Page 1 of 2 12 LastLast
Results 1 to 10 of 14

Thread: NVIDIA's VDPAU Library Updated For DRI2 Work

  1. #1
    Join Date
    Jan 2007
    Posts
    15,641

    Default NVIDIA's VDPAU Library Updated For DRI2 Work

    Phoronix: NVIDIA's VDPAU Library Updated For DRI2 Work

    NVIDIA's Aaron Plattner has announced the release of libvdpau 0.4. From November of 2008 when VDPAU was introduced to September of 2009, the Video Decode and Presentation API for Unix lived within NVIDIA's binary display driver package...

    http://www.phoronix.com/vr.php?view=NzkzNQ

  2. #2
    Join Date
    Jul 2008
    Location
    Greece
    Posts
    3,801

    Default

    Good stuff, even for people on non-NVidia hardware. At least someone seems to care about video acceleration on Linux.

  3. #3
    Join Date
    Jan 2009
    Posts
    141

    Default

    does this mean that when vdpau comes to gallium, every single piece of graphics hardware with shader processors and a gallium driver will support vdpau?

    and can vdpau be ported to arm for use on arm compatible graphics processors with a gallium driver(if/when that happens)?

    this is some exciting stuff. gallium will totally change linux forever. gallium is such a brilliant idea!

  4. #4
    Join Date
    Nov 2007
    Location
    Die trolls, die!
    Posts
    525

    Default

    I have a quesiton:
    It's great that nvidia provides an "open" video acceleration protocol. But is there a problem, that nvidia "owns" the protocol? Say open drivers implement vdpau and then nvidia changes the protocol? Or something like this?
    Thanks.

  5. #5
    Join Date
    May 2007
    Location
    Third Rock from the Sun
    Posts
    6,587

    Default

    Quote Originally Posted by bugmenot View Post
    I have a quesiton:
    It's great that nvidia provides an "open" video acceleration protocol. But is there a problem, that nvidia "owns" the protocol? Say open drivers implement vdpau and then nvidia changes the protocol? Or something like this?
    Thanks.
    Read the license.

  6. #6
    Join Date
    Oct 2009
    Posts
    2,145

    Default

    License or not, yes, this is definitely a concern.
    Nvidia came up with it, nvidia can change it.
    Nvidia can make their hardware only compatible with the "newest" iteration of it that they came up with.

    Nvidia scares the crap out of me and I *strongly resist* implementing any of their suggestions into "standard". Nvidia should be forced to comply with standards, not the other way around.

  7. #7
    Join Date
    May 2007
    Location
    Third Rock from the Sun
    Posts
    6,587

    Default

    Quote Originally Posted by droidhacker View Post
    License or not, yes, this is definitely a concern.
    Nvidia came up with it, nvidia can change it.
    Nvidia can make their hardware only compatible with the "newest" iteration of it that they came up with.

    Nvidia scares the crap out of me and I *strongly resist* implementing any of their suggestions into "standard". Nvidia should be forced to comply with standards, not the other way around.
    News flash, most standards are from corporations. Intel for example could do the same thing with VA-API. Like any other project out there a change in license can happen anytime. openGL for example was basically SGI's baby,openAL was Creatives, and then there are the crapload of other protocols and API's that were developed over the years such as USB, ATA, etc. In short your concerns are unjustified. Let's put it this way, if Redhat and freedesktop give it the thumbs up then your concerns are pretty much NIH fueled FUD.

  8. #8
    Join Date
    Nov 2007
    Location
    Die trolls, die!
    Posts
    525

    Default

    deanjo:
    Intel and ATI are known to be good for the open source ecosystem. nvidia gives a shit on it. From that perspective it trust Intel and AND way more than nvidia. It's not only a license question.

    Don't get me wrong: If the protocol is better than VA API, then take it. I am the last, who says "No, don't implement it, just because it's from nvidia". If it's the best, if there are no concerns regarding the license and if it's open and also no concerns regarding the problem I asked for, then take it.

    I'd also be interested in a developers view, if VDPAU is suitable for e.g. AMD or Intel cards, although VDPAU is developed (?) for nvidia cards. And which accel technology they contemplate. And why.

    Thx again.

  9. #9
    Join Date
    Jan 2009
    Posts
    141

    Default

    why do so many people have problems with nvidia as a company?

    what's bad about them other than being a little controlling and having a closed driver, have they ever done anything against free software?

  10. #10
    Join Date
    Oct 2007
    Location
    Under the bridge
    Posts
    2,153

    Default

    Quote Originally Posted by portets43 View Post
    why do so many people have problems with nvidia as a company?

    what's bad about them other than being a little controlling and having a closed driver, have they ever done anything against free software?
    Four reasons, mainly:

    1. they pursue an aggressive lock-in strategy instead of open standards (Physx, TWIMTBP, paying developers to remove features when running on competitors' hardware - see for example Assassin's Creed)

    2. their model numbers are intentionally misleading (continuous renames, parts with different specs sold under the same names, marketing madness). Gee, you can buy a GT360 now, which is the same as the 260, which was the same as the 9800, which was a slightly beefed up version of the 8800. Nice! Cookies for anyone who can tell what DX version and what video acceleration level each of those supports without looking it up.

    3. because they dealt the death blow to 3dfx

    4. because they are indifferent/hostile to the idea of open-source drivers

    Otherwise they produce pretty good software and hardware.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •