of samples that have passed a visibility test, and the first portion of the second graphics primitive covers at least one sample included in the plurality of samples;
determining that the third coverage mask is directed to a second render target that is ineligible to store coverage mask information; and
discarding the third coverage mask.

17. The method of claim 1 , further comprising transmitting, via a multiplexer, the first coverage mask to the color raster operations unit based on a first state of a control signal received from a streaming multiprocessor (SM).

18. The method of claim 1 , further comprising, subsequent to storing the modified second coverage mask, transmitting, via the multiplexer, color information to the color raster operations unit based on a second state of the control signal received from the SM.

19. The method of claim 1 , further comprising:
transmitting, via a multiplexer, the first coverage mask to the color raster operations unit based on a first state of a control signal received from a streaming multiprocessor (SM); and
subsequent to storing the modified second coverage mask, transmitting, via the multiplexer, color information to the color raster operations unit based on a second state of the control signal received from the SM.

20. The method of claim 19 , wherein the first coverage mask is received from a late z-raster operations unit, and the color information is received from a fragment shader.

21. A subsystem comprising:
a raster operations unit that stores coverage information in a first render target by performing the steps of:
retrieving a first coverage mask associated with a first portion of a first graphics primitive, wherein the first graphics primitive intersects a pixel that includes a plurality of samples, and the first portion of the first graphics primitive covers at least one sample included in the plurality of samples;
retrieving, from the first render target, a second coverage mask associated with a second portion of the first graphics primitive;
performing a logical operation based on the first coverage mask and the second coverage mask to generate a modified second coverage mask; and
storing, via a color raster operations unit, the modified second coverage mask in a data field in the first render target at a first location associated with the pixel; and
rendering at least one of the first portion of the first graphics primitive and the second portion of the first graphics primitive into a second rendering target based on the modified second coverage mask.

22. The subsystem of claim 21 , wherein the raster operations unit further performs the step of detecting that the first render target is eligible for storing coverage information.

23. The subsystem of claim 21 , wherein the data field is selected from a plur

Assignees

• Nvidia Corp

Inventors

• Lum Eric B
• Bastos Rui
• Duluk Jr Jerome F
• Moreton Henry Packard
• Uralsky Yury Y

Classifications

• G06T15/405

  using Z-buffer · CPC title

• G06T15/00 Primary

  Three-dimensional [3D] image rendering · CPC title

• G06T15/005 Primary

  General purpose rendering architectures · CPC title