| |||||||||||||
Poulsbo mess casts a shadow on Intel's Moblin project (ars technica)Poulsbo mess casts a shadow on Intel's Moblin project (ars technica)Posted Dec 12, 2009 4:45 UTC (Sat) by drag (subscriber, #31333)In reply to: Poulsbo mess casts a shadow on Intel's Moblin project (ars technica) by kragil Parent article: Poulsbo mess casts a shadow on Intel's Moblin project (ars technica)
> (Intel, what were they thinking?)
They needed a low power, low cost, high performance video chipset for the
So licensing the Powervr core was probably much cheaper then designing a
-------------
Keep in mind that this is a common problem with All embedded Linux stuff.
So if there was a actual open source driver then it would be fantastic as
(Log in to post comments)
Not only PowerVR Posted Dec 12, 2009 9:40 UTC (Sat) by man_ls (subscriber, #15091) [Link] Fine, they licensed PowerVR because they were cheap sods. But I believe the driver sucks in the 2D department as well? People would not complain too much otherwise, we have grown used to having poor 3D.
Not only PowerVR Posted Dec 12, 2009 10:27 UTC (Sat) by drag (subscriber, #31333) [Link] I wouldn't be surprised if it sucked all around.It's sad because the hardware is obviously useful and fairly powerful for what it is, but it's crippled by proprietary drivers. For example it has support, in Linux, for accelerating decoding and encoding of video that would make my Core2Duo machine choke. Just see the different demos. Do a Google Video search for 'Power SGX Demos' and you can see all sorts of fabulous performance that that hardware is capable of on all sorts of different platforms. Accelerated video playback, accelerated encoding, accelerated GL shaders, video games, fancy GL-driven UI, etc etc. All of them running on hardware that is tiny, very efficient, very cheap, and running Linux. It actually does seem like top notch stuff. If it was not for the state of the driver the combination of Linux + GMA500 + Atom would be a slam dunk for Intel. Moblin would be dirt cheap to use and blow the doors off of any other x86 platform in terms of efficiency and price for the performance. Of course right now I would not touch any GMA500 hardware with a ten foot pole, when it comes to purchasing decisions. It sucks. And ultimately it is going to hurt people's efforts for Linux on ARM 'smartbooks' and cell phones much more then it is going to hurt Intel.
Not only PowerVR Posted Dec 12, 2009 13:14 UTC (Sat) by kragil (subscriber, #34373) [Link]
Interesting. I can see how some stupid Intel execs could think that. And I know that a reliance on 100% FOSS would make Chrome OS computers on ARM impossible.
The current situation just sucks all around. ARM has only PowerVR and Tegra ... if only AMD/ATI hadn't given up on ARM GFX.
At the moment Google is the only tiny hope..
Not only PowerVR Posted Dec 12, 2009 22:59 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link]
Situation is not so hopeless. The proposed plan for Poulsbo is workable.
They're planning to create open source DRM, kernel modesetting and X.org 2D driver. So you'll have fully open source 2D accelerated desktop.
3D will be a closed driver based on Gallium3D. I.e. only a fairly small amount of userspace code will be closed. And in principle it won't be hard to reverse engineer it.
Not the perfect situation. But not completely doom&gloomy.
Not only PowerVR Posted Dec 13, 2009 2:13 UTC (Sun) by kragil (subscriber, #34373) [Link]
Very interesting. Do you have any current links?
From what I read (on the internet) that plan did not really materialize (was abandoned)
Not only PowerVR Posted Dec 13, 2009 3:42 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]
I mostly get info from reading IRC channels (dri-devel, etc.). As far as I understand, this plan is still a work-in-progress and certainly is not abandoned.
Also, Phoronix: http://www.phoronix.com/scan.php?page=news_item&px=Nz...
|
|||||||||||||