diff options
| author | Loren Merritt <lorenm@u.washington.edu> | 2011-02-12 11:48:16 +0000 |
|---|---|---|
| committer | Janne Grunau <janne-ffmpeg@jannau.net> | 2011-02-13 15:36:39 +0100 |
| commit | e6b1ed693ae4098e6b9eabf938fc31ec0b09b120 (patch) | |
| tree | 8afa645b29c35cde5a286b29cbab58fbbed4c1ed /libavcodec/x86/fft.c | |
| parent | 3c33c0e26325f189dbf7f212f8e9042feb83bfb0 (diff) | |
FFT: factor a shuffle out of the inner loop and merge it into fft_permute.
6% faster SSE FFT on Conroe, 2.5% on Penryn.
Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
Diffstat (limited to 'libavcodec/x86/fft.c')
| -rw-r--r-- | libavcodec/x86/fft.c | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/libavcodec/x86/fft.c b/libavcodec/x86/fft.c index 771b1e6649..5ca341d5e4 100644 --- a/libavcodec/x86/fft.c +++ b/libavcodec/x86/fft.c @@ -30,6 +30,7 @@ av_cold void ff_fft_init_mmx(FFTContext *s) s->imdct_half = ff_imdct_half_sse; s->fft_permute = ff_fft_permute_sse; s->fft_calc = ff_fft_calc_sse; + s->fft_permutation = FF_FFT_PERM_SWAP_LSBS; } else if (has_vectors & AV_CPU_FLAG_3DNOWEXT && HAVE_AMD3DNOWEXT) { /* 3DNowEx for K7 */ s->imdct_calc = ff_imdct_calc_3dn2; |