On 04/09/2013 06:57 PM, Christophe Gisquet wrote: > 233 to 107 cycles on Arrandale and Win64. > Replacing the multiplication by s_m[m] by a pand and a pxor with > appropriate vectors is slower. Unrolling is a 15 cycles win. > A SSE version was 4 cycles slower. > --- > libavcodec/aacsbrdata.h | 6 ++- > libavcodec/x86/sbrdsp.asm | 110 > +++++++++++++++++++++++++++++++++++++++++++ > libavcodec/x86/sbrdsp_init.c | 16 +++++++ > 3 files changed, 131 insertions(+), 1 deletion(-)
Looks ok to me, although I don't really know about the PIC part... -Justin _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
