Skip to content

F
ffmpeg.wasm-core
Commit 652c59b9 authored by Loren Merritt's avatar Loren Merritt
Browse files
Options

use ff_fft_calc_3dn2 instead of sse on a K8. 

1.5% faster vorbis.

Originally committed as revision 6024 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent 1e4ecf26
Hide whitespace changes
Inline Side-by-side
Showing with 3 additions and 7 deletions
libavcodec/fft.c
View file @ 652c59b9
......@@ -68,17 +68,13 @@ int ff_fft_init(FFTContext *s, int nbits, int inverse)
if (has_vectors & MM_3DNOWEXT)
s->imdct_calc = ff_imdct_calc_3dn2;
if (has_vectors & MM_3DNOWEXT)
/* 3DNowEx for Athlon(XP) */
/* 3DNowEx for K7/K8 */
s->fft_calc = ff_fft_calc_3dn2;
else if (has_vectors & MM_3DNOW)
/* 3DNow! for K6-2/3 */
s->fft_calc = ff_fft_calc_3dn;
if (has_vectors & MM_SSE2)
/* SSE for P4/K8 */
s->fft_calc = ff_fft_calc_sse;
else if ((has_vectors & MM_SSE) &&
s->fft_calc == ff_fft_calc_c)
/* SSE for P3 */
else if (has_vectors & MM_SSE)
/* SSE for P3/P4 */
s->fft_calc = ff_fft_calc_sse;
#else /* HAVE_MMX */
if (has_vectors & MM_ALTIVEC)
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment