Announcement

Collapse
No announcement yet.

10 Reasons Linux Gamers Might Want To Pass On The NVIDIA RTX 20 Series

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

  • #61
    Originally posted by Leopard View Post
    3-) For Linux side: Compability of Nvidia's driver is really good while Mesa is not good on that term.
    Are you sure ? There was some truth to that a couple of years ago but don't think that is the case today.
    Test signature

    Comment


    • #62
      Originally posted by rene View Post
      What has Nvidia to hide?
      Perhaps support for AdaptiveSync in Pascal hardware, requiring just some registers to be set by the driver to make cards work with AdaptiveSync-enabled displays, invalidating the need for G-Sync tax and possibly causing outrage when it goes public.

      Comment


      • #63
        Originally posted by L_A_G View Post

        Do I actually have to spell this out for you?

        Nothing of what you've said in this thread is new or interesting, instead everything you've said has been talked to death years ago. Yes, Nvidia doesn't release specs for their hardware like a lot of people would want them to, causing the development of open source drivers for their hardware to require way more reverse engineering that it should and let's not even go into the firmware signing that completely prevents open source developers from releasing their own drivers on new hardware before Nvidia releases the signatures.

        Your dead horse beating really does absolutely nothing!
        I think Rene likes to run various vintage hardware also on various CPU architectures, so for him something like drivers being opensource is essential.

        There are people who run Radeons with opensource driver on else than x86 arches, where nvidia blob or whatever just blobby blobs can't work of course
        Last edited by dungeon; 05 September 2018, 07:16 PM.

        Comment


        • #64
          Originally posted by reavertm View Post

          Perhaps support for AdaptiveSync in Pascal hardware, requiring just some registers to be set by the driver to make cards work with AdaptiveSync-enabled displays, invalidating the need for G-Sync tax and possibly causing outrage when it goes public.
          NVIDIA locking things down via drivers is kinda what their all about, I don't think there will be much outrage. Also from what I understand you need a AMD GPU/APU to pull that trick off.

          Comment


          • #65
            Originally posted by dungeon View Post
            There are people who run Radeons with opensource driver on else than x86 arches, where nvidia blob or whatever just blobby blobs can't work of course
            You are mostly right, except there is at least ARMv7 driver: https://www.nvidia.co.uk/Download/dr...x/137468/en-uk

            Comment


            • #66
              Originally posted by reavertm View Post

              Perhaps support for AdaptiveSync in Pascal hardware, requiring just some registers to be set by the driver to make cards work with AdaptiveSync-enabled displays, invalidating the need for G-Sync tax and possibly causing outrage when it goes public.
              Well actually that's probably the way it is. After all, gaming laptops equipped with Nvidia GPUs and a "GSync" display don't actually have any fancy "GSync" hardware in the display. It's all done on the GPU over the eDP link to the display.

              And since eDP support for power-save methods on laptop and tablet panels is where Adaptive Sync comes from, it seems pretty obvious to me that Nvidia could easily support it.

              Comment


              • #67
                Originally posted by rene View Post
                Just totally wrong, users choosing Linux for an open system should exactly care about this. If they want closed source they can go to Windows to start with. Open source is about so much more, review ability, security, fixing bugs or implementing new quirks and features yourself, long term maintenance, freedom and such ...
                Everyone has its priorities. Yours is to review the code of the kernel and fix bugs and implement new features your self. Actually have your ever done any of that? Me, I develop software for a living and I use Linux exclusively simply because it is the best unix like system available. Open source for me is like a gift and an extra. If I get my tools to be open source then great but my priorities have higher value for reliability, performance and as a programmer also proper architecture. Nvidia products excel at these values. When I know that Nvidia can reuse 90% of their code across platforms that they have built with their best developers and battle tested I am more confident that this driver works (and it does) and I get the value I paid for.

                Comment


                • #68
                  Originally posted by zoomblab View Post

                  Everyone has its priorities. Yours is to review the code of the kernel and fix bugs and implement new features your self. Actually have your ever done any of that? Me, I develop software for a living and I use Linux exclusively simply because it is the best unix like system available. Open source for me is like a gift and an extra. If I get my tools to be open source then great but my priorities have higher value for reliability, performance and as a programmer also proper architecture. Nvidia products excel at these values. When I know that Nvidia can reuse 90% of their code across platforms that they have built with their best developers and battle tested I am more confident that this driver works (and it does) and I get the value I paid for.
                  Of course I have "review the code of the kernel and fix bugs and implement new features your self" since 1998 (https://www.youtube.com/watch?v=a1YkPtfC4LI&t=160s), I run a whole Linux distribution: https://t2sde.org

                  With AMD and Intel you can reuse 100% of your code across platforms, ..! ;-)

                  IMHO when people have other priorities they should refrain from promoting and defending a company, when the system developers clearly say for a long time that this is not acceptable: https://www.youtube.com/watch?v=IVpOyKCNZYw

                  Comment


                  • #69
                    Originally posted by dungeon View Post

                    I think Rene likes to run various vintage hardware also on various CPU architectures, so for him something like drivers being opensource is essential.

                    There are people who run Radeons with opensource driver on else than x86 arches, where nvidia blob or whatever just blobby blobs can't work of course
                    I also run state of the art hardware for my daily business (https://exactcode.com) but I can not film my daily contract and product work for obvious reasons. Back in the day when my Apple G5 was new and state of the art, and I only purchased it for Linux PowerPC development, I choose the cheapest, entry level Nvidia 6600LE crap knowing I will not be able to use any kind of graphic acceleration for years to come. I believe I run it a year with an unaccelerated frame buffer, until basic 2d acceleration was reverse engineered or something like that. Would I have nvidia GPUs in the cloud infrastructure we work with? Hell NO! PS: Ironically I will soon continue hacking on the PS3/RSX driver: https://www.youtube.com/watch?v=TIm2pMXS5Q4 - would it not be nice to have the SPEC for that? ;-) I actually recorded the last S3/Trio/Virge video just for the gpu driver development education, and establishing some terms and information as instalment for next months PS3 hacking, ..!
                    Last edited by rene; 06 September 2018, 04:28 AM.

                    Comment


                    • #70
                      Originally posted by rene View Post
                      Of course I have "review the code of the kernel and fix bugs and implement new features your self" since 1998 (https://www.youtube.com/watch?v=a1YkPtfC4LI&t=160s), I run a whole Linux distribution: https://t2sde.org

                      With AMD and Intel you can reuse 100% of your code across platforms, ..! ;-)

                      IMHO when people have other priorities they should refrain from promoting and defending a company, when the system developers clearly say for a long time that this is not acceptable: https://www.youtube.com/watch?v=IVpOyKCNZYw
                      Ah, I have seen your unboxing video of the Thinkpad 470 :-) Anyway, like I said, your priorities are such that the availability of the driver code is of great value for you. You have to understand though that your point of view is distant enough from that of a typical user, and also you that can't tell others to refrain from talking out their opinions. My arguments stand perfectly from the point of view of someone that is not kernel programmer and does not want to become one. To me a graphics card product is a bundle of hardware (the card) and software (the driver). I don't need to know the actual code of the driver pretty much as I don't need to know the actual design of chip. All I care about is that these 2 components of the product implement certain interfaces (e.g. PCI for the card, OpenGL for the driver, a GUI panel for me) so that they can plug seamlessly into the system and that these interfaces work reliably and optimally. That holds of course as long as the manufacturer takes the responsibility of supporting these components, e.g. Nvidia maintaining their driver so that it works correctly with the operating systems that they support, which for me is what should be happening since they know their product design better than anyone else and of course have the insensitive to keep their customer base satisfied.
                      Last edited by zoomblab; 06 September 2018, 05:34 AM.

                      Comment

                      Working...
                      X