ffmpeg 使用 nvidia GPU 拆帧,但是 GPU 的利用率很低怎么办?

我的 GPU 是 nvidia tesla m60

视频信息如下:

╰─➤  ffmpeg -i resource/XinJiChuanYue.mp4                                                                                                          
ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 11 (Ubuntu 11.2.0-19ubuntu1)
  configuration: --prefix=/usr --extra-version=0ubuntu0.22.04.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzimg --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-librsvg --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 70.100 / 56. 70.100
  libavcodec     58.134.100 / 58.134.100
  libavformat    58. 76.100 / 58. 76.100
  libavdevice    58. 13.100 / 58. 13.100
  libavfilter     7.110.100 /  7.110.100
  libswscale      5.  9.100 /  5.  9.100
  libswresample   3.  9.100 /  3.  9.100
  libpostproc    55.  9.100 / 55.  9.100
Guessed Channel Layout for Input Stream #0.1 : 5.1
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'resource/XinJiChuanYue.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.7.100
  Duration: 02:49:09.75, start: 0.000000, bitrate: 1716 kb/s
  Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1280x720 [SAR 1:1 DAR 16:9], 1201 kb/s, 23.98 fps, 23.98 tbr, 24k tbn, 47.95 tbc (default)
    Metadata:
      handler_name    : VideoHandler
      vendor_id       : [0][0][0][0]
  Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, 5.1, fltp, 379 kb/s (default)
    Metadata:
      handler_name    : SoundHandler
      vendor_id       : [0][0][0][0]
  Stream #0:2(und): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 127 kb/s
    Metadata:
      handler_name    : SoundHandler
      vendor_id       : [0][0][0][0]
At least one output file must be specified

ffmpeg 的 hwaccels 如下:

╰─➤  ffmpeg -hwaccels                                                                                                                                                                                                                                                                                   1 ↵
ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 11 (Ubuntu 11.2.0-19ubuntu1)
  configuration: --prefix=/usr --extra-version=0ubuntu0.22.04.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzimg --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-librsvg --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 70.100 / 56. 70.100
  libavcodec     58.134.100 / 58.134.100
  libavformat    58. 76.100 / 58. 76.100
  libavdevice    58. 13.100 / 58. 13.100
  libavfilter     7.110.100 /  7.110.100
  libswscale      5.  9.100 /  5.  9.100
  libswresample   3.  9.100 /  3.  9.100
  libpostproc    55.  9.100 / 55.  9.100
Hardware acceleration methods:
vdpau
cuda
vaapi
qsv
drm
opencl

我使用下面的命令调用 ffmpeg + nvidia GPU 做视频拆帧,但是发现 GPU 利用率很低
ffmpeg -hwaccel cuda -i "resource/XinJiChuanYue.mp4" -vf fps=1 -threads 16 -vsync 0 -q:v 1 -f image2 ./snapshot/out_%05d.jpg

但我把上面的命令运行起来的时候:

╰─➤  nvidia-smi
Thu Mar  9 02:01:35 2023       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.85.12    Driver Version: 525.85.12    CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla M60           Off  | 00000000:83:00.0 Off |                  352 |
| N/A   33C    P0    50W / 150W |    174MiB /  7680MiB |     13%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  Tesla M60           Off  | 00000000:84:00.0 Off |                  352 |
| N/A   36C    P8    14W / 150W |      3MiB /  7680MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A    106690      C   ffmpeg                            171MiB |
+-----------------------------------------------------------------------------+

图片.png

GPU 利用率只有 13% 左右
图片.png

在使用 GPU 的同时,CPU 也在工作,但是只利用了一个核多一点

图片.png

拆帧的倍速只有 30+

图片.png

怎么让 GPU 满负荷工作来最大化加速呢?

这里有一个保存图片到磁盘的操作,但是磁盘不是瓶颈,因为磁盘是一个 SSD

如果我不使用 GPU,只适用 CPU,发现 CPU 都比 GPU 快一些

ffmpeg -i "resource/XinJiChuanYue.mp4" -vf fps=1 -threads 0 -vsync 0 -q:v 1 -f image2 ./snapshot/out_%05d.jpg

这个时候,倍速是 40+,比 GPU 的 30+ 快

图片.png

这个时候,CPU 利用了 10 个核心的样子

图片.png

GPU 没有参与,符合预期
图片.png

阅读 2.2k
撰写回答
你尚未登录,登录后可以
  • 和开发者交流问题的细节
  • 关注并接收问题和回答的更新提醒
  • 参与内容的编辑和改进,让解决方法与时俱进
推荐问题