本文整理匯總了C++中BYTWJ函數的典型用法代碼示例。如果您正苦於以下問題:C++ BYTWJ函數的具體用法?C++ BYTWJ怎麽用?C++ BYTWJ使用的例子?那麽, 這裏精選的函數代碼示例或許可以為您提供幫助。
在下文中一共展示了BYTWJ函數的15個代碼示例,這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚,您的評價將有助於係統推薦出更棒的C++代碼示例。
示例1: DVK
static const R *t1fv_3(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 4)) {
V T1, T3, T5, T6, T2, T4, T7, T8;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[0]), T2);
T4 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 2]), T4);
T6 = VADD(T3, T5);
ST(&(x[0]), VADD(T1, T6), dist, &(x[0]));
T7 = VFNMS(LDK(KP500000000), T6, T1);
T8 = VBYI(VMUL(LDK(KP866025403), VSUB(T5, T3)));
ST(&(x[WS(ios, 2)]), VSUB(T7, T8), dist, &(x[0]));
ST(&(x[WS(ios, 1)]), VADD(T7, T8), dist, &(x[WS(ios, 1)]));
}
END_SIMD();
return W;
}
示例2: t1fuv_4
static void t1fuv_4(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(4, rs)) {
V T1, T8, T3, T6, T7, T2, T5;
T1 = LD(&(x[0]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T8 = BYTWJ(&(W[TWVL * 4]), T7);
T2 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T3 = BYTWJ(&(W[TWVL * 2]), T2);
T5 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T6 = BYTWJ(&(W[0]), T5);
{
V T4, T9, Ta, Tb;
T4 = VSUB(T1, T3);
T9 = VBYI(VSUB(T6, T8));
ST(&(x[WS(rs, 1)]), VSUB(T4, T9), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(T4, T9), ms, &(x[WS(rs, 1)]));
Ta = VADD(T1, T3);
Tb = VADD(T6, T8);
ST(&(x[WS(rs, 2)]), VSUB(Ta, Tb), ms, &(x[0]));
ST(&(x[0]), VADD(Ta, Tb), ms, &(x[0]));
}
}
}
VLEAVE();
}
示例3: t1fuv_3
static void t1fuv_3(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, T4;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T4 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
{
V T3, T5, T8, T6, T7;
T3 = BYTWJ(&(W[0]), T2);
T5 = BYTWJ(&(W[TWVL * 2]), T4);
T8 = VMUL(LDK(KP866025403), VSUB(T5, T3));
T6 = VADD(T3, T5);
T7 = VFNMS(LDK(KP500000000), T6, T1);
ST(&(x[0]), VADD(T1, T6), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VFMAI(T8, T7), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VFNMSI(T8, T7), ms, &(x[0]));
}
}
}
示例4: DVK
static const R *t1fv_6(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 10)) {
V T4, Ti, Te, Tk, T9, Tj, T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[TWVL * 4]), T2);
T4 = VSUB(T1, T3);
Ti = VADD(T1, T3);
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Tc = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Td = BYTWJ(&(W[0]), Tc);
Te = VSUB(Tb, Td);
Tk = VADD(Tb, Td);
}
{
V T6, T8, T5, T7;
T5 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T6 = BYTWJ(&(W[TWVL * 2]), T5);
T7 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
T8 = BYTWJ(&(W[TWVL * 8]), T7);
T9 = VSUB(T6, T8);
Tj = VADD(T6, T8);
}
{
V Th, Tf, Tg, Tn, Tl, Tm;
Th = VBYI(VMUL(LDK(KP866025403), VSUB(Te, T9)));
Tf = VADD(T9, Te);
Tg = VFNMS(LDK(KP500000000), Tf, T4);
ST(&(x[WS(ios, 3)]), VADD(T4, Tf), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(Tg, Th), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VSUB(Tg, Th), dist, &(x[WS(ios, 1)]));
Tn = VBYI(VMUL(LDK(KP866025403), VSUB(Tk, Tj)));
Tl = VADD(Tj, Tk);
Tm = VFNMS(LDK(KP500000000), Tl, Ti);
ST(&(x[0]), VADD(Ti, Tl), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(Tm, Tn), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VSUB(Tm, Tn), dist, &(x[0]));
}
}
END_SIMD();
return W;
}
示例5: t1fuv_6
static void t1fuv_6(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 10)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 10), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, Ta, Tc, T5, T7;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V T3, Tb, Td, T6, T8;
T3 = BYTWJ(&(W[TWVL * 4]), T2);
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Td = BYTWJ(&(W[0]), Tc);
T6 = BYTWJ(&(W[TWVL * 2]), T5);
T8 = BYTWJ(&(W[TWVL * 8]), T7);
{
V Ti, T4, Tk, Te, Tj, T9;
Ti = VADD(T1, T3);
T4 = VSUB(T1, T3);
Tk = VADD(Tb, Td);
Te = VSUB(Tb, Td);
Tj = VADD(T6, T8);
T9 = VSUB(T6, T8);
{
V Tl, Tn, Tf, Th, Tm, Tg;
Tl = VADD(Tj, Tk);
Tn = VMUL(LDK(KP866025403), VSUB(Tk, Tj));
Tf = VADD(T9, Te);
Th = VMUL(LDK(KP866025403), VSUB(Te, T9));
ST(&(x[0]), VADD(Ti, Tl), ms, &(x[0]));
Tm = VFNMS(LDK(KP500000000), Tl, Ti);
ST(&(x[WS(rs, 3)]), VADD(T4, Tf), ms, &(x[WS(rs, 1)]));
Tg = VFNMS(LDK(KP500000000), Tf, T4);
ST(&(x[WS(rs, 2)]), VFNMSI(Tn, Tm), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(Tn, Tm), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VFNMSI(Th, Tg), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(Th, Tg), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
示例6: t1fv_5
static void t1fv_5(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 8)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(5, rs)) {
V Tc, Tg, Th, T5, Ta, Td;
Tc = LD(&(x[0]), ms, &(x[0]));
{
V T2, T9, T4, T7;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T2 = BYTWJ(&(W[0]), T1);
T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = BYTWJ(&(W[TWVL * 4]), T8);
T3 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T4 = BYTWJ(&(W[TWVL * 6]), T3);
T6 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T7 = BYTWJ(&(W[TWVL * 2]), T6);
}
Tg = VSUB(T2, T4);
Th = VSUB(T7, T9);
T5 = VADD(T2, T4);
Ta = VADD(T7, T9);
Td = VADD(T5, Ta);
}
ST(&(x[0]), VADD(Tc, Td), ms, &(x[0]));
{
V Ti, Tj, Tf, Tk, Tb, Te;
Ti = VBYI(VFMA(LDK(KP951056516), Tg, VMUL(LDK(KP587785252), Th)));
Tj = VBYI(VFNMS(LDK(KP587785252), Tg, VMUL(LDK(KP951056516), Th)));
Tb = VMUL(LDK(KP559016994), VSUB(T5, Ta));
Te = VFNMS(LDK(KP250000000), Td, Tc);
Tf = VADD(Tb, Te);
Tk = VSUB(Te, Tb);
ST(&(x[WS(rs, 1)]), VSUB(Tf, Ti), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VSUB(Tk, Tj), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 4)]), VADD(Ti, Tf), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(Tj, Tk), ms, &(x[0]));
}
}
}
VLEAVE();
}
示例7: DVK
static const R *t1fv_6(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 10), MAKE_VOLATILE_STRIDE(ios)) {
V T1, T2, Ta, Tc, T5, T7;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Ta = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
Tc = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T5 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T7 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
{
V T3, Tb, Td, T6, T8;
T3 = BYTWJ(&(W[TWVL * 4]), T2);
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Td = BYTWJ(&(W[0]), Tc);
T6 = BYTWJ(&(W[TWVL * 2]), T5);
T8 = BYTWJ(&(W[TWVL * 8]), T7);
{
V Ti, T4, Tk, Te, Tj, T9;
Ti = VADD(T1, T3);
T4 = VSUB(T1, T3);
Tk = VADD(Tb, Td);
Te = VSUB(Tb, Td);
Tj = VADD(T6, T8);
T9 = VSUB(T6, T8);
{
V Tl, Tn, Tf, Th, Tm, Tg;
Tl = VADD(Tj, Tk);
Tn = VMUL(LDK(KP866025403), VSUB(Tk, Tj));
Tf = VADD(T9, Te);
Th = VMUL(LDK(KP866025403), VSUB(Te, T9));
ST(&(x[0]), VADD(Ti, Tl), dist, &(x[0]));
Tm = VFNMS(LDK(KP500000000), Tl, Ti);
ST(&(x[WS(ios, 3)]), VADD(T4, Tf), dist, &(x[WS(ios, 1)]));
Tg = VFNMS(LDK(KP500000000), Tf, T4);
ST(&(x[WS(ios, 2)]), VFNMSI(Tn, Tm), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VFMAI(Tn, Tm), dist, &(x[0]));
ST(&(x[WS(ios, 5)]), VFNMSI(Th, Tg), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VFMAI(Th, Tg), dist, &(x[WS(ios, 1)]));
}
}
}
}
return W;
}
示例8: q1fv_2
static void q1fv_2(R *ri, R *ii, const R *W, stride rs, stride vs, INT mb, INT me, INT ms)
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(vs)) {
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[0]), VSUB(T1, T2));
T4 = LD(&(x[WS(vs, 1)]), ms, &(x[WS(vs, 1)]));
T5 = LD(&(x[WS(vs, 1) + WS(rs, 1)]), ms, &(x[WS(vs, 1) + WS(rs, 1)]));
T6 = BYTWJ(&(W[0]), VSUB(T4, T5));
ST(&(x[WS(vs, 1)]), T3, ms, &(x[WS(vs, 1)]));
ST(&(x[WS(vs, 1) + WS(rs, 1)]), T6, ms, &(x[WS(vs, 1) + WS(rs, 1)]));
ST(&(x[0]), VADD(T1, T2), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VADD(T4, T5), ms, &(x[WS(rs, 1)]));
}
}
示例9: LD
static const R *q1fv_2(R *ri, R *ii, const R *W, stride is, stride vs, INT m, INT dist)
{
INT i;
R *x;
x = ri;
for (i = 0; i < m; i = i + VL, x = x + (VL * dist), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(vs)) {
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(is, 1)]), dist, &(x[WS(is, 1)]));
T3 = BYTWJ(&(W[0]), VSUB(T1, T2));
T4 = LD(&(x[WS(vs, 1)]), dist, &(x[WS(vs, 1)]));
T5 = LD(&(x[WS(vs, 1) + WS(is, 1)]), dist, &(x[WS(vs, 1) + WS(is, 1)]));
T6 = BYTWJ(&(W[0]), VSUB(T4, T5));
ST(&(x[WS(vs, 1)]), T3, dist, &(x[WS(vs, 1)]));
ST(&(x[WS(vs, 1) + WS(is, 1)]), T6, dist, &(x[WS(vs, 1) + WS(is, 1)]));
ST(&(x[0]), VADD(T1, T2), dist, &(x[0]));
ST(&(x[WS(is, 1)]), VADD(T4, T5), dist, &(x[WS(is, 1)]));
}
return W;
}
示例10: t1fuv_7
static void t1fuv_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 12)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 12), MAKE_VOLATILE_STRIDE(rs)) {
V T1, Tg, Tj, T6, Ti, Tb, Tk, Tp, To;
T1 = LD(&(x[0]), ms, &(x[0]));
{
V Td, Tf, Tc, Te;
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[TWVL * 4]), Tc);
Te = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tf = BYTWJ(&(W[TWVL * 6]), Te);
Tg = VADD(Td, Tf);
Tj = VSUB(Tf, Td);
}
{
V T3, T5, T2, T4;
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[0]), T2);
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 10]), T4);
T6 = VADD(T3, T5);
Ti = VSUB(T5, T3);
}
{
V T8, Ta, T7, T9;
T7 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T8 = BYTWJ(&(W[TWVL * 2]), T7);
T9 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Ta = BYTWJ(&(W[TWVL * 8]), T9);
Tb = VADD(T8, Ta);
Tk = VSUB(Ta, T8);
}
ST(&(x[0]), VADD(T1, VADD(T6, VADD(Tb, Tg))), ms, &(x[0]));
Tp = VBYI(VFMA(LDK(KP433883739), Ti, VFNMS(LDK(KP781831482), Tk, VMUL(LDK(KP974927912), Tj))));
To = VFMA(LDK(KP623489801), Tb, VFNMS(LDK(KP222520933), Tg, VFNMS(LDK(KP900968867), T6, T1)));
ST(&(x[WS(rs, 4)]), VSUB(To, Tp), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VADD(To, Tp), ms, &(x[WS(rs, 1)]));
{
V Tl, Th, Tn, Tm;
Tl = VBYI(VFNMS(LDK(KP781831482), Tj, VFNMS(LDK(KP433883739), Tk, VMUL(LDK(KP974927912), Ti))));
Th = VFMA(LDK(KP623489801), Tg, VFNMS(LDK(KP900968867), Tb, VFNMS(LDK(KP222520933), T6, T1)));
ST(&(x[WS(rs, 5)]), VSUB(Th, Tl), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VADD(Th, Tl), ms, &(x[0]));
Tn = VBYI(VFMA(LDK(KP781831482), Ti, VFMA(LDK(KP974927912), Tk, VMUL(LDK(KP433883739), Tj))));
Tm = VFMA(LDK(KP623489801), T6, VFNMS(LDK(KP900968867), Tg, VFNMS(LDK(KP222520933), Tb, T1)));
ST(&(x[WS(rs, 6)]), VSUB(Tm, Tn), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VADD(Tm, Tn), ms, &(x[WS(rs, 1)]));
}
}
}
示例11: BEGIN_SIMD
static const R *q1fv_2(R *ri, R *ii, const R *W, stride is, stride vs, int m, int dist)
{
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = 0; i < m; i = i + VL, x = x + (VL * dist), W = W + (TWVL * 2)) {
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(is, 1)]), dist, &(x[WS(is, 1)]));
T3 = BYTWJ(&(W[0]), VSUB(T1, T2));
T4 = LD(&(x[WS(vs, 1)]), dist, &(x[WS(vs, 1)]));
T5 = LD(&(x[WS(vs, 1) + WS(is, 1)]), dist, &(x[WS(vs, 1) + WS(is, 1)]));
T6 = BYTWJ(&(W[0]), VSUB(T4, T5));
ST(&(x[WS(vs, 1)]), T3, dist, &(x[WS(vs, 1)]));
ST(&(x[WS(vs, 1) + WS(is, 1)]), T6, dist, &(x[WS(vs, 1) + WS(is, 1)]));
ST(&(x[0]), VADD(T1, T2), dist, &(x[0]));
ST(&(x[WS(is, 1)]), VADD(T4, T5), dist, &(x[WS(is, 1)]));
}
END_SIMD();
return W;
}
示例12: LD
static const R *t2fv_2(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(ios)) {
V T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[0]), T2);
ST(&(x[WS(ios, 1)]), VSUB(T1, T3), dist, &(x[WS(ios, 1)]));
ST(&(x[0]), VADD(T1, T3), dist, &(x[0]));
}
return W;
}
示例13: t1fuv_2
static void t1fuv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(2, rs)) {
V T1, T3, T2;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[0]), T2);
ST(&(x[WS(rs, 1)]), VSUB(T1, T3), ms, &(x[WS(rs, 1)]));
ST(&(x[0]), VADD(T1, T3), ms, &(x[0]));
}
}
VLEAVE();
}
示例14: BEGIN_SIMD
static const R *t1fv_2(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 2)) {
V T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[0]), T2);
ST(&(x[WS(ios, 1)]), VSUB(T1, T3), dist, &(x[WS(ios, 1)]));
ST(&(x[0]), VADD(T1, T3), dist, &(x[0]));
}
END_SIMD();
return W;
}
示例15: t1fuv_9
static void t1fuv_9(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP939692620, +0.939692620785908384054109277324731469936208134);
DVK(KP296198132, +0.296198132726023843175338011893050938967728390);
DVK(KP852868531, +0.852868531952443209628250963940074071936020296);
DVK(KP173648177, +0.173648177666930348851716626769314796000375677);
DVK(KP556670399, +0.556670399226419366452912952047023132968291906);
DVK(KP766044443, +0.766044443118978035202392650555416673935832457);
DVK(KP642787609, +0.642787609686539326322643409907263432907559884);
DVK(KP663413948, +0.663413948168938396205421319635891297216863310);
DVK(KP984807753, +0.984807753012208059366743024589523013670643252);
DVK(KP150383733, +0.150383733180435296639271897612501926072238258);
DVK(KP342020143, +0.342020143325668733044099614682259580763083368);
DVK(KP813797681, +0.813797681349373692844693217248393223289101568);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 16)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 16), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T6, TA, Tt, Tf, Ts, Tw, Tn, Tv;
T1 = LD(&(x[0]), ms, &(x[0]));
{
V T3, T5, T2, T4;
T2 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[TWVL * 4]), T2);
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 10]), T4);
T6 = VADD(T3, T5);
TA = VMUL(LDK(KP866025403), VSUB(T5, T3));
}
{
V T9, Td, Tb, T8, Tc, Ta, Te;
T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T9 = BYTWJ(&(W[0]), T8);
Tc = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[TWVL * 12]), Tc);
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Tt = VSUB(Td, Tb);
Te = VADD(Tb, Td);
Tf = VADD(T9, Te);
Ts = VFNMS(LDK(KP500000000), Te, T9);
}
{
V Th, Tl, Tj, Tg, Tk, Ti, Tm;
Tg = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Th = BYTWJ(&(W[TWVL * 2]), Tg);
Tk = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tl = BYTWJ(&(W[TWVL * 14]), Tk);
Ti = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tj = BYTWJ(&(W[TWVL * 8]), Ti);
Tw = VSUB(Tl, Tj);
Tm = VADD(Tj, Tl);
Tn = VADD(Th, Tm);
Tv = VFNMS(LDK(KP500000000), Tm, Th);
}
{
V Tq, T7, To, Tp;
Tq = VBYI(VMUL(LDK(KP866025403), VSUB(Tn, Tf)));
T7 = VADD(T1, T6);
To = VADD(Tf, Tn);
Tp = VFNMS(LDK(KP500000000), To, T7);
ST(&(x[0]), VADD(T7, To), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VADD(Tp, Tq), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 6)]), VSUB(Tp, Tq), ms, &(x[0]));
}
{
V TI, TB, TC, TD, Tu, Tx, Ty, Tr, TH;
TI = VBYI(VSUB(VFNMS(LDK(KP342020143), Tv, VFNMS(LDK(KP150383733), Tt, VFNMS(LDK(KP984807753), Ts, VMUL(LDK(KP813797681), Tw)))), TA));
TB = VFNMS(LDK(KP642787609), Ts, VMUL(LDK(KP663413948), Tt));
TC = VFNMS(LDK(KP984807753), Tv, VMUL(LDK(KP150383733), Tw));
TD = VADD(TB, TC);
Tu = VFMA(LDK(KP766044443), Ts, VMUL(LDK(KP556670399), Tt));
Tx = VFMA(LDK(KP173648177), Tv, VMUL(LDK(KP852868531), Tw));
Ty = VADD(Tu, Tx);
Tr = VFNMS(LDK(KP500000000), T6, T1);
TH = VFMA(LDK(KP173648177), Ts, VFNMS(LDK(KP296198132), Tw, VFNMS(LDK(KP939692620), Tv, VFNMS(LDK(KP852868531), Tt, Tr))));
ST(&(x[WS(rs, 7)]), VSUB(TH, TI), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VADD(TH, TI), ms, &(x[0]));
{
V Tz, TE, TF, TG;
Tz = VADD(Tr, Ty);
TE = VBYI(VADD(TA, TD));
ST(&(x[WS(rs, 8)]), VSUB(Tz, TE), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VADD(TE, Tz), ms, &(x[WS(rs, 1)]));
TF = VFMA(LDK(KP866025403), VSUB(TB, TC), VFNMS(LDK(KP500000000), Ty, Tr));
TG = VBYI(VADD(TA, VFNMS(LDK(KP500000000), TD, VMUL(LDK(KP866025403), VSUB(Tx, Tu)))));
ST(&(x[WS(rs, 5)]), VSUB(TF, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 4)]), VADD(TF, TG), ms, &(x[0]));
}
}
}
}
VLEAVE();
}