Microsoft Patents GPU-accelerated Window Manager

[sykozis]

[citation][nom]tofu2go[/nom]Wait a minute... GPU accelerated window manager? 2003 patent? Mac OS X has had Quartz Extreme, which added GPU acceleration to the window manager, since version 10.2 released in 2002.[/citation]
Until 2006, "GPU accelerated" simply meant that the GPU processed the graphical data necessary to display the window. This patent refers to a method in which the GPU executes the data for the window manager as a GPGPU function. Prior to 2006, no GPU was capable of handling the data necessary to make this approach possible.

 

[sykozis]

[citation][nom]meradz[/nom]I didn't see a link to the patent itself and this article can be misleading. People may be getting all hyped up about this or that OS doing it first for nothing.I honestly do not think they can get a patent for a GPU accelerated app or interface since it has already been done and can easily be challenged. If anything, they would patent the way it is being accomplished.[/citation]
Actually, MS filed for the first patent on GPGPU based "gpu accelerated window manager".....and did so prior to it even becoming possible to implement.

 

[meradz]

[citation][nom]back_by_demand[/nom]Three words...Slide To Unlock[/citation]

That is considered a specific implementation in the method to unlock the phone and can be patented.

 

[amk-aka-Phantom]

That is considered a specific implementation in the method to unlock the phone and can be patented.

WTF?! GTFO troll!!! What an outrage! It's an instinctive gesture! It's no more specific than washing your hands!

 

[Guest]

let me guess, next year they'll be extorting Android phone vendors for licensing fees because Linux violates this patent, and has, in fact, been violating it 5 years before MS had it, and 8 years before MS patented it.

To all of the Microtards: You can't patent something that somebody else already invented, nor should you be able to. The same idiots who say "oh noez, without proper IP laws, nobody will want to innovate", apparently have the gall to suggest it's OK for Linux to invent something, MS to copy it, then patent it, and then say that Linux can't do it anymore. Wouldn't you say that discourages innovation?

 

[soundthinking]

[citation][nom]amk-aka-Phantom[/nom]What if there are NO new ones? Ever thought of that?[/citation]
Then you wait...It's the same thing as waiting for a massive fix for the Windows drivers. I do agree though, the drivers have one flaw and that is that they cannot keep up with new versions of the X server. Perhaps some better luck with Wayland (google it)?

 

[fazers_on_stun]

You make a very good point. I've looked up the patient referred to in the story and found it here:

Google Patents link

Apparently the patent was "filed" in 2003 and was not published until 2005. So yes, Apple not only had the idea before Microsoft, they even went as far as implementing it before Microsoft tried to patent the concept.

Heh, OK - I looked up the actual parent patent on the patent office patent search page: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&u=%2Fnetahtml%2FPTO%2Fsearch-adv.htm&r=2&p=1&f=G&l=50&d=PTXT&S1=%22compositing+desktop+window+manager%22&OS="compositing+desktop+window+manager"&RS="compositing+desktop+window+manager"

That said, I did not read the patent, so for all we know everyone might be getting worked up over nothing. It is quite possible that the Microsoft patent is for something very specific, not simply GPU accelerated window management.

Looks like it was filed 10/23/2003 and the first claim reads:

1. A computer implemented method for rendering a desktop window in a graphical user interface of an operating system shell, comprising: receiving, at a composting desktop window manager (CDWM), application content from advanced applications in a bottom-to-top order, to display the application content received in a bottom-to-top order in windows corresponding to the advanced applications in the graphical user interface; receiving, at a desktop window manager (DWM), application content from legacy applications in a top-to-bottom order to display the application content received in a top-to-bottom order in windows corresponding to the legacy application in the graphical user interface; stripping out application content received from the legacy applications; converting the stripped application content to a graphical representation; switching between the CDWM and the DWM to render the advanced application content and legacy application content based on hardware or power constraints in a computer system displaying the graphical user interface; and displaying at least a portion of the application content in the windows, the windows having translucent frame portions."

I'm no expert but this looks far narrower or more specific than the claims of the child patent.

I also note that the patent examiner looked at some Apple references:

Author: Apple, Title: Apple introduces "Jaguar": the Next Major Release of Mac OS X; http://www.apple.com/pr/library/2002/jul/17jaguar.html; Jul. 17, 2002; pp. 1-3. cited by examiner .
Author: David Morgenstern, Title: Under the desktop: Prospecting for Quartz in Mac OS X; http://www.creativepro.com/story/feature/17439.html?origin=story; Aug. 22, 2002; pp. 1-4. cited by examiner .
Author: Moki, Title: Aqua help in Nvidia GeForce 4 [Archive]--Applelnsider; http://forums.appleinsider.com/archive/index.php/t-1122.html; Jan. 28, 2002; p. 1. cited by examiner .
Author: John Siracusa, Title: Mac OS X 10.2 Jaguar; http://arstechnica.com/reviews/os/macosx-10.2.ars/8; Sep. 5, 2002; pp. 1-5. cited by examiner .
Author: Shawn Erickson, Title: Screenshot PDF; http://www.omnigroup.com/mailman/archive/macosx-talk/2002-July/071171.htm- l; Jul. 30, 2002, p. 1-6. cited by examiner .
Author: Torrey Lyons, Title: Re: MacOS X; http://www.xfree86.org/pipermail/forum/2003-July/003741.html; Jul. 9, 2003; p. 1-3. cited by examiner.
Lindberg, "2D Graphics Using Quartz", http://oops.se/cgi-bin/wiki?action=browse&diff=1&id=WwdcRapport2001/TwoDG- raphicsUsingQuartz, May 22, 2001, pp. 2. cited by examiner .
Lipton, "QuickDraw GX for Postscript programmers", http://www.mactech.com/articles/develop/issue.sub.--15/051-070.sub.--Lipt- on.sub.--final.html, Aug. 19, 2000, pp. 17. cited by examiner .
Apple Computers, "About the Mac OS X Printing System"; Dec. 11, 2002, pp. 41. cited by examiner .
Detailed View, Products & Services, printed from http://www.sun.com/jsp.sub.--utils/ScreenShotPopup.jsp?title=Rotate%20Win- dow&base=/software/products/p... printed on Feb. 23 2005, 5 pages. cited by other .
Project Looking Glass by Sun Microsystems, printed from http://www.sun.com/software/looking.sub.--glass on Nov. 8, 2004, 9 pages. cited by other .
Apple--Mac OS X--Features--Expose, printed from http://www.apple.com/macosx/features/expose/ on Nov. 2, 2004, 2 pages. cited by other.

Anyway, the above shows that the examiner did consider a lot of MacOS or OSX references before making his decision as to whether or not the patent should be granted. Like you implied, people should look at all the facts before being so quick to judge.

 

[randomizer]

[citation][nom]amk-aka-Phantom[/nom]3) nVidia - using one myself and no matter whether I use an open-source or proprietary driver, the boot screen doesn't work; why? Modern nVidia/AMD cards' frame buffer is not supported properly or some BS like that, none of the 9000 "fixes" found on the net work.[/citation]
Hang on, when did the fact that NVIDIA and AMD choose not to support Kernel Mode Setting have anything to do with their drivers' ability to provide support for compositing window managers? You are tying two unrelated issues together. Compositing doesn't require KMS, but your bootsplash does (unless you use crappy workarounds). The Windows drivers support KMS.

 

[amk-aka-Phantom]

Hang on, when did the fact that NVIDIA and AMD choose not to support Kernel Mode Setting have anything to do with their drivers' ability to provide support for compositing window managers? You are tying two unrelated issues together. Compositing doesn't require KMS, but your bootsplash does (unless you use crappy workarounds). The Windows drivers support KMS.

*Applauds* You got it! As does everyone in the news comments...

Seriously, how long do you expect people to comment directly on the article subject?



And if nVidia and AMD for some reason don't support KMS under Linux... well - I think they have good reasons for that.

 

[randomizer]

[citation][nom]amk-aka-Phantom[/nom]*Applauds* You got it! As does everyone in the news comments...Seriously, how long do you expect people to comment directly on the article subject?[/citation]
It would be nice for at least one page of comments. My issue is more that you used this as evidence of Linux not supporting GPUs properly; a statement which you made in response to a comment saying that Linux has supported GPU-accelerated window managers for a long time.

[citation][nom]amk-aka-Phantom[/nom]And if nVidia and AMD for some reason don't support KMS under Linux... well - I think they have good reasons for that.[/citation]
Perhaps they do, but I couldn't for the life of me guess why. After all, the FOSS driver developers have managed to do it, and in the case of nouveau they did it with exactly zero support from NVIDIA.

 

[Guest]

lol the Amiga was doing this back in 1985 using its dedicated graphics hardware , the Blitter chip to speed up rendering of its windows in workbench.......The same principle take advantage of the the graphics hardware to process and shift data around much faster then the CPU and using DMA reduce overhead on cpu even further.

Lots of prior art I would have thought, linux has been doing this for years too.

Sooner patents go the better

 

[amk-aka-Phantom]

It would be nice for at least one page of comments. My issue is more that you used this as evidence of Linux not supporting GPUs properly; a statement which you made in response to a comment saying that Linux has supported GPU-accelerated window managers for a long time.


Perhaps they do, but I couldn't for the life of me guess why. After all, the FOSS driver developers have managed to do it, and in the case of nouveau they did it with exactly zero support from NVIDIA.

1) Yeah, I did, because I don't like Linux fanboys showing off with some minor things while their favorite OS can't get the simplest things right. I like Linux for more "serious" tasks, the $h!t you can do there is absolutely crazy, I have machines at work that run like 10 different services at once (VPN, PXE install server, LTSP, web hosting, update cacher, etc.) and it takes very little effort to set it all up (ONCE you know how, lol!) and services don't interfere with each other, it's really great... BUT Linux shouldn't try to throw out fancy things until they get the basics of home desktop right.

2) I can figure it out easily... AMD/nVidia don't bother to input effort into making this work under Linux since no one pays them for that.

 

[randomizer]

[citation][nom]amk-aka-Phantom[/nom]I can figure it out easily... AMD/nVidia don't bother to input effort into making this work under Linux since no one pays them for that.[/citation]
Well they keep putting resources into updating ancient drivers (at least NVIDIA does), so I don't know why adding support for KMS has never been on the agenda. I don't think it would require an enormous amount of effort. After all, most of the code for their drivers is the same across all platforms.