Announcement

Collapse
No announcement yet.

Firefox 103 Better Handles High Refresh Displays, WebGL Performance Fix On NVIDIA Driver

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

  • #11
    I wonder why the release notes are so lackluster, why are there not more major JavaScript, WASM, HTML, CSS, rendering performance improvements more often? https://www.youtube.com/watch?v=Sqk0jnDIdgk

    Comment


    • #12
      You would think this new fix would also help for FreeBSD + Nvidia proprietary (Linux) drivers. But on FreeBSD 12.3 I don't see any difference in performance between Firefox 102 and 103 in the Basemark Web 3.0 benchmark. I've run some benchmarks and Basemark Web 3.0 is one of the benchmarks where FreeBSD scores worst:

      FreeBSD 12.3 + Firefox 102
      WebXPRT 4 112
      Kraken 1.1 1774.6 ms (lower is better)
      Speedometer 2.0 62.7
      Octane 2.0 16317
      Basemark Web 3.0 141.44
      SilverBench P3425
      JetStream 2 48368
      Gimp start time 4.5s the first time
      LibreOffice start time very slow, maybe because I installed thousands of fonts
      Google ping latency 13.4 ms
      RAM XFCE and ZFS 810MB

      Clear Linux + Firefox 97
      WebXPRT 4 111
      Kraken 1.1 1799.9 ms
      Speedometer 2.0 74
      Octane 2.0 16625
      Basemark Web 3.0 272
      SilverBench P2216
      JetStream 2 62751
      Gimp start time 7s the first time
      LibreOffice start time 3s or 4s the first time
      Google ping latency 13.55 ms
      RAM Gnome 910MB

      FreeBSD opens the game 0 A.D. and the advanced spreadsheet app Gnumeric both in exactly 1 second on my FreeBSD system. But I didn't test the launch times of this software in Clear Linux.

      I did this Basemark test again today and it gives almost exactly the same result on the new Firefox. As you can see, FreeBSD scores worst in Basemark Web 3.0 relative to Clear Linux, and as far as I know this is a benchmark that places a lot of importance on WebGL 2.0 testing.

      Comment


      • #13
        I have benchmarked all these browsers and this is the general conclusion: Konqueror is faster in the graphic tasks, Chromium is narrowly faster in the 'non -graphic' tasks.
        Konqueror is effectively more than 20% faster in Basemark Web 3.0 and the same for MotionMark. In all other benchmarks, Chromium is faster but usually very narrow and there is simply nowhere where Chromium can hit a 'big' difference with Konqueror.
        Epiphany is faster than Chromium in Kraken 1.1 but Epiphany fails in MotionMark.
        Firefox is faster than Chromium in 'WebXPRT 4' which I did not expect. Firefox still performs normal in Speedometer, but in many things it is the slowest browser, especially in Basemark Web 3.0
        Konqueror, Epiphany and Falkon cannot run the 'WebXPRT 4' benchmark, but we can run WebXPRT 3 without problems.

        The biggest problem with Konqueror, Epiphany and Falkon is their audio support for YouTube. Specifically with this website they have all three audio problems. Konqueror even ensures that your entire session no longer has audio in any app. Falkon does play audio from SOUNDCLOUD but not from YouTube.

        From Epiphany, Konqueror and Falkon I have to say that Falkon is the most useful at the moment (by far the most stable) and Falkon is generally certainly 'fast enough', very often faster than Firefox, and very fast in JetStream 2. And faster than Chromium in MotionMark. I could use Falkon as the daily driver from now on.

        Then there is also a special result that I have to mention from Chromium: https://i.ibb.co/S0PYLxg/Screenshot-...9-13-45-00.png
        So more than 93 000 in JetStream 2. My hardware is an i3-3240 + 4GB RAM @1600MHZ + NVIDIA GTX 650 + EVO 850 500GB. You have to compare it to these results:


        '125 587 On A 5.0GHZ Core-I7 8700K, DDR4-3600 CL18 on Google Chrome Version 73.0.3683.86 (Official Build) (64-bit)' ''
        So three years ago a heavily overclocked i7-8700K was only exactly 34.8% faster in some cases in the JetStream 2 benchmark !!

        Met de Core i5 12600K, Core i7 12700K en topmodel Core i9 12900K moet Intel eindelijk weer het vuur na aan de schenen van concurrent AMD kunnen leggen.

        AMD Ryzen 2600 was only 37.4% faster so many years ago, and that is a fairly decent Hexa Core, with faster RAM, 3x more cores, of which every core is faster than my cores.

        How much more powerful is the new M1 Ultra, available as an optional extra in the new Mac Studio desktop? Take a look at our early test results to see if you should spring for the $2,000 upgrade.

        A MacBook Air 2019 was less than 10% faster in JetStream 2.
        And this has never been 'cheap' hardware: https://www.apple.com/nl/shop/buy-mac/macbook-air/met-m2‑Chip
        JetStream 2 is rather a synthetic benchmark that says less about real performance than, for example, speedometer, but it remains impressive.

        What is perhaps even more speaking in this link from PCMAG, my current system scores much higher in certain browsers than this MacBook Air (2019).
        In fact, by using Falkon or Konqueror I get a higher result in 'Basemark Web 3.0' than the MacBook Pro 13-inch (2019).
        The latter is actually inexplicable because the MacBook Pro 13-inch (2019) is much more recent and much more expensive hardware.

        To conclude, Firefox is one of the most stable browsers that works best on FreeBSD but it is currently the slowest browser. If you want speed and stability you'd better choose Chromium or Falkon right now.

        Comment

        Working...
        X