OpenGL desktop rendering on Linux...

Comments

# Troy Phillips said on March 29, 2005 7:12 PM:

In this regard, as far as I am concerned, both Linux and Windows are still playing catch up with Mac OS X. Quartz Extreme is delivering this sort of eye candy today!

# Concerned Cross Platform Developer said on March 29, 2005 8:14 PM:

I would say that Luminocity has surpassed what was demonstrated at WinHEC 2 years ago adding the fact that Luminocity can run on old video cards like the Intel i830. As for Longhorn being "loaded" with the ability to do hardware accelerated compositing/rendering, I have yet to see any marked progress in the Longhorn preview releases made available to developers; what was shown at WinHEC were only demonstrations for WinHEC. It's amazing that the group working on Luminocity and Cario have been able to produce such a great implementation and make the technology available for developers to try out now.

# Kam VedBrat said on March 29, 2005 8:41 PM:

Regarding demonstrating progress - you're right. We haven't shipped yet, so it's hard to point at our progress from the perspective of using it on the Desktop. What we have shipped though, is two public developer previews of "Avalon" which do allow folks in the developer community to try things out today. Stay tuned - we'll have something great to show on the desktop soon enough as well.

# Steve said on March 29, 2005 10:32 PM:

Not a very fair comparison - GTK+ will be using Cairo and thus have access to these OpenGl rendering techniques in time for the Gnome 2.12 release at the end of Q3 2005. The effects engine currently being played with in Luminosity has a pretty good likelyhood of being in the base Metacity Window Manager in the same timeframe. The Gnome developers haven't committed to the changes to Metacity this release though, so its possible the Metacity improvements will appear at the end of Q1 next year.
<p />
Best case scenario, desktop Linux has advanced rendering a year before Longhorn. Worst case, it still beats Longhorn to release. Those estimates assume that Longhorn is released on time, which isn't particularly likely.
<p />
All of this ignores the fact that these pre-optimized demos will run on crap hardware from 2001 and Longhorn definitely won't. It also ignores that fact that all the Mac users in the audience are yawning wondering why we're still yapping about stuff they had years ago.
<p />
I'm not attacking Windows - I'm just saying that its behind the curve on advanced desktop rendering.

# KdeFan said on March 29, 2005 11:55 PM:

Speaking of a linux desktop, specifically window managers like KDE or GNOME, by everyone's opinion (except Explorer developers perhaps) they much more customizable and flexlible than Windows shell. Besides that, both are built using component models as well (GNOME is built on CORBA for example), which makes Explorer playing catch up (again?).

# Frans Bouma said on March 30, 2005 12:32 AM:

"there's an odd title for an msdn blog post... and yet, I'm linking to it anyway... It's nice to see the linux community catches up with what we demonstrated on Windows at WinHEC two years ago."
When will MS finally stop spreading sillyness like this? Not only did MS play the catch-up game for years to OpenGL (and still has some catching up to do in some areas), MacOS X showed 3D powered desktop renderings way before MS ever came up with this 'inventive' idea.

And before you come up with "But MacOS X came out around the same date as we demo-ed our code", take a look at some old SGI workstations of at least 10 years old and the OS it ran and what it did with windows and the desktop. Is the new X11 environment catching up? Perhaps, but not to MS.

I like a lot of the MS software, but I'm really getting tired of all the 'inventive' labels on almost everything produced in the MS coding labs.

FB, C# MVP

# Kam VedBrat said on March 30, 2005 12:43 AM:

I think it's unfair to say that there's no innovation here. We are doing some thigns in Longhorn that are inventive in the sense that no-one's ever done them before, I can't really talk about those yet.

That said, I believe innovation can happen in different ways. Sometimes an innovation comes in the form of something new that no one's ever seen before. Other times, its comes in the form of something very familiar, but in a manner no one's ever done before.

In this case, we're doing both. To a certain degree, we are playing catch-up in terms of the "eye-candy" factor of a hardware accellerated desktop. At the same time, everything is easy in the shallow end of the pool. There is some very serious thinking that goes into doing this in a way that maintains compatibility with the millions of Windows applications out there.

# Jonesy said on March 30, 2005 1:15 AM:

I guess you Macophiles do realize that OSX is only vector based upto a certain point. All your shiny gel buttons are bitmaps, Quartz is more akin to Postscript with live compositing. When you use say Exsposé, it is simply scaling bitmaps.

In Longhorn everything is vector based, rasterization is only performed at the point of drawing to the screen by the graphics hardware, no nasty scaled bitmaps here :-)

So, who is leading who....

# n4cer said on March 30, 2005 1:48 AM:

Longhorn isn't exactly Microsoft's first attempt or showing of technologies in this area.
There have been MSR projects and a few public demos involving using GPUs to accelerate drawing and composition of the Windows desktop since the 1990s.

RE: Longhorn High GPU Recommendation

From currently available information, this is a recommendation based upon the features (OS and GPU) they want to support and the quality levels they want to achieve. Some of the various Windows OS features can be partially or fully accelerated on lesser hardware, however, there will be GPU features only available on upcoming hardware that will increase rendering efficiancy and/or quality, and provide necessary hardware support to accelerate the entire pipeline for some features.

ClearType text rendering is a good example of a Longhorn feature that will scale (in quality and efficiancy) from software rendering, to partially accelerated and with enhanced quality using the GPUs shaders, to fully accelerated and highest quality (requiring hardware support most GPUs will only have in the next generation).

# Andrew Shebanow said on March 30, 2005 9:50 AM:

RE: "We are doing some thigns in Longhorn that are inventive in the sense that no-one's ever done them before, I can't really talk about those yet."

Well, that may be so, but it is exceptionally bad form to crow about Linux "catching up" to Longhorn when:

a) Longhorn is currently demoware (and will remain so at least until Beta 1 ships to developers)
b) Longhorn's ship date is at least 12-15 months away
c) you've only demoed functionality that is "catching up" to other already shipping platforms

Keep in mind that if Luminosity ships before Longhorn (which seems highly likely), than it is Longhorn that is catching up, no matter how many demos you do first.

Sorry to beat on you, but you are showing the classic arrogance and tunnel vision that Microsoft engineers are infamous for, and I say that as a former Microsoft employee...

# Scott said on March 30, 2005 10:11 AM:

Apple did it!

http://www.tvtome.com/tvtome/servlet/GuidePageServlet/showid-344/epid-170460/

# Kam VedBrat said on March 30, 2005 1:29 PM:

Gotta respond to Andrew's comments above...

It's true that LH has only shipped to a very small community of people and in that sense it is "not real" yet.

I must admit a bit of ignorance on this, but to my knowledge the Luminocity stuff is also not integrated into any Linux distribution as a default desktop configuration either. Sure you can go get the code and hack it together, but without someone standing by it an saying "It will definitely work end-to-end this way if you do this" I'd argue it's just as much demo-ware as what we demonstrated at WinHEC two years ago.

If Luminostiy ships first, yes we will be catching up. I think we'll move ahead quickly though...

# Adam Hill said on March 30, 2005 8:54 PM:

I challenge someone on the Avalon team to show off a physics based, spring-mass jiggling window frame at Tech-Ed.

# Concerned Cross Platform Developer said on March 31, 2005 8:20 PM:

"I guess you Macophiles do realize that OSX is only vector based upto a certain point. All your shiny gel buttons are bitmaps, Quartz is more akin to Postscript with live compositing. When you use say Exsposé, it is simply scaling bitmaps."

Actually, Exposé is not simply scaling bitmaps. It uses Quartz Extreme which provides device-independent and resolution-independent rendering of anti-aliased text, bitmap images and vector graphics, converting each window into a texture, then sending it to the graphics card to render on screen. Try playing a movie and hold down the shift key before using the Exposé feature. You can see the effect in slow motion while the video is still playing.

# Andrew Shebanow said on April 1, 2005 10:13 AM:

I agree: neither Longhorn nor Luminosity has shipped yet -they're BOTH demoware at this point. That's why I made the comment about betting who would ship first.

# on said on April 2, 2005 3:49 AM:

I don't really understand why it's important. Novice users won't switch to linux because of Luminosity, business users don't need special desktops like Aero. <br>I read a lot of stupid trolling. Sometimes i feel zealots think that Windows is worse than others. But i think similiar human developers write the code for Windows or OSX.. it's the same. And all OSs are based on the same 30-year-old code. <br>It's really ridiculous that some ppl try to make new OS wars.

# on said on April 2, 2005 3:50 AM:

Oh, and Kam, i'm really waiting for your answer :-) Please, please don't forget it! Thanks.

# Kam VedBrat said on April 22, 2006 4:31 AM:

I came across Sean's blog while figuring out how to register with Technocrati... He recently got around...

# Kam VedBrat OpenGL desktop rendering on Linux | Best Eye Cream said on June 9, 2009 3:54 PM:

PingBack from http://besteyecreamsite.info/story.php?id=1159