int signal_energy_nodc(int *input,unsigned int length) {
int i;
int temp,temp2;
register __m64 mm0,mm1,mm2,mm3;
__m64 *in = (__m64 *)input;
#ifdef MAIN
short *printb;
#endif
mm0 = _m_pxor(mm0,mm0);
mm3 = _m_pxor(mm3,mm3);
for (i=0;i<length>>1;i++) {
mm1 = in[i];
mm2 = mm1;
mm1 = _m_pmaddwd(mm1,mm1);// SIMD complex multiplication
mm1 = _m_psradi(mm1,shift);
mm0 = _m_paddd(mm0,mm1);
// temp2 = mm0;
// printf("%d %d\n",((int *)&in[i])[0],((int *)&in[i])[1]);
// printb = (short *)&mm2;
// printf("mm2 %d : %d %d %d %d\n",i,printb[0],printb[1],printb[2],printb[3]);
}
/*
#ifdef MAIN
printb = (short *)&mm3;
printf("%d %d %d %d\n",printb[0],printb[1],printb[2],printb[3]);
#endif
*/
mm1 = mm0;
mm0 = _m_psrlqi(mm0,32);
mm0 = _m_paddd(mm0,mm1);
temp = _m_to_int(mm0);
temp/=length;
temp<<=shift; // this is the average of x^2
#ifdef MAIN
printf("E x^2 = %d\n",temp);
#endif
_mm_empty();
_m_empty();
return((temp>0)?temp:1);
}
void main()
{
a = _m_pxor( b, c );
printf( "m1="AS_QWORD"\n"
"m2="AS_QWORD"\n"
"mm="AS_QWORD"\n",
b, c, a );
}
void sha384Process(register sha384Param* sp)
{
#ifdef OPTIMIZE_SSE2
# if defined(_MSC_VER) || defined (__INTEL_COMPILER)
static const __m64 MASK = { 0x00FF00FF00FF00FF00 };
# elif defined(__GNUC__)
static const __m64 MASK = { 0x00FF00FF, 0x00FF00FF };
# else
# error
# endif
__m64 a, b, c, d, e, f, g, h, temp;
register __m64 *w;
register const __m64 *k;
register byte t;
w = (__m64*) sp->data;
t = 16;
while (t--)
{
temp = *w;
*(w++) = _m_pxor(
_mm_slli_si64(_m_pshufw(_m_pand(temp, MASK), 27), 8),
_m_pshufw(_m_pand(_mm_srli_si64(temp, 8), MASK), 27)
);
}
t = 64;
while (t--)
{
temp = _mm_add_si64(_mm_add_si64(sig1(w[-2]), w[-7]), _mm_add_si64(sig0(w[-15]), w[-16]));
*(w++) = temp;
}
w = (__m64*) sp->h;
a = w[0]; b = w[1]; c = w[2]; d = w[3];
e = w[4]; f = w[5]; g = w[6]; h = w[7];
w = (__m64*) sp->data;
k = (__m64*) SHA2_64BIT_K;
#else
register uint64_t a, b, c, d, e, f, g, h, temp;
register uint64_t *w;
register const uint64_t *k;
register byte t;
# if WORDS_BIGENDIAN
w = sp->data + 16;
# else
w = sp->data;
t = 16;
while (t--)
{
temp = swapu64(*w);
*(w++) = temp;
}
# endif
t = 64;
while (t--)
{
temp = sig1(w[-2]) + w[-7] + sig0(w[-15]) + w[-16];
*(w++) = temp;
}
w = sp->data;
a = sp->h[0]; b = sp->h[1]; c = sp->h[2]; d = sp->h[3];
e = sp->h[4]; f = sp->h[5]; g = sp->h[6]; h = sp->h[7];
k = SHA2_64BIT_K;
#endif
ROUND(a,b,c,d,e,f,g,h,w[ 0],k[ 0]);
ROUND(h,a,b,c,d,e,f,g,w[ 1],k[ 1]);
ROUND(g,h,a,b,c,d,e,f,w[ 2],k[ 2]);
ROUND(f,g,h,a,b,c,d,e,w[ 3],k[ 3]);
ROUND(e,f,g,h,a,b,c,d,w[ 4],k[ 4]);
ROUND(d,e,f,g,h,a,b,c,w[ 5],k[ 5]);
ROUND(c,d,e,f,g,h,a,b,w[ 6],k[ 6]);
ROUND(b,c,d,e,f,g,h,a,w[ 7],k[ 7]);
ROUND(a,b,c,d,e,f,g,h,w[ 8],k[ 8]);
ROUND(h,a,b,c,d,e,f,g,w[ 9],k[ 9]);
ROUND(g,h,a,b,c,d,e,f,w[10],k[10]);
ROUND(f,g,h,a,b,c,d,e,w[11],k[11]);
ROUND(e,f,g,h,a,b,c,d,w[12],k[12]);
ROUND(d,e,f,g,h,a,b,c,w[13],k[13]);
ROUND(c,d,e,f,g,h,a,b,w[14],k[14]);
ROUND(b,c,d,e,f,g,h,a,w[15],k[15]);
ROUND(a,b,c,d,e,f,g,h,w[16],k[16]);
ROUND(h,a,b,c,d,e,f,g,w[17],k[17]);
ROUND(g,h,a,b,c,d,e,f,w[18],k[18]);
ROUND(f,g,h,a,b,c,d,e,w[19],k[19]);
ROUND(e,f,g,h,a,b,c,d,w[20],k[20]);
ROUND(d,e,f,g,h,a,b,c,w[21],k[21]);
ROUND(c,d,e,f,g,h,a,b,w[22],k[22]);
ROUND(b,c,d,e,f,g,h,a,w[23],k[23]);
ROUND(a,b,c,d,e,f,g,h,w[24],k[24]);
ROUND(h,a,b,c,d,e,f,g,w[25],k[25]);
ROUND(g,h,a,b,c,d,e,f,w[26],k[26]);
ROUND(f,g,h,a,b,c,d,e,w[27],k[27]);
ROUND(e,f,g,h,a,b,c,d,w[28],k[28]);
ROUND(d,e,f,g,h,a,b,c,w[29],k[29]);
ROUND(c,d,e,f,g,h,a,b,w[30],k[30]);
ROUND(b,c,d,e,f,g,h,a,w[31],k[31]);
ROUND(a,b,c,d,e,f,g,h,w[32],k[32]);
ROUND(h,a,b,c,d,e,f,g,w[33],k[33]);
ROUND(g,h,a,b,c,d,e,f,w[34],k[34]);
ROUND(f,g,h,a,b,c,d,e,w[35],k[35]);
ROUND(e,f,g,h,a,b,c,d,w[36],k[36]);
ROUND(d,e,f,g,h,a,b,c,w[37],k[37]);
ROUND(c,d,e,f,g,h,a,b,w[38],k[38]);
ROUND(b,c,d,e,f,g,h,a,w[39],k[39]);
ROUND(a,b,c,d,e,f,g,h,w[40],k[40]);
ROUND(h,a,b,c,d,e,f,g,w[41],k[41]);
ROUND(g,h,a,b,c,d,e,f,w[42],k[42]);
ROUND(f,g,h,a,b,c,d,e,w[43],k[43]);
ROUND(e,f,g,h,a,b,c,d,w[44],k[44]);
ROUND(d,e,f,g,h,a,b,c,w[45],k[45]);
ROUND(c,d,e,f,g,h,a,b,w[46],k[46]);
ROUND(b,c,d,e,f,g,h,a,w[47],k[47]);
ROUND(a,b,c,d,e,f,g,h,w[48],k[48]);
ROUND(h,a,b,c,d,e,f,g,w[49],k[49]);
ROUND(g,h,a,b,c,d,e,f,w[50],k[50]);
ROUND(f,g,h,a,b,c,d,e,w[51],k[51]);
ROUND(e,f,g,h,a,b,c,d,w[52],k[52]);
ROUND(d,e,f,g,h,a,b,c,w[53],k[53]);
ROUND(c,d,e,f,g,h,a,b,w[54],k[54]);
ROUND(b,c,d,e,f,g,h,a,w[55],k[55]);
ROUND(a,b,c,d,e,f,g,h,w[56],k[56]);
ROUND(h,a,b,c,d,e,f,g,w[57],k[57]);
ROUND(g,h,a,b,c,d,e,f,w[58],k[58]);
ROUND(f,g,h,a,b,c,d,e,w[59],k[59]);
ROUND(e,f,g,h,a,b,c,d,w[60],k[60]);
ROUND(d,e,f,g,h,a,b,c,w[61],k[61]);
ROUND(c,d,e,f,g,h,a,b,w[62],k[62]);
ROUND(b,c,d,e,f,g,h,a,w[63],k[63]);
ROUND(a,b,c,d,e,f,g,h,w[64],k[64]);
ROUND(h,a,b,c,d,e,f,g,w[65],k[65]);
ROUND(g,h,a,b,c,d,e,f,w[66],k[66]);
ROUND(f,g,h,a,b,c,d,e,w[67],k[67]);
ROUND(e,f,g,h,a,b,c,d,w[68],k[68]);
ROUND(d,e,f,g,h,a,b,c,w[69],k[69]);
ROUND(c,d,e,f,g,h,a,b,w[70],k[70]);
ROUND(b,c,d,e,f,g,h,a,w[71],k[71]);
ROUND(a,b,c,d,e,f,g,h,w[72],k[72]);
ROUND(h,a,b,c,d,e,f,g,w[73],k[73]);
ROUND(g,h,a,b,c,d,e,f,w[74],k[74]);
ROUND(f,g,h,a,b,c,d,e,w[75],k[75]);
ROUND(e,f,g,h,a,b,c,d,w[76],k[76]);
ROUND(d,e,f,g,h,a,b,c,w[77],k[77]);
ROUND(c,d,e,f,g,h,a,b,w[78],k[78]);
ROUND(b,c,d,e,f,g,h,a,w[79],k[79]);
#ifdef OPTIMIZE_SSE2
w = (__m64*) sp->h;
w[0] = _mm_add_si64(w[0], a);
w[1] = _mm_add_si64(w[1], b);
w[2] = _mm_add_si64(w[2], c);
w[3] = _mm_add_si64(w[3], d);
w[4] = _mm_add_si64(w[4], e);
w[5] = _mm_add_si64(w[5], f);
w[6] = _mm_add_si64(w[6], g);
w[7] = _mm_add_si64(w[7], h);
_mm_empty();
#else
sp->h[0] += a;
sp->h[1] += b;
sp->h[2] += c;
sp->h[3] += d;
sp->h[4] += e;
sp->h[5] += f;
sp->h[6] += g;
sp->h[7] += h;
#endif
}
int signal_energy(int *input,unsigned int length) {
int i;
int temp,temp2;
register __m64 mm0,mm1,mm2,mm3;
__m64 *in = (__m64 *)input;
#ifdef MAIN
short *printb;
#endif
mm0 = _m_pxor(mm0,mm0);
mm3 = _m_pxor(mm3,mm3);
for (i=0;i<length>>1;i++) {
mm1 = in[i];
mm2 = mm1;
mm1 = _m_pmaddwd(mm1,mm1);
mm1 = _m_psradi(mm1,shift);// shift any 32 bits blocs of the word by the value shift
mm0 = _m_paddd(mm0,mm1);// add the two 64 bits words 4 bytes by 4 bytes
// temp2 = mm0;
// printf("%d %d\n",((int *)&temp2)[0],((int *)&temp2)[1]);
// printb = (short *)&mm2;
// printf("mm2 %d : %d %d %d %d\n",i,printb[0],printb[1],printb[2],printb[3]);
mm2 = _m_psrawi(mm2,shift_DC);
mm3 = _m_paddw(mm3,mm2);// add the two 64 bits words 2 bytes by 2 bytes
// printb = (short *)&mm3;
// printf("mm3 %d : %d %d %d %d\n",i,printb[0],printb[1],printb[2],printb[3]);
}
/*
#ifdef MAIN
printb = (short *)&mm3;
printf("%d %d %d %d\n",printb[0],printb[1],printb[2],printb[3]);
#endif
*/
mm1 = mm0;
mm0 = _m_psrlqi(mm0,32);
mm0 = _m_paddd(mm0,mm1);
temp = _m_to_int(mm0);
temp/=length;
temp<<=shift; // this is the average of x^2
// now remove the DC component
mm2 = _m_psrlqi(mm3,32);
mm2 = _m_paddw(mm2,mm3);
mm2 = _m_pmaddwd(mm2,mm2);
temp2 = _m_to_int(mm2);
temp2/=(length*length);
temp2<<=(2*shift_DC);
#ifdef MAIN
printf("E x^2 = %d\n",temp);
#endif
temp -= temp2;
#ifdef MAIN
printf("(E x)^2=%d\n",temp2);
#endif
_mm_empty();
_m_empty();
return((temp>0)?temp:1);
}
