diff options
Diffstat (limited to 'gst-libs/gst/idct')
| -rw-r--r-- | gst-libs/gst/idct/fastintidct.c | 26 | ||||
| -rw-r--r-- | gst-libs/gst/idct/floatidct.c | 6 | ||||
| -rw-r--r-- | gst-libs/gst/idct/ieeetest.c | 60 | ||||
| -rw-r--r-- | gst-libs/gst/idct/intidct.c | 108 | ||||
| -rw-r--r-- | gst-libs/gst/idct/mmx32idct.c | 572 |
5 files changed, 387 insertions, 385 deletions
diff --git a/gst-libs/gst/idct/fastintidct.c b/gst-libs/gst/idct/fastintidct.c index 9bb1436d..d4caaf38 100644 --- a/gst-libs/gst/idct/fastintidct.c +++ b/gst-libs/gst/idct/fastintidct.c @@ -45,17 +45,17 @@ /* this code assumes >> to be a two's-complement arithmetic */ /* right shift: (-2)>>1 == -1 , (-3)>>1 == -2 */ -#define W1 2841 /* 2048*sqrt(2)*cos(1*pi/16) */ -#define W2 2676 /* 2048*sqrt(2)*cos(2*pi/16) */ -#define W3 2408 /* 2048*sqrt(2)*cos(3*pi/16) */ -#define W5 1609 /* 2048*sqrt(2)*cos(5*pi/16) */ -#define W6 1108 /* 2048*sqrt(2)*cos(6*pi/16) */ -#define W7 565 /* 2048*sqrt(2)*cos(7*pi/16) */ +#define W1 2841 /* 2048*sqrt(2)*cos(1*pi/16) */ +#define W2 2676 /* 2048*sqrt(2)*cos(2*pi/16) */ +#define W3 2408 /* 2048*sqrt(2)*cos(3*pi/16) */ +#define W5 1609 /* 2048*sqrt(2)*cos(5*pi/16) */ +#define W6 1108 /* 2048*sqrt(2)*cos(6*pi/16) */ +#define W7 565 /* 2048*sqrt(2)*cos(7*pi/16) */ #include "dct.h" /* private data */ -static short iclip[1024]; /* clipping table */ +static short iclip[1024]; /* clipping table */ static short *iclp; /* private prototypes */ @@ -80,13 +80,13 @@ idctrow (blk) /* shortcut */ if (!((x1 = blk[4] << 11) | (x2 = blk[6]) | (x3 = blk[2]) | - (x4 = blk[1]) | (x5 = blk[7]) | (x6 = blk[5]) | (x7 = blk[3]))) { + (x4 = blk[1]) | (x5 = blk[7]) | (x6 = blk[5]) | (x7 = blk[3]))) { blk[0] = blk[1] = blk[2] = blk[3] = blk[4] = blk[5] = blk[6] = blk[7] = - blk[0] << 3; + blk[0] << 3; return; } - x0 = (blk[0] << 11) + 128; /* for proper rounding in the fourth stage */ + x0 = (blk[0] << 11) + 128; /* for proper rounding in the fourth stage */ /* first stage */ x8 = W7 * (x4 + x5); @@ -143,10 +143,10 @@ idctcol (blk) /* shortcut */ if (!((x1 = (blk[8 * 4] << 8)) | (x2 = blk[8 * 6]) | (x3 = blk[8 * 2]) | - (x4 = blk[8 * 1]) | (x5 = blk[8 * 7]) | (x6 = blk[8 * 5]) | (x7 = - blk[8 * 3]))) { + (x4 = blk[8 * 1]) | (x5 = blk[8 * 7]) | (x6 = blk[8 * 5]) | (x7 = + blk[8 * 3]))) { blk[8 * 0] = blk[8 * 1] = blk[8 * 2] = blk[8 * 3] = blk[8 * 4] = - blk[8 * 5] = blk[8 * 6] = blk[8 * 7] = iclp[(blk[8 * 0] + 32) >> 6]; + blk[8 * 5] = blk[8 * 6] = blk[8 * 7] = iclp[(blk[8 * 0] + 32) >> 6]; return; } diff --git a/gst-libs/gst/idct/floatidct.c b/gst-libs/gst/idct/floatidct.c index 0fa1e830..65e01d40 100644 --- a/gst-libs/gst/idct/floatidct.c +++ b/gst-libs/gst/idct/floatidct.c @@ -66,7 +66,7 @@ gst_idct_init_float_idct () scale = (freq == 0) ? sqrt (0.125) : 0.5; for (time = 0; time < 8; time++) gst_idct_float_c[freq][time] = - scale * cos ((PI / 8.0) * freq * (time + 0.5)); + scale * cos ((PI / 8.0) * freq * (time + 0.5)); } } @@ -85,7 +85,7 @@ gst_idct_float_idct (block) partial_product = 0.0; for (k = 0; k < 8; k++) - partial_product += gst_idct_float_c[k][j] * block[8 * i + k]; + partial_product += gst_idct_float_c[k][j] * block[8 * i + k]; tmp[8 * i + j] = partial_product; } @@ -98,7 +98,7 @@ gst_idct_float_idct (block) partial_product = 0.0; for (k = 0; k < 8; k++) - partial_product += gst_idct_float_c[k][i] * tmp[8 * k + j]; + partial_product += gst_idct_float_c[k][i] * tmp[8 * k + j]; v = (int) floor (partial_product + 0.5); block[8 * i + j] = (v < -256) ? -256 : ((v > 255) ? 255 : v); diff --git a/gst-libs/gst/idct/ieeetest.c b/gst-libs/gst/idct/ieeetest.c index d26181c1..7af101e6 100644 --- a/gst-libs/gst/idct/ieeetest.c +++ b/gst-libs/gst/idct/ieeetest.c @@ -52,10 +52,10 @@ main (int argc, char **argv) int i, j; double max, total; int method; - DCTELEM block[DCTSIZE2]; /* random source data */ - DCTELEM refcoefs[DCTSIZE2]; /* coefs from reference FDCT */ - DCTELEM refout[DCTSIZE2]; /* output from reference IDCT */ - DCTELEM testout[DCTSIZE2]; /* output from test IDCT */ + DCTELEM block[DCTSIZE2]; /* random source data */ + DCTELEM refcoefs[DCTSIZE2]; /* coefs from reference FDCT */ + DCTELEM refout[DCTSIZE2]; /* output from reference IDCT */ + DCTELEM testout[DCTSIZE2]; /* output from test IDCT */ GstIDCT *idct; guint64 tscstart, tscmin = ~0, tscmax = 0; guint64 tscstop; @@ -96,9 +96,9 @@ main (int argc, char **argv) /* clip */ for (i = 0; i < DCTSIZE2; i++) { if (refcoefs[i] < -2048) - refcoefs[i] = -2048; + refcoefs[i] = -2048; else if (refcoefs[i] > 2047) - refcoefs[i] = 2047; + refcoefs[i] = 2047; } /* perform reference IDCT */ @@ -107,17 +107,17 @@ main (int argc, char **argv) /* clip */ for (i = 0; i < DCTSIZE2; i++) { if (refout[i] < -256) - refout[i] = -256; + refout[i] = -256; else if (refout[i] > 255) - refout[i] = 255; + refout[i] = 255; } /* perform test IDCT */ if (GST_IDCT_TRANSPOSE (idct)) { for (j = 0; j < DCTSIZE; j++) { - for (i = 0; i < DCTSIZE; i++) { - testout[i * DCTSIZE + j] = refcoefs[j * DCTSIZE + i]; - } + for (i = 0; i < DCTSIZE; i++) { + testout[i * DCTSIZE + j] = refcoefs[j * DCTSIZE + i]; + } } } else { memcpy (testout, refcoefs, sizeof (DCTELEM) * DCTSIZE2); @@ -135,9 +135,9 @@ main (int argc, char **argv) /* clip */ for (i = 0; i < DCTSIZE2; i++) { if (testout[i] < -256) - testout[i] = -256; + testout[i] = -256; else if (testout[i] > 255) - testout[i] = 255; + testout[i] = 255; } /* accumulate error stats */ @@ -147,9 +147,9 @@ main (int argc, char **argv) sumerrs[i] += err; sumsqerrs[i] += err * err; if (err < 0) - err = -err; + err = -err; if (maxerr[i] < err) - maxerr[i] = err; + maxerr[i] = err; } if (curiter % 100 == 99) { @@ -290,7 +290,7 @@ dct_init (void) for (b = 0; b < 8; b++) { tmp = cos ((double) ((a + a + 1) * b) * (3.14159265358979323846 / 16.0)); if (b == 0) - tmp /= sqrt (2.0); + tmp /= sqrt (2.0); coslu[a][b] = tmp * 0.5; } } @@ -307,11 +307,11 @@ ref_fdct (DCTELEM block[8][8]) for (u = 0; u < 8; u++) { tmp = 0.0; for (y = 0; y < 8; y++) { - tmp2 = 0.0; - for (x = 0; x < 8; x++) { - tmp2 += (double) block[y][x] * coslu[x][u]; - } - tmp += coslu[y][v] * tmp2; + tmp2 = 0.0; + for (x = 0; x < 8; x++) { + tmp2 += (double) block[y][x] * coslu[x][u]; + } + tmp += coslu[y][v] * tmp2; } res[v][u] = tmp; } @@ -321,9 +321,9 @@ ref_fdct (DCTELEM block[8][8]) for (u = 0; u < 8; u++) { tmp = res[v][u]; if (tmp < 0.0) { - x = -((int) (0.5 - tmp)); + x = -((int) (0.5 - tmp)); } else { - x = (int) (tmp + 0.5); + x = (int) (tmp + 0.5); } block[v][u] = (DCTELEM) x; } @@ -342,11 +342,11 @@ ref_idct (DCTELEM block[8][8]) for (x = 0; x < 8; x++) { tmp = 0.0; for (v = 0; v < 8; v++) { - tmp2 = 0.0; - for (u = 0; u < 8; u++) { - tmp2 += (double) block[v][u] * coslu[x][u]; - } - tmp += coslu[y][v] * tmp2; + tmp2 = 0.0; + for (u = 0; u < 8; u++) { + tmp2 += (double) block[v][u] * coslu[x][u]; + } + tmp += coslu[y][v] * tmp2; } res[y][x] = tmp; } @@ -356,9 +356,9 @@ ref_idct (DCTELEM block[8][8]) for (u = 0; u < 8; u++) { tmp = res[v][u]; if (tmp < 0.0) { - x = -((int) (0.5 - tmp)); + x = -((int) (0.5 - tmp)); } else { - x = (int) (tmp + 0.5); + x = (int) (tmp + 0.5); } block[v][u] = (DCTELEM) x; } diff --git a/gst-libs/gst/idct/intidct.c b/gst-libs/gst/idct/intidct.c index 42f0ac84..d2945348 100644 --- a/gst-libs/gst/idct/intidct.c +++ b/gst-libs/gst/idct/intidct.c @@ -51,7 +51,7 @@ */ #if DCTSIZE != 8 -Sorry, this code only copes with 8 x8 DCTs. /* deliberate syntax err */ +Sorry, this code only copes with 8 x8 DCTs. /* deliberate syntax err */ #endif /* * A 2-D IDCT can be done by 1-D IDCT on each row followed by 1-D IDCT @@ -93,7 +93,7 @@ Sorry, this code only copes with 8 x8 DCTs. /* deliberate syntax err */ #define PASS1_BITS 2 #else #define CONST_BITS 13 -#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ +#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ #endif #define ONE ((INT32) 1) #define CONST_SCALE (ONE << CONST_BITS) @@ -106,18 +106,18 @@ Sorry, this code only copes with 8 x8 DCTs. /* deliberate syntax err */ * (With a reasonable C compiler, you can just rely on the FIX() macro...) */ #if CONST_BITS == 13 -#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ -#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ -#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ -#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ -#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ -#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ -#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ -#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ -#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ -#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ -#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ -#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ +#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ +#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ +#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ +#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ +#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ +#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ +#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ +#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ +#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ +#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ +#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ +#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ #else #define FIX_0_298631336 FIX(0.298631336) #define FIX_0_390180644 FIX(0.390180644) @@ -148,14 +148,14 @@ Sorry, this code only copes with 8 x8 DCTs. /* deliberate syntax err */ * NB: for 12-bit samples, a full 32-bit multiplication will be needed. */ #ifdef EIGHT_BIT_SAMPLES -#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ +#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ #define MULTIPLY(var,const) (((INT16) (var)) * ((INT16) (const))) #endif -#ifdef SHORTxLCONST_32 /* known to work with Microsoft C 6.0 */ +#ifdef SHORTxLCONST_32 /* known to work with Microsoft C 6.0 */ #define MULTIPLY(var,const) (((INT16) (var)) * ((INT32) (const))) #endif #endif -#ifndef MULTIPLY /* default definition */ +#ifndef MULTIPLY /* default definition */ #define MULTIPLY(var,const) ((var) * (const)) #endif /* @@ -186,7 +186,7 @@ gst_idct_int_idct (DCTBLOCK data) */ if ((dataptr[1] | dataptr[2] | dataptr[3] | dataptr[4] | - dataptr[5] | dataptr[6] | dataptr[7]) == 0) { + dataptr[5] | dataptr[6] | dataptr[7]) == 0) { /* AC terms all zero */ DCTELEM dcval = (DCTELEM) (dataptr[0] << PASS1_BITS); @@ -199,7 +199,7 @@ gst_idct_int_idct (DCTBLOCK data) dataptr[6] = dcval; dataptr[7] = dcval; - dataptr += DCTSIZE; /* advance pointer to next row */ + dataptr += DCTSIZE; /* advance pointer to next row */ continue; } @@ -234,16 +234,16 @@ gst_idct_int_idct (DCTBLOCK data) z2 = tmp1 + tmp2; z3 = tmp0 + tmp2; z4 = tmp1 + tmp3; - z5 = MULTIPLY (z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ + z5 = MULTIPLY (z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ - tmp0 = MULTIPLY (tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ - tmp1 = MULTIPLY (tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ - tmp2 = MULTIPLY (tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ - tmp3 = MULTIPLY (tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */ - z1 = MULTIPLY (z1, -FIX_0_899976223); /* sqrt(2) * (c7-c3) */ - z2 = MULTIPLY (z2, -FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ - z3 = MULTIPLY (z3, -FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ - z4 = MULTIPLY (z4, -FIX_0_390180644); /* sqrt(2) * (c5-c3) */ + tmp0 = MULTIPLY (tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ + tmp1 = MULTIPLY (tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ + tmp2 = MULTIPLY (tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ + tmp3 = MULTIPLY (tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */ + z1 = MULTIPLY (z1, -FIX_0_899976223); /* sqrt(2) * (c7-c3) */ + z2 = MULTIPLY (z2, -FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ + z3 = MULTIPLY (z3, -FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ + z4 = MULTIPLY (z4, -FIX_0_390180644); /* sqrt(2) * (c5-c3) */ z3 += z5; z4 += z5; @@ -264,7 +264,7 @@ gst_idct_int_idct (DCTBLOCK data) dataptr[3] = (DCTELEM) DESCALE (tmp13 + tmp0, CONST_BITS - PASS1_BITS); dataptr[4] = (DCTELEM) DESCALE (tmp13 - tmp0, CONST_BITS - PASS1_BITS); - dataptr += DCTSIZE; /* advance pointer to next row */ + dataptr += DCTSIZE; /* advance pointer to next row */ } /* Pass 2: process columns. */ @@ -283,8 +283,8 @@ gst_idct_int_idct (DCTBLOCK data) #ifndef NO_ZERO_COLUMN_TEST if ((dataptr[DCTSIZE * 1] | dataptr[DCTSIZE * 2] | dataptr[DCTSIZE * 3] | - dataptr[DCTSIZE * 4] | dataptr[DCTSIZE * 5] | dataptr[DCTSIZE * 6] | - dataptr[DCTSIZE * 7]) == 0) { + dataptr[DCTSIZE * 4] | dataptr[DCTSIZE * 5] | dataptr[DCTSIZE * 6] | + dataptr[DCTSIZE * 7]) == 0) { /* AC terms all zero */ DCTELEM dcval = (DCTELEM) DESCALE ((INT32) dataptr[0], PASS1_BITS + 3); @@ -297,7 +297,7 @@ gst_idct_int_idct (DCTBLOCK data) dataptr[DCTSIZE * 6] = dcval; dataptr[DCTSIZE * 7] = dcval; - dataptr++; /* advance pointer to next column */ + dataptr++; /* advance pointer to next column */ continue; } #endif @@ -313,11 +313,11 @@ gst_idct_int_idct (DCTBLOCK data) tmp3 = z1 + MULTIPLY (z2, FIX_0_765366865); tmp0 = - ((INT32) dataptr[DCTSIZE * 0] + - (INT32) dataptr[DCTSIZE * 4]) << CONST_BITS; + ((INT32) dataptr[DCTSIZE * 0] + + (INT32) dataptr[DCTSIZE * 4]) << CONST_BITS; tmp1 = - ((INT32) dataptr[DCTSIZE * 0] - - (INT32) dataptr[DCTSIZE * 4]) << CONST_BITS; + ((INT32) dataptr[DCTSIZE * 0] - + (INT32) dataptr[DCTSIZE * 4]) << CONST_BITS; tmp10 = tmp0 + tmp3; tmp13 = tmp0 - tmp3; @@ -337,16 +337,16 @@ gst_idct_int_idct (DCTBLOCK data) z2 = tmp1 + tmp2; z3 = tmp0 + tmp2; z4 = tmp1 + tmp3; - z5 = MULTIPLY (z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ + z5 = MULTIPLY (z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ - tmp0 = MULTIPLY (tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ - tmp1 = MULTIPLY (tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ - tmp2 = MULTIPLY (tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ - tmp3 = MULTIPLY (tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */ - z1 = MULTIPLY (z1, -FIX_0_899976223); /* sqrt(2) * (c7-c3) */ - z2 = MULTIPLY (z2, -FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ - z3 = MULTIPLY (z3, -FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ - z4 = MULTIPLY (z4, -FIX_0_390180644); /* sqrt(2) * (c5-c3) */ + tmp0 = MULTIPLY (tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ + tmp1 = MULTIPLY (tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ + tmp2 = MULTIPLY (tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ + tmp3 = MULTIPLY (tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */ + z1 = MULTIPLY (z1, -FIX_0_899976223); /* sqrt(2) * (c7-c3) */ + z2 = MULTIPLY (z2, -FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ + z3 = MULTIPLY (z3, -FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ + z4 = MULTIPLY (z4, -FIX_0_390180644); /* sqrt(2) * (c5-c3) */ z3 += z5; z4 += z5; @@ -359,22 +359,22 @@ gst_idct_int_idct (DCTBLOCK data) /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ dataptr[DCTSIZE * 0] = (DCTELEM) DESCALE (tmp10 + tmp3, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 7] = (DCTELEM) DESCALE (tmp10 - tmp3, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 1] = (DCTELEM) DESCALE (tmp11 + tmp2, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 6] = (DCTELEM) DESCALE (tmp11 - tmp2, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 2] = (DCTELEM) DESCALE (tmp12 + tmp1, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 5] = (DCTELEM) DESCALE (tmp12 - tmp1, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 3] = (DCTELEM) DESCALE (tmp13 + tmp0, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); dataptr[DCTSIZE * 4] = (DCTELEM) DESCALE (tmp13 - tmp0, - CONST_BITS + PASS1_BITS + 3); + CONST_BITS + PASS1_BITS + 3); - dataptr++; /* advance pointer to next column */ + dataptr++; /* advance pointer to next column */ } } diff --git a/gst-libs/gst/idct/mmx32idct.c b/gst-libs/gst/idct/mmx32idct.c index cd191f0c..d5bbbdb9 100644 --- a/gst-libs/gst/idct/mmx32idct.c +++ b/gst-libs/gst/idct/mmx32idct.c @@ -117,15 +117,15 @@ qword ptr equ mword ptr */ #endif #include <mmx.h>
-#define BITS_INV_ACC 4 /*; 4 or 5 for IEEE
*/ +#define BITS_INV_ACC 4 /*; 4 or 5 for IEEE
*/ /* 5 yields higher accuracy, but lessens dynamic range on the input matrix
*/ #define SHIFT_INV_ROW (16 - BITS_INV_ACC)
-#define SHIFT_INV_COL (1 + BITS_INV_ACC +14 ) /* changed from Intel's val)
*/ +#define SHIFT_INV_COL (1 + BITS_INV_ACC +14 ) /* changed from Intel's val)
*/ /*#define SHIFT_INV_COL (1 + BITS_INV_ACC )
*/ #define RND_INV_ROW (1 << (SHIFT_INV_ROW-1))
#define RND_INV_COL (1 << (SHIFT_INV_COL-1))
-#define RND_INV_CORR (RND_INV_COL - 1) /*; correction -1.0 and round
*/ +#define RND_INV_CORR (RND_INV_COL - 1) /*; correction -1.0 and round
*/ /*#define RND_INV_ROW (1024 * (6 - BITS_INV_ACC)) //; 1 << (SHIFT_INV_ROW-1)
*/ /*#define RND_INV_COL (16 * (BITS_INV_ACC - 3)) //; 1 << (SHIFT_INV_COL-1)
*/ @@ -158,12 +158,12 @@ const static long r_inv_row[2] = { RND_INV_ROW, RND_INV_ROW }; /*const static short ocos_4_16[4] = {23170, 23170, 23170, 23170 }; //cos * (2<<15) + 0.5
*/ /*concatenated table, for forward DCT transformation
*/ -const static short tg_all_16[] = {
13036, 13036, 13036, 13036, /* tg * (2<<16) + 0.5
*/ - 27146, 27146, 27146, 27146, /*tg * (2<<16) + 0.5
*/ - -21746, -21746, -21746, -21746, /* tg * (2<<16) + 0.5
*/ - -19195, -19195, -19195, -19195, /*cos * (2<<16) + 0.5
*/ +const static short tg_all_16[] = {
13036, 13036, 13036, 13036, /* tg * (2<<16) + 0.5
*/ + 27146, 27146, 27146, 27146, /*tg * (2<<16) + 0.5
*/ + -21746, -21746, -21746, -21746, /* tg * (2<<16) + 0.5
*/ + -19195, -19195, -19195, -19195, /*cos * (2<<16) + 0.5
*/ 23170, 23170, 23170, 23170 -}; /*cos * (2<<15) + 0.5
*/ +}; /*cos * (2<<15) + 0.5
*/ #define tg_1_16 (tg_all_16 + 0)
@@ -233,51 +233,51 @@ IF _MMX ; MMX code ;=============================================================================
/*; Table for rows 0,4 - constants are multiplied by cos_4_16
*/ -const short tab_i_04[] = {
16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ - 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ - 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ - -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ - 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ - 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ - 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ +const short tab_i_04[] = {
16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ + 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ + 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ + -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ + 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ + 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ + 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ -22725, 19266, -12873, -22725 -}; /*w31 w29 w27 w25
*/ +}; /*w31 w29 w27 w25
*/ /*; Table for rows 1,7 - constants are multiplied by cos_1_16
*/ -const short tab_i_17[] = {
22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ - 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ - 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ - -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ - 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ - 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ - 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ +const short tab_i_17[] = {
22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ + 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ + 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ + -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ + 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ + 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ + 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ -31521, 26722, -17855, -31521 -}; /* w31 w29 w27 w25
*/ +}; /* w31 w29 w27 w25
*/ /*; Table for rows 2,6 - constants are multiplied by cos_2_16
*/ -const short tab_i_26[] = {
21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ - 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ - 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ - -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ - 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ - 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ - 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ +const short tab_i_26[] = {
21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ + 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ + 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ + -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ + 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ + 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ + 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ -29692, 25172, -16819, -29692 -}; /* ;w31 w29 w27 w25
*/ +}; /* ;w31 w29 w27 w25
*/ /*; Table for rows 3,5 - constants are multiplied by cos_3_16
*/ -const short tab_i_35[] = {
19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ - 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ - 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ - -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ - 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ - 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ - 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ +const short tab_i_35[] = {
19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ + 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ + 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ + -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ + 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ + 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ + 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ -26722, 22654, -15137, -26722 -}; /*; w31 w29 w27 w25
*/ +}; /*; w31 w29 w27 w25
*/ */
/* CONCATENATED TABLE, rows 0,1,2,3,4,5,6,7 (in order )
*/ @@ -286,109 +286,109 @@ const short tab_i_35[] = {
19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 /*
*/ static const short tab_i_01234567[] = {
/*row0, this row is required
*/ - 16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ - 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ - 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ - -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ - 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ - 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ - 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ - -22725, 19266, -12873, -22725, /*w31 w29 w27 w25
*/ + 16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ + 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ + 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ + -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ + 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ + 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ + 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ + -22725, 19266, -12873, -22725, /*w31 w29 w27 w25
*/ /* the rest of these rows (1-7), aren't used !
*/ /*row1
*/ - 22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ - 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ - 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ - -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ - 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ - 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ - 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ - -31521, 26722, -17855, -31521, /* w31 w29 w27 w25
*/ + 22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ + 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ + 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ + -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ + 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ + 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ + 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ + -31521, 26722, -17855, -31521, /* w31 w29 w27 w25
*/ /*row2
*/ - 21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ - 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ - 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ - -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ - 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ - 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ - 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ - -29692, 25172, -16819, -29692, /* ;w31 w29 w27 w25
*/ + 21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ + 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ + 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ + -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ + 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ + 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ + 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ + -29692, 25172, -16819, -29692, /* ;w31 w29 w27 w25
*/ /*row3
*/ - 19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ - 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ - 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ - -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ - 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ - 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ - 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ - -26722, 22654, -15137, -26722, /*; w31 w29 w27 w25
*/ + 19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ + 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ + 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ + -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ + 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ + 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ + 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ + -26722, 22654, -15137, -26722, /*; w31 w29 w27 w25
*/ /*row4
*/ - 16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ - 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ - 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ - -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ - 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ - 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ - 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ - -22725, 19266, -12873, -22725, /*w31 w29 w27 w25
*/ + 16384, 16384, 16384, -16384, /* ; movq-> w06 w04 w02 w00
*/ + 21407, 8867, 8867, -21407, /* w07 w05 w03 w01
*/ + 16384, -16384, 16384, 16384, /*; w14 w12 w10 w08
*/ + -8867, 21407, -21407, -8867, /*; w15 w13 w11 w09
*/ + 22725, 12873, 19266, -22725, /*; w22 w20 w18 w16
*/ + 19266, 4520, -4520, -12873, /*; w23 w21 w19 w17
*/ + 12873, 4520, 4520, 19266, /*; w30 w28 w26 w24
*/ + -22725, 19266, -12873, -22725, /*w31 w29 w27 w25
*/ /*row5
*/ - 19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ - 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ - 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ - -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ - 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ - 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ - 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ - -26722, 22654, -15137, -26722, /*; w31 w29 w27 w25
*/ + 19266, 19266, 19266, -19266, /*; movq-> w06 w04 w02 w00
*/ + 25172, 10426, 10426, -25172, /*; w07 w05 w03 w01
*/ + 19266, -19266, 19266, 19266, /*; w14 w12 w10 w08
*/ + -10426, 25172, -25172, -10426, /*; w15 w13 w11 w09
*/ + 26722, 15137, 22654, -26722, /*; w22 w20 w18 w16
*/ + 22654, 5315, -5315, -15137, /*; w23 w21 w19 w17
*/ + 15137, 5315, 5315, 22654, /*; w30 w28 w26 w24
*/ + -26722, 22654, -15137, -26722, /*; w31 w29 w27 w25
*/ /*row6
*/ - 21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ - 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ - 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ - -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ - 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ - 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ - 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ - -29692, 25172, -16819, -29692, /* ;w31 w29 w27 w25
*/ + 21407, 21407, 21407, -21407, /* ; movq-> w06 w04 w02 w00
*/ + 27969, 11585, 11585, -27969, /* ; w07 w05 w03 w01
*/ + 21407, -21407, 21407, 21407, /* ; w14 w12 w10 w08
*/ + -11585, 27969, -27969, -11585, /* ;w15 w13 w11 w09
*/ + 29692, 16819, 25172, -29692, /* ;w22 w20 w18 w16
*/ + 25172, 5906, -5906, -16819, /* ;w23 w21 w19 w17
*/ + 16819, 5906, 5906, 25172, /* ;w30 w28 w26 w24
*/ + -29692, 25172, -16819, -29692, /* ;w31 w29 w27 w25
*/ /*row7
*/ - 22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ - 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ - 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ - -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ - 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ - 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ - 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ + 22725, 22725, 22725, -22725, /* ; movq-> w06 w04 w02 w00
*/ + 29692, 12299, 12299, -29692, /* ; w07 w05 w03 w01
*/ + 22725, -22725, 22725, 22725, /*; w14 w12 w10 w08
*/ + -12299, 29692, -29692, -12299, /*; w15 w13 w11 w09
*/ + 31521, 17855, 26722, -31521, /*; w22 w20 w18 w16
*/ + 26722, 6270, -6270, -17855, /*; w23 w21 w19 w17
*/ + 17855, 6270, 6270, 26722, /*; w30 w28 w26 w24
*/ -31521, 26722, -17855, -31521 -}; /* w31 w29 w27 w25
*/ +}; /* w31 w29 w27 w25
*/ -#define INP eax /* pointer to (short *blk)
*/ -#define OUT ecx /* pointer to output (temporary store space qwTemp[])
*/ -#define TABLE ebx /* pointer to tab_i_01234567[]
*/ +#define INP eax /* pointer to (short *blk)
*/ +#define OUT ecx /* pointer to output (temporary store space qwTemp[])
*/ +#define TABLE ebx /* pointer to tab_i_01234567[]
*/ #define round_inv_row edx
#define round_inv_col edx
-#define ROW_STRIDE 8 /* for 8x8 matrix transposer
*/ +#define ROW_STRIDE 8 /* for 8x8 matrix transposer
*/ /* private variables and functions
*/ /*temporary storage space, 8x8 of shorts
*/ -
__inline static void idct_mmx32_rows (short *blk); /* transform rows
*/ -__inline static void idct_mmx32_cols (short *blk); /* transform "columns"
*/ +
__inline static void idct_mmx32_rows (short *blk); /* transform rows
*/ +__inline static void idct_mmx32_cols (short *blk); /* transform "columns"
*/ - /* the "column" transform actually transforms rows, it is
*/ - /* identical to the row-transform except for the ROUNDING
*/ - /* and SHIFTING coefficients.
*/ + /* the "column" transform actually transforms rows, it is
*/ + /* identical to the row-transform except for the ROUNDING
*/ + /* and SHIFTING coefficients.
*/
static void
idct_mmx32_rows (short *blk) -{ /* transform all 8 rows of 8x8 iDCT block
*/ +{ /* transform all 8 rows of 8x8 iDCT block
*/
int x;
short qwTemp[64];
short *out = &qwTemp[0]; @@ -402,58 +402,58 @@ idct_mmx32_rows (short *blk) /*
*/ /* 2) transpose the matrix (which was stored in qwTemp[])
*/ /* qwTemp[] -> [8x8 matrix transpose] -> blk[]
*/ -
for (x = 0; x < 8; x++) { /* transform one row per iteration
*/ - movq_m2r (*(inptr), mm0); /* 0 ; x3 x2 x1 x0
*/ -
movq_m2r (*(inptr + 4), mm1); /* 1 ; x7 x6 x5 x4
*/ - movq_r2r (mm0, mm2); /* 2 ; x3 x2 x1 x0
*/ -
movq_m2r (*(tab_i_01234567), mm3); /* 3 ; w06 w04 w02 w00
*/ - punpcklwd_r2r (mm1, mm0); /* x5 x1 x4 x0
*/ +
for (x = 0; x < 8; x++) { /* transform one row per iteration
*/ + movq_m2r (*(inptr), mm0); /* 0 ; x3 x2 x1 x0
*/ +
movq_m2r (*(inptr + 4), mm1); /* 1 ; x7 x6 x5 x4
*/ + movq_r2r (mm0, mm2); /* 2 ; x3 x2 x1 x0
*/ +
movq_m2r (*(tab_i_01234567), mm3); /* 3 ; w06 w04 w02 w00
*/ + punpcklwd_r2r (mm1, mm0); /* x5 x1 x4 x0
*/ - /* ----------
*/ - movq_r2r (mm0, mm5); /* 5 ; x5 x1 x4 x0
*/ - punpckldq_r2r (mm0, mm0); /* x4 x0 x4 x0
*/ -
movq_m2r (*(tab_i_01234567 + 4), mm4); /* 4 ; w07 w05 w03 w01
*/ - punpckhwd_r2r (mm1, mm2); /* 1 ; x7 x3 x6 x2
*/ -
pmaddwd_r2r (mm0, mm3); /* x4*w06+x0*w04 x4*w02+x0*w00
*/ - movq_r2r (mm2, mm6); /* 6 ; x7 x3 x6 x2
*/ -
movq_m2r (*(tab_i_01234567 + 16), mm1); /* 1 ; w22 w20 w18 w16
*/ - punpckldq_r2r (mm2, mm2); /* x6 x2 x6 x2
*/ -
pmaddwd_r2r (mm2, mm4); /* x6*w07+x2*w05 x6*w03+x2*w01
*/ - punpckhdq_r2r (mm5, mm5); /* x5 x1 x5 x1
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 8), mm0); /* x4*w14+x0*w12 x4*w10+x0*w08
*/ - punpckhdq_r2r (mm6, mm6); /* x7 x3 x7 x3
*/ -
movq_m2r (*(tab_i_01234567 + 20), mm7); /* 7 ; w23 w21 w19 w17
*/ - pmaddwd_r2r (mm5, mm1); /* x5*w22+x1*w20 x5*w18+x1*w16
*/ -
paddd_m2r (*(r_inv_row), mm3); /* +rounder
*/ - pmaddwd_r2r (mm6, mm7); /* x7*w23+x3*w21 x7*w19+x3*w17
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 12), mm2); /* x6*w15+x2*w13 x6*w11+x2*w09
*/ - paddd_r2r (mm4, mm3); /* 4 ; a1=sum(even1) a0=sum(even0)
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 24), mm5); /* x5*w30+x1*w28 x5*w26+x1*w24
*/ - movq_r2r (mm3, mm4); /* 4 ; a1 a0
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 28), mm6); /* x7*w31+x3*w29 x7*w27+x3*w25
*/ - paddd_r2r (mm7, mm1); /* 7 ; b1=sum(odd1) b0=sum(odd0)
*/ -
paddd_m2r (*(r_inv_row), mm0); /* +rounder
*/ - psubd_r2r (mm1, mm3); /* a1-b1 a0-b0
*/ -
psrad_i2r (SHIFT_INV_ROW, mm3); /* y6=a1-b1 y7=a0-b0
*/ - paddd_r2r (mm4, mm1); /* 4 ; a1+b1 a0+b0
*/ -
paddd_r2r (mm2, mm0); /* 2 ; a3=sum(even3) a2=sum(even2)
*/ - psrad_i2r (SHIFT_INV_ROW, mm1); /* y1=a1+b1 y0=a0+b0
*/ -
paddd_r2r (mm6, mm5); /* 6 ; b3=sum(odd3) b2=sum(odd2)
*/ - movq_r2r (mm0, mm4); /* 4 ; a3 a2
*/ -
paddd_r2r (mm5, mm0); /* a3+b3 a2+b2
*/ - psubd_r2r (mm5, mm4); /* 5 ; a3-b3 a2-b2
*/ -
psrad_i2r (SHIFT_INV_ROW, mm4); /* y4=a3-b3 y5=a2-b2
*/ - psrad_i2r (SHIFT_INV_ROW, mm0); /* y3=a3+b3 y2=a2+b2
*/ -
packssdw_r2r (mm3, mm4); /* 3 ; y6 y7 y4 y5
*/ -
packssdw_r2r (mm0, mm1); /* 0 ; y3 y2 y1 y0
*/ - movq_r2r (mm4, mm7); /* 7 ; y6 y7 y4 y5
*/ -
psrld_i2r (16, mm4); /* 0 y6 0 y4
*/ -
movq_r2m (mm1, *(out)); /* 1 ; save y3 y2 y1 y0
*/ - pslld_i2r (16, mm7); /* y7 0 y5 0
*/ -
por_r2r (mm4, mm7); /* 4 ; y7 y6 y5 y4
*/ + /* ----------
*/ + movq_r2r (mm0, mm5); /* 5 ; x5 x1 x4 x0
*/ + punpckldq_r2r (mm0, mm0); /* x4 x0 x4 x0
*/ +
movq_m2r (*(tab_i_01234567 + 4), mm4); /* 4 ; w07 w05 w03 w01
*/ + punpckhwd_r2r (mm1, mm2); /* 1 ; x7 x3 x6 x2
*/ +
pmaddwd_r2r (mm0, mm3); /* x4*w06+x0*w04 x4*w02+x0*w00
*/ + movq_r2r (mm2, mm6); /* 6 ; x7 x3 x6 x2
*/ +
movq_m2r (*(tab_i_01234567 + 16), mm1); /* 1 ; w22 w20 w18 w16
*/ + punpckldq_r2r (mm2, mm2); /* x6 x2 x6 x2
*/ +
pmaddwd_r2r (mm2, mm4); /* x6*w07+x2*w05 x6*w03+x2*w01
*/ + punpckhdq_r2r (mm5, mm5); /* x5 x1 x5 x1
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 8), mm0); /* x4*w14+x0*w12 x4*w10+x0*w08
*/ + punpckhdq_r2r (mm6, mm6); /* x7 x3 x7 x3
*/ +
movq_m2r (*(tab_i_01234567 + 20), mm7); /* 7 ; w23 w21 w19 w17
*/ + pmaddwd_r2r (mm5, mm1); /* x5*w22+x1*w20 x5*w18+x1*w16
*/ +
paddd_m2r (*(r_inv_row), mm3); /* +rounder
*/ + pmaddwd_r2r (mm6, mm7); /* x7*w23+x3*w21 x7*w19+x3*w17
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 12), mm2); /* x6*w15+x2*w13 x6*w11+x2*w09
*/ + paddd_r2r (mm4, mm3); /* 4 ; a1=sum(even1) a0=sum(even0)
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 24), mm5); /* x5*w30+x1*w28 x5*w26+x1*w24
*/ + movq_r2r (mm3, mm4); /* 4 ; a1 a0
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 28), mm6); /* x7*w31+x3*w29 x7*w27+x3*w25
*/ + paddd_r2r (mm7, mm1); /* 7 ; b1=sum(odd1) b0=sum(odd0)
*/ +
paddd_m2r (*(r_inv_row), mm0); /* +rounder
*/ + psubd_r2r (mm1, mm3); /* a1-b1 a0-b0
*/ +
psrad_i2r (SHIFT_INV_ROW, mm3); /* y6=a1-b1 y7=a0-b0
*/ + paddd_r2r (mm4, mm1); /* 4 ; a1+b1 a0+b0
*/ +
paddd_r2r (mm2, mm0); /* 2 ; a3=sum(even3) a2=sum(even2)
*/ + psrad_i2r (SHIFT_INV_ROW, mm1); /* y1=a1+b1 y0=a0+b0
*/ +
paddd_r2r (mm6, mm5); /* 6 ; b3=sum(odd3) b2=sum(odd2)
*/ + movq_r2r (mm0, mm4); /* 4 ; a3 a2
*/ +
paddd_r2r (mm5, mm0); /* a3+b3 a2+b2
*/ + psubd_r2r (mm5, mm4); /* 5 ; a3-b3 a2-b2
*/ +
psrad_i2r (SHIFT_INV_ROW, mm4); /* y4=a3-b3 y5=a2-b2
*/ + psrad_i2r (SHIFT_INV_ROW, mm0); /* y3=a3+b3 y2=a2+b2
*/ +
packssdw_r2r (mm3, mm4); /* 3 ; y6 y7 y4 y5
*/ +
packssdw_r2r (mm0, mm1); /* 0 ; y3 y2 y1 y0
*/ + movq_r2r (mm4, mm7); /* 7 ; y6 y7 y4 y5
*/ +
psrld_i2r (16, mm4); /* 0 y6 0 y4
*/ +
movq_r2m (mm1, *(out)); /* 1 ; save y3 y2 y1 y0
*/ + pslld_i2r (16, mm7); /* y7 0 y5 0
*/ +
por_r2r (mm4, mm7); /* 4 ; y7 y6 y5 y4
*/ - /* begin processing row 1
*/ - movq_r2m (mm7, *(out + 4)); /* 7 ; save y7 y6 y5 y4
*/ + /* begin processing row 1
*/ + movq_r2m (mm7, *(out + 4)); /* 7 ; save y7 y6 y5 y4
*/
inptr += 8;
out += 8;
} @@ -477,31 +477,31 @@ idct_mmx32_rows (short *blk) out = &qwTemp[0];
movq_m2r (*(out + ROW_STRIDE * 0), mm0);
movq_m2r (*(out + ROW_STRIDE * 1), mm1); -
movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ +
movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/
movq_m2r (*(out + ROW_STRIDE * 2), mm2); -
punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ +
punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/
movq_m2r (*(out + ROW_STRIDE * 3), mm3); -
punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/ +
punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/
movq_r2r (mm2, mm6); -
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ -
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ - movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ +
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ +
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ + movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/
inptr = blk; -
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ -
movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ - punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ -
movq_r2m (mm0, *(inptr + ROW_STRIDE * 0)); /* store row 1
*/ - punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ +
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ +
movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ + punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ +
movq_r2m (mm0, *(inptr + ROW_STRIDE * 0)); /* store row 1
*/ + punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ /* begin reading next quadrant (lower-right)
*/ movq_m2r (*(out + ROW_STRIDE * 4 + 4), mm0); -
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ -
movq_r2m (mm4, *(inptr + ROW_STRIDE * 2)); /* store row 3
*/ - movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ -
movq_r2m (mm1, *(inptr + ROW_STRIDE * 1)); /* store row 2
*/ +
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ +
movq_r2m (mm4, *(inptr + ROW_STRIDE * 2)); /* store row 3
*/ + movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ +
movq_r2m (mm1, *(inptr + ROW_STRIDE * 1)); /* store row 2
*/
movq_m2r (*(out + ROW_STRIDE * 5 + 4), mm1); -
movq_r2m (mm3, *(inptr + ROW_STRIDE * 3)); /* store row 4
*/ - punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ +
movq_r2m (mm3, *(inptr + ROW_STRIDE * 3)); /* store row 4
*/ + punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ /* 2) transpose lower-right quadrant
*/ @@ -512,28 +512,28 @@ idct_mmx32_rows (short *blk)
movq_m2r (*(out + ROW_STRIDE * 6 + 4), mm2); /* punpcklwd mm0, mm1; // mm0 = [ 0 4 1 5]
*/ - punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/ + punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/
movq_m2r (*(out + ROW_STRIDE * 7 + 4), mm3);
movq_r2r (mm2, mm6); -
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ - movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ -
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ - movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ -
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ -
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ - ; /* slot
*/ -
movq_r2m (mm0, *(inptr + ROW_STRIDE * 4 + 4)); /* store row 1
*/ - punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ +
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ + movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ +
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ + movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ +
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ +
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ + ; /* slot
*/ +
movq_r2m (mm0, *(inptr + ROW_STRIDE * 4 + 4)); /* store row 1
*/ + punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/
movq_m2r (*(out + ROW_STRIDE * 4), mm0); -
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ -
movq_r2m (mm4, *(inptr + ROW_STRIDE * 6 + 4)); /* store row 3
*/ - movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ -
movq_r2m (mm1, *(inptr + ROW_STRIDE * 5 + 4)); /* store row 2
*/ - ; /* slot
*/ +
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ +
movq_r2m (mm4, *(inptr + ROW_STRIDE * 6 + 4)); /* store row 3
*/ + movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ +
movq_r2m (mm1, *(inptr + ROW_STRIDE * 5 + 4)); /* store row 2
*/ + ; /* slot
*/
movq_m2r (*(out + ROW_STRIDE * 5), mm1); -
; /* slot
*/ -
movq_r2m (mm3, *(inptr + ROW_STRIDE * 7 + 4)); /* store row 4
*/ - punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ +
; /* slot
*/ +
movq_r2m (mm3, *(inptr + ROW_STRIDE * 7 + 4)); /* store row 4
*/ + punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ /* 3) transpose lower-left
*/ /* movq mm0, qword ptr [OUT + ROW_STRIDE * 4 ]
*/ @@ -543,28 +543,28 @@ idct_mmx32_rows (short *blk)
movq_m2r (*(out + ROW_STRIDE * 6), mm2); /* punpcklwd mm0, mm1; // mm0 = [ 0 4 1 5]
*/ - punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/ + punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/
movq_m2r (*(out + ROW_STRIDE * 7), mm3);
movq_r2r (mm2, mm6); -
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ - movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ -
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ - movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ -
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ -
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ - ; /*slot
*/ -
movq_r2m (mm0, *(inptr + ROW_STRIDE * 0 + 4)); /* store row 1
*/ - punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ +
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ + movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ +
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ + movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ +
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ +
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ + ; /*slot
*/ +
movq_r2m (mm0, *(inptr + ROW_STRIDE * 0 + 4)); /* store row 1
*/ + punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ /* begin reading next quadrant (upper-right)
*/ movq_m2r (*(out + ROW_STRIDE * 0 + 4), mm0); -
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ -
movq_r2m (mm4, *(inptr + ROW_STRIDE * 2 + 4)); /* store row 3
*/ - movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ -
movq_r2m (mm1, *(inptr + ROW_STRIDE * 1 + 4)); /* store row 2
*/ +
punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ +
movq_r2m (mm4, *(inptr + ROW_STRIDE * 2 + 4)); /* store row 3
*/ + movq_r2r (mm0, mm4); /* mm4 = copy of row1[A B C D]
*/ +
movq_r2m (mm1, *(inptr + ROW_STRIDE * 1 + 4)); /* store row 2
*/ movq_m2r (*(out + ROW_STRIDE * 1 + 4), mm1); -
movq_r2m (mm3, *(inptr + ROW_STRIDE * 3 + 4)); /* store row 4
*/ - punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ +
movq_r2m (mm3, *(inptr + ROW_STRIDE * 3 + 4)); /* store row 4
*/ + punpcklwd_r2r (mm1, mm0); /* mm0 = [ 0 4 1 5]
*/ /* 2) transpose lower-right quadrant
*/ @@ -575,28 +575,28 @@ idct_mmx32_rows (short *blk)
movq_m2r (*(out + ROW_STRIDE * 2 + 4), mm2); /* punpcklwd mm0, mm1; // mm0 = [ 0 4 1 5]
*/ - punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/ + punpckhwd_r2r (mm1, mm4); /* mm4 = [ 2 6 3 7]
*/
movq_m2r (*(out + ROW_STRIDE * 3 + 4), mm3);
movq_r2r (mm2, mm6); -
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ - movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ -
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ - movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ -
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ -
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ - ; /* slot
*/ -
movq_r2m (mm0, *(inptr + ROW_STRIDE * 4)); /* store row 1
*/ - punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ -
movq_r2m (mm1, *(inptr + ROW_STRIDE * 5)); /* store row 2
*/ - punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ -
movq_r2m (mm4, *(inptr + ROW_STRIDE * 6)); /* store row 3
*/ - ; /* slot
*/ -
movq_r2m (mm3, *(inptr + ROW_STRIDE * 7)); /* store row 4
*/ - ; /* slot
*/ +
punpcklwd_r2r (mm3, mm2); /* mm2 = [ 8 12 9 13]
*/ + movq_r2r (mm0, mm1); /* mm1 = [ 0 4 1 5]
*/ +
punpckhwd_r2r (mm3, mm6); /* mm6 = 10 14 11 15]
*/ + movq_r2r (mm4, mm3); /* mm3 = [ 2 6 3 7]
*/ +
punpckldq_r2r (mm2, mm0); /* final result mm0 = row1 [0 4 8 12]
*/ +
punpckhdq_r2r (mm2, mm1); /* mm1 = final result mm1 = row2 [1 5 9 13]
*/ + ; /* slot
*/ +
movq_r2m (mm0, *(inptr + ROW_STRIDE * 4)); /* store row 1
*/ + punpckldq_r2r (mm6, mm4); /* final result mm4 = row3 [2 6 10 14]
*/ +
movq_r2m (mm1, *(inptr + ROW_STRIDE * 5)); /* store row 2
*/ + punpckhdq_r2r (mm6, mm3); /* final result mm3 = row4 [3 7 11 15]
*/ +
movq_r2m (mm4, *(inptr + ROW_STRIDE * 6)); /* store row 3
*/ + ; /* slot
*/ +
movq_r2m (mm3, *(inptr + ROW_STRIDE * 7)); /* store row 4
*/ + ; /* slot
*/
}
static void
idct_mmx32_cols (short *blk) -{ /* transform all 8 cols of 8x8 iDCT block
*/ +{ /* transform all 8 cols of 8x8 iDCT block
*/
int x;
short *inptr = blk; @@ -612,58 +612,58 @@ idct_mmx32_cols (short *blk) /*
*/ /* 2) transpose the matrix (which was stored in qwTemp[])
*/ /* qwTemp[] -> [8x8 matrix transpose] -> blk[]
*/ -
for (x = 0; x < 8; x++) { /* transform one row per iteration
*/ -
movq_m2r (*(inptr), mm0); /* 0 ; x3 x2 x1 x0
*/ -
movq_m2r (*(inptr + 4), mm1); /* 1 ; x7 x6 x5 x4
*/ - movq_r2r (mm0, mm2); /* 2 ; x3 x2 x1 x0
*/ -
movq_m2r (*(tab_i_01234567), mm3); /* 3 ; w06 w04 w02 w00
*/ - punpcklwd_r2r (mm1, mm0); /* x5 x1 x4 x0
*/ +
for (x = 0; x < 8; x++) { /* transform one row per iteration
*/ +
movq_m2r (*(inptr), mm0); /* 0 ; x3 x2 x1 x0
*/ +
movq_m2r (*(inptr + 4), mm1); /* 1 ; x7 x6 x5 x4
*/ + movq_r2r (mm0, mm2); /* 2 ; x3 x2 x1 x0
*/ +
movq_m2r (*(tab_i_01234567), mm3); /* 3 ; w06 w04 w02 w00
*/ + punpcklwd_r2r (mm1, mm0); /* x5 x1 x4 x0
*/ /* ----------
*/ - movq_r2r (mm0, mm5); /* 5 ; x5 x1 x4 x0
*/ - punpckldq_r2r (mm0, mm0); /* x4 x0 x4 x0
*/ -
movq_m2r (*(tab_i_01234567 + 4), mm4); /* 4 ; w07 w05 w03 w01
*/ - punpckhwd_r2r (mm1, mm2); /* 1 ; x7 x3 x6 x2
*/ -
pmaddwd_r2r (mm0, mm3); /* x4*w06+x0*w04 x4*w02+x0*w00
*/ - movq_r2r (mm2, mm6); /* 6 ; x7 x3 x6 x2
*/ -
movq_m2r (*(tab_i_01234567 + 16), mm1); /* 1 ; w22 w20 w18 w16
*/ - punpckldq_r2r (mm2, mm2); /* x6 x2 x6 x2
*/ -
pmaddwd_r2r (mm2, mm4); /* x6*w07+x2*w05 x6*w03+x2*w01
*/ - punpckhdq_r2r (mm5, mm5); /* x5 x1 x5 x1
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 8), mm0); /* x4*w14+x0*w12 x4*w10+x0*w08
*/ - punpckhdq_r2r (mm6, mm6); /* x7 x3 x7 x3
*/ -
movq_m2r (*(tab_i_01234567 + 20), mm7); /* 7 ; w23 w21 w19 w17
*/ - pmaddwd_r2r (mm5, mm1); /* x5*w22+x1*w20 x5*w18+x1*w16
*/ -
paddd_m2r (*(r_inv_col), mm3); /* +rounder
*/ - pmaddwd_r2r (mm6, mm7); /* x7*w23+x3*w21 x7*w19+x3*w17
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 12), mm2); /* x6*w15+x2*w13 x6*w11+x2*w09
*/ - paddd_r2r (mm4, mm3); /* 4 ; a1=sum(even1) a0=sum(even0)
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 24), mm5); /* x5*w30+x1*w28 x5*w26+x1*w24
*/ - movq_r2r (mm3, mm4); /* 4 ; a1 a0
*/ -
pmaddwd_m2r (*(tab_i_01234567 + 28), mm6); /* x7*w31+x3*w29 x7*w27+x3*w25
*/ - paddd_r2r (mm7, mm1); /* 7 ; b1=sum(odd1) b0=sum(odd0)
*/ -
paddd_m2r (*(r_inv_col), mm0); /* +rounder
*/ - psubd_r2r (mm1, mm3); /* a1-b1 a0-b0
*/ -
psrad_i2r (SHIFT_INV_COL, mm3); /* y6=a1-b1 y7=a0-b0
*/ - paddd_r2r (mm4, mm1); /* 4 ; a1+b1 a0+b0
*/ -
paddd_r2r (mm2, mm0); /* 2 ; a3=sum(even3) a2=sum(even2)
*/ - psrad_i2r (SHIFT_INV_COL, mm1); /* y1=a1+b1 y0=a0+b0
*/ -
paddd_r2r (mm6, mm5); /* 6 ; b3=sum(odd3) b2=sum(odd2)
*/ - movq_r2r (mm0, mm4); /* 4 ; a3 a2
*/ -
paddd_r2r (mm5, mm0); /* a3+b3 a2+b2
*/ - psubd_r2r (mm5, mm4); /* 5 ; a3-b3 a2-b2
*/ -
psrad_i2r (SHIFT_INV_COL, mm4); /* y4=a3-b3 y5=a2-b2
*/ - psrad_i2r (SHIFT_INV_COL, mm0); /* y3=a3+b3 y2=a2+b2
*/ -
packssdw_r2r (mm3, mm4); /* 3 ; y6 y7 y4 y5
*/ -
packssdw_r2r (mm0, mm1); /* 0 ; y3 y2 y1 y0
*/ - movq_r2r (mm4, mm7); /* 7 ; y6 y7 y4 y5
*/ -
psrld_i2r (16, mm4); /* 0 y6 0 y4
*/ -
movq_r2m (mm1, *(inptr)); /* 1 ; save y3 y2 y1 y0
*/ - pslld_i2r (16, mm7); /* y7 0 y5 0
*/ -
por_r2r (mm4, mm7); /* 4 ; y7 y6 y5 y4
*/ + movq_r2r (mm0, mm5); /* 5 ; x5 x1 x4 x0
*/ + punpckldq_r2r (mm0, mm0); /* x4 x0 x4 x0
*/ +
movq_m2r (*(tab_i_01234567 + 4), mm4); /* 4 ; w07 w05 w03 w01
*/ + punpckhwd_r2r (mm1, mm2); /* 1 ; x7 x3 x6 x2
*/ +
pmaddwd_r2r (mm0, mm3); /* x4*w06+x0*w04 x4*w02+x0*w00
*/ + movq_r2r (mm2, mm6); /* 6 ; x7 x3 x6 x2
*/ +
movq_m2r (*(tab_i_01234567 + 16), mm1); /* 1 ; w22 w20 w18 w16
*/ + punpckldq_r2r (mm2, mm2); /* x6 x2 x6 x2
*/ +
pmaddwd_r2r (mm2, mm4); /* x6*w07+x2*w05 x6*w03+x2*w01
*/ + punpckhdq_r2r (mm5, mm5); /* x5 x1 x5 x1
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 8), mm0); /* x4*w14+x0*w12 x4*w10+x0*w08
*/ + punpckhdq_r2r (mm6, mm6); /* x7 x3 x7 x3
*/ +
movq_m2r (*(tab_i_01234567 + 20), mm7); /* 7 ; w23 w21 w19 w17
*/ + pmaddwd_r2r (mm5, mm1); /* x5*w22+x1*w20 x5*w18+x1*w16
*/ +
paddd_m2r (*(r_inv_col), mm3); /* +rounder
*/ + pmaddwd_r2r (mm6, mm7); /* x7*w23+x3*w21 x7*w19+x3*w17
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 12), mm2); /* x6*w15+x2*w13 x6*w11+x2*w09
*/ + paddd_r2r (mm4, mm3); /* 4 ; a1=sum(even1) a0=sum(even0)
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 24), mm5); /* x5*w30+x1*w28 x5*w26+x1*w24
*/ + movq_r2r (mm3, mm4); /* 4 ; a1 a0
*/ +
pmaddwd_m2r (*(tab_i_01234567 + 28), mm6); /* x7*w31+x3*w29 x7*w27+x3*w25
*/ + paddd_r2r (mm7, mm1); /* 7 ; b1=sum(odd1) b0=sum(odd0)
*/ +
paddd_m2r (*(r_inv_col), mm0); /* +rounder
*/ + psubd_r2r (mm1, mm3); /* a1-b1 a0-b0
*/ +
psrad_i2r (SHIFT_INV_COL, mm3); /* y6=a1-b1 y7=a0-b0
*/ + paddd_r2r (mm4, mm1); /* 4 ; a1+b1 a0+b0
*/ +
paddd_r2r (mm2, mm0); /* 2 ; a3=sum(even3) a2=sum(even2)
*/ + psrad_i2r (SHIFT_INV_COL, mm1); /* y1=a1+b1 y0=a0+b0
*/ +
paddd_r2r (mm6, mm5); /* 6 ; b3=sum(odd3) b2=sum(odd2)
*/ + movq_r2r (mm0, mm4); /* 4 ; a3 a2
*/ +
paddd_r2r (mm5, mm0); /* a3+b3 a2+b2
*/ + psubd_r2r (mm5, mm4); /* 5 ; a3-b3 a2-b2
*/ +
psrad_i2r (SHIFT_INV_COL, mm4); /* y4=a3-b3 y5=a2-b2
*/ + psrad_i2r (SHIFT_INV_COL, mm0); /* y3=a3+b3 y2=a2+b2
*/ +
packssdw_r2r (mm3, mm4); /* 3 ; y6 y7 y4 y5
*/ +
packssdw_r2r (mm0, mm1); /* 0 ; y3 y2 y1 y0
*/ + movq_r2r (mm4, mm7); /* 7 ; y6 y7 y4 y5
*/ +
psrld_i2r (16, mm4); /* 0 y6 0 y4
*/ +
movq_r2m (mm1, *(inptr)); /* 1 ; save y3 y2 y1 y0
*/ + pslld_i2r (16, mm7); /* y7 0 y5 0
*/ +
por_r2r (mm4, mm7); /* 4 ; y7 y6 y5 y4
*/ - /* begin processing row 1
*/ - movq_r2m (mm7, *(inptr + 4)); /* 7 ; save y7 y6 y5 y4
*/ + /* begin processing row 1
*/ + movq_r2m (mm7, *(inptr + 4)); /* 7 ; save y7 y6 y5 y4
*/
inptr += 8;
} @@ -679,10 +679,12 @@ gst_idct_mmx32_idct (short *blk) { /* 1) iDCT row transformation
*/ - idct_mmx32_rows (blk); /* 1) transform iDCT row, and transpose
*/ + idct_mmx32_rows (blk); /* 1) transform iDCT row, and transpose
*/ /* 2) iDCT column transformation
*/ - idct_mmx32_cols (blk); /* 2) transform iDCT row, and transpose
*/ -
emms (); /* restore processor state
*/ + idct_mmx32_cols (blk); /* 2) transform iDCT row, and transpose
*/ +
emms (); /* restore processor state
*/ /* all done
*/ -}
+} + +
|