comparison dsputil.c @ 6385:40fbc878ce3f libavcodec

pseudo-simd add_bytes and diff_bytes 2x faster than scalar in 32bit, 4x faster in 64bit (as opposed to 8x in mmx)

6384:0a403ade8c81

/* pngdec.c */
void ff_add_png_paeth_prediction(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp);

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

const uint8_t ff_zigzag_direct[64] = {
    0,   1,  8, 16,  9,  2,  3, 10,
    17, 24, 32, 25, 18, 11,  4,  5,
    12, 19, 26, 33, 40, 48, 41, 34,

{
    memset(blocks, 0, sizeof(DCTELEM)*6*64);
}

static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
    int i;
    for(i=0; i+7<w; i+=8){
        dst[i+0] += src[i+0];
        dst[i+1] += src[i+1];
        dst[i+2] += src[i+2];
        dst[i+3] += src[i+3];
        dst[i+4] += src[i+4];
        dst[i+5] += src[i+5];
        dst[i+6] += src[i+6];
        dst[i+7] += src[i+7];
    }
    for(; i<w; i++)
        dst[i+0] += src[i+0];
}

static void add_bytes_l2_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
    int i;
    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
        long a = *(long*)(src1+i);
        long b = *(long*)(src2+i);
        *(long*)(dst+i) = ((a&0x7f7f7f7f7f7f7f7fL) + (b&0x7f7f7f7f7f7f7f7fL)) ^ ((a^b)&0x8080808080808080L);
    }
    for(; i<w; i++)
        dst[i] = src1[i]+src2[i];
}

static void diff_bytes_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
    int i;
    for(i=0; i+7<w; i+=8){
        dst[i+0] = src1[i+0]-src2[i+0];
        dst[i+1] = src1[i+1]-src2[i+1];
        dst[i+2] = src1[i+2]-src2[i+2];
        dst[i+3] = src1[i+3]-src2[i+3];
        dst[i+4] = src1[i+4]-src2[i+4];
        dst[i+5] = src1[i+5]-src2[i+5];
        dst[i+6] = src1[i+6]-src2[i+6];
        dst[i+7] = src1[i+7]-src2[i+7];
    }
    for(; i<w; i++)
        dst[i+0] = src1[i+0]-src2[i+0];
}

6385:40fbc878ce3f

/* pngdec.c */
void ff_add_png_paeth_prediction(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp);

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

static const unsigned long pb_7f = 0x7f7f7f7f7f7f7f7fUL;
static const unsigned long pb_80 = 0x8080808080808080UL;

const uint8_t ff_zigzag_direct[64] = {
    0,   1,  8, 16,  9,  2,  3, 10,
    17, 24, 32, 25, 18, 11,  4,  5,
    12, 19, 26, 33, 40, 48, 41, 34,

{
    memset(blocks, 0, sizeof(DCTELEM)*6*64);
}

static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
    long i;
    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
        long a = *(long*)(src+i);
        long b = *(long*)(dst+i);
        *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
    }
    for(; i<w; i++)
        dst[i+0] += src[i+0];
}

static void add_bytes_l2_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
    long i;
    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
        long a = *(long*)(src1+i);
        long b = *(long*)(src2+i);
        *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
    }
    for(; i<w; i++)
        dst[i] = src1[i]+src2[i];
}

static void diff_bytes_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
    long i;
#ifndef HAVE_FAST_UNALIGNED
    if((long)src2 & (sizeof(long)-1)){
    for(i=0; i+7<w; i+=8){
        dst[i+0] = src1[i+0]-src2[i+0];
        dst[i+1] = src1[i+1]-src2[i+1];
        dst[i+2] = src1[i+2]-src2[i+2];
        dst[i+3] = src1[i+3]-src2[i+3];
        dst[i+4] = src1[i+4]-src2[i+4];
        dst[i+5] = src1[i+5]-src2[i+5];
        dst[i+6] = src1[i+6]-src2[i+6];
        dst[i+7] = src1[i+7]-src2[i+7];
    }
    }else
#endif
    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
        long a = *(long*)(src1+i);
        long b = *(long*)(src2+i);
        *(long*)(dst+i) = ((a|pb_80) - (b&pb_7f)) ^ ((a^b^pb_80)&pb_80);
    }
    for(; i<w; i++)
        dst[i+0] = src1[i+0]-src2[i+0];
}