Page 1 of 2 12 LastLast
Results 1 to 10 of 15

Thread: Canonical To Not Enable UXA, Too Problematic

  1. #1
    Join Date
    Jan 2007
    Posts
    14,561

    Default Canonical To Not Enable UXA, Too Problematic

    Phoronix: Canonical To Not Enable UXA, Too Problematic

    In August of last year Intel had introduced the UMA Acceleration Architecture (commonly referred to as UXA). UXA is based upon the very common EXA acceleration API but it handles the pixmap management using GEM objects...

    http://www.phoronix.com/vr.php?view=NzE3MQ

  2. #2
    Join Date
    Sep 2008
    Posts
    12

    Default

    I, for one, am glad they're making a quality control decision if the UXA isn't up to snuff.

  3. #3
    Join Date
    Mar 2008
    Posts
    573

    Default

    Wise decision.

  4. #4
    Join Date
    Mar 2009
    Posts
    18

    Default What about Nvidia..

    Hi! About nvidia and ati drivers. Will they be using UXA someday?

  5. #5
    Join Date
    Jul 2007
    Posts
    403

    Default

    No, NVIDIA and ATI drivers don't need to use UXA, they can get all the benefits that UXA offers using only the EXA API, simply by replacing the "dumb" memory allocation system in the backend with GEM.

  6. #6
    Join Date
    Oct 2007
    Posts
    15

    Default

    Funny, I geat all kinds of graphical corruptions with UXA disabled in Jaunty. First, I thought that it's related to Qt 4.5, but as soon as I changed my xorg.conf to use UXA, all corruptions went away and I got a major performace boost for free.

  7. #7
    Join Date
    Jan 2008
    Posts
    206

    Default good descision

    I guess this is also a message to the intel guys that shipping alpha/beta quality software as stable is not in the interest of the distributors and causing a _lot_ of pain for end-users.

  8. #8
    Join Date
    Oct 2008
    Posts
    151

    Default

    Quote Originally Posted by Linuxhippy View Post
    I guess this is also a message to the intel guys that shipping alpha/beta quality software as stable is not in the interest of the distributors and causing a _lot_ of pain for end-users.
    I haven't been paying attention to what claims Intel has made, but part of an open source development process is that alpha/beta software is developed out in the open and for use at your own risk. There's really no code too poor to release, as long it's communicated that something is "highly experimental" and not ready for use. Problems only happen when developers call one thing and the code is something completely different *cough*KDE4*cough*.

  9. #9
    Join Date
    Mar 2009
    Posts
    30

    Default

    Quote Originally Posted by piquadrat View Post
    Funny, I geat all kinds of graphical corruptions with UXA disabled in Jaunty. First, I thought that it's related to Qt 4.5, but as soon as I changed my xorg.conf to use UXA, all corruptions went away and I got a major performace boost for free.
    Interesting; which intel GMA is that? For me (855) it hung and gave some corruptions. I also had some with 2.6.1, but those are gone since .3

  10. #10
    Join Date
    Aug 2008
    Location
    Tokyo, Japan
    Posts
    36

    Default

    Quote Originally Posted by piquadrat View Post
    Funny, I geat all kinds of graphical corruptions with UXA disabled in Jaunty. First, I thought that it's related to Qt 4.5, but as soon as I changed my xorg.conf to use UXA, all corruptions went away and I got a major performace boost for free.
    Does

    Option "EXAOptimizeMigration" "off"

    help with EXA?

Posting Permissions

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