Recherche avancée

Sur d’autres sites (4968)

• avcodec/cuviddec : fix AV1 decoding error

  24 juin 2022, par Zhao Zhili

  avcodec/cuviddec : fix AV1 decoding error
  cuvidParseVideoData only supports pure OBUs, it reports an unknown
  
  error with AV1CodecConfigurationRecord. Check whether extradata
  
  is AV1CodecConfigurationRecord and skip the first 4 bytes to fix
  
  the issue.
  The bug is revealed in ffmpeg cmd since 45e3b6a68 and ffd1316e.
  Signed-off-by : Timo Rothenpieler <timo@rothenpieler.org>
  

  • [DH] libavcodec/cuviddec.c

• Build ffmpeg under Cygwin : Invalid relocation. Offset at address doesn't fit into 32 bits

  11 août 2022, par synner

  Trying to build ffmpeg-5.1 under Cygwin, "natively" i.e to be dependant under the cygwin DLL, not necessarily cross-compiling for pure Windows.

  


  I built separately a lot of the dependant libs.
Some of the libs were provided by the cygwin repo naturally.

  


  [ /usr/local/src/ffmpeg-5.1]λ ./configure --disable-shared --enable-static --extra-cflags="-I/usr/local/include" --extra-ldflags="-L/usr/local/lib" --enable-gpl --enable-nonfree --enable-openssl --enable-gcrypt --enable-libass --enable-libcaca --enable-libcelt --enable-libcdio --enable-libdav1d --enable-libfdk-aac --enable-libfontconfig --enable-libfreetype --enable-libgsm --enable-libmp3lame --enable-libopenjpeg --enable-libopus --enable-libsoxr --enable-libssh --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxvid --enable-libxml2 --enable-libzmq --enable-pic --disable-avdevice

install prefix            /usr/local
source path               .
C compiler                gcc
C library                 newlib
ARCH                      x86 (generic)
big-endian                no
runtime cpu detection     yes
standalone assembly       yes
x86 assembler             nasm
MMX enabled               yes
MMXEXT enabled            yes
3DNow! enabled            yes
3DNow! extended enabled   yes
SSE enabled               yes
SSSE3 enabled             yes
AESNI enabled             yes
AVX enabled               yes
AVX2 enabled              yes
AVX-512 enabled           yes
AVX-512ICL enabled        yes
XOP enabled               yes
FMA3 enabled              yes
FMA4 enabled              yes
i686 features enabled     yes
CMOV is fast              yes
EBX available             yes
EBP available             yes
debug symbols             yes
strip symbols             yes
optimize for size         no
optimizations             yes
static                    yes
shared                    no
postprocessing support    yes
network support           yes
threading support         pthreads
safe bitstream reader     yes
texi2html enabled         no
perl enabled              yes
pod2man enabled           yes
makeinfo enabled          yes
makeinfo supports HTML    yes
xmllint enabled           yes

...


  


  [ /usr/local/src/ffmpeg-5.1]λ ./ffmpeg -version
Cygwin runtime failure: /usr/local/src/ffmpeg-5.1/ffmpeg.exe: Invalid relocation.  Offset 0x7ffd7778b627 at address 0x100a92895 doesn't fit into 32 bits


  


  Suggestions for build options / root cause ?

  


• x86/tx_float : implement inverse MDCT AVX2 assembly

  3 septembre 2022, par Lynne

  x86/tx_float : implement inverse MDCT AVX2 assembly
  This commit implements an iMDCT in pure assembly.
  This is capable of processing any mod-8 transforms, rather than just
  
  power of two, but since power of two is all we have assembly for
  
  currently, that's what's supported.
  
  It would really benefit if we could somehow use the C code to decide
  
  which function to jump into, but exposing function labels from assebly
  
  into C is anything but easy.
  
  The post-transform loop could probably be improved.
  This was somewhat annoying to write, as we must support arbitrary
  
  strides during runtime. There's a fast branch for stride == 4 bytes
  
  and a slower one which uses vgatherdps.
  Zen 3 benchmarks for stride == 4 for old (av_imdct_half) vs new (av_tx) :
  128pt :
  
     2811 decicycles in         av_tx (imdct),16775916 runs,   1300 skips
  
     3082 decicycles in         av_imdct_half,16776751 runs,    465 skips
  256pt :
  
     4920 decicycles in         av_tx (imdct),16775820 runs,   1396 skips
  
     5378 decicycles in         av_imdct_half,16776411 runs,    805 skips
  512pt :
  
     9668 decicycles in         av_tx (imdct),16775774 runs,   1442 skips
  
    10626 decicycles in         av_imdct_half,16775647 runs,   1569 skips
  1024pt :
  
    19812 decicycles in         av_tx (imdct),16777144 runs,     72 skips
  
    23036 decicycles in         av_imdct_half,16777167 runs,     49 skips
  

  • [DH] libavutil/tx.c
  • [DH] libavutil/tx_priv.h
  • [DH] libavutil/x86/tx_float.asm
  • [DH] libavutil/x86/tx_float_init.c