[FFmpeg-devel] [PATCH v2] aacenc: add SIMD optimizations for abs_pow34 and quantization
henrik at gramner.com
Sun Oct 9 16:39:47 EEST 2016
On Sun, Oct 9, 2016 at 2:15 PM, Rostislav Pehlivanov
<atomnuker at gmail.com> wrote:
> +cglobal aac_quantize_bands, 6, 6, 6, out, in, scaled, size, is_signed, maxval, Q34, rounding
Now that this function is SSE2 you should explicitly use
floating-point instructions to avoid bypass delays from transitioning
between int and float domains.
E.g. movaps instead of mova and shufps instead of SPLATD. Also andps
instead of pand.
More information about the ffmpeg-devel