Announcement

Collapse
No announcement yet.

Linux 5.11 Drops AMD Zen Voltage/Current Reporting Over Lack Of Documentation

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

  • #21
    Originally posted by sandy8925 View Post
    Yeah, I agree. Nothing's really changed from a decade ago. Some shit is kind of supported, some of it works, some of it is very buggy and won't be fixed for several months (or ever) and also, every tiny thing is "proprietary information that could destroy the company" so it won't be documented or supported.
    Complaining about/to agd5f is missing the point - he works long hours negotiating approvals (as I did before he took over as IP lead) to release GPU-related information following processes that took almost a decade to build into the organization, and which many companies still don't have.

    Without all that the situation is exactly like you say - every tiny thing is considered "proprietary information that could destroy the company" by management so it doesn't get documented or supported without a massive amount of effort. That should not be a surprise to you and it's in no way an AMD-only thing - if anything AMD's default mindset is less paranoid than that of most companies.

    We don't have the same people or processes on the CPU side yet - we lost them during the layoffs a few years back and it's one of the gaps we still need to close.

    Originally posted by sandy8925 View Post
    Funny, the my initially said they won't support UVD on Linux due to DRM concerns, but it seems that DRM on the GPU only started with TMZ support, so that claim from them sounds like utter crap.
    I don't understand. I initially said that you (customers) should assume that we will not be able to support UVD on Linux because of DRM concerns on Windows (it's the same HW so if we expose something for Linux that could compromise DRM on Windows we have a big problem).

    Challenges exposing UVD on Linux had nothing to do with DRM on Linux, which as you say is just starting to show up recently.
    Last edited by bridgman; 22 December 2020, 07:49 PM.
    Test signature

    Comment


    • #22
      Originally posted by duby229 View Post
      Stating that it's not exactly critical implies that you don't care, and that after you admit that it would be nice to have... That whole last sentence should have been left out. You should have just stated that there is no one to do it at the moment (and stated a reason why no-one can do it at the moment) and left it at that...
      Not critical means it's not a top priority, not that it isn't important. It was stated that there is a lack of available devs to allocate time towards this atm, they've got other work that's considered higher priority. AFAIK they have devs that focus on linux support, so presumably that would mean they're still working on something useful, just higher priority, no idea what that task is but if we knew you'd probably also find someone upset if it was considered low priority.

      Critical also means that shit can hit the fan if it's not addressed, and that the impact could affect a wide audience of their product(s). Power info is useful and a nice to have, it's not that it doesn't have benefits that can be valuable and important, it's just that it's low priority in the sense that it's not going to prevent the CPU from working. I don't manage servers, but I assume that EPYC and ThreadRipper models are a market AMD cares quite a bit about and you'll not find as many of those customers messing with power settings or running those things on batteries like a consumer with a laptop. Consumers are important too I'm sure, but the amount of linux ones statistically may not be enough to justify whatever projected cost/effort working on this feature would require vs the ROI on higher priority tasks.

      Personally, I'm confused with the issue being a lack of free talent available. This is AMD, they're unable to allocate some funds towards hiring a new dev or sponsoring a third-party to tackle it? It's the one area that I think Intel laptops can handle better by extending battery life (I think PSR support for AMD only became available this year or last? That's been supported by Intel for almost a decade? with PSR2 being a successor).

      ---

      TL;DR: agd5f probably does care, and would like to see it worked on as he expressed. Just that his hands are tied on prioritizing it. Best he can perhaps do is express demand/interest from it from some vocal online linux users, but that's unlikely to be sufficient to justify prioritizing over whatever is currently considered more worthwhile to work on.

      Comment


      • #23
        AMD position is: they earn 60% with linux enterprise cloud/server and 39% with windows Desktop
        and only 1% linux desktop market share.
        bridgman already told us that they already spend 100% more on linux desktop than the marketshare of 1% would have if they do not count in the linux people in the tech support and decision makers are all linux lovers. means they spend on linux like it has 2% market share instead of 1%.

        how much %%% market share do we need to have a driver-gui like in windows or Voltage/Current Reporting

        ??? i think 5%-10%---

        i think AMD also agree that according to steam statistic linux people spend much more money on hardware than the average windows people for example average windows vram is 1gb and average linux vram is 8GB... linux people also have more cores and more mhz and more ram and so one ans so one.

        and i think these companies really do not understand the market because imagine the linux-desktop support would be better the linux people would not stay on old hardware just because the new hardware has no compatibility with linux and any new released hardware need 6-12 month to get a usefull support level. means in fact linux people would buy even more highend if the support would be in place.

        also amd should spend more money on legacy product support to make amd used hardware more valuable to make sure the brand is well in place and people can sell old hardware on ebay to spend the money from this on new hardware.

        the market right now is overheating have fun find any 6800 6800XT or 6900 i really watch daily and there is non to buy. so i still use my vega64 because the upgrade to 5700xt would only 10% faster and a 6800xt would make it 107% faster.

        market is overheaded if they can not produce enough new products they really should improve the software/driver for legacy old hardware. this would be the only way to make the situation better than now. a good example for this would be VGA video output for the GCN 1.0 and 1.1 gpus. to switch to the AMDGPU driver for this old hardware to bring vulkan support to the masses.

        but just tell people to buy new hardware does not work out right now because even if you try hard and you pay high price like 850€ for a 6800 you hardly get any.

        if you ever ask why linux only has 1% desktop market share? well the answer could be the lag of desktop linux hardware support features from amd and the lag of legacy hardware driver supports from amd...
        Last edited by qarium; 22 December 2020, 08:07 PM.
        Phantom circuit Sequence Reducer Dyslexia

        Comment


        • #24
          Look at the number of entitled pricks in this thread...

          We should be grateful that most AMD hardware are now capable of working properly in Linux with proper upstream drivers, and yet all I see are complaints about AMD treating Linux users as second-class citizens.

          If AMD really treated Linux users as second class citizens, they won't never have bothered to rewrite and release the amdgpu driver in Linux-native code, and we'd still be using radeon and radeonsi.

          Comment


          • #25
            Originally posted by Sonadow View Post
            Look at the number of entitled pricks in this thread...

            We should be grateful that most AMD hardware are now capable of working properly in Linux with proper upstream drivers, and yet all I see are complaints about AMD treating Linux users as second-class citizens.

            If AMD really treated Linux users as second class citizens, they won't never have bothered to rewrite and release the amdgpu driver in Linux-native code, and we'd still be using radeon and radeonsi.
            Didn't AMD not release Vulkan drivers for us the same day they did for Windows?
            Where's Ryzenmaster for Linux if we're not 2nd class?

            I think it's understandable why we are not first class with our market share, but I don't see the benefit of denying it.


            Originally posted by bridgman View Post
            We don't have the same people or processes on the CPU side yet - we lost them during the layoffs a few years back and it's one of the gaps we still need to close.
            I know you're from ATI, but can you somehow help jumpstart at least that discussion?
            Last edited by geearf; 22 December 2020, 09:07 PM.

            Comment


            • #26
              Originally posted by polarathene View Post

              Not critical means it's not a top priority, not that it isn't important. It was stated that there is a lack of available devs to allocate time towards this atm, they've got other work that's considered higher priority. AFAIK they have devs that focus on linux support, so presumably that would mean they're still working on something useful, just higher priority, no idea what that task is but if we knew you'd probably also find someone upset if it was considered low priority.

              Critical also means that shit can hit the fan if it's not addressed, and that the impact could affect a wide audience of their product(s). Power info is useful and a nice to have, it's not that it doesn't have benefits that can be valuable and important, it's just that it's low priority in the sense that it's not going to prevent the CPU from working. I don't manage servers, but I assume that EPYC and ThreadRipper models are a market AMD cares quite a bit about and you'll not find as many of those customers messing with power settings or running those things on batteries like a consumer with a laptop. Consumers are important too I'm sure, but the amount of linux ones statistically may not be enough to justify whatever projected cost/effort working on this feature would require vs the ROI on higher priority tasks.

              Personally, I'm confused with the issue being a lack of free talent available. This is AMD, they're unable to allocate some funds towards hiring a new dev or sponsoring a third-party to tackle it? It's the one area that I think Intel laptops can handle better by extending battery life (I think PSR support for AMD only became available this year or last? That's been supported by Intel for almost a decade? with PSR2 being a successor).

              ---

              TL;DR: agd5f probably does care, and would like to see it worked on as he expressed. Just that his hands are tied on prioritizing it. Best he can perhaps do is express demand/interest from it from some vocal online linux users, but that's unlikely to be sufficient to justify prioritizing over whatever is currently considered more worthwhile to work on.
              Thank you for the level headed response.

              So I have basically two issues, the first one is that Marketshare on linux is not as low as for whatever reason people claim it is. There is plenty of evidence that it is just over 2%, ok so that leaves 98% to other OSes sure, but it's still higher than most people claim. Additionally you can't just look at it from a percentage, you have to look at it from a raw numbers perspective and 2% is still millions of users. And Linux users tend to buy higher end products with bigger margins. And additional to that those same millions of high end users are also the same ones working professionally with enterprise linux for a living.

              I'm absolutely positive Linux users contribute a much larger percentage of AMD's profit than gets recognized by most people. Especially now that we have really good graphics drivers to use.

              The second problem that I have is that Alex has expressed multiple times now that he doesn't really care about linux support. And you might say who cares if he doesn't care, except he has a really high level position where caring about linux support is critical. You might also say that he never did express that he doesn't care, but come on, read between lines, what he definitely did say was that proper reporting of critical metrics isn't critical. If that doesn't mean he doesn't he doesn't care, then what does it mean?

              Comment


              • #27
                Originally posted by Sonadow View Post
                Look at the number of entitled pricks in this thread...

                We should be grateful that most AMD hardware are now capable of working properly in Linux with proper upstream drivers, and yet all I see are complaints about AMD treating Linux users as second-class citizens.

                If AMD really treated Linux users as second class citizens, they won't never have bothered to rewrite and release the amdgpu driver in Linux-native code, and we'd still be using radeon and radeonsi.
                You don't even know what you're saying. First of all radeon -is- a native linux driver, second amdgpu -is- a fork of radeon but with a more sophisticated winsys (for newer hardware). Third radeonsi is a opengl mesa driver for -both- of those kernel drivers.
                Last edited by duby229; 22 December 2020, 10:23 PM.

                Comment


                • #28
                  Originally posted by sandy8925 View Post

                  Well in Linux 5.10, Vcore reports a constant 1.6 V for some reason, while Vsoc is reporting the correct voltage. This wasn't the case before Linux 5.10
                  ​​​​​
                  That's strange, I am on 5.9, so yeah.

                  fenixex
                  As for values on my particular system, they are correct for sure, not only because I disable PB anyway, but also because it doesn't boost that high at all (in terms of clocks, but it does in terms of voltage = hence, better to disable it completely since it is next to useless).

                  Comment


                  • #29
                    Originally posted by sandy8925 View Post

                    Cut the bullshit. This is extreme important information, especially given AMDs weird boost behaviour. Using a CPU like the 3900X to it's full extent, requires the ability to see voltage, current, temperature etc. from userspace. This is very much critical functionality.

                    If this support was missing on Windows, AMD would get terrible reviews. So cut the pure bullshit.

                    I get that you can't criticize your employer, doesn't mean you have to come in here and defend every shitty thing they do.
                    Exactly, if Alex said this on a popular hardware review sites forum like Anandtech, he'd get flamed so hard. It would be so bad.

                    Comment


                    • #30
                      Originally posted by sandy8925 View Post
                      Cut the bullshit. This is extreme important information, especially given AMDs weird boost behaviour. Using a CPU like the 3900X to it's full extent, requires the ability to see voltage, current, temperature etc. from userspace. This is very much critical functionality.
                      All this information is useless for the vast majority of users. The only thing they need to know is that their processors are working as they are expected to. So get off your damn pedestal thinking that YOU decide what is critical information and what isn't.

                      I use a 2990WX headless and I don't give one shit about voltage, current, temperature or other bullshit 'parameters'. All it needs to do is to run FFMPEG or compile daily Chromium builds with all 64 threads in use.

                      People are getting too damn entitled here. AMD should just drop all their open source support this instant, abandon every single open source driver they have ever worked on and cease adding support to any new products in the existing OSS drivers to show the users who's boss. Let the so-called community pick up the pieces themselves if they think they are so great, like what is going on with Nouveau.
                      Last edited by Sonadow; 22 December 2020, 10:51 PM.

                      Comment

                      Working...
                      X