videolan/ffmpeg
1
Fork 0
mirror of https://git.videolan.org/git/ffmpeg.git synced 2024-10-05 01:51:53 +02:00
Code Releases Activity

removing ALTIVEC_USE_REFERENCE_C_CODE, since has no use anymore

Browse Source 
Originally committed as revision 6606 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
Luca Barbato 2006-10-09 18:29:46 +00:00
parent 1c2a417f6a
commit b420448e38
10 changed files with 0 additions and 425 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
configure vendored
View File

@ -1789,8 +1789,6 @@ fi
if test "$altivec" = "yes" ; then
echo "TARGET_ALTIVEC=yes" >> config.mak
echo "#define HAVE_ALTIVEC 1" >> $TMPH
echo "// Enable the next line to use the reference C code instead of AltiVec" >> $TMPH
echo "// #define ALTIVEC_USE_REFERENCE_C_CODE 1" >> $TMPH
if test "$_altivec_h" = "yes" ; then
echo "#define HAVE_ALTIVEC_H 1" >> $TMPH
else

268
libavcodec/ppc/dsputil_altivec.c
View File

@ -618,21 +618,6 @@ void diff_pixels_altivec(DCTELEM *restrict block, const uint8_t *s1,
}
void add_bytes_altivec(uint8_t *dst, uint8_t *src, int w) {
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int i;
for(i=0; i+7<w; i++){
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];
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char vdst, vsrc;
@ -649,30 +634,12 @@ void add_bytes_altivec(uint8_t *dst, uint8_t *src, int w) {
{
dst[i] = src[i];
}
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 16) == 0) */
void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_put_pixels16_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int i;
POWERPC_PERF_START_COUNT(altivec_put_pixels16_num, 1);
for(i=0; i<h; i++) {
*((uint32_t*)(block)) = LD32(pixels);
*((uint32_t*)(block+4)) = LD32(pixels+4);
*((uint32_t*)(block+8)) = LD32(pixels+8);
*((uint32_t*)(block+12)) = LD32(pixels+12);
pixels+=line_size;
block +=line_size;
}
POWERPC_PERF_STOP_COUNT(altivec_put_pixels16_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register vector unsigned char pixelsv1, pixelsv2;
register vector unsigned char pixelsv1B, pixelsv2B;
register vector unsigned char pixelsv1C, pixelsv2C;
@ -722,8 +689,6 @@ POWERPC_PERF_START_COUNT(altivec_put_pixels16_num, 1);
}
#endif
POWERPC_PERF_STOP_COUNT(altivec_put_pixels16_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 16) == 0) */
@ -731,23 +696,6 @@ POWERPC_PERF_STOP_COUNT(altivec_put_pixels16_num, 1);
void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_avg_pixels16_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int i;
POWERPC_PERF_START_COUNT(altivec_avg_pixels16_num, 1);
for(i=0; i<h; i++) {
op_avg(*((uint32_t*)(block)),LD32(pixels));
op_avg(*((uint32_t*)(block+4)),LD32(pixels+4));
op_avg(*((uint32_t*)(block+8)),LD32(pixels+8));
op_avg(*((uint32_t*)(block+12)),LD32(pixels+12));
pixels+=line_size;
block +=line_size;
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels16_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register vector unsigned char pixelsv1, pixelsv2, pixelsv, blockv;
register vector unsigned char perm = vec_lvsl(0, pixels);
int i;
@ -766,37 +714,12 @@ POWERPC_PERF_START_COUNT(altivec_avg_pixels16_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels16_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 8) == 0) */
void avg_pixels8_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_avg_pixels8_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int i;
POWERPC_PERF_START_COUNT(altivec_avg_pixels8_num, 1);
for (i = 0; i < h; i++) {
*((uint32_t *) (block)) =
(((*((uint32_t *) (block))) |
((((const struct unaligned_32 *) (pixels))->l))) -
((((*((uint32_t *) (block))) ^
((((const struct unaligned_32 *) (pixels))->
l))) & 0xFEFEFEFEUL) >> 1));
*((uint32_t *) (block + 4)) =
(((*((uint32_t *) (block + 4))) |
((((const struct unaligned_32 *) (pixels + 4))->l))) -
((((*((uint32_t *) (block + 4))) ^
((((const struct unaligned_32 *) (pixels +
4))->
l))) & 0xFEFEFEFEUL) >> 1));
pixels += line_size;
block += line_size;
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register vector unsigned char pixelsv1, pixelsv2, pixelsv, blockv;
int i;
@ -832,52 +755,12 @@ POWERPC_PERF_START_COUNT(altivec_avg_pixels8_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 8) == 0) */
void put_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_put_pixels8_xy2_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int j;
POWERPC_PERF_START_COUNT(altivec_put_pixels8_xy2_num, 1);
for (j = 0; j < 2; j++) {
int i;
const uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
const uint32_t b =
(((const struct unaligned_32 *) (pixels + 1))->l);
uint32_t l0 =
(a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
uint32_t h0 =
((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
uint32_t l1, h1;
pixels += line_size;
for (i = 0; i < h; i += 2) {
uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
l1 = (a & 0x03030303UL) + (b & 0x03030303UL);
h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
a = (((const struct unaligned_32 *) (pixels))->l);
b = (((const struct unaligned_32 *) (pixels + 1))->l);
l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
} pixels += 4 - line_size * (h + 1);
block += 4 - line_size * h;
}
POWERPC_PERF_STOP_COUNT(altivec_put_pixels8_xy2_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char
pixelsv1, pixelsv2,
@ -948,51 +831,12 @@ POWERPC_PERF_START_COUNT(altivec_put_pixels8_xy2_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_put_pixels8_xy2_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 8) == 0) */
void put_no_rnd_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_put_no_rnd_pixels8_xy2_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int j;
POWERPC_PERF_START_COUNT(altivec_put_no_rnd_pixels8_xy2_num, 1);
for (j = 0; j < 2; j++) {
int i;
const uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
const uint32_t b =
(((const struct unaligned_32 *) (pixels + 1))->l);
uint32_t l0 =
(a & 0x03030303UL) + (b & 0x03030303UL) + 0x01010101UL;
uint32_t h0 =
((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
uint32_t l1, h1;
pixels += line_size;
for (i = 0; i < h; i += 2) {
uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
l1 = (a & 0x03030303UL) + (b & 0x03030303UL);
h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
a = (((const struct unaligned_32 *) (pixels))->l);
b = (((const struct unaligned_32 *) (pixels + 1))->l);
l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x01010101UL;
h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
} pixels += 4 - line_size * (h + 1);
block += 4 - line_size * h;
}
POWERPC_PERF_STOP_COUNT(altivec_put_no_rnd_pixels8_xy2_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char
pixelsv1, pixelsv2,
@ -1064,51 +908,12 @@ POWERPC_PERF_START_COUNT(altivec_put_no_rnd_pixels8_xy2_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_put_no_rnd_pixels8_xy2_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 16) == 0) */
void put_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_put_pixels16_xy2_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int j;
POWERPC_PERF_START_COUNT(altivec_put_pixels16_xy2_num, 1);
for (j = 0; j < 4; j++) {
int i;
const uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
const uint32_t b =
(((const struct unaligned_32 *) (pixels + 1))->l);
uint32_t l0 =
(a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
uint32_t h0 =
((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
uint32_t l1, h1;
pixels += line_size;
for (i = 0; i < h; i += 2) {
uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
l1 = (a & 0x03030303UL) + (b & 0x03030303UL);
h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
a = (((const struct unaligned_32 *) (pixels))->l);
b = (((const struct unaligned_32 *) (pixels + 1))->l);
l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
} pixels += 4 - line_size * (h + 1);
block += 4 - line_size * h;
}
POWERPC_PERF_STOP_COUNT(altivec_put_pixels16_xy2_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char
pixelsv1, pixelsv2, pixelsv3, pixelsv4;
@ -1185,51 +990,12 @@ POWERPC_PERF_START_COUNT(altivec_put_pixels16_xy2_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_put_pixels16_xy2_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* next one assumes that ((line_size % 16) == 0) */
void put_no_rnd_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_put_no_rnd_pixels16_xy2_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int j;
POWERPC_PERF_START_COUNT(altivec_put_no_rnd_pixels16_xy2_num, 1);
for (j = 0; j < 4; j++) {
int i;
const uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
const uint32_t b =
(((const struct unaligned_32 *) (pixels + 1))->l);
uint32_t l0 =
(a & 0x03030303UL) + (b & 0x03030303UL) + 0x01010101UL;
uint32_t h0 =
((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
uint32_t l1, h1;
pixels += line_size;
for (i = 0; i < h; i += 2) {
uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
l1 = (a & 0x03030303UL) + (b & 0x03030303UL);
h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
a = (((const struct unaligned_32 *) (pixels))->l);
b = (((const struct unaligned_32 *) (pixels + 1))->l);
l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x01010101UL;
h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) =
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL);
pixels += line_size;
block += line_size;
} pixels += 4 - line_size * (h + 1);
block += 4 - line_size * h;
}
POWERPC_PERF_STOP_COUNT(altivec_put_no_rnd_pixels16_xy2_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char
pixelsv1, pixelsv2, pixelsv3, pixelsv4;
@ -1307,7 +1073,6 @@ POWERPC_PERF_START_COUNT(altivec_put_no_rnd_pixels16_xy2_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_put_no_rnd_pixels16_xy2_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
int hadamard8_diff8x8_altivec(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h){
@ -1727,38 +1492,6 @@ static void vorbis_inverse_coupling_altivec(float *mag, float *ang,
void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
POWERPC_PERF_DECLARE(altivec_avg_pixels8_xy2_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int j;
POWERPC_PERF_START_COUNT(altivec_avg_pixels8_xy2_num, 1);
for (j = 0; j < 2; j++) {
int i;
const uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
const uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
uint32_t l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
uint32_t h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
uint32_t l1, h1;
pixels += line_size;
for (i = 0; i < h; i += 2) {
uint32_t a = (((const struct unaligned_32 *) (pixels))->l);
uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l);
l1 = (a & 0x03030303UL) + (b & 0x03030303UL);
h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) = rnd_avg32(*((uint32_t *) block), h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL));
pixels += line_size;
block += line_size;
a = (((const struct unaligned_32 *) (pixels))->l);
b = (((const struct unaligned_32 *) (pixels + 1))->l);
l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL;
h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2);
*((uint32_t *) block) = rnd_avg32(*((uint32_t *) block), h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL));
pixels += line_size;
block += line_size;
} pixels += 4 - line_size * (h + 1);
block += 4 - line_size * h;
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_xy2_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsavg;
register vector unsigned char blockv, temp1, temp2, blocktemp;
@ -1821,7 +1554,6 @@ POWERPC_PERF_START_COUNT(altivec_avg_pixels8_xy2_num, 1);
}
POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_xy2_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
void dsputil_init_altivec(DSPContext* c, AVCodecContext *avctx)

4
libavcodec/ppc/dsputil_altivec.h
View File

@ -67,10 +67,6 @@ void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size,
#define vcii(a,b,c,d) (const vector float){FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d}
#endif
#else /* HAVE_ALTIVEC */
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
#error "I can't use ALTIVEC_USE_REFERENCE_C_CODE if I don't use HAVE_ALTIVEC"
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
#endif /* HAVE_ALTIVEC */
#endif /* _DSPUTIL_ALTIVEC_ */

4
libavcodec/ppc/dsputil_ppc.c
View File

@ -291,11 +291,7 @@ void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx)
{
c->idct_put = idct_put_altivec;
c->idct_add = idct_add_altivec;
#ifndef ALTIVEC_USE_REFERENCE_C_CODE
c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
c->idct_permutation_type = FF_NO_IDCT_PERM;
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
}

7
libavcodec/ppc/fdct_altivec.c
View File

@ -198,12 +198,6 @@ static vector float fdctconsts[3] = {
void fdct_altivec(int16_t *block)
{
POWERPC_PERF_DECLARE(altivec_fdct, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
POWERPC_PERF_START_COUNT(altivec_fdct, 1);
void ff_jpeg_fdct_islow(int16_t *block);
ff_jpeg_fdct_islow(block);
POWERPC_PERF_STOP_COUNT(altivec_fdct, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
vector signed short *bp;
vector float *cp;
vector float b00, b10, b20, b30, b40, b50, b60, b70;
@ -494,7 +488,6 @@ POWERPC_PERF_STOP_COUNT(altivec_fdct, 1);
/* }}} */
POWERPC_PERF_STOP_COUNT(altivec_fdct, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
/* vim:set foldmethod=marker foldlevel=0: */

79
libavcodec/ppc/fft_altivec.c
View File

@ -65,83 +65,6 @@
void ff_fft_calc_altivec(FFTContext *s, FFTComplex *z)
{
POWERPC_PERF_DECLARE(altivec_fft_num, s->nbits >= 6);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
int ln = s->nbits;
int j, np, np2;
int nblocks, nloops;
register FFTComplex *p, *q;
FFTComplex *exptab = s->exptab;
int l;
FFTSample tmp_re, tmp_im;
POWERPC_PERF_START_COUNT(altivec_fft_num, s->nbits >= 6);
np = 1 << ln;
/* pass 0 */
p=&z[0];
j=(np >> 1);
do {
BF(p[0].re, p[0].im, p[1].re, p[1].im,
p[0].re, p[0].im, p[1].re, p[1].im);
p+=2;
} while (--j != 0);
/* pass 1 */
p=&z[0];
j=np >> 2;
if (s->inverse) {
do {
BF(p[0].re, p[0].im, p[2].re, p[2].im,
p[0].re, p[0].im, p[2].re, p[2].im);
BF(p[1].re, p[1].im, p[3].re, p[3].im,
p[1].re, p[1].im, -p[3].im, p[3].re);
p+=4;
} while (--j != 0);
} else {
do {
BF(p[0].re, p[0].im, p[2].re, p[2].im,
p[0].re, p[0].im, p[2].re, p[2].im);
BF(p[1].re, p[1].im, p[3].re, p[3].im,
p[1].re, p[1].im, p[3].im, -p[3].re);
p+=4;
} while (--j != 0);
}
/* pass 2 .. ln-1 */
nblocks = np >> 3;
nloops = 1 << 2;
np2 = np >> 1;
do {
p = z;
q = z + nloops;
for (j = 0; j < nblocks; ++j) {
BF(p->re, p->im, q->re, q->im,
p->re, p->im, q->re, q->im);
p++;
q++;
for(l = nblocks; l < np2; l += nblocks) {
CMUL(tmp_re, tmp_im, exptab[l].re, exptab[l].im, q->re, q->im);
BF(p->re, p->im, q->re, q->im,
p->re, p->im, tmp_re, tmp_im);
p++;
q++;
}
p += nloops;
q += nloops;
}
nblocks = nblocks >> 1;
nloops = nloops << 1;
} while (nblocks != 0);
POWERPC_PERF_STOP_COUNT(altivec_fft_num, s->nbits >= 6);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
#ifdef CONFIG_DARWIN
register const vector float vczero = (const vector float)(0.);
#else
@ -244,6 +167,4 @@ POWERPC_PERF_START_COUNT(altivec_fft_num, s->nbits >= 6);
} while (nblocks != 0);
POWERPC_PERF_STOP_COUNT(altivec_fft_num, s->nbits >= 6);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}

28
libavcodec/ppc/gmc_altivec.c
View File

@ -34,32 +34,6 @@
void gmc1_altivec(uint8_t *dst /* align 8 */, uint8_t *src /* align1 */, int stride, int h, int x16, int y16, int rounder)
{
POWERPC_PERF_DECLARE(altivec_gmc1_num, GMC1_PERF_COND);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
const int A=(16-x16)*(16-y16);
const int B=( x16)*(16-y16);
const int C=(16-x16)*( y16);
const int D=( x16)*( y16);
int i;
POWERPC_PERF_START_COUNT(altivec_gmc1_num, GMC1_PERF_COND);
for(i=0; i<h; i++)
{
dst[0]= (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + rounder)>>8;
dst[1]= (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + rounder)>>8;
dst[2]= (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + rounder)>>8;
dst[3]= (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + rounder)>>8;
dst[4]= (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5] + rounder)>>8;
dst[5]= (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6] + rounder)>>8;
dst[6]= (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7] + rounder)>>8;
dst[7]= (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8] + rounder)>>8;
dst+= stride;
src+= stride;
}
POWERPC_PERF_STOP_COUNT(altivec_gmc1_num, GMC1_PERF_COND);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
const unsigned short __attribute__ ((aligned(16))) rounder_a[8] =
{rounder, rounder, rounder, rounder,
rounder, rounder, rounder, rounder};
@ -169,6 +143,4 @@ POWERPC_PERF_START_COUNT(altivec_gmc1_num, GMC1_PERF_COND);
}
POWERPC_PERF_STOP_COUNT(altivec_gmc1_num, GMC1_PERF_COND);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}

14
libavcodec/ppc/idct_altivec.c
View File

@ -169,12 +169,6 @@ static const_vector_s16_t constants[5] = {
void idct_put_altivec(uint8_t* dest, int stride, vector_s16_t* block)
{
POWERPC_PERF_DECLARE(altivec_idct_put_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
void simple_idct_put(uint8_t *dest, int line_size, int16_t *block);
simple_idct_put(dest, stride, (int16_t*)block);
POWERPC_PERF_STOP_COUNT(altivec_idct_put_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
vector_u8_t tmp;
#ifdef POWERPC_PERFORMANCE_REPORT
@ -197,18 +191,11 @@ POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
COPY (dest, vx7)
POWERPC_PERF_STOP_COUNT(altivec_idct_put_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
void idct_add_altivec(uint8_t* dest, int stride, vector_s16_t* block)
{
POWERPC_PERF_DECLARE(altivec_idct_add_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1);
void simple_idct_add(uint8_t *dest, int line_size, int16_t *block);
simple_idct_add(dest, stride, (int16_t*)block);
POWERPC_PERF_STOP_COUNT(altivec_idct_add_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
vector_u8_t tmp;
vector_s16_t tmp2, tmp3;
vector_u8_t perm0;
@ -246,6 +233,5 @@ POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1);
ADD (dest, vx7, perm1)
POWERPC_PERF_STOP_COUNT(altivec_idct_add_num, 1);
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}

15
libavcodec/ppc/mpegvideo_altivec.c
View File

@ -554,19 +554,6 @@ POWERPC_PERF_START_COUNT(altivec_dct_unquantize_h263_num, 1);
nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ];
}
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
for(;i<=nCoeffs;i++) {
level = block[i];
if (level) {
if (level < 0) {
level = level * qmul - qadd;
} else {
level = level * qmul + qadd;
}
block[i] = level;
}
}
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
{
register const_vector signed short vczero = (const_vector signed short)vec_splat_s16(0);
short __attribute__ ((aligned(16))) qmul8[] =
@ -645,7 +632,5 @@ POWERPC_PERF_START_COUNT(altivec_dct_unquantize_h263_num, 1);
block[0] = backup_0;
}
}
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
POWERPC_PERF_STOP_COUNT(altivec_dct_unquantize_h263_num, nCoeffs == 63);
}

4
libavcodec/ppc/mpegvideo_ppc.c
View File

@ -48,11 +48,7 @@ void MPV_common_init_ppc(MpegEncContext *s)
{
s->dsp.idct_put = idct_put_altivec;
s->dsp.idct_add = idct_add_altivec;
#ifndef ALTIVEC_USE_REFERENCE_C_CODE
s->dsp.idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
s->dsp.idct_permutation_type = FF_NO_IDCT_PERM;
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}
}