Mercurial > libavcodec.hg
comparison x86/dsputil_mmx.c @ 10964:abb3b23bda35 libavcodec
Implement an sse version of scalarproduct_float().
| author | alexc |
|---|---|
| date | Fri, 22 Jan 2010 23:07:58 +0000 |
| parents | 34a65026fa06 |
| children | 001eb7e3e2d3 |
comparison
equal
deleted
inserted
replaced
| 10963:81033a080136 | 10964:abb3b23bda35 |
|---|---|
| 2508 int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8); | 2508 int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8); |
| 2509 void ff_snow_inner_add_yblock_mmx(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, | 2509 void ff_snow_inner_add_yblock_mmx(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, |
| 2510 int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8); | 2510 int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8); |
| 2511 | 2511 |
| 2512 | 2512 |
| 2513 float ff_scalarproduct_float_sse(const float *v1, const float *v2, int order); | |
| 2514 | |
| 2513 void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) | 2515 void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx) |
| 2514 { | 2516 { |
| 2515 mm_flags = mm_support(); | 2517 mm_flags = mm_support(); |
| 2516 | 2518 |
| 2517 if (avctx->dsp_mask) { | 2519 if (avctx->dsp_mask) { |
| 2963 c->vector_fmul_window = vector_fmul_window_sse; | 2965 c->vector_fmul_window = vector_fmul_window_sse; |
| 2964 c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse; | 2966 c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse; |
| 2965 c->vector_clipf = vector_clipf_sse; | 2967 c->vector_clipf = vector_clipf_sse; |
| 2966 c->float_to_int16 = float_to_int16_sse; | 2968 c->float_to_int16 = float_to_int16_sse; |
| 2967 c->float_to_int16_interleave = float_to_int16_interleave_sse; | 2969 c->float_to_int16_interleave = float_to_int16_interleave_sse; |
| 2970 #if HAVE_YASM | |
| 2971 c->scalarproduct_float = ff_scalarproduct_float_sse; | |
| 2972 #endif | |
| 2968 } | 2973 } |
| 2969 if(mm_flags & FF_MM_3DNOW) | 2974 if(mm_flags & FF_MM_3DNOW) |
| 2970 c->vector_fmul_add = vector_fmul_add_3dnow; // faster than sse | 2975 c->vector_fmul_add = vector_fmul_add_3dnow; // faster than sse |
| 2971 if(mm_flags & FF_MM_SSE2){ | 2976 if(mm_flags & FF_MM_SSE2){ |
| 2972 c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse2; | 2977 c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_sse2; |