Announcement

Collapse
No announcement yet.

Dav1d 0.6 AV1 Video Decoder Delivering Healthy Improvements For Intel + AMD Processors

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

  • Dav1d 0.6 AV1 Video Decoder Delivering Healthy Improvements For Intel + AMD Processors

    Phoronix: Dav1d 0.6 AV1 Video Decoder Delivering Healthy Improvements For Intel + AMD Processors

    Given this week's release of dav1d 0.6, here are some fresh benchmarks of this open-source AV1 video decoder on a few different Intel and AMD systems so far...

    http://www.phoronix.com/scan.php?pag...AV1-Benchmarks

  • #2
    I'd be curious to see benchmarks on ARM, i3, or Ryzen 3 systems too. For video decode, those are the kinds of platforms that reach the widest audience, so if one of them isn't capable of keeping up with a decent frame rate, either the codec needs work or the platform is too slow and should be avoided.

    Comment


    • #3
      Originally posted by schmidtbag View Post
      I'd be curious to see benchmarks on ARM, i3, or Ryzen 3 systems too. For video decode, those are the kinds of platforms that reach the widest audience, so if one of them isn't capable of keeping up with a decent frame rate, either the codec needs work or the platform is too slow and should be avoided.
      Did test dav1d 0.5.2 vs 0.6(Odroid N2 - Arm64/aarch64 - 4xA73(1.8Ghz) 2xA53(1.9Ghz))
      https://code.videolan.org/videolan/dav1d/issues/15#note_52552

      Wanted to make the test on my Odroid N2 with the Phoronix Test Suite but something is strange. It does take a long time between the runs. Will look if i can find the problem and upload data later

      EDIT: seems like it does after every run >Parsing Search Key: "fps"< for every frame on 1 core (all 8929frames)
      Result Parsing Search Key: "fps"

      Result Line: Decoded 8779 8929 frames (98.3%) - 52.74 23.98 fps (2.20x)
      Don't think it should work that way.
      Did kill it after 22minutes and it had only looked at 2000/8929frames
      Manjaro arm phoronix-test-suite-9.4.1.tar.gz
      bug report https://github.com/phoronix-test-sui...les/issues/133
      Last edited by Toggleton; 03-11-2020, 12:51 PM.

      Comment


      • #4
        Intel Core 2 Quad [email protected] to test ssse3 /sse2 Launch Date Q1'07

        https://openbenchmarking.org/result/...PT39&obr_sor=y

        Comment


        • #5
          Odroid-N2(Amlogic S922X)- Arm64/aarch64 - 4xA73(1.8Ghz) 2xA53(1.9Ghz)
          https://openbenchmarking.org/result/...PT34&obr_sor=y

          Comment


          • #6
            Originally posted by Toggleton View Post
            Odroid-N2(Amlogic S922X)- Arm64/aarch64 - 4xA73(1.8Ghz) 2xA53(1.9Ghz)
            https://openbenchmarking.org/result/...PT34&amp;obr_sor=y
            Were you aware that with PTS you can stack your test results if you have more than 1 board to compare against the baseline?

            When you ran your test with the Core 2 Quad (result # 2003112-VE-2003101PT39) and ran that result # against your Odroid N2, you would have had a combined result file with both CPU types. In case you weren't aware.

            Thanks for sharing your personal results.

            Comment

            Working...
            X