[FFmpeg-user] ffmpeg performance scaling on multicore cpus with libx264

[Andreas Zwicker]

I am using ffmpeg to convert mythtv dvb-s2 streams to matroska files, 
using libx264. Can somebody explain me the difference between specifiyng 
-stream 0 and not ? Tries ffmpeg to autoscale the number of encoding 
threads to the number of cores?

My cmdline, vmstat shows 25% idle time on Intel Q9400 quadcore:
/usr/local/bin/ffmpeg -loglevel info -threads 0 -copyinkf -i 
/mnt/store.1/tv/3002_20120507062300.mpg -y -vcodec libx264 -vpre 
720p25-bluray -crf 19 -c:a copy -sn -ss 145.7400 -t 1438.0600 -map 0:0 
-map 0:2 -f matroska /mnt/store.3/convert/3002_20120507062300.mpg_1.mkv

ffmpeg -version:
ffmpeg version 0.10.3
built on May  8 2012 09:35:19 with gcc 4.4.5
configuration: --prefix=/usr/local --extra-cflags='-march=native -O3 
-Wall' --cc='ccache cc' --enable-static --disable-shared 
--enable-libmp3lame --enable-gpl --enable-nonfree 
--disable-decoder=libdirac --enable-libvorbis --enable-pthreads 
--enable-libfaac --enable-libxvid --enable-postproc --enable-x11grab 
--enable-libgsm --enable-libtheora --enable-libopencore-amrnb 
--enable-libopencore-amrwb --enable-libx264 --enable-libspeex 
--enable-nonfree --enable-libschroedinger 
--disable-encoder=libschroedinger --enable-version3 --enable-libopenjpeg 
--enable-libvpx --enable-librtmp --enable-avfilter --enable-libfreetype 
--enable-libvo-aacenc --disable-decoder=amrnb --enable-libvo-amrwbenc 
--enable-libaacplus --libdir=/usr/lib/i386-linux-gnu --disable-vda 
--enable-libdc1394 --disable-altivec --disable-armv5te --disable-armv6 
--disable-vis --disable-debug
libavutil      51. 35.100 / 51. 35.100
libavcodec     53. 61.100 / 53. 61.100
libavformat    53. 32.100 / 53. 32.100
libavdevice    53.  4.100 / 53.  4.100
libavfilter     2. 61.100 /  2. 61.100
libswscale      2.  1.100 /  2.  1.100
libswresample   0.  6.100 /  0.  6.100
libpostproc    52.  0.100 / 52.  0.100