[Libav-user] av_hwdevice_ctx_create fails when using AV_HWDEVICE_TYPE_CUDA

[Strahinja Radman]

> On 5. 7. 2022., at 11:10, Santosh Kamble <skamble at cemtrexlabs.com> wrote:
> 
> Helo,
> I am trying to decode a video stream in C++ application using ffmpeg cuda hardware acceleration.
> My windows machine has GPU is GTX1060  Update driver version 512.59
> 
> 
> err = av_hwdevice_ctx_create(&hw_device_ctx, AV_HWDEVICE_TYPE_CUDA,
> NULL, NULL, 0) 
> 
> Function fails with errorcode -12
> but when I pass AV_HWDEVICE_TYPE_DXVA2, it succeeds.
> 
> ----------------------------
> My observations about ffmpeg cuda compiled tool :
>   I observed that cuda based image decoding and video decoding is working fine using ffmpeg tool as below example works and use GPU for decoding purpose.
> ex1. ffmpeg -hwaccel cuda -i "D:\myvideo.avi"  output.mp4
> ex2. ffmpeg -hwaccel cuda -i D:\myvideo.mp4 -y ./tests/%04d.jpg
> 
> also the tool shows it supports cuda hardware acceleration.
> command:  ffmpeg.exe -hwaccels
> output: ffmpeg version 4.4.git Copyright (c) 2000-2022 the FFmpeg developers
>   built with Microsoft (R) C/C++ Optimizing Compiler Version 19.29.30140 for x64
>   configuration: --enable-nonfree --enable-shared --enable-ffnvcodec --enable-cuda-llvm --enable-gpl --disable-debug --enable-cuda-nvcc --enable-libnpp --toolchain=msvc --extra-cflags=-I../nv_sdk --extra-ldflags='-libpath:../nv_sdk' --nvccflags='-gencode arch=compute_60,code=sm_60 -O2'
>   libavutil      57. 24.101 / 57. 24.101
>   libavcodec     59. 28.100 / 59. 28.100
>   libavformat    59. 23.100 / 59. 23.100
>   libavdevice    59.  6.100 / 59.  6.100
>   libavfilter     8. 38.100 /  8. 38.100
>   libswscale      6.  6.100 /  6.  6.100
>   libswresample   4.  6.100 /  4.  6.100
>   libpostproc    56.  5.100 / 56.  5.100
> Hardware acceleration methods:
> cuda
> dxva2
> d3d11va
> 
> 
> -----------------------------------------
> 
> 
> So it looks like the compiled libraries are fine. I am missing some other thing.
> Can someone help on this cuda based decoding using ffmpeg.
> 
> 
> Thanks
> Santosh
> _______________________________________________
> Libav-user mailing list
> Libav-user at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/libav-user
> 
> To unsubscribe, visit link above, or email
> libav-user-request at ffmpeg.org with subject "unsubscribe”.


Hello Santosh,

You can try this piece of code in order to get the type by name or search for available types for your GPU.

    type = av_hwdevice_find_type_by_name("cuda");
    if (type == AV_HWDEVICE_TYPE_NONE)
    {
        LOG(LOG_LEVEL_ERROR) << "Device type " << "cuda" << " is not supported.";
        LOG(LOG_LEVEL_ERROR) << "Available device types:";
        while ((type = av_hwdevice_iterate_types(type)) != AV_HWDEVICE_TYPE_NONE)
            LOG(LOG_LEVEL_INFO) << av_hwdevice_get_type_name(type);
    }