Page 3 of 6 FirstFirst 12345 ... LastLast
Results 21 to 30 of 55

Thread: When'll AMD Opensource Drivers be feature-complete for Evergreen chips

  1. #21
    Join Date
    Jun 2012
    Location
    newdelhi, india
    Posts
    24

    Default

    Quote Originally Posted by bridgman View Post
    Can you be a bit more specific about the "so poor and messy" issues you're seeing with the radeon driver?
    Thanks Bridgman if you can suggest me some workarounds. Here are the poor and messy part of radeon oss on my e-350 apu (hp dm1z, kubuntu 12.10).

    Poor: HD video playing sucks on oss radeon/linux - on win7/catalyst it's multifold better. I can say the performance is comparable to the Intel IGPs before Sandy Bridge - somewhat like g45 and g41 chips.
    Messy: Occasionally I can't connect it with my HDMI. Even when I change/share screens from "1366x768" (my laptop res.) it's not seamless, whereas win7/catalyst manages quite well, adapts as per the display in question.

    Is it just me? Or others are having such problems?

  2. #22
    Join Date
    Jul 2010
    Posts
    443

    Default

    HD video performance is known - the open source drivers do not have GPU-based video decoding, so this is being done on the CPU, which likely isn't up to the task.

  3. #23
    Join Date
    Jan 2012
    Location
    Moscow
    Posts
    58

    Default

    One feature I've never seen mentioned anywhere is support for 3D displays. Nothern Islands cards are supposed to support them on Windows over HDMI, but I've never seen any report of that working or not working on Linux, neither with fglrx nor with r600g. Am I missing something?

  4. #24
    Join Date
    Nov 2010
    Posts
    100

    Default

    Quote Originally Posted by bridgman View Post
    Quite the opposite, in fact -- WIP usually means Work *is* In Progress
    When you're sitting at a computer all day, you get used to seeing progress bars. Maybe instead of WIP, we can add those circular loading animations that never end.

  5. #25
    Join Date
    Apr 2010
    Posts
    1,946

    Default

    Quote Originally Posted by renkin View Post
    When you're sitting at a computer all day, you get used to seeing progress bars. Maybe instead of WIP, we can add those circular loading animations that never end.
    We could use [===]
    [=--] - WIP, work started
    [==-] - WIP, code present, untested
    [===] - WIP, code present, beta

    If you want more detailed progress bars..

  6. #26
    Join Date
    Jul 2010
    Posts
    443

    Default

    Quote Originally Posted by crazycheese View Post
    We could use [===]
    [=--] - WIP, work started
    [==-] - WIP, code present, untested
    [===] - WIP, code present, beta
    I give it a few weeks before we also get:

    [-=-] - WIP, untested and not sure if I started in the right way
    [-==] - WIP, beta, but probably needs a rewrite
    [--=] - WIP, I found this code and it seems to work, but I don't know why or how
    [==|] - WIP, basically works, but I've hit a roadblock
    [==/] - WIP, I think I see a way around the roadblock

    :-)

  7. #27
    Join Date
    Sep 2011
    Location
    Rio de Janeiro
    Posts
    192

    Default

    Quote Originally Posted by duby229 View Post
    Again I can't help but disagree with you. Intels one and only goal with their graphics drivers are to provide somewhat useable support. I mean they still arent going to gallium and have no plans to do so. Using their drivers requires a different configuration than everything else. They are moving more and more out of sync with the rest of the community everyday. They have done very little to improve the state of the graphics stack .

    The bottom line is that Intel doesnt care about linux. As long as their hardware is somewhat functional. They have gone out of their way to do as little as possible.
    Do you hosnestly expect a "for profit" company to help out its competitors? Intel's drivers strategy seems spot on for me. I would probably do the same if I were in their position. The code is already out there, if you want to reimplement it on gallium, go on. radeon and noveau already benefit greatly from their work, there is no reason for intel to help out even more.

    There is something we have to understand regarding oss/linux. We cannot expect for profit companies to write open and free software and compete with microsoft for no profit at all. Unfortunatelly, no one has yet come up with a business model that makes desktop linux economically viable. Canonical operates at a loss. To me, Valve seems to be the one closest to actually acomplishing that. They are also the ones who will most benefit from a top notch linux graphics stack, since it will mean more sales on the steambox. If they feel that porting intel's driver to gallium would benefit them, they would be free to do that.

    In my mind, there would certainly be benefits for valve if the 3 drivers were all oss and feature complete, since they could have more control over the platform their clients develop for. But they would have to consider the time and cost involved, which we can only estimate...

    In a nutshell, i believe intel is more forward looking, not the great linux messiah. AMD almost ignores linux, and I suppose we can agree that such position will bite them in the ass in the sometime in the future.

  8. #28
    Join Date
    Jan 2013
    Posts
    1

    Default

    * Hybrid Graphics - lots of work on this over the last year, mostly by airlied

    I observe that things had been working less and less
    smoothly as the kernel version had been progressing from 3.4 to 3.7. The version 3.8.rc3 now fails to switch to
    ati card and start Xorg


    I have problem to use ati card on my muxless hybrid graphics laptop.
    This is hp envy-1210nr laptop with ATI/Intel cards
    fglrx is unusable for my laptop SHAME to AMD and HP (http://ati.cchtml.com/show_bug.cgi?id=276)!
    I used to use vga_switcheroo to switch off ati card and to switch it on when I need to suspend/hibernate or to use the
    display and hdmi ports for presentation. Note that for 3.4 and 3.5 versions of kernel
    there was a nice patch that allowed the comfortable use of the vga_switcheroo
    (http://us.generation-nt.com/answer/p...207270831.html)
    Unfortunately kernel hackers rejected it. So in the kernel version 3.6 I had to start the vga_switcheroo related scripts manually.
    It was not as comfortable as earlier, however it did work from the odinary user. For the kernel 3.7 the situation was changed,
    so I have to use sudo everytime to suspend/hibernate. Finally, 3.8.rc3 has possibility to switch on/off the power for the AMD card but
    no possibility to start Xorg on it (https://bugs.mageia.org/show_bug.cgi?id=8718).

  9. #29
    Join Date
    Oct 2007
    Location
    Toronto-ish
    Posts
    7,285

    Default

    Quote Originally Posted by Figueiredo View Post
    AMD almost ignores linux, and I suppose we can agree that such position will bite them in the ass in the sometime in the future.
    I believe AMD invests more in Linux relative to the size of the company than Intel does.

    What "position" you are talking about -- being smaller ?
    Last edited by bridgman; 01-17-2013 at 10:12 AM.

  10. #30
    Join Date
    Dec 2007
    Posts
    2,279

    Default

    Quote Originally Posted by manmath View Post
    Messy: Occasionally I can't connect it with my HDMI. Even when I change/share screens from "1366x768" (my laptop res.) it's not seamless, whereas win7/catalyst manages quite well, adapts as per the display in question.
    What does this mean? Comes up in a mode you don't like? dualhead vs. clone? something else? The driver just generates an event when a monitor is plugged or unplugged. It's up to your desktop environment to decide what to do with that event.

Posting Permissions

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