支持nvidia GPU 的硬件编解码的ffmpeg编译记要
中间目录:out
1、x264
下载x264-stable.zip
unzip x264-stable.zip
cd x264-stable
./configure --enable-shared --enable-static
make
make DESTDIR=../out install
2、安装nv-codec-headers
git clone https://git.videolan.org/git/ffmpeg/nv-codec-headers.git
cd nv-codec-headers
make
sudo make isntall
3、ffmpeg
下载 FFmpeg-release-5.0.zip
unzip FFmpeg-release-5.0.zip
cd FFmpeg-release-5.0/
./configure --enable-shared --enable-static --enable-debug --enable-libx264 --enable-cuda --enable-cuvid --enable-gpl --disable-optimizations --disable-stripping --enable-libfreetype --extra-cflags=-I../out/usr/local/include --extra-ldflags=-L../out/usr/local/lib
make -j8
make DESTDIR=../out install
4、验证
cd out/usr/local
export LD_LIBRARY_PATH='pwd'/lib:$LD_LIBRARY_PATH
./bin/ffmpeg -r 1 -c:v h264_cuvid -i qiaohu.mp4 -f rawvideo 1.yuv
5、容器docker内执行
可能不需要加上主宿机上的驱动和cuda版本库,加上 -e NVIDIA_DRIVER_CAPABILITIES=compute,utility,video
libcuda.so
libcuda.so.1
libcuda.so.xxx.xx
libnvcuvid.so
libnvcuvid.so.1
libnvcuvid.so.xxx.xx.so
6、容器启动加上参数
-e NVIDIA_DRIVER_CAPABILITIES=compute,utility,video
本文内容由网友自发贡献,版权归原作者所有,本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容,请联系:hwhale#tublm.com(使用前将#替换为@)