1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

new HD x264 video encoding benchmark

Discussion in 'Overclocking & Cooling' started by graysky, Mar 15, 2008.

  1. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Techarp is hosting a new HD x264 benchmark. Have a look and run it on your system to help populate the data table.

    Basically, you run the test encode of an HD sample (1280x720) and it will report back frames-per-second values for your machine @ it's clock/overclock level. You can run it at your stock settings and at your overclock settings to see how your machine compares to others in the database.

    My goal is to have a representative set of data for many different chips and chipsets. Please just report your results here in this thread. I will keep the data at that url to keep things simple.

    Thanks all and enjoy!

    The following image will be updated automatically as new data comes in. It doesn't display any results, but it does show the number of data points collected so far and how they break down by CPU type:

    [​IMG]
     
    Last edited: Mar 18, 2008
  2. regan1985 New Member

    Joined:
    Jun 7, 2006
    Messages:
    1,451 (0.46/day)
    Thanks Received:
    23
    Location:
    Coventry UNI England
    all i get is this???

    [​IMG]
     
  3. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Damn... are you running vista 64 by chance? If so, have a look at your:

    c:\Program Files(x86)\AviSynth 2.5

    Inside there, is there a directory called "plugins"? If so, does it contain DGDecode.dll at all?
     
  4. regan1985 New Member

    Joined:
    Jun 7, 2006
    Messages:
    1,451 (0.46/day)
    Thanks Received:
    23
    Location:
    Coventry UNI England
    no i am running vista 32 with sp1 beta

    but i have looked in that location and no it doesnt contain that file hope this helps
     
  5. regan1985 New Member

    Joined:
    Jun 7, 2006
    Messages:
    1,451 (0.46/day)
    Thanks Received:
    23
    Location:
    Coventry UNI England
    after copying the file into the plugings in AviSynth it worked fine. here is my 1st result on my everyday use
    with vista 32 sp1 beta


    [​IMG]

    [​IMG]
     
  6. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Cool man, thanks for the data.
     
  7. exodusprime1337

    exodusprime1337

    Joined:
    Nov 27, 2007
    Messages:
    2,188 (0.83/day)
    Thanks Received:
    342
    [​IMG]

    here's mine
     
  8. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Cool man, thanks... what vista flavor is it (32 or 64)?
     
  9. exodusprime1337

    exodusprime1337

    Joined:
    Nov 27, 2007
    Messages:
    2,188 (0.83/day)
    Thanks Received:
    342
    it's windows xp not vista at all
     
  10. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Sorry man, I saw regan1985's screenshot when I asked you that... XP 32-bit or 64-bit?
     
  11. exodusprime1337

    exodusprime1337

    Joined:
    Nov 27, 2007
    Messages:
    2,188 (0.83/day)
    Thanks Received:
    342
    32 bit buddy
     
  12. lemonadesoda

    lemonadesoda

    Joined:
    Aug 30, 2006
    Messages:
    6,289 (2.04/day)
    Thanks Received:
    972
    Great to have an easy to use encoding benchmark! Good stuff. BUT

    1./ The benchmark is way too thorough... takes much to long to do multiple runs of the same encoding. Quite honestly, this isnt necessary.
    2./ The benchmark results are silly. Why report the data density of the media file (ie the kb/s of the media stream)? What should be reported is the ENCODING SPEED.
    3./ Just a summary benchmark INDEX would be handy
    4./ Wheres the checksum for "honest results" checking?

    [quote="ICE-QUAD", Windows 2003 SP2 32-bit, Intel 865PE, FSB 1066, Q6600 @2.4, 2Gb DDR @400, 2-3-3-6, X800SE AGP]
    ---------- RUN1PASS1.LOG
    encoded 1442 frames, 46.10 fps, 3905.42 kb/s

    ---------- RUN2PASS1.LOG
    encoded 1442 frames, 46.10 fps, 3905.42 kb/s

    ---------- RUN3PASS1.LOG
    encoded 1442 frames, 46.03 fps, 3905.42 kb/s

    ---------- RUN4PASS1.LOG
    encoded 1442 frames, 45.85 fps, 3905.42 kb/s

    ---------- RUN1PASS2.LOG
    encoded 1442 frames, 13.61 fps, 3952.85 kb/s

    ---------- RUN2PASS2.LOG
    encoded 1442 frames, 13.62 fps, 3952.85 kb/s

    ---------- RUN3PASS2.LOG
    encoded 1442 frames, 13.63 fps, 3952.85 kb/s

    ---------- RUN4PASS2.LOG
    encoded 1442 frames, 13.60 fps, 3952.85 kb/s[/quote]

    P.S. variable results, because I'm browsing whilst this benchmark is running.

    LINK http://www.techarp.com/x264_Benchmark/hd/HD_benchmark_results_pub.xls
     
    Last edited: Mar 23, 2008
  13. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Can you give me a few more details on your hardware?

    It's a q6600 but what is the multiplier and FSB? Also what speed is your memory running and what are the timings? Finally, what chipset is your MB?

    Thanks for the input. Let me try to answer your four questions:

    1) I wanted multiples for the runs to insure an accurate result. I think an n=3 is just good science.

    2) The data density gets reported automatically by x264.exe; I'm not actually using it, just the FPS values for each pass which allows the calculation of the total encoding speed (have a look at the tables on the techarp link).

    3) Good idea! I will put something together... what do you think should be kept from the larger tables in an index? CPU, Core Speed, Total Encode Time?

    4) I would loved to have incorporated a data integrity check, but sadly, I don't have the programming skills to make it work. I do check the data against itself for like processors and if a data point clearly stands out from the pack, I delete it. I am happy to say that of the over 280 results of my first "SDTV" x264 benchmark (used a 720x480 clip), I only found one such suspect data point.
     
  14. lemonadesoda

    lemonadesoda

    Joined:
    Aug 30, 2006
    Messages:
    6,289 (2.04/day)
    Thanks Received:
    972
    I'm Q6600 at stock on intel 865PE.

    1) I understand the reason... but for the sake of everyones short life... do n=1. People can always re-run if they want a second opinion. Alternatively, make n= user input. But then only report the AVERAGE of the n runs.

    2) I see, ok

    3) INDEX = total encode time (average if n>1)

    4) OK, but you could do a very rudimentary checksum calculation on the INDEX figure.

    Otherwise, great work.

    PS. I didnt even look at the mpg file... perhaps I'll go that now! LOL
     
  15. lemonadesoda

    lemonadesoda

    Joined:
    Aug 30, 2006
    Messages:
    6,289 (2.04/day)
    Thanks Received:
    972
    Just looked... oh. Boring. But I did notice the sound was lost on the MP4. Is that deliberate? Does the utility only render the video portion? Or do I have a broken MP4 codec?
     
  16. VulkanBros

    VulkanBros

    Joined:
    Jan 31, 2005
    Messages:
    1,406 (0.39/day)
    Thanks Received:
    328
    Location:
    The Pico Mundo Grill
    Here are mine ( Vista 64-bit )
     

    Attached Files:

    Crunching for Team TPU
  17. CrAsHnBuRnXp

    CrAsHnBuRnXp

    Joined:
    Oct 19, 2007
    Messages:
    5,640 (2.12/day)
    Thanks Received:
    702
    Where can I find the DGDecode.dll file to copy into the plugin folder?
     
  18. VulkanBros

    VulkanBros

    Joined:
    Jan 31, 2005
    Messages:
    1,406 (0.39/day)
    Thanks Received:
    328
    Location:
    The Pico Mundo Grill
    Here it is
     

    Attached Files:

    Crunching for Team TPU
  19. DanTheBanjoman Señor Moderator

    Joined:
    May 20, 2004
    Messages:
    10,553 (2.70/day)
    Thanks Received:
    1,383
    Ran it on the background, CPU load was ~60% during the run.
     
  20. CrAsHnBuRnXp

    CrAsHnBuRnXp

    Joined:
    Oct 19, 2007
    Messages:
    5,640 (2.12/day)
    Thanks Received:
    702
    But mine isnt in there. Im talking about where I can get the file so I can put it there. :)
     
  21. lemonadesoda

    lemonadesoda

    Joined:
    Aug 30, 2006
    Messages:
    6,289 (2.04/day)
    Thanks Received:
    972
    Ooooh, nice Dan. Me also wantie Dual Quad Xeons :)

    Thats a great demo of the power of 8.

    WAIT. Somethings wrong. Look at those kb/s numbers. That's the video density. PASS 1 should ALWAYS BE THE SAME (same for PASS2). What's going on? Why is your run giving you different figures? Broken FPU? One of those cores got a math bug?
     
  22. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    @lemonaidsoda - a few other questions: what is the running speed of your RAM and what o/s are you running (like windows xp 32-bit or 64-bit)?

    Yeah, since it's a measure of pure video, I did not included an audio track.
     
  23. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    The dll is in the "test" directory of the benchmark. It should have been automatically copied into your %programfiles%\avisynth 2.5\plugins or if running 64-bit o/s your %programfiles(x86)%\avisynth 2.5\plugins

    For some reason this isn't happening on some vista machines. Regan1985's for example, and now your machine. Can you do me a favor to help troubleshoot this?

    Please post the full path to the location of your avisynth directory.
     
  24. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Thanks for the data! Just to be sure I got it right... is your memory running @ 1,066 MHz or just rated at that speed? CPU-Z can tell you (memory tab) what it's running at.
     
  25. graysky

    Joined:
    Jun 11, 2007
    Messages:
    192 (0.07/day)
    Thanks Received:
    81
    Thanks for the results! Since your clock speed is 2.28 GHz, I guessing you're running @ 6x380? Also, when you said it only ran @ 60 %, was that just for the 1st pass or did the 2nd pass as well only run @ 60 %?
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page