Announcement

Collapse
No announcement yet.

20-Way GPU Gaming Comparison With March 2020 Linux Drivers

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

  • 20-Way GPU Gaming Comparison With March 2020 Linux Drivers

    Phoronix: 20-Way GPU Gaming Comparison With March 2020 Linux Drivers

    Given the continuously evolving state of the open-source Radeon Linux graphics drivers in particular, here are fresh AMD Radeon vs. NVIDIA GeForce Linux gaming benchmarks with the latest Linux graphics drivers as we begin March 2020. Besides the latest NVIDIA 440.64 Linux driver, on the Radeon side was Mesa 20.1-devel paired with the Linux 5.5.5 kernel and also having the RADV ACO back-end enabled.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Typo:

    Originally posted by phoronix View Post
    the oerformance out of this Windows game on Linux is still very good for both drivers.

    Comment


    • #3
      When looking trough top benchmarks with min/max specifed, you can easily see that Radeon MIN is usually highest and very often it doubles the top card. And min is the most important factor for gaming. MAX can be interesting only from benchmarking point of view ;-)

      Comment


      • #4
        phoronix, say, is there a chance you could do a full laptop GPU comparison as well? Both iGPUs and dGPU, current-gen, and last-gen, with the data for a low-end desktop card for reference? Bonus points for multiple implementations -- there's at least two different (TDP) versions of the MX250 out there, for example.

        Comment


        • #5
          Originally posted by fallenguru View Post
          phoronix, say, is there a chance you could do a full laptop GPU comparison as well? Both iGPUs and dGPU, current-gen, and last-gen, with the data for a low-end desktop card for reference? Bonus points for multiple implementations -- there's at least two different (TDP) versions of the MX250 out there, for example.
          Nope, don't have many laptops and nearly all of them are iGPU not to mention many other differences between them.
          Michael Larabel
          https://www.michaellarabel.com/

          Comment


          • #6
            Vulkan results look a lot better now with AMD than it used to
            The only benchmark where i feel that AMD doesn't perform as well as it should is Unigine Heaven v4.0 since AMD has such good OpenGL drivers.
            Michael Thanks for writing out if it's OpenGL, Vulkan, Steamplay, DXVK and so on, it makes it so much easier to comprehend the results.

            Comment


            • #7
              Michael you should include 95 percentile. Lately, there is a vulkan layer that is able to calculate and provide those statistics. It is called mango hud. Perhaps, you could integrate it in PTS?

              Comment


              • #8
                Some of those OpenGL vs Vulkan benchmarks beg the question: Is AMD better or Nvidia just worse in those extreme scenarios?

                Anyhow, very good results. Parity with Windows on mesa+aco on top of very stable drivers paint a very good picture for gaming on Linux. And while it won't happen immediately, we will see an influx of users because of this on Linux. As always: Big thanks to the devs out there making this possible!

                Comment


                • #9
                  these benchmarks also needs an overall power consumption graphic at the end of the article.

                  Comment


                  • #10
                    Since when has the Valley benchmark been a game, It screws the Geometric Mean to boot.

                    Comment

                    Working...
                    X