[Ffmpeg-devel] benchmark of different CABAC routines

Guillaume Poirier gpoirier
Wed Oct 11 15:39:19 CEST 2006


Hi,

P4 Northwood 2.40GHz gcc (GCC) 3.3.5 (Debian 1:3.3.5-13) --march=pentium4

BRANCHLESS_CABAC_DECODER, CMOV_IS_FAST 1
7609 dezicycles in decode_residual, 2095132 runs, 2020 skipsbits/s
dup=0 drop=0

BRANCHLESS_CABAC_DECODER, CMOV_IS_FAST 0
7608 dezicycles in decode_residual, 2095081 runs, 2071 skipsbits/s
dup=0 drop=0

undef BRANCHLESS_CABAC_DECODER, CMOV_IS_FAST 1
7404 dezicycles in decode_residual, 2095574 runs, 1578 skipsbits/s
dup=0 drop=0


undef BRANCHLESS_CABAC_DECODER, CMOV_IS_FAST 0
7405 dezicycles in decode_residual, 2095498 runs, 1654 skipsbits/s
dup=0 drop=0


ARCH_X86_DISABLED, BRANCHLESS_CABAC_DECODER
7457 dezicycles in decode_residual, 2095028 runs, 2124 skipsbits/s
dup=0 drop=0

ARCH_X86_DISABLED, undef BRANCHLESS_CABAC_DECODER
7244 dezicycles in decode_residual, 2095453 runs, 1699 skipsbits/s
dup=0 drop=0

So on P4 Northwood, the fastest version is the unoptimized C...

I shall test on Prescott/Smithfield later if I have some time

Guillaume




More information about the ffmpeg-devel mailing list