IP Library Granted Patent US 9,607,407
Granted Patent B2
US 9,607,407 · App. 13/732,313 · Granted Mar 28, 2017

Variable-width differential memory compression

Inventors: Jonathan Dunaisky (Fort Collins, CO); David Kirk McAllister (Holladay, UT); William Craig McKnight (Harvest, AL)
Assignee: NVIDIA CORPORATION
G06T9/00H04N19/426
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 9,607,407
App. No.
13/732,313
Granted
Mar 28, 2017
Kind
B2
Abstract

A method, in one embodiment, can include performing difference transformation of image samples. In addition, the method can also include performing length selection. Furthermore; the method can include performing packing that includes utilizing varying sized bit fields to produce a compressed representation.

Claims (35)

1. A method comprising:

a graphics processor performing difference transformation of image samples;

said graphics processor performing length selection; and

said graphics processor performing packing that comprises utilizing varying sized bit fields to produce a compression representation that comprises a first and second compressed representations, said first compressed representation is packed with its bit fields descending from a first end of said compression representation and said second compressed representation is packed with its bit fields ascending from a second end of said compression representation.

2. The method of claim 1 , wherein said performing packing comprises using a lowest channel size.

3. The method of claim 1 , further comprising:

said graphics processor performing channel decorrelation of image color channels.

4. The method of claim 1 , wherein said performing length selection comprising determining raw length of each difference of said performing difference transformation of image samples.

5. The method of claim 1 , wherein said performing length selection comprising determining maximum lengths of differences for a pair of samples.

6. The method of claim 1 , wherein said performing length selection comprising determining total size of differences and lengths in a channel using lengths by pairs.

7. The method of claim 1 , wherein said performing length selection comprising determining total size of all differences and lengths in a channel using a maximum length.

8. A system comprising:

a graphics processor;

a frame buffer coupled to said graphics processor;

said graphics processor for performing a method comprising:

performing difference transformation of image samples;

performing length selection; and

performing packing that comprises utilizing varying sized bit fields to produce a compression representation that comprises a first and second compressed representations, said first compressed representation is packed with its bit fields descending from a first end of said compression representation and said second compressed representation is packed with its bit fields ascending from a second end of said compression representation.

9. The system of claim 8 , wherein said method further comprising:

performing a success check of a portion of an image.

10. The system of claim 8 , wherein said image samples were compressed before said performing difference transformation.

11. The system of claim 8 , wherein said performing length selection further comprises length index encoding.

12. The system of claim 8 , wherein said performing length selection further comprises performing alpha uniform encoding.

13. The system of claim 8 , wherein said performing packing further comprises encoding a fast-clear.

14. The system of claim 8 , wherein said image samples include data selected from a group consisting of color, depth, and stencil.

15. A method comprising:

a graphics processor performing difference transformation of image samples;

said graphics processor performing length selection which comprises performing length encoding; and

said graphics processor performing packing that comprises utilizing varying sized bit fields to produce a compression representation that comprises a first and second compressed representations, said first compressed representation is packed with its bit fields descending from a first end of said compression representation and said second compressed representation is packed with its bit fields ascending from a second end of said compression representation.

16. The method of claim 15 , wherein said performing packing further comprises using a bit field to describe an anchor sample's packing.

17. The method of claim 15 , further comprising:

said graphics processor performing channel decorrelation of image color channels.

18. The method of claim 17 , wherein said performing length selection further comprises using a single maximum length for just red and blue channels.

19. The method of claim 15 , wherein said performing length selection further comprises using a single maximum length for just red, green, and blue channels.

20. The method of claim 15 , wherein said performing packing further comprises a field to describe which fixed length set to use.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 9, 2016
From: DUNAISKY, JONATHAN; MCALLISTER, DAVID KIRK; MCKNIGHT, CRAIG
To: NVIDIA CORPORATION
Reel/Frame 039387/0944 →
Continuity (1)
Related Publication 20140184612A1 · Jul 3, 2014