Announcement

Collapse
No announcement yet.

Digging Deeper Into AMD's UVD Code Drop

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

  • #81
    Come on, leave Bridgman a break. Firmware initializes hardware bits. Mark this very well. You don't need its sources to *USE* your GPU, in the *FOSS* world. You were given almost the entire chip (except PM), programming model, and you are still unsatisfied.

    Comment


    • #82
      Originally posted by bridgman View Post
      Other than giving away functionality for free that vendors charge extra for today, and losing a big pile of money in the process, what alternatives are you (I mean "everyone posting here", not just you) suggesting ?
      Selling each piece of hardware at its full potential. Artificial limitations in firmware are the issue.

      It's dishonest to sell the exact same hw with crippled firmware; if the hardware were physically crippled, then it would be ok. Whether this comes from natural breakage like defects in some shader group, or from designing a lower-end product in the first place doesn't matter. E-fuses on the other hand I would count in the dishonest category, the hw came full-featured from the factory.

      I suppose the point of contention is the current practise of segmenting the area with physically the same product. Even if it may get more $$$ to the producer, most consumers see it as wrong. Best case it would get banned by the consumer protection laws, since obviously nobody will voluntarily cut their income stream.

      Comment


      • #83
        Originally posted by bridgman View Post
        I would say no, because we are selling the hardware/software combination and those *are* different. If you strip both products down to the bare chips, ignore the driver and ignore fused-off features you can say that the major components are the same or similar, but you could say the same about a VW Toureg and a Porsche Cayenne in similar configurations.
        I think this is where we disagree. I would say it's like selling two identicle Porsches, but the shift gate on one of them has 5th gear blocked off.

        Comment


        • #84
          I haven't heard a viable alternative yet.

          Since vendors generally can not afford to cut their income dramatically while maintaining all the same R&D costs, that implies the resulting cost of a "single SKU with all features" product would be considerably *higher* than the current low end offering. Would this really be an improvement ?

          Saying "oh yes it would be better if we got the high end features at the low end price and HW vendors should be forced to give us that and give up the extra revenues" is not really a valid option. You don't *really* get something for nothing despite what generations of marketers try to tell you.
          Test signature

          Comment


          • #85
            Originally posted by benmoran View Post
            I think this is where we disagree. I would say it's like selling two identicle Porsches, but the shift gate on one of them has 5th gear blocked off.
            I don't think your analogy really works other than sounding good (and maybe aligning with an emotional response)... it suggests little or no extra cost for development & qualification on that extra cost feature (5th gear), which is not the case at all for most of the real world product scenarios we are discussing here.

            Then again it's arguably my fault for drifting into car analogies

            Let's run with your analogy for a minute anyways. Finance says you need an average $100K selling price to cover costs and make a small profit. Rather than selling a single model at $100,000 you sell the 4-speed version for $80,000 (with an appropriate final drive ratio) and the 5-speed for $120,000, appealing to two different groups of customers. I'm assuming equal per-model sales for simplicity.

            All is well, except a group of people on a popular car forum think there's something wrong with the practice and that you should be forced to offer only a single model with all 5 speeds enabled, and assume they would get that model for $80,000. What do you tell them ?
            Last edited by bridgman; 09 April 2013, 07:41 AM.
            Test signature

            Comment


            • #86
              Originally posted by bridgman View Post
              "single SKU with all features".
              You keep on banging that drum. But that is a strawman. What you say is not demanded by the FSF. They want only that the firmware is free software.

              The issue of users taking control of the functions that their hardware has was only brought up in response to your claim
              Originally posted by bridgman View Post
              Just to be clear, you're saying that microcode permanently stored on the chip is "better" than the same microcode loaded at boot, right ? I realize this is a commonly held position but nobody has ever been able to rationalize it other than by saying "Stallman says" or "if the driver doesn't have to load it then we don't care if it's non-free".
              Which is not the truth.

              Comment


              • #87
                Originally posted by chithanh View Post
                You keep on banging that drum. But that is a strawman. What you say is not demanded by the FSF. They want only that the firmware is free software.
                I'm not bringing it up myself, I'm responding to other people's posts about hypothetical scenarios. We stopped talking about FSF a lot of posts ago.

                To your "this is not the truth" comment (which I don't agree with, obviously), someone mentioned that a possible reason for writable microcode being "worse" than ROM'ed microcode was using different microcode to provide different feature sets. That led to the following :

                "what AMD should do is have different drivers for different features"
                "that's what we do"
                "that's evil and bad"

                ... and that's what the discussion was about. Try to keep up
                Test signature

                Comment


                • #88
                  Non-free firmware in ROM is better for the user because it makes it harder for the vendor to do market segmentation. Market segmentation is used to capture consumer surplus, ie. make users pay more money.

                  The FSF would probably agree that differentiating in drivers is totally ok, as long as these drivers are free software.

                  Comment


                  • #89
                    Having firmware in ROM actually makes it easier to do market segmentation.

                    If you do segmentation via free drivers then you don't really have effective segmentation.
                    Test signature

                    Comment


                    • #90
                      Of course it can be effective. You don't need to provide support for custom user configurations. Digium and others do this very successfully.

                      Free drivers and firmware in ROM will level the playing field between manufacturer and customer.

                      Comment

                      Working...
                      X