Announcement

Collapse
No announcement yet.

The Open-Source ATI Driver Is Becoming A Lot Faster

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

  • #51
    Originally posted by marek View Post
    We very well know what he was talking about and he said nothing new to us. Some of us used to be hobbyist or professional game/3D graphics/CUDA developers. We know our stuff. It's also pretty obvious that he is a coder with a good deal of skills.
    Hi, sorry I didn't mean to be disrespectful! It's just that he mentioned that a lot, even professionals, make these mistakes. I'm really happy that you and the rest are aware of the best design for the future.

    Much respect!

    Comment


    • #52
      Originally posted by L33F3R View Post
      raise your hand if you bout an r300 card in the last 5 years.
      I bought 2, an x1950 pro agp card and an x1900 xtx pcie, both second hand and therefore quite cheap.

      Bought them both based on the progress of the open ati drivers and have never regretted it.

      Comment


      • #53
        A Bit Of A Misrepresentation, As Usual

        Originally posted by DeepDayze View Post
        It is so sad to see people like gordboy posting his fglrx fanboi drivel. No wonder the IQ drops whenever he comes into a thread and spouts some meaningless dribble-laden idiocy.

        The FOSS ATI drivers are coming along nicely and there's hope yet for my old Thinkpad T42 with the Radeon Mobility 7500 chip. The devs should be commended for their work.
        I must be the only fglrx fanboy who isn't using fglrx. It certainly makes you wonder about the rest of the meaningless and unnecessary insults.

        I'm not going to bother answering any of the rest of the idiocy, except to point out that the article was clearly dishonest. There is NO WAY that the figures presented support the hype both in the precis, and in the actual article.

        The bit that really made me wonder exactly what was going on, was this:

        "some major performance boosts for the open-source ATI driver as it nears the level of performance of the proprietary Catalyst driver"

        There is something very sad about a bunch of self-appointed experts taking it upon themselves to try to discredit well-founded criticism. But it seems to be par for the course here.

        Comment


        • #54
          Originally posted by Drago View Post
          Very interesting read indeed. I think this MESA multi-threading topic needs more talking and less coding for now. Just to clear things out.
          it seems that many readers might like to watch this MIT video lecture on threading algorithms etc to help get a better understanding, he did use the word "paradigm" but dont let that put you off ,the video is good non the less .

          This is the thirteenth post in an article series about MIT's lecture course "Introduction to Algorithms." In this post I will review lectures twenty and twenty-one on parallel algorithms. These lectures cover the basics of multithreaded programming and multithreaded algorithms. Lecture twenty begins with a good...

          "This is the thirteenth post in an article series about MIT's lecture course "Introduction to Algorithms." In this post I will review lectures twenty and twenty-one on parallel algorithms. These lectures cover the basics of multithreaded programming and multithreaded algorithms.

          Lecture twenty begins with a good overview of multithreaded programming paradigm, introduces to various concepts of parallel programming and at the end talks about the Cilk programming language.

          Lecture twenty-one implements several multithreaded algorithms, such as nxn matrix addition, nxn matrix multiplication, and parallel merge-sort. It then goes into great detail to analyze the running time and parallelism of these algorithms."

          Comment


          • #55
            Originally posted by popper View Post
            it seems that many readers might like to watch this MIT video lecture on threading algorithms etc to help get a better understanding, he did use the word "paradigm" but dont let that put you off ,the video is good non the less .

            This is the thirteenth post in an article series about MIT's lecture course "Introduction to Algorithms." In this post I will review lectures twenty and twenty-one on parallel algorithms. These lectures cover the basics of multithreaded programming and multithreaded algorithms. Lecture twenty begins with a good...

            "This is the thirteenth post in an article series about MIT's lecture course "Introduction to Algorithms." In this post I will review lectures twenty and twenty-one on parallel algorithms. These lectures cover the basics of multithreaded programming and multithreaded algorithms.

            Lecture twenty begins with a good overview of multithreaded programming paradigm, introduces to various concepts of parallel programming and at the end talks about the Cilk programming language.

            Lecture twenty-one implements several multithreaded algorithms, such as nxn matrix addition, nxn matrix multiplication, and parallel merge-sort. It then goes into great detail to analyze the running time and parallelism of these algorithms."
            "(A thread is defined to be a maximal sequence of instructions not containing the parallel control statements spawn, sync, and return, not something like Java threads or Posix threads.)

            I put the most important keywords in bold. They are "spawn" and "sync".

            The keyword spawn is the parallel analog of an ordinary subroutine call. Spawn before the subroutine call in line 3 indicates that the subprocedure Fibonacci(n-1) can execute in parallel with the procedure Fibonacci(n) itself. Unlike an ordinary function call, where the parent is not resumed until after its child returns, in the case of a spawn, the parent can continue to execute in parallel with the child. In this case, the parent goes on to spawn Fibonacci(n-2). In general, the parent can continue to spawn off children, producing a high degree of parallelism.

            A procedure cannot safely use the return values of the children it has spawned until it executes a sync statement. If any of its children have not completed when it executes a sync, the procedure suspends and does not resume until all of its children have completed. When all of its children return, execution of the procedure resumes at the point immediately following the sync statement. In the Fibonacci example, the sync statement in line 5 is required before the return statement in line 6 to avoid the anomaly that would occur if x and y were summed before each had been computed.

            The spawn and sync keywords specify logical parallelism, not "actual" parallelism. That is, these keywords indicate which code may possibly execute in parallel, but what actually runs in parallel is determined by a scheduler, which maps the dynamically unfolding computation onto the available processors."

            Comment


            • #56
              Originally posted by gordboy View Post
              I must be the only fglrx fanboy who isn't using fglrx. It certainly makes you wonder about the rest of the meaningless and unnecessary insults.
              It makes me wonder about your autism and motivation for mocking. If you think, in contrast to a lot of people who use it, that it's worthless, then why do you keep checking upon it?

              There is something very sad about a bunch of self-appointed experts taking it upon themselves to try to discredit well-founded criticism.
              There's something very sad about your trolling. What part of basic math do you not understand? Gallium 0.4 != Gallium 1.0. It means that development is underway to make good FLOSS drivers. This article clearly indicates the advances made from previous r300g state of performance to the current r300g state of performance.

              What's so extremely notable here is that there are a handful FLOSS drivers devs versus a factor 100 or more fglrx devs that are still reaching respectable levels of performance given the fact and they are taking every mainstream desktop GPU upon them instead of just one.

              Comment


              • #57
                Broken Promises And More Thuggery

                Vincent

                I hate to break it to you, but you and your AMD fanboy chums are very much in the minority in the linux world. This forum has turned into an AMD love-in. The fact that bridgman is the top contributor says it all.

                Anyways, the reason for the repeated complaints is the repeated FAILURE to deliver a properly working graphics driver stack for ATI cards. I'm not just talking about performance. I'm talking about basic functionality being broken.

                I sincerely hope that Xorg *do* pull the DDX back in. That is the only way that the current crop of pathetic underperformers will ever be replaced, by the looks of it.

                The fact that Michael is bewitched by these charlatans and snake-oil merchants, is sad enough. But the fact that he seems to have an army of teenage thugs, ready to smite anyone who dares to complain, is perhaps the saddest thing of all.

                Does anyone here actually care about linux ? Or are you happy to be unpaid stooges for a sordid corporate strategy whose inevitable endgame is the attempted final destruction of linux itself ?

                Comment


                • #58
                  really? cause i'm pretty sure this forum is still quite divided, seems every topic i see on the matter has a few people complaining about how horrible AMD is and a few saying they don't have any problems.

                  regardless of what you have to say though, you'd probably be much better recieved if you tried some level of respect. calling everyone "stooges" and "teenage thugs" doesn't really impress people.

                  Comment


                  • #59
                    For a moment I feared they removed tags from vBullettin 4, fortunately they didn't
                    ## VGA ##
                    AMD: X1950XTX, HD3870, HD5870
                    Intel: GMA45, HD3000 (Core i5 2500K)

                    Comment


                    • #60
                      Originally posted by gordboy View Post
                      Vincent

                      I hate to break it to you, but you and your AMD fanboy chums are very much in the minority in the linux world. This forum has turned into an AMD love-in. The fact that bridgman is the top contributor says it all.

                      Anyways, the reason for the repeated complaints is the repeated FAILURE to deliver a properly working graphics driver stack for ATI cards. I'm not just talking about performance. I'm talking about basic functionality being broken.

                      I sincerely hope that Xorg *do* pull the DDX back in. That is the only way that the current crop of pathetic underperformers will ever be replaced, by the looks of it.

                      The fact that Michael is bewitched by these charlatans and snake-oil merchants, is sad enough. But the fact that he seems to have an army of teenage thugs, ready to smite anyone who dares to complain, is perhaps the saddest thing of all.

                      Does anyone here actually care about linux ? Or are you happy to be unpaid stooges for a sordid corporate strategy whose inevitable endgame is the attempted final destruction of linux itself ?
                      Bahahaha.

                      4.5 out of 5!

                      Comment

                      Working...
                      X