Announcement

Collapse
No announcement yet.

R800 3D mesa driver is shortly before release

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

  • Originally posted by monraaf View Post
    I remember the same thing being said about Evergreen. The ISA was mostly the same as R600/R700, just with a few new instructions added and some registers moved. Other than that the changes were mostly in the display controller part.
    Yeah, that's what we were told. We're digging a bit deeper this time

    Comment


    • Originally posted by monraaf View Post
      Let's be a little realistic about this. AMD's open source driver team is heavily understaffed, and unless they are going to do something about that
      They've been pretty explicit that their plan is for the community to get more involved, not for them to spend more money on staff.

      Comment


      • Originally posted by monraaf View Post
        I remember the same thing being said about Evergreen. The ISA was mostly the same as R600/R700, just with a few new instructions added and some registers moved.
        That's still mostly true, but the devil is in those little details. The driver came up much faster, but sorting out the little changes was the tough part. The documentation still has to go through IP review regardless.

        Comment


        • Originally posted by elanthis View Post
          They've been pretty explicit that their plan is for the community to get more involved, not for them to spend more money on staff.
          To be precise, what we're saying is that:

          - with the *current* level of staffing and community involvement there has been much more progress than is immediately obvious, as a consequence of (a) catching up on several years of new hardware introduction and (b) community focus on moving to a new architecture for the driver stack rather than adding features & performance to the old architecture

          - with the *current* level of staffing and community involvement we are approaching the point where we can be considered "caught up" and are working entirely on new HW support rather than backfilling for old HW and working on architectural changes

          - with the *current* level of staffing and community involvement we expect that you will see more timely support for new hardware

          Comment


          • Originally posted by bridgman View Post
            - with the *current* level of staffing and community involvement we expect that you will see more timely support for new hardware
            Once you have caught up to support hardware near the release date, is that what you expect AMD employees will be focusing on? Or will they also be able to help implement new features as well? (I'm thinking figuring out how to accelerate 4.0 tesselation on ATI hardware, for example).

            Comment


            • Both, although doing enough to at least get the existing level of functionality running on new hardware should probably be the first priority.

              Comment


              • Originally posted by trapxvi View Post
                This has nothing to do with typed vs. untyped lambda calculus. If you want to write a type signature for emulate, it would be State⊗Integer → State⊗Integer. In general, typed lambda calculus is less expressive than untyped lambda calculus, since the latter can obviously express any well-typed expression. In this case, the typedness of the lambda calculus you're using has nothing to do with this.



                No values are changing at all. He's expressing an algorithm by recursion. At cycle zero, the machine is in a known initial state. To this we apply the emulation operation which maps the old state to a new state and increments the number of cycles. If you wanted to expand this out, you'd have:
                emulate(emulate(emulate(...emulate(initial_state, 0))...)))
                alternatively, in a language like C, you'd have:

                cycle_count = 0;
                state = initial_state;
                while(1) state = emulate(state, &cycle_count)); // emulate increments cycle_count
                These are completely equivalent computations. In both cases, there's a serial data dependency: you can't compute the next state until you know the current state. Expressing this in lambda calculus or a purely functional style does not in any way remove this fundamental computational property. It would be like claiming that Fermat's last theorem is true in a base-12 number system but not in base-10: it's a mathematical property that's completely independent of a particular representation.

                If you claim to have a faster-than-serial way to solve this emulation problem, congratulations, you've solved the halting problem and disproved one of the most fundamental theorems in computer science.

                Quad, I respect you for your ability to find interesting info. I'm not sure if you're trolling or just confused, but you need to give this up.

                i found something: http://www.heise.de/tp/blogs/3/148161


                this means "P!=NP" is maybe not valid because no one check this Theory

                this also means a Quantum computer maybe can turn an P!=NP into P=NP

                so if no one can valid "P!=NP" up to 100% then a speed up by parralism can be unlimited-

                Comment


                • @Qari: Real people talk about real hardware available at the time of the talk or in a reasonable period of time. You are talking about highly theoretical stuff while you apparently have no clue of real hardware.

                  Comment


                  • Originally posted by bridgman View Post
                    Actually no, I'm saying you're wrong today

                    CPUs are directly connected to the memory controller, but (a) can use HT links to access memory connected to another CPU, and (b) can respond to access requests over HT from attached IGPs and other peripherals. Normal CPU memory accesses do not go through HT.
                    ok i'm wrong...

                    simple question: can the APU(gpu) fusion access over a ht link into ram to another fusion cpu in a multisocket system ? to enable functions like Crossfire?

                    Comment


                    • Originally posted by Lynxeye View Post
                      @Qari: Real people talk about real hardware available at the time of the talk or in a reasonable period of time. You are talking about highly theoretical stuff while you apparently have no clue of real hardware.
                      what if you need 'no clue' to be free to think free and without blocking limitations ?

                      Comment


                      • Q, P=NP has nothing to do at all with halting problem which in term has nothing to do with P=NC.
                        P=NP is about whether a certain set of problems are solvable with polynomial time (might be possible to be done, who knows).
                        Halting problem is whether you can have a machine that can draw the conclusion that an algorithm will loop indefinitely (can't be done with a Turing machine ie any currently thinkable computer).
                        P=NC is the one about parallelism. Don't mix them up.

                        Comment


                        • Originally posted by nanonyme View Post
                          Q, P=NP has nothing to do at all with halting problem which in term has nothing to do with P=NC.
                          P=NP is about whether a certain set of problems are solvable with polynomial time (might be possible to be done, who knows).
                          Halting problem is whether you can have a machine that can draw the conclusion that an algorithm will loop indefinitely (can't be done with a Turing machine ie any currently thinkable computer).
                          P=NC is the one about parallelism. Don't mix them up.
                          the halting problem is only valid on determined systems but quantum computers are not Determinism systems...

                          if you build an Sedenion-dimension-quantum-computer you get an high probability of the result in one of the dimensions on the first cycle.

                          Comment


                          • what if you need 'no clue' to be free to think free and without blocking limitations ?
                            If your premise is false, you can prove *anything*. You seem to be advocating the idea that knowing less about a subject makes you more likely to succeed in it.

                            Any Lynxeye's comments regarding classical vs quantum computing still stand. Nothing is added to a discussion by saying "If you build this highly theoretical computer then it might solve this problem but it might not."

                            Comment


                            • Originally posted by archibald View Post
                              If your premise is false, you can prove *anything*. You seem to be advocating the idea that knowing less about a subject makes you more likely to succeed in it.

                              Any Lynxeye's comments regarding classical vs quantum computing still stand. Nothing is added to a discussion by saying "If you build this highly theoretical computer then it might solve this problem but it might not."
                              there is one theoretical solution for the waiting for an future calculating problem.

                              random an number and you get a tiny chance to get the same result.

                              on an Sedenion based system you can have 16 dimensions of random numbers in the same time.

                              you only need to valid the result in the future after the calculation.

                              means if you calculate more than 1 problem you can get results faster by checking multiple speculativ versions in the same time and get the result faster than on an normal way.

                              Comment


                              • Originally posted by Qaridarium View Post
                                You mean this paper: http://www.scribd.com/doc/35539144/pnp12pt

                                this means "P!=NP" is maybe not valid because no one check this Theory
                                The paper has already been proved wrong

                                this also means a Quantum computer maybe can turn an P!=NP into P=NP
                                This has something to do with classes. Quantum computers can only divert superposition into 1 or 0. They cannot turn the mathematic of formulas upside down xD

                                Originally posted by Qaridarium View Post
                                what if you need 'no clue' to be free to think free and without blocking limitations ?
                                You actualy managed to say something truly awesome. Why did you hack Qaridariums account?

                                Comment

                                Working...
                                X