Docker NVidia/cuda映像中的NVidia驱动程序库

我想在基于官方

nvidia/cuda

图像的容器中运行ffmpeg和cuvid hw加速解码。Ffmpeg无法找到libnvcuvid.so，尽管有所有必需的cuda LIB。 容器的

ldconfig-p | grep libnv

输出：

libnvrtc.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc.so
libnvrtc-builtins.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc-builtins.so
libnvidia-ptxjitcompiler.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.1
libnvidia-opencl.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1
libnvidia-ml.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1
libnvidia-fatbinaryloader.so.390.12 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-fatbinaryloader.so.390.12
libnvidia-compiler.so.390.12 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-compiler.so.390.12
libnvidia-cfg.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-cfg.so.1
libnvgraph.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvgraph.so
libnvblas.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvblas.so
libnvToolsExt.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvToolsExt.so

我应该从主机上复制

libnvcuvid.so

？如果底层驱动程序版本发生变化，它不会崩溃吗？

我找到了答案。只需传递env变量

env NVIDIA_DRIVER_CAPABILITIES video，compute，utility

或

-e NVIDIA_DRIVER_CAPABILITIES=compute，utility，video

。现在我有了所有必需的libs

ldconfig-p | grep libnv

：

libnvrtc.so.9.0 (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc.so.9.0
libnvrtc.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc.so
libnvrtc-builtins.so.9.0 (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc-builtins.so.9.0
libnvrtc-builtins.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvrtc-builtins.so
libnvidia-ptxjitcompiler.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.1
libnvidia-opencl.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1
libnvidia-ml.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1
libnvidia-fatbinaryloader.so.390.30 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-fatbinaryloader.so.390.30
libnvidia-encode.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-encode.so.1
libnvidia-compiler.so.390.30 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-compiler.so.390.30
libnvidia-cfg.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-cfg.so.1
libnvgraph.so.9.0 (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvgraph.so.9.0
libnvgraph.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvgraph.so
libnvcuvid.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvcuvid.so.1
libnvblas.so.9.0 (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvblas.so.9.0
libnvblas.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvblas.so
libnvToolsExt.so.1 (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvToolsExt.so.1
libnvToolsExt.so (libc6,x86-64) => /usr/local/cuda-9.0/targets/x86_64-linux/lib/libnvToolsExt.so

---

您确定编译ffmpeg时使用了

--启用cuda

--启用cuvid

选项吗？至于有关的

ldd

。我相信ffmpeg库使用运行时绑定（dynload），因此这些lib可能不会出现在

ldd

列表中。@thekamilz该代码在主机系统上使用硬件解码时非常有效。我按照以下方式加载解码器

avcodec\u按名称（“h264\u cuvid”）

。虽然我只使用

--enable cuvid

编译了ffmpeg，但是

--enable cuda

是强制性的吗？也许我也应该尝试一下……是的，请尝试一下，我假设您已经在系统中安装了Nvidia视频编解码器SDK：