github/mpv
1
Fork 0
mirror of https://github.com/mpv-player/mpv synced 2024-10-14 11:54:36 +02:00
Code Releases Activity

runtime cpu detection for the idct

Browse Source 
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@3580 b3059339-0415-0410-9bf9-f77b7e298cf2
This commit is contained in:
michael 2001-12-18 04:00:29 +00:00
parent 1db3d17b8a
commit 6682ec830b
3 changed files with 228 additions and 166 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
liba52/Makefile
View File

@ -19,7 +19,7 @@ $(LIBNAME): $(OBJS)
$(AR) r $(LIBNAME) $(OBJS)
test: $(LIBNAME) test.c
$(CC) $(CFLAGS) test.c -o test -L. -la52
$(CC) $(CFLAGS) test.c ../cpudetect.c -o test -L. -la52 -lm
test2: $(LIBNAME) test.c
$(CC) $(CFLAGS) test.c -o test2 -L../libac3 -lac3 -L. -la52

380
liba52/imdct.c
View File

@ -19,9 +19,12 @@
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* SSE optimizations from Michael Niedermayer (michaelni@gmx.at)
*/
#include "config.h"
#include "../cpudetect.h"
#include <math.h>
#include <stdio.h>
@ -72,7 +75,7 @@ static uint8_t bit_reverse_256[] = {
0x03, 0x23, 0x13, 0x33, 0x0b, 0x2b, 0x1b, 0x3b,
0x07, 0x27, 0x17, 0x37, 0x0f, 0x2f, 0x1f, 0x3f};
#ifdef HAVE_SSE
#ifdef ARCH_X86
// NOTE: SSE needs 16byte alignment or it will segfault
static complex_t __attribute__((aligned(16))) buf[128];
static float __attribute__((aligned(16))) sseSinCos1a[256];
@ -185,36 +188,6 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
/* 512 IMDCT with source and dest data in 'data' */
#ifdef HAVE_SSE
/* Pre IFFT complex multiply plus IFFT cmplx conjugate */
/* Bit reversed shuffling */
asm volatile(
"xorl %%esi, %%esi \n\t"
"leal bit_reverse_512, %%eax \n\t"
"movl $1008, %%edi \n\t"
"pushl %%ebp \n\t" //use ebp without telling gcc
".balign 16 \n\t"
"1: \n\t"
"movaps (%0, %%esi), %%xmm0 \n\t"
"movaps (%0, %%edi), %%xmm1 \n\t"
"shufps $0xA0, %%xmm0, %%xmm0 \n\t"
"shufps $0x5F, %%xmm1, %%xmm1 \n\t"
"mulps sseSinCos1a(%%esi), %%xmm0 \n\t"
"mulps sseSinCos1b(%%esi), %%xmm1 \n\t"
"addps %%xmm1, %%xmm0 \n\t"
"movzbl (%%eax), %%edx \n\t"
"movzbl 1(%%eax), %%ebp \n\t"
"movlps %%xmm0, (%1, %%edx,8) \n\t"
"movhps %%xmm0, (%1, %%ebp,8) \n\t"
"addl $16, %%esi \n\t"
"addl $2, %%eax \n\t" // avoid complex addressing for P4 crap
"subl $16, %%edi \n\t"
" jnc 1b \n\t"
"popl %%ebp \n\t"//no we didnt touch ebp *g*
:: "b" (data), "c" (buf)
: "%esi", "%edi", "%eax", "%edx"
);
#else
/* Pre IFFT complex multiply plus IFFT cmplx conjugate */
for( i=0; i < 128; i++) {
/* z[i] = (X[256-2*i-1] + j * X[2*i]) * (xcos1[i] + j * xsin1[i]) ; */
@ -228,7 +201,6 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
if (k < i)
swap_cmplx(&buf[i],&buf[k]);
}
#endif
/* FFT Merge */
@ -258,7 +230,209 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
}
*/
#ifdef HAVE_SSE
for(i = 0; i < 128; i += 2) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+1].real;
tmp_b_i = buf[i+1].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+1].real = tmp_a_r - tmp_b_r;
buf[i+1].imag = tmp_a_i - tmp_b_i;
}
// Note w[1]={{1,0}, {0,-1}}
for(i = 0; i < 128; i += 4) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+2].real;
tmp_b_i = buf[i+2].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+2].real = tmp_a_r - tmp_b_r;
buf[i+2].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+1].real;
tmp_a_i = buf[i+1].imag;
tmp_b_r = buf[i+3].imag;
tmp_b_i = buf[i+3].real;
buf[i+1].real = tmp_a_r + tmp_b_r;
buf[i+1].imag = tmp_a_i - tmp_b_i;
buf[i+3].real = tmp_a_r - tmp_b_r;
buf[i+3].imag = tmp_a_i + tmp_b_i;
}
for(i = 0; i < 128; i += 8) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+4].real;
tmp_b_i = buf[i+4].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+4].real = tmp_a_r - tmp_b_r;
buf[i+4].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[1+i].real;
tmp_a_i = buf[1+i].imag;
tmp_b_r = (buf[i+5].real + buf[i+5].imag) * w[2][1].real;
tmp_b_i = (buf[i+5].imag - buf[i+5].real) * w[2][1].real;
buf[1+i].real = tmp_a_r + tmp_b_r;
buf[1+i].imag = tmp_a_i + tmp_b_i;
buf[i+5].real = tmp_a_r - tmp_b_r;
buf[i+5].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+2].real;
tmp_a_i = buf[i+2].imag;
tmp_b_r = buf[i+6].imag;
tmp_b_i = - buf[i+6].real;
buf[i+2].real = tmp_a_r + tmp_b_r;
buf[i+2].imag = tmp_a_i + tmp_b_i;
buf[i+6].real = tmp_a_r - tmp_b_r;
buf[i+6].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+3].real;
tmp_a_i = buf[i+3].imag;
tmp_b_r = (buf[i+7].real - buf[i+7].imag) * w[2][3].imag;
tmp_b_i = (buf[i+7].imag + buf[i+7].real) * w[2][3].imag;
buf[i+3].real = tmp_a_r + tmp_b_r;
buf[i+3].imag = tmp_a_i + tmp_b_i;
buf[i+7].real = tmp_a_r - tmp_b_r;
buf[i+7].imag = tmp_a_i - tmp_b_i;
}
for (m=3; m < 7; m++) {
two_m = (1 << m);
two_m_plus_one = two_m<<1;
for(i = 0; i < 128; i += two_m_plus_one) {
for(k = 0; k < two_m; k++) {
int p = k + i;
int q = p + two_m;
tmp_a_r = buf[p].real;
tmp_a_i = buf[p].imag;
tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag;
tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag;
buf[p].real = tmp_a_r + tmp_b_r;
buf[p].imag = tmp_a_i + tmp_b_i;
buf[q].real = tmp_a_r - tmp_b_r;
buf[q].imag = tmp_a_i - tmp_b_i;
}
}
}
/* Post IFFT complex multiply plus IFFT complex conjugate*/
for( i=0; i < 128; i++) {
/* y[n] = z[n] * (xcos1[n] + j * xsin1[n]) ; */
tmp_a_r = buf[i].real;
tmp_a_i = -1.0 * buf[i].imag;
buf[i].real =(tmp_a_r * xcos1[i]) - (tmp_a_i * xsin1[i]);
buf[i].imag =(tmp_a_r * xsin1[i]) + (tmp_a_i * xcos1[i]);
}
data_ptr = data;
delay_ptr = delay;
window_ptr = imdct_window;
/* Window and convert to real valued signal */
for(i=0; i< 64; i++) {
*data_ptr++ = -buf[64+i].imag * *window_ptr++ + *delay_ptr++ + bias;
*data_ptr++ = buf[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias;
}
for(i=0; i< 64; i++) {
*data_ptr++ = -buf[i].real * *window_ptr++ + *delay_ptr++ + bias;
*data_ptr++ = buf[128-i-1].imag * *window_ptr++ + *delay_ptr++ + bias;
}
/* The trailing edge of the window goes into the delay line */
delay_ptr = delay;
for(i=0; i< 64; i++) {
*delay_ptr++ = -buf[64+i].real * *--window_ptr;
*delay_ptr++ = buf[64-i-1].imag * *--window_ptr;
}
for(i=0; i<64; i++) {
*delay_ptr++ = buf[i].imag * *--window_ptr;
*delay_ptr++ = -buf[128-i-1].real * *--window_ptr;
}
}
#ifdef ARCH_X86
void
imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
{
int i,k;
int p,q;
int m;
int two_m;
int two_m_plus_one;
sample_t tmp_a_i;
sample_t tmp_a_r;
sample_t tmp_b_i;
sample_t tmp_b_r;
sample_t *data_ptr;
sample_t *delay_ptr;
sample_t *window_ptr;
/* 512 IMDCT with source and dest data in 'data' */
/* Pre IFFT complex multiply plus IFFT cmplx conjugate */
/* Bit reversed shuffling */
asm volatile(
"xorl %%esi, %%esi \n\t"
"leal bit_reverse_512, %%eax \n\t"
"movl $1008, %%edi \n\t"
"pushl %%ebp \n\t" //use ebp without telling gcc
".balign 16 \n\t"
"1: \n\t"
"movaps (%0, %%esi), %%xmm0 \n\t"
"movaps (%0, %%edi), %%xmm1 \n\t"
"shufps $0xA0, %%xmm0, %%xmm0 \n\t"
"shufps $0x5F, %%xmm1, %%xmm1 \n\t"
"mulps sseSinCos1a(%%esi), %%xmm0 \n\t"
"mulps sseSinCos1b(%%esi), %%xmm1 \n\t"
"addps %%xmm1, %%xmm0 \n\t"
"movzbl (%%eax), %%edx \n\t"
"movzbl 1(%%eax), %%ebp \n\t"
"movlps %%xmm0, (%1, %%edx,8) \n\t"
"movhps %%xmm0, (%1, %%ebp,8) \n\t"
"addl $16, %%esi \n\t"
"addl $2, %%eax \n\t" // avoid complex addressing for P4 crap
"subl $16, %%edi \n\t"
" jnc 1b \n\t"
"popl %%ebp \n\t"//no we didnt touch ebp *g*
:: "b" (data), "c" (buf)
: "%esi", "%edi", "%eax", "%edx"
);
/* FFT Merge */
/* unoptimized variant
for (m=1; m < 7; m++) {
if(m)
two_m = (1 << m);
else
two_m = 1;
two_m_plus_one = (1 << (m+1));
for(i = 0; i < 128; i += two_m_plus_one) {
for(k = 0; k < two_m; k++) {
p = k + i;
q = p + two_m;
tmp_a_r = buf[p].real;
tmp_a_i = buf[p].imag;
tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag;
tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag;
buf[p].real = tmp_a_r + tmp_b_r;
buf[p].imag = tmp_a_i + tmp_b_i;
buf[q].real = tmp_a_r - tmp_b_r;
buf[q].imag = tmp_a_i - tmp_b_i;
}
}
}
*/
// Note w[0][0]={1,0}
asm volatile(
"xorps %%xmm1, %%xmm1 \n\t"
@ -279,21 +453,8 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
:: "g" (buf), "r" (buf + 128)
: "%esi"
);
#else
for(i = 0; i < 128; i += 2) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+1].real;
tmp_b_i = buf[i+1].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+1].real = tmp_a_r - tmp_b_r;
buf[i+1].imag = tmp_a_i - tmp_b_i;
}
#endif
// Note w[1]={{1,0}, {0,-1}}
#ifdef HAVE_SSE
asm volatile(
"movaps ps111_1, %%xmm7 \n\t" // 1,1,1,-1
"movl %0, %%esi \n\t"
@ -314,28 +475,7 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
:: "g" (buf), "r" (buf + 128)
: "%esi"
);
#else
for(i = 0; i < 128; i += 4) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+2].real;
tmp_b_i = buf[i+2].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+2].real = tmp_a_r - tmp_b_r;
buf[i+2].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+1].real;
tmp_a_i = buf[i+1].imag;
tmp_b_r = buf[i+3].imag;
tmp_b_i = buf[i+3].real;
buf[i+1].real = tmp_a_r + tmp_b_r;
buf[i+1].imag = tmp_a_i - tmp_b_i;
buf[i+3].real = tmp_a_r - tmp_b_r;
buf[i+3].imag = tmp_a_i + tmp_b_i;
}
#endif
#ifdef HAVE_SSE
/*
Note sseW2+0={1,1,sqrt(2),sqrt(2))
Note sseW2+16={0,0,sqrt(2),-sqrt(2))
@ -380,44 +520,7 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
:: "g" (buf), "r" (buf + 128)
: "%esi"
);
#else
for(i = 0; i < 128; i += 8) {
tmp_a_r = buf[i].real;
tmp_a_i = buf[i].imag;
tmp_b_r = buf[i+4].real;
tmp_b_i = buf[i+4].imag;
buf[i].real = tmp_a_r + tmp_b_r;
buf[i].imag = tmp_a_i + tmp_b_i;
buf[i+4].real = tmp_a_r - tmp_b_r;
buf[i+4].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[1+i].real;
tmp_a_i = buf[1+i].imag;
tmp_b_r = (buf[i+5].real + buf[i+5].imag) * w[2][1].real;
tmp_b_i = (buf[i+5].imag - buf[i+5].real) * w[2][1].real;
buf[1+i].real = tmp_a_r + tmp_b_r;
buf[1+i].imag = tmp_a_i + tmp_b_i;
buf[i+5].real = tmp_a_r - tmp_b_r;
buf[i+5].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+2].real;
tmp_a_i = buf[i+2].imag;
tmp_b_r = buf[i+6].imag;
tmp_b_i = - buf[i+6].real;
buf[i+2].real = tmp_a_r + tmp_b_r;
buf[i+2].imag = tmp_a_i + tmp_b_i;
buf[i+6].real = tmp_a_r - tmp_b_r;
buf[i+6].imag = tmp_a_i - tmp_b_i;
tmp_a_r = buf[i+3].real;
tmp_a_i = buf[i+3].imag;
tmp_b_r = (buf[i+7].real - buf[i+7].imag) * w[2][3].imag;
tmp_b_i = (buf[i+7].imag + buf[i+7].real) * w[2][3].imag;
buf[i+3].real = tmp_a_r + tmp_b_r;
buf[i+3].imag = tmp_a_i + tmp_b_i;
buf[i+7].real = tmp_a_r - tmp_b_r;
buf[i+7].imag = tmp_a_i - tmp_b_i;
}
#endif
#ifdef HAVE_SSE
for (m=3; m < 7; m++) {
two_m = (1 << m);
two_m_plus_one = two_m<<1;
@ -452,28 +555,6 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
);
}
#else
for (m=3; m < 7; m++) {
two_m = (1 << m);
two_m_plus_one = two_m<<1;
for(i = 0; i < 128; i += two_m_plus_one) {
for(k = 0; k < two_m; k++) {
int p = k + i;
int q = p + two_m;
tmp_a_r = buf[p].real;
tmp_a_i = buf[p].imag;
tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag;
tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag;
buf[p].real = tmp_a_r + tmp_b_r;
buf[p].imag = tmp_a_i + tmp_b_i;
buf[q].real = tmp_a_r - tmp_b_r;
buf[q].imag = tmp_a_i - tmp_b_i;
}
}
}
#endif
/* Post IFFT complex multiply plus IFFT complex conjugate*/
for( i=0; i < 128; i++) {
@ -489,7 +570,6 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
window_ptr = imdct_window;
/* Window and convert to real valued signal */
#ifdef HAVE_SSE
asm volatile(
"xorl %%edi, %%edi \n\t" // 0
"xorl %%esi, %%esi \n\t" // 0
@ -516,14 +596,7 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
data_ptr+=128;
delay_ptr+=128;
// window_ptr+=128;
#else
for(i=0; i< 64; i++) {
*data_ptr++ = -buf[64+i].imag * *window_ptr++ + *delay_ptr++ + bias;
*data_ptr++ = buf[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias;
}
#endif
#ifdef HAVE_SSE
asm volatile(
"movl $1024, %%edi \n\t" // 512
"xorl %%esi, %%esi \n\t" // 0
@ -549,17 +622,10 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
);
data_ptr+=128;
// window_ptr+=128;
#else
for(i=0; i< 64; i++) {
*data_ptr++ = -buf[i].real * *window_ptr++ + *delay_ptr++ + bias;
*data_ptr++ = buf[128-i-1].imag * *window_ptr++ + *delay_ptr++ + bias;
}
#endif
/* The trailing edge of the window goes into the delay line */
delay_ptr = delay;
#ifdef HAVE_SSE
asm volatile(
"xorl %%edi, %%edi \n\t" // 0
"xorl %%esi, %%esi \n\t" // 0
@ -581,14 +647,7 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
);
delay_ptr+=128;
// window_ptr-=128;
#else
for(i=0; i< 64; i++) {
*delay_ptr++ = -buf[64+i].real * *--window_ptr;
*delay_ptr++ = buf[64-i-1].imag * *--window_ptr;
}
#endif
#ifdef HAVE_SSE
asm volatile(
"movl $1024, %%edi \n\t" // 1024
"xorl %%esi, %%esi \n\t" // 0
@ -608,13 +667,8 @@ imdct_do_512(sample_t data[],sample_t delay[], sample_t bias)
:: "r" (buf), "r" (delay_ptr)
: "%esi", "%edi"
);
#else
for(i=0; i<64; i++) {
*delay_ptr++ = buf[i].imag * *--window_ptr;
*delay_ptr++ = -buf[128-i-1].real * *--window_ptr;
}
#endif
}
#endif //arch_x86
void
imdct_do_256(sample_t data[],sample_t delay[],sample_t bias)
@ -756,14 +810,15 @@ void imdct_init (uint32_t mm_accel)
{
int i, j, k;
fprintf (stderr, "No accelerated IMDCT transform found\n");
if(gCpuCaps.hasSSE) fprintf (stderr, "Using SSE optimized IMDCT transform\n");
else fprintf (stderr, "No accelerated IMDCT transform found\n");
/* Twiddle factors to turn IFFT into IMDCT */
for (i = 0; i < 128; i++) {
xcos1[i] = -cos ((M_PI / 2048) * (8 * i + 1));
xsin1[i] = -sin ((M_PI / 2048) * (8 * i + 1));
}
#ifdef HAVE_SSE
#ifdef ARCH_X86
for (i = 0; i < 128; i++) {
sseSinCos1a[2*i+0]= -xsin1[i];
sseSinCos1a[2*i+1]= -xcos1[i];
@ -785,7 +840,7 @@ void imdct_init (uint32_t mm_accel)
w[i][k].imag = sin (-M_PI * k / j);
}
}
#ifdef HAVE_SSE
#ifdef ARCH_X86
for (i = 1; i < 7; i++) {
j = 1 << i;
for (k = 0; k < j; k+=2) {
@ -828,9 +883,10 @@ void imdct_init (uint32_t mm_accel)
sseWindow[384 + 2*i+0]= imdct_window[126 - 2*i+1];
sseWindow[384 + 2*i+1]= -imdct_window[126 - 2*i+0];
}
#endif
imdct_512 = imdct_do_512;
#endif // arch_x86
if(gCpuCaps.hasSSE) imdct_512 = imdct_do_512_sse;
else imdct_512 = imdct_do_512;
imdct_256 = imdct_do_256;
}
}

12
liba52/test.c
View File

@ -11,6 +11,7 @@
#include <inttypes.h>
#include "a52.h"
#include "../cpudetect.h"
static sample_t * samples;
static a52_state_t state;
@ -46,6 +47,11 @@ int bit_rate=0;
long long t, sum=0, min=256*256*256*64;
#endif
FILE *temp= stdout;
stdout= stderr; //EVIL HACK FIXME
GetCpuCaps(&gCpuCaps);
stdout= temp;
samples = a52_init (accel);
if (samples == NULL) {
fprintf (stderr, "A52 init failed\n");
@ -81,7 +87,7 @@ ENDTIMING
buf_size=0;
// decode:
flags=A52_STEREO; // A52_DOLBY // A52_2F2R // A52_3F2R | A52_LFE
flags=A52_STEREO; //A52_STEREO; // A52_DOLBY // A52_2F2R // A52_3F2R | A52_LFE
channels=2;
flags |= A52_ADJUST_LEVEL;
@ -92,15 +98,15 @@ ENDTIMING
// a52_dynrng (&state, NULL, NULL); // disable dynamic range compensation
STARTTIMING
a52_resample_init(flags,channels);
s16 = out_buf;
for (i = 0; i < 6; i++) {
STARTTIMING
if (a52_block (&state, samples))
{ fprintf(stderr,"error at sampling\n"); break; }
ENDTIMING
// float->int + channels interleaving:
s16+=a52_resample(samples,s16);
ENDTIMING
}
#ifdef TIMING
if(sum<min) min=sum;