comparison dsputil.c @ 3248:7aa9f80e7954 libavcodec

mmx implementation of 3-point GMC. (5x faster than C)

3247:ac7bfc9bf1a4

        dst+= stride;
        src+= stride;
    }
}

static void gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
                  int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height)
{
    int y, vx, vy;
    const int s= 1<<shift;

    c->put_signed_pixels_clamped = put_signed_pixels_clamped_c;
    c->add_pixels_clamped = add_pixels_clamped_c;
    c->add_pixels8 = add_pixels8_c;
    c->add_pixels4 = add_pixels4_c;
    c->gmc1 = gmc1_c;
    c->gmc = gmc_c;
    c->clear_blocks = clear_blocks_c;
    c->pix_sum = pix_sum_c;
    c->pix_norm1 = pix_norm1_c;

    /* TODO [0] 16  [1] 8 */

3248:7aa9f80e7954

        dst+= stride;
        src+= stride;
    }
}

void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
                  int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height)
{
    int y, vx, vy;
    const int s= 1<<shift;

    c->put_signed_pixels_clamped = put_signed_pixels_clamped_c;
    c->add_pixels_clamped = add_pixels_clamped_c;
    c->add_pixels8 = add_pixels8_c;
    c->add_pixels4 = add_pixels4_c;
    c->gmc1 = gmc1_c;
    c->gmc = ff_gmc_c;
    c->clear_blocks = clear_blocks_c;
    c->pix_sum = pix_sum_c;
    c->pix_norm1 = pix_norm1_c;

    /* TODO [0] 16  [1] 8 */