Announcement

Collapse
No announcement yet.

XWayland Gets Initial Support For EGLStreams To Support NVIDIA's Driver

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

  • #51
    Originally posted by cybertraveler View Post
    [*]Do you have a GeForce 7 series card? Well "Nvidia no longer supports drivers for your card. This means that these drivers do not support the current Xorg version." So guess what: your options are now:
    You must be kidding. Geforce 7 are GPUs presented in 2005. Same time as Radeon x1*00 series. Do you realy need a modern hiperf driver for OpenGL 2.1 card?

    Comment


    • #52
      Originally posted by cj.wijtmans View Post

      Whats the difference between hardware forcing you to run certain software and software forcing you to run certain hardware? You are the other side of the shekel mate.
      Forking this!

      Comment


      • #53
        When i read all the comments i always get the feeling people do odd comparisons on linux and nvidia. Nvidia is a huge company, for sure, but linux is a big thing too. Don't consider just the desktop, consider all the other usecases. It's quite a fact, that nvidia needs to present working drivers. It's not that linux has to got the route of nvidia.

        This is what happened:
        Nvidia cheated on linux, once again, blocking progress. They did this by pushing their own mediocre, semi-standardized stuff, not agreeing on all others. They don't care. They want to be assholes with wrong policies.

        What helped them:
        Resistance from the people that know better isn't organized. Nvidia has paid people and economical partnerships, which worked to get their stuff in. It took quite a while, but nvidia won. In a fair and ideal world, this should never have happened. And if linux people would all be lazy like that, AMDs DC Code would have been merged on day 1. Or intel wouldn't have removed MIR support from their drivers long ago, they tried to do the right thing.

        I still think this should be stopped. I think this should never be merged. Nvidia will react, if they really need to. They need to. Their policies are orientated strong backwards. Maybe someone should start a petition.

        Comment


        • #54
          Originally posted by Khrundel View Post
          You must be kidding. Geforce 7 are GPUs presented in 2005. Same time as Radeon x1*00 series. Do you realy need a modern hiperf driver for OpenGL 2.1 card?
          I need a driver that makes my old card work at its best AND I need to keep my system up to date at the same time. I won't be playing the latest Hitman on that rig, obviously.

          Now the radeon driver can drive the x1*00 series with the upcoming 4.16 kernel, the upcoming Mesa 18.x and all the rest at full speed, and it does so out of the box. What about the nvidia driver with a GeForce 7800 GTX?
          Last edited by lucrus; 08 February 2018, 07:07 PM.

          Comment


          • #55
            Originally posted by Hibbelharry View Post
            ........
            This is what happened:
            Nvidia cheated on linux, once again, blocking progress. They did this by pushing their own mediocre, semi-standardized stuff, not agreeing on all others. They don't care. They want to be assholes with wrong policies.
            .....
            Not true. NVIDIA is asking for certain features which have to do with creating buffers optimized in format and size for a variety of use cases. EGL streams is just one option which was favored initially, and NVIDIA probably didn’t expect so much resistance to that proposal. NVIDIA would be fine with those features added to GBM, or with some new tech designed in collaboration.

            Comment


            • #56
              Originally posted by cybertraveler View Post
              I expect the setup experience is going to be messy and limiting for almost all the less popular distros, but even popular distros like Arch have special steps that users have to follow:
              https://wiki.archlinux.org/index.php/NVIDIA

              So I'm not just considering the short term immediate needs of a new GNU/Linux user. I want to set them on a path which will likely give them good results now and for years to come.

              ARCH SUPPORT
              Arch is popular yes, but if you know anything about the community and way about going to install it(excluding things like Antergos, we're talking specifically about Arch here), you're generally expected to have technical competency that you understand there will be more effort involved and knowledge required, thankfully they've got great Arch Wiki for documentation. If you go onto their forums or subreddit, you'll find them not too friendly if you ask anything "noobish" that demonstrates little effort/knowledge about whatever you're having trouble with.

              That's quite a contrast from more friendlier distros, it's not likely those you're referring to Linux are going to start off here. If they wanted to use Arch with less headaches, they'd quite easily come across Manjaro, all the points you raise for Arch become redundant here). This will further be the case with Solus driver solution further developing and being adopted. Your more than likely to have issues with things like proprietary wireless driver support, and this isn't something that only plagues Arch. I'm pretty sure I've seen an open-source one that was good until a few kernel releases ago, then something changed and it's no good but the vendor isn't too interested or something in addressing it.

              LEGACY HARDWARE SUPPORT
              Geforce 7 is pre 2008(I remember getting into Linux around then with a Geforce 8 card). Your users aren't going to be running this series of GPU are they? Your advice tends to be towards hardware purchases IIRC, which would defninitely rule that out. You can argue that the support may be removed for other cards in future, but realistically, how long ago did that support get removed since that GPU was released? Are your users highly likely to still be using the system with that GPU when that time comes and have an issue with it impacting software support? This isn't uncommon, even if you ignore how this happens on other OS/platform(I have to deal with it on my budget Android phone stuck on 4.2.2 and apps excluding features or completely ceasing to function, like my personal banking app later this month).

              A user with an old enough machine and hardware isn't going to have a great time as their open-source software progressed and became heavier on resources a decade later, depends what they were using but lets say they were a fan of Gnome and wanted to continue with the improvements/revisions, if they had hardware back then with 512MB of RAM are they going to have a good time with Gnome Shell? I think it's fair to understand that over time hardware can be replaced, choices can be changed(eg stay with older software or migrate to more efficient options like alternative DE/WM), otherwise they can just accept the negative impact of time on their preferences.

              I'm not entirely sure what support is like now for GPUs back then on AMD end that only worked well with proprietary driver, I'm guessing as long as they're in the r600g support they're still good?(no idea what year range of cards that covers) EDIT: according to an earlier post, AMD cards are still well supported, so kudos to AMD here.

              KERNEL/VENDOR SUPPORT
              As far as your issues with kernel support or complications with installing, go onto the AUR, get the dkms package, no special commands to run, when kernel updates, the nvidia driver updates to work with it, even with custom kernels. nvidia will continue to provide support anyhow, they make money from linux users, not gaming sector per se, professional market uses their cards for workstations and compute work where Linux is a preferrable OS for certain companies/industries, usually for these users, AMD is not an option, as stated this will get better in the near future.

              ISSUES NOT ONLY APPLY TO NVIDIA
              If you had just stated AMD was good to go with, when the vega products came out, you'd have issues that you bring up with nvidia, counter that however you like, it'd still have caused some confusion to the users(Oh so AMD is good but not all AMD work properly?), you could have migrated a user to Linux prior, then they purchase a new laptop and get confused why their AMD vega GPU isn't working for months(I thought AMD didn't have these problems? they think in frustration). Fun fact, some distros like Manjaro in 2016 held back xorg server version because of problems with the AMD driver transition support, moving to the newer version would break open-source driver support for many AMD users iirc, nvidia users were completely fine. This lasted something like 6-12 months. But no, this can only happen with nvidia hardware amirite?

              I understand compute support might not be that important to average desktop user, but it can be quite important for those who do need this, like a majority of professionals in my industry. I look forward to AMD's progress throughout the year, I intend to buy AMD myself in future as my next GPU, I just wanted to tackle the somewhat biased points you were raising of AMD being a solid choice where issues aren't a problem like nvidia. You make a bigger deal about nvidia issues than they actually are.
              Last edited by polarathene; 09 February 2018, 01:23 AM.

              Comment


              • #57
                Originally posted by polarathene View Post
                you still have to run a command to have some apps use the dGPU while everything else runs on iGPU don't you?
                Edit Steam shortcut once, mind you.

                Originally posted by Khrundel View Post
                You must be kidding. Geforce 7 are GPUs presented in 2005. Same time as Radeon x1*00 series. Do you realy need a modern hiperf driver for OpenGL 2.1 card?
                KWin 5 crash on legacy nvidia driver. Also tearing was fixed since 340 series, while legacy driver for 7xxx is 304, so I do need modern driver for this hardware. Note that Intel and AMD hardware does not have such issues.

                Originally posted by polarathene View Post
                if they had hardware back then with 512MB of RAM are they going to have a good time with Gnome Shell?
                ...the fsck you talking about? Most of this machines is upgraded to 2-4 GB RAM, I have two laptops with GeForce 7300/7400 Go with 3 GB RAM as far I remember. They kicking just fine for web-browsing, office work, playback presentations and video to projector, etc. There is zero reasons to replace it besides nvidia being dick to it's own customers (which fortunately teaching them to never go with nvidia hardware again).

                Comment


                • #58
                  Originally posted by Hibbelharry View Post
                  This is what happened:
                  Nvidia cheated on linux, once again, blocking progress. They did this by pushing their own mediocre, semi-standardized stuff, not agreeing on all others. They don't care. They want to be assholes with wrong policies.
                  Why this is always like a war, with resistance, betraying "stay your ground" and all these propaganda crap?
                  There is a word "freedom" which means nobody may force anybody to do something without consent. I bet you like this word, but right now you're damning Nvidia for not obeying blindly to other people's decision. Which decision, I think, can be traced down to 3 random guys within Red Hat, who decided not to care about anything except DRI/Mesa stack for a time, and then forget to think again. And now anyone who doesn't like this decision is an enemy of whole mankind. Now NVidia should drop their own OpenGL/Vulkan/CUDA/OpenCL implementation and stick with mediocre Mesa/Nouveau solution just because xorg developers want to drop old code.
                  Freedom doesn't work that way. If your want something you should give something in return. If you planning something, which requires other people cooperation, you shouldn't plan other people will pay for everything or you may find them not so eager to comply. Wayland team have made bad design decision: in a world where NVidia holds a half or all discrete GPU market with their custom opengl implementation they forgot to add an attach point for nvidia driver. Now they have to decide if they will proceed without NVidia or will fix wayland to become compatible.

                  Comment


                  • #59
                    Originally posted by RussianNeuroMancer View Post
                    Edit Steam shortcut once, mind you.

                    ...the fsck you talking about? Most of this machines is upgraded to 2-4 GB RAM
                    IIRC running dGPU command for Steam doesn't run subsequent launched games on the dGPU, instead you had to add these additions to each steam games launch command. Maybe that's changed since, but I do remember this being a thing on my 2014 laptop. That and games aren't the only things that need to use the dGPU, I've had plenty of other software too

                    I was talking about hardware that had low RAM in the 512MB-1GB RAM on the system, upgrading isn't always an option either because sourcing compatible RAM isn't available at the time anymore or because it was soldered on, I am aware for some models you can upgrade or add additional RAM, that wasn't the point I was trying to make here It was that regardless of open-source or proprietary choices for your drivers, the evolution of software hasn't been maintaining resource usage, I pointed to Gnome as a good example of that. On a 2GB RAM machine Linux can run alright, I have KDE on an old laptop with ATI(AMD) graphics and 2GB RAM, it uses about 450MB on boot, or over 1GB with Gnome or Deepin IIRC.

                    Nvidia doesn't care too much for desktop market on Linux as much as it does in the business sector AFIAK. Workstations where far more expensive cards are purchased or the super computers with large arrays of their GPUs in use, that's what gets nvidia love :P RedHat being commercially driven, needs to meet it's customers needs whom more than likely still go with nvidia atm. Or is there more $ being put towards nvidia by desktop users wanting things supported in a certain way? Commercial market just want shit to work, proprietary or not, they'll put their money with what works best to their needs and support. In time AMD may very well become more dominant there, nothing against that. I don't think boycotting nvidia on the desktop end of things is going to have much of an impact though.

                    If you want nvidia to shuv it, you don't block efforts to make their hardware work with Linux ecosystem. Instead it's far better to just provide superior solutions with your open-source and better supported hardware like AMD. Doesn't require any negativity, doesn't require saying "avoid nvidia because XYZ", it becomes more of a "Choose AMD because of XYZ". Let the differences speak for themselves :\ Get professional market to prefer AMD(Assuming it meets their needs, which atm it still isn't perfectly able to). Nvidia can continue to do it's thing and people can accept patches and merge stuff, no one has to develop anything themselves to support nvidia(eg kwin), nvidia can be responsible for managing any development to keep them relevant, if non-nvidia devs don't want to assist, that is their choice and something nvidia would have to address by making changes on their end(as far as I can tell they're trying to make efforts to but still get shat on for some reason).

                    Something doesn't neccessarily have to be better and have more positives to be the most preferred/favoured, that might not make much sense to you, but I've seen it many many times between competing services and users. You're better off understanding why that happens, and when a transition from one to another happens enmasse, why and what was preventing it prior?

                    Oh and Intel and AMD have had plenty of issues for me in the past I remember several systems that were Intel based and running off their iGPU alone where I had a multitude of problems, was around 2016-2017(haswell, broadwell, skylake). With AMD I remember performance and glitches/crashes, the AMD stuff was when I used them last around 2010 sometime before the GPU died/burned out(HD4770). While we're at it, plenty of issues even with recent AMD CPUs(ThreadRipper 1950x). It's great that you don't experience issues so much, but it doesn't really matter what vendor you go with imo, they all have pro/cons.

                    Comment


                    • #60
                      Originally posted by RussianNeuroMancer View Post
                      KWin 5 crash on legacy nvidia driver. Also tearing was fixed since 340 series, while legacy driver for 7xxx is 304, so I do need modern driver for this hardware. Note that Intel and AMD hardware does not have such issues.
                      Please, don't tell me about AMD support. I had radeon 5870 3 years ago, I remember how I had to wait for weeks after Ubuntu upgrade before somebody (not AMD) will fix DKMS part of their last fglrx driver. Wel, I could use opensource driver and mesa and have opengl3 instead of 4.4.
                      AMD couldn't give adequate support when my card was 4 years old. Geforce 7 cards are about 12 years old now. Compare yourself.

                      Comment

                      Working...
                      X