comparison vp8dsp.c @ 12194:80b142c2e9f7 libavcodec

Change function prototypes for width=8 inner and mbedge loopfilter functions so that it does both U and V planes at the same time. This will have speed advantages when using SSE2 (or higher) optimizations, since we can do both the U and V rows together in a single xmm register. This also renames filter16 to filter16y and filter8 to filter8uv so that it's more obvious what each function is used for.

12193:0a63bed2a00e

    p[ 0*stride] = cm[q0 - a0];
    p[ 1*stride] = cm[q1 - a1];
    p[ 2*stride] = cm[q2 - a2];
}

#define LOOP_FILTER(dir, size, stridea, strideb) \
static void vp8_ ## dir ## _loop_filter ## size ## _c(uint8_t *dst, int stride,\
                                     int flim_E, int flim_I, int hev_thresh)\
{\
    int i;\
\
    for (i = 0; i < size; i++)\

            else\
                filter_mbedge(dst+i*stridea, strideb);\
        }\
}\
\
static void vp8_ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst, int stride,\
                                      int flim_E, int flim_I, int hev_thresh)\
{\
    int i;\
\
    for (i = 0; i < size; i++)\

            else \
                filter_common(dst+i*stridea, strideb, 0);\
        }\
}

LOOP_FILTER(v, 16, 1, stride)
LOOP_FILTER(h, 16, stride, 1)
LOOP_FILTER(v,  8, 1, stride)
LOOP_FILTER(h,  8, stride, 1)

static void vp8_v_loop_filter_simple_c(uint8_t *dst, int stride, int flim)
{
    int i;

{
    dsp->vp8_luma_dc_wht = vp8_luma_dc_wht_c;
    dsp->vp8_idct_add    = vp8_idct_add_c;
    dsp->vp8_idct_dc_add = vp8_idct_dc_add_c;

    dsp->vp8_v_loop_filter16 = vp8_v_loop_filter16_c;
    dsp->vp8_h_loop_filter16 = vp8_h_loop_filter16_c;
    dsp->vp8_v_loop_filter8  = vp8_v_loop_filter8_c;
    dsp->vp8_h_loop_filter8  = vp8_h_loop_filter8_c;

    dsp->vp8_v_loop_filter16_inner = vp8_v_loop_filter16_inner_c;
    dsp->vp8_h_loop_filter16_inner = vp8_h_loop_filter16_inner_c;
    dsp->vp8_v_loop_filter8_inner  = vp8_v_loop_filter8_inner_c;
    dsp->vp8_h_loop_filter8_inner  = vp8_h_loop_filter8_inner_c;

    dsp->vp8_v_loop_filter_simple = vp8_v_loop_filter_simple_c;
    dsp->vp8_h_loop_filter_simple = vp8_h_loop_filter_simple_c;

    VP8_MC_FUNC(0, 16);

12194:80b142c2e9f7

    p[ 0*stride] = cm[q0 - a0];
    p[ 1*stride] = cm[q1 - a1];
    p[ 2*stride] = cm[q2 - a2];
}

#define LOOP_FILTER(dir, size, stridea, strideb, maybe_inline) \
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _c(uint8_t *dst, int stride,\
                                     int flim_E, int flim_I, int hev_thresh)\
{\
    int i;\
\
    for (i = 0; i < size; i++)\

            else\
                filter_mbedge(dst+i*stridea, strideb);\
        }\
}\
\
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst, int stride,\
                                      int flim_E, int flim_I, int hev_thresh)\
{\
    int i;\
\
    for (i = 0; i < size; i++)\

            else \
                filter_common(dst+i*stridea, strideb, 0);\
        }\
}

LOOP_FILTER(v, 16, 1, stride,)
LOOP_FILTER(h, 16, stride, 1,)

#define UV_LOOP_FILTER(dir, stridea, strideb) \
LOOP_FILTER(dir, 8, stridea, strideb, av_always_inline) \
static void vp8_ ## dir ## _loop_filter8uv_c(uint8_t *dstU, uint8_t *dstV, int stride,\
                                      int fE, int fI, int hev_thresh)\
{\
  vp8_ ## dir ## _loop_filter8_c(dstU, stride, fE, fI, hev_thresh);\
  vp8_ ## dir ## _loop_filter8_c(dstV, stride, fE, fI, hev_thresh);\
}\
static void vp8_ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU, uint8_t *dstV, int stride,\
                                      int fE, int fI, int hev_thresh)\
{\
  vp8_ ## dir ## _loop_filter8_inner_c(dstU, stride, fE, fI, hev_thresh);\
  vp8_ ## dir ## _loop_filter8_inner_c(dstV, stride, fE, fI, hev_thresh);\
}

UV_LOOP_FILTER(v, 1, stride)
UV_LOOP_FILTER(h, stride, 1)

static void vp8_v_loop_filter_simple_c(uint8_t *dst, int stride, int flim)
{
    int i;

{
    dsp->vp8_luma_dc_wht = vp8_luma_dc_wht_c;
    dsp->vp8_idct_add    = vp8_idct_add_c;
    dsp->vp8_idct_dc_add = vp8_idct_dc_add_c;

    dsp->vp8_v_loop_filter16y = vp8_v_loop_filter16_c;
    dsp->vp8_h_loop_filter16y = vp8_h_loop_filter16_c;
    dsp->vp8_v_loop_filter8uv = vp8_v_loop_filter8uv_c;
    dsp->vp8_h_loop_filter8uv = vp8_h_loop_filter8uv_c;

    dsp->vp8_v_loop_filter16y_inner = vp8_v_loop_filter16_inner_c;
    dsp->vp8_h_loop_filter16y_inner = vp8_h_loop_filter16_inner_c;
    dsp->vp8_v_loop_filter8uv_inner = vp8_v_loop_filter8uv_inner_c;
    dsp->vp8_h_loop_filter8uv_inner = vp8_h_loop_filter8uv_inner_c;

    dsp->vp8_v_loop_filter_simple = vp8_v_loop_filter_simple_c;
    dsp->vp8_h_loop_filter_simple = vp8_h_loop_filter_simple_c;

    VP8_MC_FUNC(0, 16);