Benchmarks: CUDA H.264 vs x264 (Stanley posted on July 3rd, 2009 )

Hardware platform:

  • CPU: Intel Core2 Quad Q6600 (2.4Ghz/4 cores)
  • GPU: nVidia GTX 275 (896M/448bit/240 SP)
  • Memory: 4GB

Software platform:

  • OS: Windows XP x64
  • Transcoder: MediaCoder 0.7.1.4475
  • x264 command line options:  –no-psnr –no-ssim –keyint 250 –min-keyint 25 –me hex –merange 16    –non-deterministic –aq-mode 1 –aq-strength 1.0 –b-adapt 1 –ref 1 –subme 6 –psy-rd 1:0 –bframes 3  –trellis 1 –weightb –direct auto –bitrate $(VideoBitrate) –qcomp 0.6  -o “$(DestFile)”

Test 1:

  • Input: H.264/9Mbps/1920×800/23.97fps
  • Output: H.264/2Mbps/1920×800/23.97fps
  • Content duration: 60 seconds
  • Results:
    • x264: 57.4s/25.05fps/1.04x
    • CUDA: 23.5s/61.19fps/2.55x

Test 2:

  • Input: MPEG-2/5Mbps/720×480/29.97fps/Interlaced
  • Output: H.264/1Mbps/720×480/29.97fps/Progressive
  • Content duration: 60 seconds
  • Results:
    • x264: 19.5s/92.48fps/3.08x
    • CUDA: 6.3s/284.29fps/9.48x

Test 3:

  • Input: WMV/600Kbps/352×240/25fps
  • Output: H.264/500Kbps/352×240/25fps
  • Content duration: 60 seconds
  • Results:
    • x264: 5s/299.10fps/11.96x
    • CUDA: 2.9s/510.55fps/20.42x

 

COMMENTS: 9 Comments »

9 Responses

  1. St Devious says:

    Is the Nvidia CUDA encoder enabled again ?

  2. stanley says:

    Yes, in 4470.

  3. St Devious says:

    4450 only available on download. what happened to 4470 ?

  4. kumi says:

    http://forum.doom9.org/showthread.php?p=1304124#post1304124

    A recent thread on doom9.org, titled “CUDA H.264 vs x264 Speed and Image Quality Benchmarks”, compares x264’s ultrafast preset to CUDA.

    Results?

    CUDA GPU H.264 – 23.5s 114.4 FPS
    x264 preset ultrafast – 30s 90.8 FPS

    x264’s image quality was far, far superior.

  5. stanley says:

    Did you read that thread carefully? It shows x264 generates lower quality, though it is able to generate far better quality with a lot slower.

  6. kumi says:

    You’re right, I was looking at the wrong image.

  7. stanley says:

    I made this mistake too at the first look. 😉

  8. solarpower says:

    我咋就没觉得有多大区别呢

  9. pakmenu says:

    about that website and the comparison: it’s not saying anything: If anybody noticed the bitrate in x264 peaks at 2229kbps! to make an average of 900kbps…Just think: to have scenes at 2230kbps and yet have an average of 900kbps, you have to have a lot of scenes having only let’s say only 300kbps.
    The cuda clip peaks at 1776kbps, which makes for less variable bitrate, thus MORE bitrate at low motion, low contracst scenes. to have a peak at 1780 you can have low mation scenes at maybe 500kbps, 60-80% more…

    Anyway: comparing a FRAME and not knowing the FRAMESIZE the encoder used…. is pointless! a fair comarison would look at one whole GOP structure (I frame and dependent b/p frames) and have both encoders set to encode the same size for all the frames in the GOP, then look at the quality.

    The MORE variability a bitrate has, the more will be allocated at high contrast, high movement scenes and LESS bitrate to low contrast, low motion scenes. It looks like the picture grabbed is extremely dark, low contrast and probably not a lot of mation.
    This means x264 would have allocated MORE bitrate to better lighted and higher motion scenes.
    since it’s not known how the rest of the clip looks, the rest of the clip might look better for x264!

    Anyway when looking only at FRAME comparrison, the SIZE of THE FRAME should be taken EQUAL to be fair in comparrison!
    (the framesize can be looked up when enabeling OSD in FFDshow.)

    Indeed a pointless and unfair comparison!