Attachment 15418 Details for Bug 49193 – патч

[patch] патч

cuda.patch (text/plain), 1.85 KB, created by Toxblh on 2024-01-26 03:46:19 MSK

(hide)

Description:

Filename:

MIME Type:

Creator: Toxblh

Created: 2024-01-26 03:46:19 MSK

Size: 1.85 KB

patch

obsolete

>diff --git a/.gear/ffmpeg.spec b/.gear/ffmpeg.spec
>index 72a925cf3e..2854155eeb 100644
>--- a/.gear/ffmpeg.spec
>+++ b/.gear/ffmpeg.spec
>@@ -13,7 +13,7 @@
> # License
> %def_enable gpl
> %def_enable version3
>-%def_disable nonfree
>+%def_enable nonfree
> 
> # Enable/Disable stuff
> %def_enable doc
>@@ -142,10 +142,20 @@
> 
> # nvidia cuda doesn't support arm, mips and others
> # https://developer.nvidia.com/nvidia-video-codec-sdk/download
>+# ---
>+# Describing Nvidia specific options --nvccflags:
>+# nvcc from CUDA toolkit version 11.0 or higher does not support compiling for 'compute_30' (default in ffmpeg)
>+# 52 is the minimum required for the current CUDA 11 version (Quadro M6000 , GeForce 900, GTX-970, GTX-980, GTX Titan X)
>+# https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/
>+# Anyway below 50 arch card don't support nvenc encoding hevc https://developer.nvidia.com/nvidia-video-codec-sdk (Supported devices)
>+# Nvidia docs:
>+# https://docs.nvidia.com/video-technologies/video-codec-sdk/12.1/pdf/Using_FFmpeg_with_NVIDIA_GPU_Hardware_Acceleration.pdf
> %ifarch %ix86 x86_64 aarch64 ppc64le
> %def_enable cuvid
>+%def_enable cuda
> %else
> %def_disable cuvid
>+%def_enable cuda
> %endif # cuvid
> 
> %define avdevicever 60
>@@ -251,6 +261,7 @@ BuildRequires:	yasm
> %{?_enable_vdpau:BuildRequires: libvdpau-devel}
> %{?_enable_vulkan:BuildRequires: libvulkan-devel}
> %{?_enable_cuvid:BuildRequires: nv-codec-headers}
>+%{?_enable_cuda:BuildRequires: nvidia-cuda-toolkit}
> 
> %define common_descr \
> FFmpeg is a collection of libraries and tools to process multimedia content\
>@@ -690,6 +701,11 @@ xz Changelog
> 	%{subst_enable cuvid} \
> 	--enable-hardcoded-tables \
> 	--enable-runtime-cpudetect \
>+%if_enabled cuda
>+	--enable-cuda-nvcc \
>+	--enable-libnpp \
>+	--nvccflags="-gencode arch=compute_52,code=sm_52 -O2" \
>+%endif
> %if_enabled debug
> 	--enable-debug \
> %else

Actions: View | Diff

Attachments on bug 49193: 15418 | 15419