Announcement

Collapse
No announcement yet.

NVIDIA Announces The GeForce RTX 4060 Series

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

  • #11
    Originally posted by RejectModernity View Post

    Oh wow, even Nvidia shill is pissed how shitty new nvidia cards are.
    I need to start collecting ad hominems, insults and name calling on these forums just so that people perfectly realized who Linux fans are.

    There's no other forum, social network or place on WWW where I've seen and received so much naked hostility and aggression. 4chan is simply a warm calm welcoming place after Phoronix forums. I'm sad the ignore list doesn't work here but Michael doesn't seem to care. But he sure damn cares when someone speaks the truth and he's so eager and ready to ban people for that.

    Originally posted by dimko View Post
    Its simple, don't buy it. Vote with Wallet.
    I was waiting for over half year for an upgrade.
    I realized, I can wait more. And I have FU money, to buy 7900xtx.
    But AMD and Nvidia pissed me of hard.
    I won't buy it, I'm waiting for something decent from AMD this time around but I'm not sure they are even interested in competing any longer considering the two flops called 7900 XT and XTX that are barely better in terms of performance per buck than the previous gen.

    Consumer graphics for both companies has become an after thought and Intel is not yet there.
    Last edited by avis; 18 May 2023, 01:04 PM.

    Comment


    • #12
      What a waste of time - show me the official freely-licensed drivers or shove off nvidia.

      Comment


      • #13
        Originally posted by andyprough View Post
        What a waste of time - show me the official freely-licensed drivers or shove off nvidia.
        NVIDIA Linux open GPU kernel module source. Contribute to NVIDIA/open-gpu-kernel-modules development by creating an account on GitHub.


        Under the MIT license which is a ton more permissive than GPL 2.0. A fat binary blob/firmware is also included.

        Comment


        • #14
          for linux gaming more vram is must.

          Comment


          • #15
            Originally posted by avis View Post

            NVIDIA Linux open GPU kernel module source. Contribute to NVIDIA/open-gpu-kernel-modules development by creating an account on GitHub.


            Under the MIT license which is a ton more permissive than GPL 2.0. A fat binary blob/firmware is also included.
            This is not a 'freely-licenced driver'
            This is a kernel module on top of which a 'freely-licenced driver' may sit -- the latter does not exist (at least not from nvidia)

            Compare with:
            * amdgpu (the kernel module)
            * radeonsi
            * radv
            * amdvlk

            Comment


            • #16
              Originally posted by avis View Post
              Under the MIT license which is a ton more permissive than GPL 2.0. A fat binary blob/firmware is also included.
              From the github page: "Note that the kernel modules built here must be used with GSP firmware and user-space NVIDIA GPU driver components from a corresponding 530.41.03 driver release." I may be wrong, but that's not very free-licensed sounding.

              Comment


              • #17
                Originally posted by user1 View Post

                And I was thinking Nvidia was the only remaining vendor who doesn't cheap out on PCI lanes and provides full PCI x16 on budget and mainstream cards.. looks like I was wrong. It's really infuriating because people like me who still have pci-e 3.0 motherboards will have degraded performance with PCI x8 compared to full PCI x16.
                Broooo, so true. I have fam waiting for me to get them a gpu too. 6600 xt was almost the thing, but that x8 really ruined it for a lot of systems.

                EDIT: Some of us indeed will not spend more than $300 for this, but there's no reason why we shouldn't be offered a WHOLE, though lower-spec, graphics card for our tier. We are not actually a small market at all. This new nvidia card looks like a shameless tier shift (and man the shills love to gaslight consumers) and an upsell scheme (16gb for $100 more). The GPU MFGs seem to no longer care at all about sensible/responsible people. And inb4 "APU" ..I already have very capable CPUs in plenty capable systems, but GPU is the bottleneck for most things.
                Last edited by doomie; 18 May 2023, 12:42 PM.

                Comment


                • #18
                  Originally posted by Grinness View Post

                  This is not a 'freely-licenced driver'
                  This is a kernel module on top of which a 'freely-licenced driver' may sit -- the latter does not exist (at least not from nvidia)

                  Compare with:
                  * amdgpu (the kernel module)
                  * radeonsi
                  * radv
                  * amdvlk
                  NVIDIA doesn't prevent anyone from writing user space libraries for their open kernel modules. In fact the NVK project has been exactly that.

                  The "problem" with NVIDIA is that their driver and user space are near perfect despite all the screams to the contrary here. Yeah, their are regressions, missing features here and there but NVIDIA is a commercial company which serves its commercial customers who primarily use Windows.

                  AMD has a ton of open source here and there but from what I've heard in terms of quality it's nowhere near to what NVIDIA offers.

                  In short, if you love the idea of open source, great, Intel and AMD will happily serve you.
                  If you love the idea of having a top notch driver and features, well, NVIDIA is there for you.

                  Blaming NVIDIA for not releasing a complete open source stack for an obscure OS which has zero customers on the desktop is kinda irrational IMO. Yes, I'm not joking. Show me Linux users who actually buy software and services (aside from rare Hollywood studios using NVIDIA cards for Optix/rendering).

                  At the same time NVIDIA sells literally tens of thousands H100 accelerators (priced from $15K) which primarily run headless distributed CUDA workflows. Tell me again why they need to "properly" support Linux with almost no commercial CUDA software and wildly unstable APIs.

                  Linux fans' "We want open source" is not something NVIDIA is interested in. They are in the business of making money.

                  Comment


                  • #19
                    Originally posted by L_A_G View Post
                    We start seeing significant hitching in 1440p on 8GB using RT and what does Nvidia do? Tells people to just play at 1080p and upscale with DLSS. Which would be fine if we were talking about cards priced at $300 or less, but a $400 card in 2023 should not come with 8 GB of memory. Another $100 for some more memory? Talk about pissing on people and telling them its raining.

                    Mind you, I'm saying this whose five cards have gone: GTX 670 -> GTX 970 -> GTX 1070 Ti -> RTX 3090
                    Thing is I think DLSS does take a bit more VRAM, so you'd be starting from even lower quality pre-upscale.

                    Comment


                    • #20
                      Originally posted by avis View Post


                      Consumer graphics for both companies has become an after thought and Intel is not yet there.
                      Ouch. I was thinking intel needs to save us faster. Never thought I'd hear that from myself.

                      Comment

                      Working...
                      X