Announcement

Collapse
No announcement yet.

Standard Radeon vs. fglrx

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Originally posted by RealNC View Post
    And people call *me* a troll...
    Actually, Admax88 is right, isn't he? Gsnoorky should complain at amd fglrx forum. If there's no such forum then he should ask for help rather then saying people what they should do. What's funny it seems he disappeared.
    Last edited by kraftman; 01-25-2010, 06:33 AM.

    Comment


    • #12
      We're doing direct support for stream and OpenCL issues on the amd forums (since there was no active forum at the time the amd.com forums were set up), but for graphics related questions we're hanging out here rather than trying to move everyone to amd.com.

      Comment


      • #13
        Alright purists....

        Originally posted by Smorg View Post
        These statements are incompatable. How do you suggest developers address "key issues" in drivers for which they have no control, source, or way of testing?

        Whats more, core linux projects should not have to accomodate the whims of fractious hardware vendors with ugly hacks in order to integrate binary blobs.
        Incompatible.

        OK, I did get it to work. I had to include source and unapproved updates.

        You misunderstand--I'm not saying that the "pure" teams should work on it. I merely indicate that linux developers should make sure that the "ugly" (ATi binary) team is well ready before new distro versions pop up.

        Comment


        • #14
          Help takes too long

          Originally posted by kraftman View Post
          Actually, Admax88 is right, isn't he? Gsnoorky should complain at amd fglrx forum. If there's no such forum then he should ask for help rather then saying people what they should do. What's funny it seems he disappeared.
          Eventually, I resolve these things. I'm not looking at this forum 24/7....

          I'm not saying "Pontius" should help by writing code, blindly, for ATi--nonetheless, he shouldn't be so unhelpful by his attitude concerning these issues. "He" doesn't work in the vacuum he wants us to believe he does.

          This is the "general" forum, after all. I have just a few suggestions and observations, that's all.

          Comment


          • #15
            Originally posted by pingufunkybeat View Post
            I use the open source driver on a HD 4550 and I have absolutely no idea what you're talking about.

            You're doing something terribly wrong.

            Have you considered hardware differences and monitor/card combinations?: It's not as simple as you think. I did get it to work--some kind of "trick" often is necessary. Divining the trick to use, perhaps, is often the problem. I figured the card may have been "toast" or needed extra cooling, as it uses a fanless HS. If it was toast, it likely wouldn't accept a more "advanced" driver.

            It seems that linux code indicates analog, 1280X1024 and 1024X768 monitors still are the norm. (Linux code is always behind, and, I don't see the solution to that.) An off-brand or weird sized monitor (say, AOC 15.6" or LCD TV) may be difficult to config, also. My monitor (display) icon always has a "slash" through it with the newer distros.

            It just seems that video card driver config is more tricky and difficult than in the Radeon 9800 days and earlier. Sometimes the restricted driver from distro teams doesn't work--aticonfig does then, though. In this case, the restricted driver does work after all while aticonfig doesn't....

            I haven't gotten the "plain" driver to work, with good performance, in a long time. Sometimes, it doesn't work at all--again, I have to use aticonfig to get a working driver.

            Comment


            • #16
              OK, Purist.

              Originally posted by monraaf View Post
              Wow, I've never seen such an ill-informed post about Linux outside of Ubuntu Forums.
              I didn't come here asking for help. I don't think that I phrased the post well.

              I've been fooling with Linux for about 10 years. I understand the philosophy--source code must be available for the purists to accept any routine for the core.

              On the other hand, most users with systems less than 10 years old with fairly decent video cards would agree that the standard ATi driver generally doesn't cut it--I agree that the community, in a practical sense, is hobbled with limited cooperation and resulting binaries from ATi, and with legal issues.

              It's true that Cyberlink now sells software for Linux--in order to enable legal codecs. As I understand it, Linus Torvalds doesn't like it, yet, he accepts it. Linus is a really sharp guy, but, I don't think he completely understood the ramifications of legal and proprietary code when he started toying with UNIX long ago--such issues never go away, nor, are they ever alleviated to any extent....

              The world is swirling around you, descending into a sinkhole: Somehow, you don't seem to be aware of that. I'm just noting that your Ivory Tower is beginning to look somewhat like the Leaning Tower of Pisa!

              Comment


              • #17
                gsnoorky, can you be a bit more specific about which driver you are talking about ? When you say "the plain driver" or "the standard driver" are you talking about the open source drivers (if so which ones ?) or about the proprietary driver installed then used without running aticonfig ?

                The aticonfig utility won't do anything for the open source drivers (in fact any remnants of the proprietary driver will stop the open source driver from working properly), but the installation instructions *require* the use of aticonfig before running the proprietary driver (aticonfig --initial with options for specific configurations, eg dual-head etc..).

                I suspect that part of the friction here is coming from other posters not really understanding what you are trying to say, and I suspect being more specific about the drivers would help. Talking about "standard" and "plain" is pretty ambiguous.

                Isn't the Cyberlink software all userspace ? The pushback against proprietary code is for the kernel, not for userspace code AFAIK.

                Enterprise distro vendors, at least the ones catering to 3D workstation customers, generally do wait until binary drivers are ready and work directly with hardware vendors to make sure the schedules line up. That doesn't work well for "rolling release" distros, however, or for distros which focus on bringing the newest and niftiest code to their users.
                Last edited by bridgman; 01-25-2010, 11:48 AM.

                Comment


                • #18
                  Originally posted by RealNC View Post
                  And people call *me* a troll...
                  Haha trolling? I'm not trolling.

                  If AMD wants to make a binary driver then they have to support it. Nobody in the open source world should have to deal with supporting that driver.

                  If AMD wants their driver to be supported by and incorporated into the community then they need to play by the communities rules, open source it and get it merged upstream.

                  There's no reason that kernel or X.org dev's should waste even a minute of their time fixing bugs related to binary drivers/software.

                  Could you imagine the hell that linux would be if every device manufacturer decided to release binary drivers/software as binary blobs only and tried to support them all?

                  Edit:

                  Thankfully this situation is improving with AMD's release of some documentation. Which is why I now buy AMD rather than nVIDIA when I need high performance graphics cards. Its not perfect, but if things keep going like this, maybe we can have complete support of AMD graphics cards on linux in a few years.
                  Last edited by admax88; 01-25-2010, 03:47 PM.

                  Comment


                  • #19
                    Originally posted by yotambien View Post
                    That, coming from somebody who used to run QuakeLive at 2 FPS until a couple of weeks ago, adds a new dimension to the word hypocrisy.
                    Do you know what the word "hypocrisy" means?
                    Last edited by pingufunkybeat; 01-25-2010, 10:03 PM.

                    Comment


                    • #20
                      Originally posted by admax88 View Post
                      Haha trolling? I'm not trolling.

                      If AMD wants to make a binary driver then they have to support it. Nobody in the open source world should have to deal with supporting that driver.

                      If AMD wants their driver to be supported by and incorporated into the community then they need to play by the communities rules, open source it and get it merged upstream.

                      There's no reason that kernel or X.org dev's should waste even a minute of their time fixing bugs related to binary drivers/software.

                      Could you imagine the hell that linux would be if every device manufacturer decided to release binary drivers/software as binary blobs only and tried to support them all?

                      Edit:

                      Thankfully this situation is improving with AMD's release of some documentation. Which is why I now buy AMD rather than nVIDIA when I need high performance graphics cards. Its not perfect, but if things keep going like this, maybe we can have complete support of AMD graphics cards on linux in a few years.


                      OK, I'm not saying the development groups that you mention can fix such code. As for ATi giving up the source--they never will--ATi doesn't want proprietary secrets revealed to nVidia (and vice-versa) and other competitors such as VIA and SIS.

                      "Open source driver" is the correct terminology for the driver that I believe, generally, has limited usefulness these days. As I understand it, aticonfig (if fglrx is installed) downloads, installs, and allows config of the most current driver from ATi. The "restricted" driver has been tested--at least to some extent, unlike the most current driver via aticonfig....

                      I read about MB mfrs. considering chucking the 16-bit BIOS for a more elaborate improvement (I can't remember the name of the tech--I think it was discussed in Tom's Hardware Forums.) Apple strongly supports this: Many drivers would be written into "firmware" instead of code for the OS--supposedly giving a more unified platform for developers to work on for some code.

                      I imagine that one of the key reasons Apple supports this is that, as with our beloved Linux, many expensive peripherals, previously "paperweights" (sound familiar?), later on may actually have drivers (pre-loaded) that work--independent of an OS--courtesy of firmware. Maybe video card vendors actually could write for this tech, also.

                      Some Gamers and many others are very much against this--for one thing, they would have to look up many new settings. Then, their tired, yet, familiar argument goes like this "If it's not broke, then, don't fix it!" Of course, they like gaming with Wintel....

                      If, somehow, this "Super BIOS" actually is implemented outside of Apple hardware, this makes our arguments moot, right? I'm going to move on, after answering some other posts. You guys think that I don't get your points, yet, I do. I know the difference between binary, object, and source: I developed code a long time ago for TI--you don't have to lecture me as if I'm a toddler, Dad!

                      We've been mutually distracted long enough. I noticed that someone else here gets my point. (They indicated that Enterprise version developers do consider such concerns as mine more seriously.)

                      Comment

                      Working...
                      X