comparison h264_direct.c @ 11104:bb877c9cb102 libavcodec

Detect spatial direct MBs partitioned smaller than 16x16 that can be partitioned as 16x16 (except ones changing interlacing relative to the colocated MB). 20 cycles slower during MV generation 175 cycles faster during MC

11103:3bc0125e19fe

                b= pack16to32(mv[1][0],mv[1][1]);
            }
            fill_rectangle(&h->mv_cache[0][scan8[0]], 4, 4, 8, a, 4);
            fill_rectangle(&h->mv_cache[1][scan8[0]], 4, 4, 8, b, 4);
        }else{
            for(i8=0; i8<4; i8++){
                const int x8 = i8&1;
                const int y8 = i8>>1;

                if(is_b8x8 && !IS_DIRECT(h->sub_mb_type[i8]))

                        if(FFABS(mv_col[0]) <= 1 && FFABS(mv_col[1]) <= 1){
                            if(ref[0] == 0)
                                fill_rectangle(&h->mv_cache[0][scan8[i8*4]], 2, 2, 8, 0, 4);
                            if(ref[1] == 0)
                                fill_rectangle(&h->mv_cache[1][scan8[i8*4]], 2, 2, 8, 0, 4);
                        }
                    }else{
                        int m=0;
                    for(i4=0; i4<4; i4++){
                        const int16_t *mv_col = l1mv[x8*2 + (i4&1) + (y8*2 + (i4>>1))*b4_stride];

                            m++;
                        }
                    }
                    if(!(m&3))
                        h->sub_mb_type[i8]+= MB_TYPE_16x16 - MB_TYPE_8x8;
                    }
                }
            }
        }
    }else{ /* direct temporal mv pred */
        const int *map_col_to_list0[2] = {h->map_col_to_list0[0], h->map_col_to_list0[1]};
        const int *dist_scale_factor = h->dist_scale_factor;
        int ref_offset;

11104:bb877c9cb102

                b= pack16to32(mv[1][0],mv[1][1]);
            }
            fill_rectangle(&h->mv_cache[0][scan8[0]], 4, 4, 8, a, 4);
            fill_rectangle(&h->mv_cache[1][scan8[0]], 4, 4, 8, b, 4);
        }else{
            int n=0;
            for(i8=0; i8<4; i8++){
                const int x8 = i8&1;
                const int y8 = i8>>1;

                if(is_b8x8 && !IS_DIRECT(h->sub_mb_type[i8]))

                        if(FFABS(mv_col[0]) <= 1 && FFABS(mv_col[1]) <= 1){
                            if(ref[0] == 0)
                                fill_rectangle(&h->mv_cache[0][scan8[i8*4]], 2, 2, 8, 0, 4);
                            if(ref[1] == 0)
                                fill_rectangle(&h->mv_cache[1][scan8[i8*4]], 2, 2, 8, 0, 4);
                            n+=4;
                        }
                    }else{
                        int m=0;
                    for(i4=0; i4<4; i4++){
                        const int16_t *mv_col = l1mv[x8*2 + (i4&1) + (y8*2 + (i4>>1))*b4_stride];

                            m++;
                        }
                    }
                    if(!(m&3))
                        h->sub_mb_type[i8]+= MB_TYPE_16x16 - MB_TYPE_8x8;
                    n+=m;
                    }
                }
            }
            if(!is_b8x8 && !(n&15))
                *mb_type= (*mb_type & ~(MB_TYPE_8x8|MB_TYPE_16x8|MB_TYPE_8x16|MB_TYPE_P1L0|MB_TYPE_P1L1))|MB_TYPE_16x16|MB_TYPE_DIRECT2;
        }
    }else{ /* direct temporal mv pred */
        const int *map_col_to_list0[2] = {h->map_col_to_list0[0], h->map_col_to_list0[1]};
        const int *dist_scale_factor = h->dist_scale_factor;
        int ref_offset;