Announcement

Collapse
No announcement yet.

The Major Open-Source ATI Improvements Over Two Years

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

  • #16
    Thanks to pingufunkybeat and tball for those answers.
    However, I really like Michael to write (and keep updated) something giving the comparison between open-source and catalyst, so the end-user that I am can see what areas are supported and what others needs to be improved on open-source side.

    I'll visit the link given by pingufunkybeat though, this should partially answer the point.

    Thanks.

    Comment


    • #17
      Originally posted by Gaius View Post
      Not to mention the power management which is still not as good as the one from catalyst (powerplay) (http://phoronix.com/forums/showthrea...ht=#post145519)
      Unfortunately, there hasn't been any commit in radeon_pm.c (radeon power management) for almost two years ... (http://cgit.freedesktop.org/xorg/dri...rc/radeon_pm.c)
      The power management has been moved to the kernel since the move over to KMS. The Xorg driver has nothing to do with it right now.

      Comment


      • #18
        Originally posted by Fixxer_Linux View Post
        It also would be nice to see whether or not some commercial games like UT2004, QuakeWars or Heroes of Newerth are working and if yes, with what quality of image.

        We have never any feedback on the few commercial title that exists on Linux. I'm waiting for those title to work with opensource drivers to swap from catalyst to open-source.

        Could it be possible to have such review?
        Those games are working very well, but sometimes you need the S3TC library. (libtxc_dxtn)

        Comment


        • #19
          just use 2 systems at same time, one with open drivers, for everyday work, blob for gamings, just 30 seconds to reboot

          Comment


          • #20
            Originally posted by marek View Post
            The power management has been moved to the kernel since the move over to KMS. The Xorg driver has nothing to do with it right now.
            Yes, sorry, I did not notice that...
            So instead of 2 years, it's 2 months

            Comment


            • #21
              Originally posted by 89c51 View Post
              video decoders...
              I forgot about that. I'll give it another shot at compiling all the shit that's required one of these days to see if this time I can make it work.

              Comment


              • #22
                I do remain hopeful about the progress of the oss drivers, but they need a lot more work in the power management area.

                These drivers are practically useless to any laptop users until they do something about the power management. Dynpm hardly works and still flickers all the time. Setting to low profile still runs hotter than catalyst and performance is noticeable worse in low profile all the time.

                Comment


                • #23
                  These are some great statistics for r300c/g, but when will r600g get the same kind of boost? It's been "coming" for a while now...

                  Comment


                  • #24
                    sigh

                    ATI is a lost cause. Especially since that sank a slew of cards to legacy support which hasn't seen action for years. Well, actually the Windows driver did see action. Low and behold.

                    v9.3 Ubuntu 9.04/Slackware 12.1/Fedora 8
                    http://support.amd.com/us/gpudownloa...2&lang=English


                    v10.2 support for Windows 7 64bit
                    http://support.amd.com/us/gpudownloa...3&lang=English

                    I regret ever getting rid of my tnt2 ultra 10 years ago.

                    btw 80% of catalyst is really 40% of Windows.

                    Comment


                    • #25
                      You are being too pessimistic.

                      In less than two years, new cards (r600+) went from drawing a triangle to playing Quake4. The only thing that does not work are the Unigine benchmarks (and they are close).

                      The older cards (r500-) went from being OK for desktop stuff to surpassing Catalyst in features, and often also in performance.

                      I really wonder what some of you people expect.

                      Comment


                      • #26
                        On the feature side I'd also be optimistic. Once we have reasonably stable and fast Gallium3D drivers, for both amd and nv, and they become standard for distributions, I suspect the state tracker business will take off (there are so many ideas/proposals/alpha-stage projects).
                        After all, devs are much more enthusiastic about developing stuff that can be used right away (and I don't blame 'em).

                        Comment


                        • #27
                          Originally posted by squirrl View Post
                          v9.3 Ubuntu 9.04/Slackware 12.1/Fedora 8
                          http://support.amd.com/us/gpudownloa...2&lang=English

                          v10.2 support for Windows 7 64bit
                          http://support.amd.com/us/gpudownloa...3&lang=English
                          Yeah, Windows gets all the attention and the best support. Nothing new there.

                          Originally posted by pingufunkybeat
                          I really wonder what some of you people expect.
                          Even when the OSS drivers have better performance and more features than Catalyst people are going to find something to complain about.

                          Comment


                          • #28
                            Originally posted by pingufunkybeat View Post
                            I really wonder what some of you people expect.
                            Here's my glewinfo from r600g git master.

                            See all the MISSING?

                            I expect them to become "OK".

                            And then when FooApplication (proprietary, free software, or otherwise) calls them, I want them to work. And for bonus points, it should be fast.

                            That's what I expect.

                            However, I realize that it takes time and manpower and attention to get to that state, so I'm not "complaining" so much as stating what the true end-goal should be. Just as it's unwarranted to go on an internet raging rampage because of inadequate drivers, it's equally unwarranted to sit back on your laurels, content in the bliss that is compiz or mutter, without acknowledging that, as far as we've come, there's a lot of work yet to be done.

                            Comment


                            • #29
                              Originally posted by allquixotic View Post
                              Here's my glewinfo from r600g git master.

                              See all the MISSING?

                              I expect them to become "OK".
                              Mhm. Even the nv blob produces 725 MISSING vs 1764 OK. So it's not really black and white...

                              Comment


                              • #30
                                Originally posted by allquixotic View Post
                                Here's my glewinfo from r600g git master.

                                See all the MISSING?

                                I expect them to become "OK".

                                And then when FooApplication (proprietary, free software, or otherwise) calls them, I want them to work. And for bonus points, it should be fast.

                                That's what I expect.
                                There's a good reason why many of those functions are missing. For instance, moern hardware no longer supports color tables - should drivers implement a slow software fallback for that case? What about hardware- and vendor-specific extensions?

                                What the goal should be is to make everything in OpenGL core available first, work on the compatibility profile second and, if there's nothing better left to do, add oddball extensions last. OpenGL 4.1 core is probably missing fewer than 100 entry points at this point (most of which are overloads of each other).

                                Edit: for comparison, OpenGL 4.1 compatibility has ~1000 entry points or ~2200 incl. extensions.

                                Comment

                                Working...
                                X