comparison dsputil.c @ 5737:efa3c1f9259a libavcodec

sse2 version of compute_autocorr(). 4x faster than c (somehow, even though doubles only allow 2x simd). overal flac encoding: 15-50% faster on core2, 4-11% on k8, 3-13% on p4.

5736:810067f2c33d

/* snow.c */
void ff_spatial_dwt(int *buffer, int width, int height, int stride, int type, int decomposition_count);

/* vorbis.c */
void vorbis_inverse_coupling(float *mag, float *ang, int blocksize);

uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
uint32_t ff_squareTbl[512] = {0, };

const uint8_t ff_zigzag_direct[64] = {

#endif

#ifdef CONFIG_VORBIS_DECODER
    c->vorbis_inverse_coupling = vorbis_inverse_coupling;
#endif
    c->vector_fmul = vector_fmul_c;
    c->vector_fmul_reverse = vector_fmul_reverse_c;
    c->vector_fmul_add_add = ff_vector_fmul_add_add_c;
    c->float_to_int16 = ff_float_to_int16_c;

5737:efa3c1f9259a

/* snow.c */
void ff_spatial_dwt(int *buffer, int width, int height, int stride, int type, int decomposition_count);

/* vorbis.c */
void vorbis_inverse_coupling(float *mag, float *ang, int blocksize);

/* flacenc.c */
void ff_flac_compute_autocorr(const int32_t *data, int len, int lag, double *autoc);

uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
uint32_t ff_squareTbl[512] = {0, };

const uint8_t ff_zigzag_direct[64] = {

#endif

#ifdef CONFIG_VORBIS_DECODER
    c->vorbis_inverse_coupling = vorbis_inverse_coupling;
#endif
#ifdef CONFIG_FLAC_ENCODER
    c->flac_compute_autocorr = ff_flac_compute_autocorr;
#endif
    c->vector_fmul = vector_fmul_c;
    c->vector_fmul_reverse = vector_fmul_reverse_c;
    c->vector_fmul_add_add = ff_vector_fmul_add_add_c;
    c->float_to_int16 = ff_float_to_int16_c;