Announcement

Collapse
No announcement yet.

ATI R600g Gains Mip-Map, Face Culling Support

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

  • #16
    Also, for what it's worth, lots of HP's printer backends are closed source due to 3rd party raster engines used in some printers.

    Comment


    • #17
      Originally posted by crazycheese View Post
      Yes, I think you are right, when it comes to drivers only. But with hp it works and works for all models and they release fast and opensource.
      ...
      I wonder how long brigman and the others are going to write at phoronix. For the last weeks most of the comments are like yours. Complain, complain and even more complain.

      Don't complain if...
      ... someone is developing something FOR FREE (and of course also for you) in his spare time.
      ... a company (like AMD) helps those developers with documentation and support FOR FREE (and even more with own developers)
      ... a company pays lawyers to check if its legal to publish those documentations (because there is something like lincensing - you may have heard of it)
      ... if the development isn't as fast as you would like to have it (if thats your problem get your own ass up or pay someone to do it for you)


      Cartman: [happy, for once] No, I know. You're right, Mom. I need to learn to be patient. I think I can wait three weeks for the Nintendo Wii to come out.

      Comment


      • #18
        Originally posted by mendieta View Post
        I asked that very question to bridgeman, and I hope not to misquote him, but my recollection is that some of the technology (software/hardware) used to obtain the last extra bit of performance is proprietary, and licensed by others, so AMD can't replace those blobs. That would not be a big issue for desktop usage, but workstation users spend a ton of money on their cards, they use them professionally and they need as much performance as they can get. These folks would still need binary support for the foreseable future, and that means AMD can't leave the binary drivers behind.

        Hopefully, down the road they can build hardware free of these limitation (which is what Intel does), and therefore be able to fully switch to open source. One can only hope. In the meantime, I buy their hardware as a token of appreciation.
        The reality of the situation is that opening hardware and supporting open source drivers is expensive.

        For the 'linux enthusiast', that is somebody who is foolish enough (like you and me) to run Linux on their home PC is extremely low priority.

        Both ATI and Nvidia could completely and utterly abandon all support for Linux home users and basically see no impact, AT ALL, on their business lines.

        Linux home users and gaming just not important. This is something that needs to be kept in line. I know this is going to be hard for people here to understand this, but it's the reality.

        The reason that Nvidia and ATI supports Linux is because Linux is extremely popular on high-end graphical workstations. It's used heavily in movie production, scientific visualizations, 3d graphics, etc etc.

        People think that things like Final Cut and all that is important and widely used... but that is just mickey-mouse BS. It's common for people that make television commercials and that sort of thing, but for the high-end stuff linux is the popular platform. All these tools are very expensive and proprietary and often are not sold separate from the hardware.

        These people don't think twice to drop 2K on graphics hardware if they think it will make things go faster. That sort of stuff is relatively cheap when compared to the rest of their budgets and time is money.

        THAT is why Nvidia and ATI support Linux with proprietary drivers. You and I are completely secondary.

        This is a competitive environment and drivers DO make or break their competitiveness. They likelihood that they will open their drivers or prioritize open drivers is extremely small.

        The only way it's going to happen is if the open source folks are able to prove to ATI that it will provide a competitive advantage over Nvidia and a significant number of ATI's customers start prefering the open source drivers.

        That's it.

        Comment


        • #19
          The problem with trying to replace fglrx with a fully open driver is that we are able to offer Linux users much more in the way of features and performance by sharing code across multiple OSes than we would be able to do if the Linux driver effort was limited to what the Linux client PC market could justify.

          The downside of sharing code across OSes is that most of the other OSes are proprietary and require robust DRM, so the shared driver code needs to remain proprietary as well. We are looking at opening up a few more of the Linux-specific bits (maybe some kind of X server compatibility layer similar to the kernel compatibility layer we have today) but the shared code is going to have to stay proprietary.

          The non-obvious part is that there is a "diminishing returns" effect with driver development - you have seen what a small number of skilled developers can accomplish with the open source drivers and probably think it wouldn't take that much more work to out-do the proprietary driver, but that's unfortunately not the case. Every increment in functionality or performance raises the complexity of the driver and eventually the cost of every little improvement becomes incredibly high, so you might need 50-100x the development resources to get 2x the features and 2x the performance. The key point is that a code-shared proprietary driver only has to spend that money once (or 1.3 times, or whatever) for the entire PC market rather than once for each different OS.

          I still believe that the open source drivers will get sufficiently close to fglrx features and performance that most consumer users won't care about the difference and will be very happy with the open drivers, but for some other markets (3D workstation, hardcore gaming) the difference will probably always be important.

          I don't think we will end up holding back any programming info that has much effect on overall performance. We're definitely being stingier with the documentation in the areas which we feel are more sensitive, but I expect that in the end we're going to see the open drivers be only a few percent slower (if any) as a result of information holdback.

          Nearly all of the performance delta is going to come from the fact that the open drivers are maybe 200,000 lines of clean, maintainable code while the proprietary drivers are many millions of lines of relentlessly tweaked and optimized code.

          Comment


          • #20
            So basicaly the Radeon cards can easily be stressed to the max by the FLOSS drivers, but by means of insane optimisation it can be made much more efficient? Is that why it takes so much extra people to go to such insane lenghts just to squeze every last drop of juice out of the cards?

            Comment


            • #21
              Pretty much, yeah

              Comment


              • #22
                Originally posted by Otus View Post
                I don't really care that much about games, but anyone know how close this is to running Compiz?
                it already can, but - shadows doesnt work, and screen rotatin on cube - all broken, but windows wobling

                Comment


                • #23
                  Originally posted by bridgman View Post
                  I still believe that the open source drivers will get sufficiently close to fglrx features and performance that most consumer users won't care about the difference and will be very happy with the open drivers, but for some other markets (3D workstation, hardcore gaming) the difference will probably always be important.
                  Hopefully it will make people complain less anyway.

                  Comment


                  • #24
                    So Bridgman, if your fglrx team would have the same documentation (not counting DRM) as AMD is going to give to the public, how much of a performance decreas would we be talking about? Zero?

                    Comment


                    • #25
                      It really depends how much work the community puts into the open source driver. As John mentioned before, above a certain level the amount of driver work grows exponentially as performance increases.

                      Comment


                      • #26
                        Originally posted by agd5f View Post
                        It really depends how much work the community puts into the open source driver. As John mentioned before, above a certain level the amount of driver work grows exponentially as performance increases.
                        I read that. What I wanted to know is how, if any, the possible lack of documentation is going to be of impact on the speed of the FLOSS drivers, not counting optimisation work, which is why I brought the fglrx team into the equation.

                        Comment


                        • #27
                          Originally posted by bridgman View Post
                          The downside of sharing code across OSes is that most of the other OSes are proprietary and require robust DRM, so the shared driver code needs to remain proprietary as well.
                          I'd like to thank you for engaging the community, but I find this reasoning quite puzzling.

                          The only DRM that I can think of is the one related to the playback of BluRay and other HDCP protected media, and that surely doesn't impact the 3D driver.

                          It seems you should be able to open source both fglrx and the Windows driver except for the video acceleration+DRM parts relatively easily.
                          Besides, all video DRM seems completely cracked and BluRay releases seem widely available on public BitTorrent sites, so I don't think you could possibly make the situation worse for content producers.

                          If this is not the case, what am I missing?

                          The argument of wanting to keep the driver secret to not give a free gift to competitors seems instead much more understandable. Is this the real reason?

                          However, opening some things like your AMD IL -> r600+ ISA compiler/optimizer shouldn't give any competitive advantage to Intel/nVidia since they use completely different non-VLIW architectures.

                          It would be great and very interesting to have a much more detailed and accurate rationale for why you could or cannot open each component of fglrx/Catalyst.

                          And opening fglrx and your Windows driver would have several advantages:
                          1. Having bugs actually fixed by third parties, or at least more detailed bug reports
                          2. Rock solid Linux support and dominance of the Linux market (all Linux distributions and enthusiasts would be in the position to recommend buying ATI GPUs exclusively)
                          3. More mindshare among commercial game programmers, resulting in games that work better on your hardware. This would be due to the obvious advantage of being able to find out exactly what the driver is spending CPU time on and look intimately at what 3D calls are being translated to and how the hardware actually works.
                          4. Possibly getting commercial game developers to help you in tuning your drivers for their games (without needing cumbersome NDAs), resulting in better performance on your cards compared to competitors, especially in things like multi-GPU support

                          So, for parts of the driver where DRM, third party code, or competitive advantage concerns don't apply, I think you should consider opening the code (for both Linux and Windows).

                          Comment


                          • #28
                            Originally posted by V!NCENT View Post
                            So Bridgman, if your fglrx team would have the same documentation (not counting DRM) as AMD is going to give to the public, how much of a performance decreas would we be talking about? Zero?
                            the true is the opensource one can beat the FGLRX thats because FGLRX is a userspace driver!

                            you know? a kernel space driver do have less overhead.

                            fglrx do have a usersprace vram managment radeon do have kernel based vram managment..

                            Theoretical radeon can beat FGLRX very hard IF the radeon driver becomes the 'LOVE' of more DEV's

                            Comment


                            • #29
                              Originally posted by Agdr View Post
                              It seems you should be able to open source both fglrx and the Windows driver except for the video acceleration+DRM parts relatively easily.
                              you're making the assumption that fglrx is neatly modularized with clean inter-module APIs. I wouldn't count on that.

                              Then there's the legal review. If you follow the news, you'll know how much work legal review is for the OS drivers that are mostly written from a known amount of documentation. Try to imagine sifting through a few hundred thousand lines of fglrx code to determine if it contains any licensed code that mustn't be shared, uses any patents that aren't covered outside of fglrx or tells internals about the hardware which AMD wants to keep secret.

                              When AMD/ATI's linux strategy was formed, opening up parts of fglrx was discussed. But it was deemed easier to start from the scratch.
                              The result may not be as fast as fglrx, but it's cleaner, better structured code. To me, that's more important.

                              Comment


                              • #30
                                Originally posted by Qaridarium View Post
                                the true is the opensource one can beat the FGLRX thats because FGLRX is a userspace driver!
                                I wish it was. An untrusted user space driver confined to its own little sandbox, unable to lock up my system or stealing all my precious ram.

                                Unfortunately it's not.

                                Comment

                                Working...
                                X