static void t1fuv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
V Td, TA, T4, Ta, Tk, TE, Tp, TF, TB, T9, T1, T2, Tb;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V Tg, Tn, Ti, Tl;
Tg = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tn = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tl = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
{
V T6, T8, T5, Tc;
T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
{
V T3, Th, To, Tj, Tm, T7;
T7 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[TWVL * 8]), T2);
Th = BYTWJ(&(W[TWVL * 6]), Tg);
To = BYTWJ(&(W[0]), Tn);
Tj = BYTWJ(&(W[TWVL * 16]), Ti);
Tm = BYTWJ(&(W[TWVL * 10]), Tl);
T6 = BYTWJ(&(W[TWVL * 2]), T5);
Td = BYTWJ(&(W[TWVL * 4]), Tc);
T8 = BYTWJ(&(W[TWVL * 12]), T7);
TA = VADD(T1, T3);
T4 = VSUB(T1, T3);
Ta = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tk = VSUB(Th, Tj);
TE = VADD(Th, Tj);
Tp = VSUB(Tm, To);
TF = VADD(Tm, To);
}
TB = VADD(T6, T8);
T9 = VSUB(T6, T8);
}
}
Tb = BYTWJ(&(W[TWVL * 14]), Ta);
{
V TL, TG, Tw, Tq, TC, Te;
TL = VSUB(TE, TF);
TG = VADD(TE, TF);
Tw = VSUB(Tk, Tp);
Tq = VADD(Tk, Tp);
TC = VADD(Tb, Td);
Te = VSUB(Tb, Td);
{
V TM, TD, Tv, Tf;
TM = VSUB(TB, TC);
TD = VADD(TB, TC);
Tv = VSUB(T9, Te);
Tf = VADD(T9, Te);
{
V TP, TN, TH, TJ, Tz, Tx, Tr, Tt, TI, Ts;
TP = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TL, TM));
TN = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TM, TL));
TH = VADD(TD, TG);
TJ = VSUB(TD, TG);
Tz = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tv, Tw));
Tx = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tw, Tv));
Tr = VADD(Tf, Tq);
Tt = VSUB(Tf, Tq);
ST(&(x[0]), VADD(TA, TH), ms, &(x[0]));
TI = VFNMS(LDK(KP250000000), TH, TA);
ST(&(x[WS(rs, 5)]), VADD(T4, Tr), ms, &(x[WS(rs, 1)]));
Ts = VFNMS(LDK(KP250000000), Tr, T4);
{
V TK, TO, Tu, Ty;
TK = VFNMS(LDK(KP559016994), TJ, TI);
TO = VFMA(LDK(KP559016994), TJ, TI);
Tu = VFMA(LDK(KP559016994), Tt, Ts);
Ty = VFNMS(LDK(KP559016994), Tt, Ts);
ST(&(x[WS(rs, 8)]), VFNMSI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFMAI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(Tz, Ty), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(Tz, Ty), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
VLEAVE();
}
static void n1fv_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(32, is), MAKE_VOLATILE_STRIDE(32, os)) {
V Tp, T13, Tu, TN, Tm, T14, Tv, TY, T7, T17, Ty, TT, Te, T16, Tx;
V TQ;
{
V Tn, To, TM, Ts, Tt, TL;
Tn = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
TM = VADD(Tn, To);
Ts = LD(&(xi[0]), ivs, &(xi[0]));
Tt = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
TL = VADD(Ts, Tt);
Tp = VSUB(Tn, To);
T13 = VADD(TL, TM);
Tu = VSUB(Ts, Tt);
TN = VSUB(TL, TM);
}
{
V Ti, TW, Tl, TX;
{
V Tg, Th, Tj, Tk;
Tg = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Ti = VSUB(Tg, Th);
TW = VADD(Tg, Th);
Tj = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tl = VSUB(Tj, Tk);
TX = VADD(Tj, Tk);
}
Tm = VMUL(LDK(KP707106781), VSUB(Ti, Tl));
T14 = VADD(TX, TW);
Tv = VMUL(LDK(KP707106781), VADD(Tl, Ti));
TY = VSUB(TW, TX);
}
{
V T3, TR, T6, TS;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
TR = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TS = VADD(T4, T5);
}
T7 = VFNMS(LDK(KP923879532), T6, VMUL(LDK(KP382683432), T3));
T17 = VADD(TR, TS);
Ty = VFMA(LDK(KP923879532), T3, VMUL(LDK(KP382683432), T6));
TT = VSUB(TR, TS);
}
{
V Ta, TO, Td, TP;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
TO = VADD(T8, T9);
Tb = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TP = VADD(Tb, Tc);
}
Te = VFMA(LDK(KP382683432), Ta, VMUL(LDK(KP923879532), Td));
T16 = VADD(TO, TP);
Tx = VFNMS(LDK(KP382683432), Td, VMUL(LDK(KP923879532), Ta));
TQ = VSUB(TO, TP);
}
{
V T15, T18, T19, T1a;
T15 = VADD(T13, T14);
T18 = VADD(T16, T17);
ST(&(xo[WS(os, 8)]), VSUB(T15, T18), ovs, &(xo[0]));
ST(&(xo[0]), VADD(T15, T18), ovs, &(xo[0]));
T19 = VSUB(T13, T14);
T1a = VBYI(VSUB(T17, T16));
ST(&(xo[WS(os, 12)]), VSUB(T19, T1a), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(T19, T1a), ovs, &(xo[0]));
}
{
V TV, T11, T10, T12, TU, TZ;
TU = VMUL(LDK(KP707106781), VADD(TQ, TT));
TV = VADD(TN, TU);
T11 = VSUB(TN, TU);
TZ = VMUL(LDK(KP707106781), VSUB(TT, TQ));
T10 = VBYI(VADD(TY, TZ));
T12 = VBYI(VSUB(TZ, TY));
ST(&(xo[WS(os, 14)]), VSUB(TV, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VADD(T11, T12), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VADD(TV, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VSUB(T11, T12), ovs, &(xo[0]));
}
{
V Tr, TB, TA, TC;
{
V Tf, Tq, Tw, Tz;
Tf = VSUB(T7, Te);
Tq = VSUB(Tm, Tp);
Tr = VBYI(VSUB(Tf, Tq));
TB = VBYI(VADD(Tq, Tf));
Tw = VADD(Tu, Tv);
Tz = VADD(Tx, Ty);
TA = VSUB(Tw, Tz);
TC = VADD(Tw, Tz);
}
ST(&(xo[WS(os, 7)]), VADD(Tr, TA), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 15)]), VSUB(TC, TB), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VSUB(TA, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VADD(TB, TC), ovs, &(xo[WS(os, 1)]));
}
{
V TF, TJ, TI, TK;
{
V TD, TE, TG, TH;
TD = VSUB(Tu, Tv);
TE = VADD(Te, T7);
TF = VADD(TD, TE);
TJ = VSUB(TD, TE);
TG = VADD(Tp, Tm);
TH = VSUB(Ty, Tx);
TI = VBYI(VADD(TG, TH));
TK = VBYI(VSUB(TH, TG));
}
ST(&(xo[WS(os, 13)]), VSUB(TF, TI), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VADD(TJ, TK), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VADD(TF, TI), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VSUB(TJ, TK), ovs, &(xo[WS(os, 1)]));
}
}
}
VLEAVE();
}
static void n1bv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
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 i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V Tp, Ty, Tl, TL, Tq, TE, T7, TJ, Ts, TB, Te, TK, Tr, TH, Tn;
V To;
Tn = LD(&(xi[0]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tp = VSUB(Tn, To);
Ty = VADD(Tn, To);
{
V Th, TC, Tk, TD;
{
V Tf, Tg, Ti, Tj;
Tf = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tg = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Th = VSUB(Tf, Tg);
TC = VADD(Tf, Tg);
Ti = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tj = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tk = VSUB(Ti, Tj);
TD = VADD(Ti, Tj);
}
Tl = VSUB(Th, Tk);
TL = VSUB(TD, TC);
Tq = VADD(Th, Tk);
TE = VADD(TC, TD);
}
{
V T3, Tz, T6, TA;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tz = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TA = VADD(T4, T5);
}
T7 = VSUB(T3, T6);
TJ = VSUB(Tz, TA);
Ts = VADD(T3, T6);
TB = VADD(Tz, TA);
}
{
V Ta, TF, Td, TG;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
TF = VADD(T8, T9);
Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TG = VADD(Tb, Tc);
}
Te = VSUB(Ta, Td);
TK = VSUB(TG, TF);
Tr = VADD(Ta, Td);
TH = VADD(TF, TG);
}
ST(&(xo[WS(os, 7)]), VADD(Tp, VADD(Ts, VADD(Tq, Tr))), ovs, &(xo[WS(os, 1)]));
ST(&(xo[0]), VADD(Ty, VADD(TB, VADD(TE, TH))), ovs, &(xo[0]));
{
V Tm, Tt, TQ, TP;
Tm = VBYI(VFMA(LDK(KP433883739), T7, VFNMS(LDK(KP781831482), Tl, VMUL(LDK(KP974927912), Te))));
Tt = VFMA(LDK(KP623489801), Tq, VFNMS(LDK(KP222520933), Tr, VFNMS(LDK(KP900968867), Ts, Tp)));
ST(&(xo[WS(os, 3)]), VADD(Tm, Tt), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VSUB(Tt, Tm), ovs, &(xo[WS(os, 1)]));
TQ = VBYI(VFMA(LDK(KP974927912), TJ, VFMA(LDK(KP433883739), TL, VMUL(LDK(KP781831482), TK))));
TP = VFMA(LDK(KP623489801), TH, VFNMS(LDK(KP900968867), TE, VFNMS(LDK(KP222520933), TB, Ty)));
ST(&(xo[WS(os, 12)]), VSUB(TP, TQ), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VADD(TP, TQ), ovs, &(xo[0]));
}
{
V Tu, Tv, TM, TI;
Tu = VBYI(VFMA(LDK(KP781831482), T7, VFMA(LDK(KP974927912), Tl, VMUL(LDK(KP433883739), Te))));
Tv = VFMA(LDK(KP623489801), Ts, VFNMS(LDK(KP900968867), Tr, VFNMS(LDK(KP222520933), Tq, Tp)));
ST(&(xo[WS(os, 1)]), VADD(Tu, Tv), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 13)]), VSUB(Tv, Tu), ovs, &(xo[WS(os, 1)]));
TM = VBYI(VFNMS(LDK(KP433883739), TK, VFNMS(LDK(KP974927912), TL, VMUL(LDK(KP781831482), TJ))));
TI = VFMA(LDK(KP623489801), TB, VFNMS(LDK(KP900968867), TH, VFNMS(LDK(KP222520933), TE, Ty)));
ST(&(xo[WS(os, 6)]), VSUB(TI, TM), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VADD(TI, TM), ovs, &(xo[0]));
}
{
V TO, TN, Tx, Tw;
TO = VBYI(VFMA(LDK(KP433883739), TJ, VFNMS(LDK(KP974927912), TK, VMUL(LDK(KP781831482), TL))));
TN = VFMA(LDK(KP623489801), TE, VFNMS(LDK(KP222520933), TH, VFNMS(LDK(KP900968867), TB, Ty)));
ST(&(xo[WS(os, 4)]), VSUB(TN, TO), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VADD(TN, TO), ovs, &(xo[0]));
Tx = VBYI(VFNMS(LDK(KP781831482), Te, VFNMS(LDK(KP433883739), Tl, VMUL(LDK(KP974927912), T7))));
Tw = VFMA(LDK(KP623489801), Tr, VFNMS(LDK(KP900968867), Tq, VFNMS(LDK(KP222520933), Ts, Tp)));
ST(&(xo[WS(os, 5)]), VSUB(Tw, Tx), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VADD(Tx, Tw), ovs, &(xo[WS(os, 1)]));
}
}
}
VLEAVE();
}
static void n1fv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V Ti, Ty, Tm, Tn, Tw, Tt, Tz, TA, TB, T7, Te, Tj, Tg, Th;
Tg = LD(&(xi[0]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Ti = VSUB(Tg, Th);
Ty = VADD(Tg, Th);
{
V T3, Tu, Td, Ts, T6, Tv, Ta, Tr;
{
V T1, T2, Tb, Tc;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tu = VADD(T1, T2);
Tb = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
Ts = VADD(Tb, Tc);
}
{
V T4, T5, T8, T9;
T4 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
Tv = VADD(T4, T5);
T8 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
Tr = VADD(T8, T9);
}
Tm = VSUB(T3, T6);
Tn = VSUB(Ta, Td);
Tw = VSUB(Tu, Tv);
Tt = VSUB(Tr, Ts);
Tz = VADD(Tu, Tv);
TA = VADD(Tr, Ts);
TB = VADD(Tz, TA);
T7 = VADD(T3, T6);
Te = VADD(Ta, Td);
Tj = VADD(T7, Te);
}
ST(&(xo[WS(os, 5)]), VADD(Ti, Tj), ovs, &(xo[WS(os, 1)]));
ST(&(xo[0]), VADD(Ty, TB), ovs, &(xo[0]));
{
V To, Tq, Tl, Tp, Tf, Tk;
To = VBYI(VFMA(LDK(KP951056516), Tm, VMUL(LDK(KP587785252), Tn)));
Tq = VBYI(VFNMS(LDK(KP587785252), Tm, VMUL(LDK(KP951056516), Tn)));
Tf = VMUL(LDK(KP559016994), VSUB(T7, Te));
Tk = VFNMS(LDK(KP250000000), Tj, Ti);
Tl = VADD(Tf, Tk);
Tp = VSUB(Tk, Tf);
ST(&(xo[WS(os, 1)]), VSUB(Tl, To), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VADD(Tq, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VADD(To, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VSUB(Tp, Tq), ovs, &(xo[WS(os, 1)]));
}
{
V Tx, TF, TE, TG, TC, TD;
Tx = VBYI(VFNMS(LDK(KP587785252), Tw, VMUL(LDK(KP951056516), Tt)));
TF = VBYI(VFMA(LDK(KP951056516), Tw, VMUL(LDK(KP587785252), Tt)));
TC = VFNMS(LDK(KP250000000), TB, Ty);
TD = VMUL(LDK(KP559016994), VSUB(Tz, TA));
TE = VSUB(TC, TD);
TG = VADD(TD, TC);
ST(&(xo[WS(os, 2)]), VADD(Tx, TE), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VSUB(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VSUB(TE, Tx), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(TF, TG), ovs, &(xo[0]));
}
}
}
static void t1bv_16(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 30)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(rs)) {
V TJ, T1b, TD, T1c, T17, T18, Ty, TK, T10, T11, T12, Tb, TM, T13, T14;
V T15, Tm, TN, TG, TI, TH;
TG = LD(&(x[0]), ms, &(x[0]));
TH = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
TI = BYTW(&(W[TWVL * 14]), TH);
TJ = VSUB(TG, TI);
T1b = VADD(TG, TI);
{
V TA, TC, Tz, TB;
Tz = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
TA = BYTW(&(W[TWVL * 6]), Tz);
TB = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TC = BYTW(&(W[TWVL * 22]), TB);
TD = VSUB(TA, TC);
T1c = VADD(TA, TC);
}
{
V Tp, Tw, Tr, Tu, Ts, Tx;
{
V To, Tv, Tq, Tt;
To = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tp = BYTW(&(W[TWVL * 2]), To);
Tv = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tw = BYTW(&(W[TWVL * 10]), Tv);
Tq = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
Tr = BYTW(&(W[TWVL * 18]), Tq);
Tt = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
Tu = BYTW(&(W[TWVL * 26]), Tt);
}
T17 = VADD(Tp, Tr);
T18 = VADD(Tu, Tw);
Ts = VSUB(Tp, Tr);
Tx = VSUB(Tu, Tw);
Ty = VMUL(LDK(KP707106781), VSUB(Ts, Tx));
TK = VMUL(LDK(KP707106781), VADD(Ts, Tx));
}
{
V T2, T9, T4, T7, T5, Ta;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T2 = BYTW(&(W[0]), T1);
T8 = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
T9 = BYTW(&(W[TWVL * 24]), T8);
T3 = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
T4 = BYTW(&(W[TWVL * 16]), T3);
T6 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T7 = BYTW(&(W[TWVL * 8]), T6);
}
T10 = VADD(T2, T4);
T11 = VADD(T7, T9);
T12 = VSUB(T10, T11);
T5 = VSUB(T2, T4);
Ta = VSUB(T7, T9);
Tb = VFNMS(LDK(KP382683432), Ta, VMUL(LDK(KP923879532), T5));
TM = VFMA(LDK(KP382683432), T5, VMUL(LDK(KP923879532), Ta));
}
{
V Td, Tk, Tf, Ti, Tg, Tl;
{
V Tc, Tj, Te, Th;
Tc = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
Td = BYTW(&(W[TWVL * 28]), Tc);
Tj = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tk = BYTW(&(W[TWVL * 20]), Tj);
Te = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Tf = BYTW(&(W[TWVL * 12]), Te);
Th = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Ti = BYTW(&(W[TWVL * 4]), Th);
}
T13 = VADD(Td, Tf);
T14 = VADD(Ti, Tk);
T15 = VSUB(T13, T14);
Tg = VSUB(Td, Tf);
Tl = VSUB(Ti, Tk);
Tm = VFMA(LDK(KP923879532), Tg, VMUL(LDK(KP382683432), Tl));
TN = VFNMS(LDK(KP382683432), Tg, VMUL(LDK(KP923879532), Tl));
}
{
V T1a, T1g, T1f, T1h;
{
V T16, T19, T1d, T1e;
T16 = VMUL(LDK(KP707106781), VSUB(T12, T15));
T19 = VSUB(T17, T18);
T1a = VBYI(VSUB(T16, T19));
T1g = VBYI(VADD(T19, T16));
T1d = VSUB(T1b, T1c);
T1e = VMUL(LDK(KP707106781), VADD(T12, T15));
T1f = VSUB(T1d, T1e);
T1h = VADD(T1d, T1e);
}
ST(&(x[WS(rs, 6)]), VADD(T1a, T1f), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VSUB(T1h, T1g), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VSUB(T1f, T1a), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(T1g, T1h), ms, &(x[0]));
}
{
V T1k, T1o, T1n, T1p;
{
V T1i, T1j, T1l, T1m;
T1i = VADD(T1b, T1c);
T1j = VADD(T17, T18);
T1k = VSUB(T1i, T1j);
T1o = VADD(T1i, T1j);
T1l = VADD(T10, T11);
T1m = VADD(T13, T14);
T1n = VBYI(VSUB(T1l, T1m));
T1p = VADD(T1l, T1m);
}
ST(&(x[WS(rs, 12)]), VSUB(T1k, T1n), ms, &(x[0]));
ST(&(x[0]), VADD(T1o, T1p), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VADD(T1k, T1n), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(T1o, T1p), ms, &(x[0]));
}
{
V TF, TQ, TP, TR;
{
V Tn, TE, TL, TO;
Tn = VSUB(Tb, Tm);
TE = VSUB(Ty, TD);
TF = VBYI(VSUB(Tn, TE));
TQ = VBYI(VADD(TE, Tn));
TL = VSUB(TJ, TK);
TO = VSUB(TM, TN);
TP = VSUB(TL, TO);
TR = VADD(TL, TO);
}
ST(&(x[WS(rs, 5)]), VADD(TF, TP), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VSUB(TR, TQ), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VSUB(TP, TF), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(TQ, TR), ms, &(x[WS(rs, 1)]));
}
{
V TU, TY, TX, TZ;
{
V TS, TT, TV, TW;
TS = VADD(TJ, TK);
TT = VADD(Tb, Tm);
TU = VADD(TS, TT);
TY = VSUB(TS, TT);
TV = VADD(TD, Ty);
TW = VADD(TM, TN);
TX = VBYI(VADD(TV, TW));
TZ = VBYI(VSUB(TW, TV));
}
ST(&(x[WS(rs, 15)]), VSUB(TU, TX), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(TY, TZ), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(TU, TX), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VSUB(TY, TZ), ms, &(x[WS(rs, 1)]));
}
}
}
VLEAVE();
}
static void hc2cbdftv_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 30)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(64, rs)) {
V Tf, T16, TZ, T1C, TI, T1a, TV, T1D, T1F, T1G, Ty, T19, TC, T17, TS;
V T10;
{
V T2, TD, T4, TF, Tc, Tb, Td, T6, T8, T9, T3, TE, Ta, T7, T5;
V Te, TX, TY, TG, TH, TT, TU, Tj, TM, Tw, TQ, Tn, TN, Ts, TP;
V Tg, Ti, Th, Tt, Tv, Tu, Tk, Tm, Tl, Tr, Tq, Tp, To, Tx, TA;
V TB, TO, TR;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
TD = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 7)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VCONJ(T3);
TE = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
TF = VCONJ(TE);
Tc = LD(&(Rp[WS(rs, 6)]), ms, &(Rp[0]));
Ta = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tb = VCONJ(Ta);
Td = VSUB(Tb, Tc);
T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T7 = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
T8 = VCONJ(T7);
T9 = VSUB(T6, T8);
T5 = VSUB(T2, T4);
Te = VMUL(LDK(KP707106781), VADD(T9, Td));
Tf = VADD(T5, Te);
T16 = VSUB(T5, Te);
TX = VADD(T2, T4);
TY = VADD(TD, TF);
TZ = VSUB(TX, TY);
T1C = VADD(TX, TY);
TG = VSUB(TD, TF);
TH = VMUL(LDK(KP707106781), VSUB(T9, Td));
TI = VADD(TG, TH);
T1a = VSUB(TH, TG);
TT = VADD(T6, T8);
TU = VADD(Tb, Tc);
TV = VSUB(TT, TU);
T1D = VADD(TT, TU);
Tg = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Th = LD(&(Rm[WS(rs, 6)]), -ms, &(Rm[0]));
Ti = VCONJ(Th);
Tj = VSUB(Tg, Ti);
TM = VADD(Tg, Ti);
Tt = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tu = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tv = VCONJ(Tu);
Tw = VSUB(Tt, Tv);
TQ = VADD(Tt, Tv);
Tk = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Tl = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Tm = VCONJ(Tl);
Tn = VSUB(Tk, Tm);
TN = VADD(Tk, Tm);
Tr = LD(&(Rp[WS(rs, 7)]), ms, &(Rp[WS(rs, 1)]));
Tp = LD(&(Rm[0]), -ms, &(Rm[0]));
Tq = VCONJ(Tp);
Ts = VSUB(Tq, Tr);
TP = VADD(Tq, Tr);
T1F = VADD(TM, TN);
T1G = VADD(TP, TQ);
To = VFNMS(LDK(KP382683432), Tn, VMUL(LDK(KP923879532), Tj));
Tx = VFMA(LDK(KP923879532), Ts, VMUL(LDK(KP382683432), Tw));
Ty = VADD(To, Tx);
T19 = VSUB(To, Tx);
TA = VFMA(LDK(KP382683432), Tj, VMUL(LDK(KP923879532), Tn));
TB = VFNMS(LDK(KP382683432), Ts, VMUL(LDK(KP923879532), Tw));
TC = VADD(TA, TB);
T17 = VSUB(TA, TB);
TO = VSUB(TM, TN);
TR = VSUB(TP, TQ);
TS = VMUL(LDK(KP707106781), VSUB(TO, TR));
T10 = VMUL(LDK(KP707106781), VADD(TO, TR));
}
{
V T21, T1W, T1u, T20, T1I, T1O, TK, T1S, T12, T1e, T1k, T1A, T1o, T1w, T1c;
V T1M, T1U, T1V, T1T, T1s, T1t, T1r, T1Z, T1E, T1H, T1B, T1N, Tz, TJ, T1;
V T1R, TW, T11, TL, T1d, T1i, T1j, T1h, T1z, T1m, T1n, T1l, T1v, T18, T1b;
V T15, T1L, T13, T1g, T1X, T23, T14, T1f, T1Y, T22, T1p, T1y, T1J, T1Q, T1q;
V T1x, T1K, T1P;
T1U = VADD(T1C, T1D);
T1V = VADD(T1F, T1G);
T21 = VADD(T1U, T1V);
T1T = LDW(&(W[TWVL * 14]));
T1W = VZMUL(T1T, VSUB(T1U, T1V));
T1s = VADD(Tf, Ty);
T1t = VBYI(VADD(TI, TC));
T1r = LDW(&(W[TWVL * 28]));
T1u = VZMULI(T1r, VSUB(T1s, T1t));
T1Z = LDW(&(W[0]));
T20 = VZMULI(T1Z, VADD(T1s, T1t));
T1E = VSUB(T1C, T1D);
T1H = VBYI(VSUB(T1F, T1G));
T1B = LDW(&(W[TWVL * 22]));
T1I = VZMUL(T1B, VSUB(T1E, T1H));
T1N = LDW(&(W[TWVL * 6]));
T1O = VZMUL(T1N, VADD(T1E, T1H));
Tz = VSUB(Tf, Ty);
TJ = VBYI(VSUB(TC, TI));
T1 = LDW(&(W[TWVL * 12]));
TK = VZMULI(T1, VADD(Tz, TJ));
T1R = LDW(&(W[TWVL * 16]));
T1S = VZMULI(T1R, VSUB(Tz, TJ));
TW = VBYI(VSUB(TS, TV));
T11 = VSUB(TZ, T10);
TL = LDW(&(W[TWVL * 10]));
T12 = VZMUL(TL, VADD(TW, T11));
T1d = LDW(&(W[TWVL * 18]));
T1e = VZMUL(T1d, VSUB(T11, TW));
T1i = VBYI(VADD(T1a, T19));
T1j = VADD(T16, T17);
T1h = LDW(&(W[TWVL * 4]));
T1k = VZMULI(T1h, VADD(T1i, T1j));
T1z = LDW(&(W[TWVL * 24]));
T1A = VZMULI(T1z, VSUB(T1j, T1i));
T1m = VBYI(VADD(TV, TS));
T1n = VADD(TZ, T10);
T1l = LDW(&(W[TWVL * 2]));
T1o = VZMUL(T1l, VADD(T1m, T1n));
T1v = LDW(&(W[TWVL * 26]));
T1w = VZMUL(T1v, VSUB(T1n, T1m));
T18 = VSUB(T16, T17);
T1b = VBYI(VSUB(T19, T1a));
T15 = LDW(&(W[TWVL * 20]));
T1c = VZMULI(T15, VSUB(T18, T1b));
T1L = LDW(&(W[TWVL * 8]));
T1M = VZMULI(T1L, VADD(T1b, T18));
T13 = VADD(TK, T12);
ST(&(Rp[WS(rs, 3)]), T13, ms, &(Rp[WS(rs, 1)]));
T1g = VCONJ(VSUB(T1e, T1c));
ST(&(Rm[WS(rs, 5)]), T1g, -ms, &(Rm[WS(rs, 1)]));
T1X = VADD(T1S, T1W);
ST(&(Rp[WS(rs, 4)]), T1X, ms, &(Rp[0]));
T23 = VCONJ(VSUB(T21, T20));
ST(&(Rm[0]), T23, -ms, &(Rm[0]));
T14 = VCONJ(VSUB(T12, TK));
ST(&(Rm[WS(rs, 3)]), T14, -ms, &(Rm[WS(rs, 1)]));
T1f = VADD(T1c, T1e);
ST(&(Rp[WS(rs, 5)]), T1f, ms, &(Rp[WS(rs, 1)]));
T1Y = VCONJ(VSUB(T1W, T1S));
ST(&(Rm[WS(rs, 4)]), T1Y, -ms, &(Rm[0]));
T22 = VADD(T20, T21);
ST(&(Rp[0]), T22, ms, &(Rp[0]));
T1p = VADD(T1k, T1o);
ST(&(Rp[WS(rs, 1)]), T1p, ms, &(Rp[WS(rs, 1)]));
T1y = VCONJ(VSUB(T1w, T1u));
ST(&(Rm[WS(rs, 7)]), T1y, -ms, &(Rm[WS(rs, 1)]));
T1J = VADD(T1A, T1I);
ST(&(Rp[WS(rs, 6)]), T1J, ms, &(Rp[0]));
T1Q = VCONJ(VSUB(T1O, T1M));
ST(&(Rm[WS(rs, 2)]), T1Q, -ms, &(Rm[0]));
T1q = VCONJ(VSUB(T1o, T1k));
ST(&(Rm[WS(rs, 1)]), T1q, -ms, &(Rm[WS(rs, 1)]));
T1x = VADD(T1u, T1w);
ST(&(Rp[WS(rs, 7)]), T1x, ms, &(Rp[WS(rs, 1)]));
T1K = VCONJ(VSUB(T1I, T1A));
ST(&(Rm[WS(rs, 6)]), T1K, -ms, &(Rm[0]));
T1P = VADD(T1M, T1O);
ST(&(Rp[WS(rs, 2)]), T1P, ms, &(Rp[0]));
}
}
}
VLEAVE();
}
static void n1fv_9(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP342020143, +0.342020143325668733044099614682259580763083368);
DVK(KP813797681, +0.813797681349373692844693217248393223289101568);
DVK(KP939692620, +0.939692620785908384054109277324731469936208134);
DVK(KP296198132, +0.296198132726023843175338011893050938967728390);
DVK(KP642787609, +0.642787609686539326322643409907263432907559884);
DVK(KP663413948, +0.663413948168938396205421319635891297216863310);
DVK(KP556670399, +0.556670399226419366452912952047023132968291906);
DVK(KP766044443, +0.766044443118978035202392650555416673935832457);
DVK(KP984807753, +0.984807753012208059366743024589523013670643252);
DVK(KP150383733, +0.150383733180435296639271897612501926072238258);
DVK(KP852868531, +0.852868531952443209628250963940074071936020296);
DVK(KP173648177, +0.173648177666930348851716626769314796000375677);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T5, Ts, Tj, To, Tf, Tn, Tp, Tu, Tl, Ta, Tk, Tm, Tt;
{
V T1, T2, T3, T4;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
T5 = VADD(T1, T4);
Ts = VMUL(LDK(KP866025403), VSUB(T3, T2));
Tj = VFNMS(LDK(KP500000000), T4, T1);
}
{
V Tb, Te, Tc, Td;
Tb = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Te = VADD(Tc, Td);
To = VSUB(Td, Tc);
Tf = VADD(Tb, Te);
Tn = VFNMS(LDK(KP500000000), Te, Tb);
Tp = VFMA(LDK(KP173648177), Tn, VMUL(LDK(KP852868531), To));
Tu = VFNMS(LDK(KP984807753), Tn, VMUL(LDK(KP150383733), To));
}
{
V T6, T9, T7, T8;
T6 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T9 = VADD(T7, T8);
Tl = VSUB(T8, T7);
Ta = VADD(T6, T9);
Tk = VFNMS(LDK(KP500000000), T9, T6);
Tm = VFMA(LDK(KP766044443), Tk, VMUL(LDK(KP556670399), Tl));
Tt = VFNMS(LDK(KP642787609), Tk, VMUL(LDK(KP663413948), Tl));
}
{
V Ti, Tg, Th, Tz, TA;
Ti = VBYI(VMUL(LDK(KP866025403), VSUB(Tf, Ta)));
Tg = VADD(Ta, Tf);
Th = VFNMS(LDK(KP500000000), Tg, T5);
ST(&(xo[0]), VADD(T5, Tg), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VADD(Th, Ti), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VSUB(Th, Ti), ovs, &(xo[0]));
Tz = VFMA(LDK(KP173648177), Tk, VFNMS(LDK(KP296198132), To, VFNMS(LDK(KP939692620), Tn, VFNMS(LDK(KP852868531), Tl, Tj))));
TA = VBYI(VSUB(VFNMS(LDK(KP342020143), Tn, VFNMS(LDK(KP150383733), Tl, VFNMS(LDK(KP984807753), Tk, VMUL(LDK(KP813797681), To)))), Ts));
ST(&(xo[WS(os, 7)]), VSUB(Tz, TA), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VADD(Tz, TA), ovs, &(xo[0]));
{
V Tr, Tx, Tw, Ty, Tq, Tv;
Tq = VADD(Tm, Tp);
Tr = VADD(Tj, Tq);
Tx = VFMA(LDK(KP866025403), VSUB(Tt, Tu), VFNMS(LDK(KP500000000), Tq, Tj));
Tv = VADD(Tt, Tu);
Tw = VBYI(VADD(Ts, Tv));
Ty = VBYI(VADD(Ts, VFNMS(LDK(KP500000000), Tv, VMUL(LDK(KP866025403), VSUB(Tp, Tm)))));
ST(&(xo[WS(os, 8)]), VSUB(Tr, Tw), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(Tx, Ty), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VADD(Tw, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VSUB(Tx, Ty), ovs, &(xo[WS(os, 1)]));
}
}
}
}
static void n1bv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T5, Ta, TG, TF, Ty, Tm, Ti, Tp, TJ, TI, Tx, Ts;
{
V T1, T6, T4, Tk, T9, Tl;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T2, T3, T7, T8;
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Tk = VSUB(T2, T3);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Tl = VSUB(T7, T8);
}
T5 = VFNMS(LDK(KP500000000), T4, T1);
Ta = VFNMS(LDK(KP500000000), T9, T6);
TG = VADD(T6, T9);
TF = VADD(T1, T4);
Ty = VADD(Tk, Tl);
Tm = VMUL(LDK(KP866025403), VSUB(Tk, Tl));
}
{
V Tn, Tq, Te, To, Th, Tr;
Tn = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tq = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Tc, Td, Tf, Tg;
Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Te = VSUB(Tc, Td);
To = VADD(Tc, Td);
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tg = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Th = VSUB(Tf, Tg);
Tr = VADD(Tf, Tg);
}
Ti = VMUL(LDK(KP866025403), VSUB(Te, Th));
Tp = VFNMS(LDK(KP500000000), To, Tn);
TJ = VADD(Tq, Tr);
TI = VADD(Tn, To);
Tx = VADD(Te, Th);
Ts = VFNMS(LDK(KP500000000), Tr, Tq);
}
{
V TH, TK, TL, TM;
TH = VSUB(TF, TG);
TK = VBYI(VSUB(TI, TJ));
ST(&(xo[WS(os, 3)]), VSUB(TH, TK), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VADD(TH, TK), ovs, &(xo[WS(os, 1)]));
TL = VADD(TF, TG);
TM = VADD(TI, TJ);
ST(&(xo[WS(os, 6)]), VSUB(TL, TM), ovs, &(xo[0]));
ST(&(xo[0]), VADD(TL, TM), ovs, &(xo[0]));
}
{
V Tj, Tv, Tu, Tw, Tb, Tt;
Tb = VSUB(T5, Ta);
Tj = VSUB(Tb, Ti);
Tv = VADD(Tb, Ti);
Tt = VSUB(Tp, Ts);
Tu = VBYI(VADD(Tm, Tt));
Tw = VBYI(VSUB(Tt, Tm));
ST(&(xo[WS(os, 11)]), VSUB(Tj, Tu), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VADD(Tv, Tw), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VADD(Tj, Tu), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VSUB(Tv, Tw), ovs, &(xo[WS(os, 1)]));
}
{
V Tz, TD, TC, TE, TA, TB;
Tz = VBYI(VMUL(LDK(KP866025403), VSUB(Tx, Ty)));
TD = VBYI(VMUL(LDK(KP866025403), VADD(Ty, Tx)));
TA = VADD(T5, Ta);
TB = VADD(Tp, Ts);
TC = VSUB(TA, TB);
TE = VADD(TA, TB);
ST(&(xo[WS(os, 2)]), VADD(Tz, TC), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VSUB(TE, TD), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VSUB(TC, Tz), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(TD, TE), ovs, &(xo[0]));
}
}
}
static void n1bv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T1, T6, Tc, Th, Td, Te, Ti, Tz, T4, TA, T9, Tj, Tf, Tw;
{
V T2, T3, T7, T8;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Th = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Ti = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tz = VSUB(T2, T3);
T4 = VADD(T2, T3);
TA = VSUB(T7, T8);
T9 = VADD(T7, T8);
Tj = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
}
Tf = VADD(Td, Te);
Tw = VSUB(Td, Te);
{
V T5, Tp, TJ, TB, Ta, Tq, Tk, Tx, Tg, Ts;
T5 = VADD(T1, T4);
Tp = VFNMS(LDK(KP500000000), T4, T1);
TJ = VSUB(Tz, TA);
TB = VADD(Tz, TA);
Ta = VADD(T6, T9);
Tq = VFNMS(LDK(KP500000000), T9, T6);
Tk = VADD(Ti, Tj);
Tx = VSUB(Tj, Ti);
Tg = VADD(Tc, Tf);
Ts = VFNMS(LDK(KP500000000), Tf, Tc);
{
V Tr, TF, Tb, Tn, TG, Ty, Tl, Tt;
Tr = VADD(Tp, Tq);
TF = VSUB(Tp, Tq);
Tb = VSUB(T5, Ta);
Tn = VADD(T5, Ta);
TG = VADD(Tw, Tx);
Ty = VSUB(Tw, Tx);
Tl = VADD(Th, Tk);
Tt = VFNMS(LDK(KP500000000), Tk, Th);
{
V TC, TE, TH, TL, Tu, TI, Tm, To;
TC = VMUL(LDK(KP866025403), VSUB(Ty, TB));
TE = VMUL(LDK(KP866025403), VADD(TB, Ty));
TH = VFNMS(LDK(KP866025403), TG, TF);
TL = VFMA(LDK(KP866025403), TG, TF);
Tu = VADD(Ts, Tt);
TI = VSUB(Ts, Tt);
Tm = VSUB(Tg, Tl);
To = VADD(Tg, Tl);
{
V TK, TM, Tv, TD;
TK = VFMA(LDK(KP866025403), TJ, TI);
TM = VFNMS(LDK(KP866025403), TJ, TI);
Tv = VSUB(Tr, Tu);
TD = VADD(Tr, Tu);
ST(&(xo[0]), VADD(Tn, To), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VSUB(Tn, To), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VFMAI(Tm, Tb), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VFNMSI(Tm, Tb), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VFMAI(TM, TL), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VFNMSI(TM, TL), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VFNMSI(TK, TH), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VFMAI(TK, TH), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VFNMSI(TE, TD), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VFMAI(TE, TD), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VFMAI(TC, Tv), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VFNMSI(TC, Tv), ovs, &(xo[0]));
}
}
}
}
}
}
static void t2bv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 14)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V Tl, Tq, Tg, Tr, T5, Tt, Ta, Tu, Ti, Tk, Tj;
Ti = LD(&(x[0]), ms, &(x[0]));
Tj = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tk = BYTW(&(W[TWVL * 6]), Tj);
Tl = VSUB(Ti, Tk);
Tq = VADD(Ti, Tk);
{
V Td, Tf, Tc, Te;
Tc = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Td = BYTW(&(W[TWVL * 2]), Tc);
Te = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tf = BYTW(&(W[TWVL * 10]), Te);
Tg = VSUB(Td, Tf);
Tr = VADD(Td, Tf);
}
{
V T2, T4, T1, T3;
T1 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T2 = BYTW(&(W[0]), T1);
T3 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T4 = BYTW(&(W[TWVL * 8]), T3);
T5 = VSUB(T2, T4);
Tt = VADD(T2, T4);
}
{
V T7, T9, T6, T8;
T6 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T7 = BYTW(&(W[TWVL * 12]), T6);
T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = BYTW(&(W[TWVL * 4]), T8);
Ta = VSUB(T7, T9);
Tu = VADD(T7, T9);
}
{
V Ts, Tv, Tw, Tx;
Ts = VSUB(Tq, Tr);
Tv = VBYI(VSUB(Tt, Tu));
ST(&(x[WS(rs, 6)]), VSUB(Ts, Tv), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(Ts, Tv), ms, &(x[0]));
Tw = VADD(Tq, Tr);
Tx = VADD(Tt, Tu);
ST(&(x[WS(rs, 4)]), VSUB(Tw, Tx), ms, &(x[0]));
ST(&(x[0]), VADD(Tw, Tx), ms, &(x[0]));
{
V Th, To, Tn, Tp, Tb, Tm;
Tb = VMUL(LDK(KP707106781), VSUB(T5, Ta));
Th = VBYI(VSUB(Tb, Tg));
To = VBYI(VADD(Tg, Tb));
Tm = VMUL(LDK(KP707106781), VADD(T5, Ta));
Tn = VSUB(Tl, Tm);
Tp = VADD(Tl, Tm);
ST(&(x[WS(rs, 3)]), VADD(Th, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VSUB(Tp, To), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VSUB(Tn, Th), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(To, Tp), ms, &(x[WS(rs, 1)]));
}
}
}
}
static void t2bv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 14)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, Th, Tj, T5, T7, Ta, Tc;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Th = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tj = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T5 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Ta = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
{
V T3, Ti, Tk, T6, T8, Tb, Td;
T3 = BYTW(&(W[TWVL * 6]), T2);
Ti = BYTW(&(W[TWVL * 2]), Th);
Tk = BYTW(&(W[TWVL * 10]), Tj);
T6 = BYTW(&(W[0]), T5);
T8 = BYTW(&(W[TWVL * 8]), T7);
Tb = BYTW(&(W[TWVL * 12]), Ta);
Td = BYTW(&(W[TWVL * 4]), Tc);
{
V Tq, T4, Tr, Tl, Tt, T9, Tu, Te, Tw, Ts;
Tq = VADD(T1, T3);
T4 = VSUB(T1, T3);
Tr = VADD(Ti, Tk);
Tl = VSUB(Ti, Tk);
Tt = VADD(T6, T8);
T9 = VSUB(T6, T8);
Tu = VADD(Tb, Td);
Te = VSUB(Tb, Td);
Tw = VADD(Tq, Tr);
Ts = VSUB(Tq, Tr);
{
V Tx, Tv, Tm, Tf;
Tx = VADD(Tt, Tu);
Tv = VSUB(Tt, Tu);
Tm = VSUB(T9, Te);
Tf = VADD(T9, Te);
{
V Tp, Tn, To, Tg;
ST(&(x[0]), VADD(Tw, Tx), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VSUB(Tw, Tx), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(Tv, Ts), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(Tv, Ts), ms, &(x[0]));
Tp = VFMA(LDK(KP707106781), Tm, Tl);
Tn = VFNMS(LDK(KP707106781), Tm, Tl);
To = VFMA(LDK(KP707106781), Tf, T4);
Tg = VFNMS(LDK(KP707106781), Tf, T4);
ST(&(x[WS(rs, 1)]), VFMAI(Tp, To), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(Tp, To), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFMAI(Tn, Tg), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(Tn, Tg), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
static void t1fuv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 14)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V T4, Tq, Tm, Tr, T9, Tt, Te, Tu, T1, T3, T2;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T3 = BYTWJ(&(W[TWVL * 6]), T2);
T4 = VSUB(T1, T3);
Tq = VADD(T1, T3);
{
V Tj, Tl, Ti, Tk;
Ti = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tj = BYTWJ(&(W[TWVL * 2]), Ti);
Tk = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tl = BYTWJ(&(W[TWVL * 10]), Tk);
Tm = VSUB(Tj, Tl);
Tr = VADD(Tj, Tl);
}
{
V T6, T8, T5, T7;
T5 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T6 = BYTWJ(&(W[0]), T5);
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T8 = BYTWJ(&(W[TWVL * 8]), T7);
T9 = VSUB(T6, T8);
Tt = VADD(T6, T8);
}
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Tb = BYTWJ(&(W[TWVL * 12]), Ta);
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[TWVL * 4]), Tc);
Te = VSUB(Tb, Td);
Tu = VADD(Tb, Td);
}
{
V Ts, Tv, Tw, Tx;
Ts = VADD(Tq, Tr);
Tv = VADD(Tt, Tu);
ST(&(x[WS(rs, 4)]), VSUB(Ts, Tv), ms, &(x[0]));
ST(&(x[0]), VADD(Ts, Tv), ms, &(x[0]));
Tw = VSUB(Tq, Tr);
Tx = VBYI(VSUB(Tu, Tt));
ST(&(x[WS(rs, 6)]), VSUB(Tw, Tx), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(Tw, Tx), ms, &(x[0]));
{
V Tg, To, Tn, Tp, Tf, Th;
Tf = VMUL(LDK(KP707106781), VADD(T9, Te));
Tg = VADD(T4, Tf);
To = VSUB(T4, Tf);
Th = VMUL(LDK(KP707106781), VSUB(Te, T9));
Tn = VBYI(VSUB(Th, Tm));
Tp = VBYI(VADD(Tm, Th));
ST(&(x[WS(rs, 7)]), VSUB(Tg, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(To, Tp), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(Tg, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VSUB(To, Tp), ms, &(x[WS(rs, 1)]));
}
}
}
}
static const R *t3fv_16(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(ios)) {
V T13, Tg, TY, T14, T1A, T1q, T1f, T1x, T1r, T1i, Tt, T16, TB, T1j, T1k;
V TH;
{
V T2, T8, Tu, T3;
T2 = LDW(&(W[0]));
T8 = LDW(&(W[TWVL * 2]));
Tu = LDW(&(W[TWVL * 6]));
T3 = LDW(&(W[TWVL * 4]));
{
V Ty, T1o, Tf, T1b, T7, Tr, TR, TX, T1g, Tl, To, Tw, TG, Tz, T1p;
V T1e, TC;
{
V T1, T5, Ta, Td;
T1 = LD(&(x[0]), dist, &(x[0]));
T5 = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
Ta = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
Td = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
{
V Tx, TO, TE, Tb, Tm, Tp, TN, Te, T6, TW, TP, TS;
{
V TM, T9, TL, Tc, TU, T4, TV;
TM = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
Tx = VZMULJ(T2, T8);
T9 = VZMUL(T2, T8);
TL = VZMULJ(T2, Tu);
TO = VZMULJ(T8, T3);
Tc = VZMUL(T8, T3);
TU = VZMUL(T2, T3);
T4 = VZMULJ(T2, T3);
TV = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
TE = VZMUL(Tx, T3);
Ty = VZMULJ(Tx, T3);
Tb = VZMULJ(T9, Ta);
Tm = VZMULJ(T9, T3);
Tp = VZMUL(T9, T3);
TN = VZMULJ(TL, TM);
Te = VZMULJ(Tc, Td);
T6 = VZMULJ(T4, T5);
TW = VZMULJ(TU, TV);
}
TP = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
TS = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
{
V TQ, TT, Ti, Tk, Tn, Th, Tq, Tj;
Th = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Tq = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
Tj = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
T1o = VSUB(Tb, Te);
Tf = VADD(Tb, Te);
T1b = VSUB(T1, T6);
T7 = VADD(T1, T6);
TQ = VZMULJ(TO, TP);
TT = VZMULJ(Tx, TS);
Ti = VZMULJ(T2, Th);
Tr = VZMULJ(Tp, Tq);
Tk = VZMULJ(T3, Tj);
Tn = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
{
V T1d, T1c, Tv, TF;
Tv = LD(&(x[WS(ios, 15)]), dist, &(x[WS(ios, 1)]));
TF = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
T1d = VSUB(TN, TQ);
TR = VADD(TN, TQ);
T1c = VSUB(TT, TW);
TX = VADD(TT, TW);
T1g = VSUB(Ti, Tk);
Tl = VADD(Ti, Tk);
To = VZMULJ(Tm, Tn);
Tw = VZMULJ(Tu, Tv);
TG = VZMULJ(TE, TF);
Tz = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
T1p = VSUB(T1d, T1c);
T1e = VADD(T1c, T1d);
TC = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
}
}
}
}
{
V T1h, Ts, TA, TD;
T13 = VADD(T7, Tf);
Tg = VSUB(T7, Tf);
T1h = VSUB(To, Tr);
Ts = VADD(To, Tr);
TY = VSUB(TR, TX);
T14 = VADD(TX, TR);
TA = VZMULJ(Ty, Tz);
T1A = VFMA(LDK(KP707106781), T1p, T1o);
T1q = VFNMS(LDK(KP707106781), T1p, T1o);
T1f = VFMA(LDK(KP707106781), T1e, T1b);
T1x = VFNMS(LDK(KP707106781), T1e, T1b);
TD = VZMULJ(T8, TC);
T1r = VFMA(LDK(KP414213562), T1g, T1h);
T1i = VFNMS(LDK(KP414213562), T1h, T1g);
Tt = VSUB(Tl, Ts);
T16 = VADD(Tl, Ts);
TB = VADD(Tw, TA);
T1j = VSUB(Tw, TA);
T1k = VSUB(TG, TD);
TH = VADD(TD, TG);
}
}
}
{
V T15, T19, T1l, T1s, TI, T17;
T15 = VADD(T13, T14);
T19 = VSUB(T13, T14);
T1l = VFNMS(LDK(KP414213562), T1k, T1j);
T1s = VFMA(LDK(KP414213562), T1j, T1k);
TI = VSUB(TB, TH);
T17 = VADD(TB, TH);
{
V T1y, T1t, T1B, T1m;
T1y = VADD(T1r, T1s);
T1t = VSUB(T1r, T1s);
T1B = VSUB(T1l, T1i);
T1m = VADD(T1i, T1l);
{
V T18, T1a, TJ, TZ;
T18 = VADD(T16, T17);
T1a = VSUB(T17, T16);
TJ = VADD(Tt, TI);
TZ = VSUB(TI, Tt);
{
V T1u, T1w, T1z, T1D;
T1u = VFNMS(LDK(KP923879532), T1t, T1q);
T1w = VFMA(LDK(KP923879532), T1t, T1q);
T1z = VFNMS(LDK(KP923879532), T1y, T1x);
T1D = VFMA(LDK(KP923879532), T1y, T1x);
{
V T1n, T1v, T1C, T1E;
T1n = VFNMS(LDK(KP923879532), T1m, T1f);
T1v = VFMA(LDK(KP923879532), T1m, T1f);
T1C = VFNMS(LDK(KP923879532), T1B, T1A);
T1E = VFMA(LDK(KP923879532), T1B, T1A);
ST(&(x[WS(ios, 12)]), VFNMSI(T1a, T19), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VFMAI(T1a, T19), dist, &(x[0]));
ST(&(x[0]), VADD(T15, T18), dist, &(x[0]));
ST(&(x[WS(ios, 8)]), VSUB(T15, T18), dist, &(x[0]));
{
V T10, T12, TK, T11;
T10 = VFNMS(LDK(KP707106781), TZ, TY);
T12 = VFMA(LDK(KP707106781), TZ, TY);
TK = VFNMS(LDK(KP707106781), TJ, Tg);
T11 = VFMA(LDK(KP707106781), TJ, Tg);
ST(&(x[WS(ios, 1)]), VFNMSI(T1w, T1v), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 15)]), VFMAI(T1w, T1v), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VFMAI(T1u, T1n), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 9)]), VFNMSI(T1u, T1n), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 3)]), VFMAI(T1E, T1D), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 13)]), VFNMSI(T1E, T1D), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 11)]), VFMAI(T1C, T1z), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VFNMSI(T1C, T1z), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 14)]), VFNMSI(T12, T11), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VFMAI(T12, T11), dist, &(x[0]));
ST(&(x[WS(ios, 10)]), VFMAI(T10, TK), dist, &(x[0]));
ST(&(x[WS(ios, 6)]), VFNMSI(T10, TK), dist, &(x[0]));
}
}
}
}
}
}
}
return W;
}
static const R *t3fv_16(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(ios)) {
V T4, T5, T6, To, T1, Ty, T7, T8, TO, TV, Te, Tp, TB, TH, Ts;
T4 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 2]));
T6 = VZMULJ(T4, T5);
To = VZMUL(T4, T5);
T1 = LDW(&(W[TWVL * 6]));
Ty = VZMULJ(T4, T1);
T7 = LDW(&(W[TWVL * 4]));
T8 = VZMULJ(T6, T7);
TO = VZMUL(T5, T7);
TV = VZMULJ(T4, T7);
Te = VZMUL(T6, T7);
Tp = VZMULJ(To, T7);
TB = VZMULJ(T5, T7);
TH = VZMUL(T4, T7);
Ts = VZMUL(To, T7);
{
V TY, T1f, TR, T1g, T1q, T1r, TL, TZ, T1l, T1m, T1n, Ti, T12, T1i, T1j;
V T1k, Tw, T11, TU, TX, TW;
TU = LD(&(x[0]), dist, &(x[0]));
TW = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
TX = VZMULJ(TV, TW);
TY = VSUB(TU, TX);
T1f = VADD(TU, TX);
{
V TN, TQ, TM, TP;
TM = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
TN = VZMULJ(To, TM);
TP = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
TQ = VZMULJ(TO, TP);
TR = VSUB(TN, TQ);
T1g = VADD(TN, TQ);
}
{
V TA, TJ, TD, TG, TE, TK;
{
V Tz, TI, TC, TF;
Tz = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
TA = VZMULJ(Ty, Tz);
TI = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
TJ = VZMULJ(TH, TI);
TC = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
TD = VZMULJ(TB, TC);
TF = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
TG = VZMULJ(T6, TF);
}
T1q = VADD(TA, TD);
T1r = VADD(TG, TJ);
TE = VSUB(TA, TD);
TK = VSUB(TG, TJ);
TL = VMUL(LDK(KP707106781), VSUB(TE, TK));
TZ = VMUL(LDK(KP707106781), VADD(TK, TE));
}
{
V T3, Tg, Ta, Td, Tb, Th;
{
V T2, Tf, T9, Tc;
T2 = LD(&(x[WS(ios, 15)]), dist, &(x[WS(ios, 1)]));
T3 = VZMULJ(T1, T2);
Tf = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
Tg = VZMULJ(Te, Tf);
T9 = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
Ta = VZMULJ(T8, T9);
Tc = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Td = VZMULJ(T5, Tc);
}
T1l = VADD(T3, Ta);
T1m = VADD(Td, Tg);
T1n = VSUB(T1l, T1m);
Tb = VSUB(T3, Ta);
Th = VSUB(Td, Tg);
Ti = VFNMS(LDK(KP923879532), Th, VMUL(LDK(KP382683432), Tb));
T12 = VFMA(LDK(KP923879532), Tb, VMUL(LDK(KP382683432), Th));
}
{
V Tk, Tu, Tm, Tr, Tn, Tv;
{
V Tj, Tt, Tl, Tq;
Tj = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Tk = VZMULJ(T4, Tj);
Tt = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
Tu = VZMULJ(Ts, Tt);
Tl = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
Tm = VZMULJ(T7, Tl);
Tq = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
Tr = VZMULJ(Tp, Tq);
}
T1i = VADD(Tk, Tm);
T1j = VADD(Tr, Tu);
T1k = VSUB(T1i, T1j);
Tn = VSUB(Tk, Tm);
Tv = VSUB(Tr, Tu);
Tw = VFMA(LDK(KP382683432), Tn, VMUL(LDK(KP923879532), Tv));
T11 = VFNMS(LDK(KP382683432), Tv, VMUL(LDK(KP923879532), Tn));
}
{
V T1p, T1v, T1u, T1w;
{
V T1h, T1o, T1s, T1t;
T1h = VSUB(T1f, T1g);
T1o = VMUL(LDK(KP707106781), VADD(T1k, T1n));
T1p = VADD(T1h, T1o);
T1v = VSUB(T1h, T1o);
T1s = VSUB(T1q, T1r);
T1t = VMUL(LDK(KP707106781), VSUB(T1n, T1k));
T1u = VBYI(VADD(T1s, T1t));
T1w = VBYI(VSUB(T1t, T1s));
}
ST(&(x[WS(ios, 14)]), VSUB(T1p, T1u), dist, &(x[0]));
ST(&(x[WS(ios, 6)]), VADD(T1v, T1w), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VADD(T1p, T1u), dist, &(x[0]));
ST(&(x[WS(ios, 10)]), VSUB(T1v, T1w), dist, &(x[0]));
}
{
V T1z, T1D, T1C, T1E;
{
V T1x, T1y, T1A, T1B;
T1x = VADD(T1f, T1g);
T1y = VADD(T1r, T1q);
T1z = VADD(T1x, T1y);
T1D = VSUB(T1x, T1y);
T1A = VADD(T1i, T1j);
T1B = VADD(T1l, T1m);
T1C = VADD(T1A, T1B);
T1E = VBYI(VSUB(T1B, T1A));
}
ST(&(x[WS(ios, 8)]), VSUB(T1z, T1C), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(T1D, T1E), dist, &(x[0]));
ST(&(x[0]), VADD(T1z, T1C), dist, &(x[0]));
ST(&(x[WS(ios, 12)]), VSUB(T1D, T1E), dist, &(x[0]));
}
{
V TT, T15, T14, T16;
{
V Tx, TS, T10, T13;
Tx = VSUB(Ti, Tw);
TS = VSUB(TL, TR);
TT = VBYI(VSUB(Tx, TS));
T15 = VBYI(VADD(TS, Tx));
T10 = VADD(TY, TZ);
T13 = VADD(T11, T12);
T14 = VSUB(T10, T13);
T16 = VADD(T10, T13);
}
ST(&(x[WS(ios, 7)]), VADD(TT, T14), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 15)]), VSUB(T16, T15), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 9)]), VSUB(T14, TT), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(T15, T16), dist, &(x[WS(ios, 1)]));
}
{
V T19, T1d, T1c, T1e;
{
V T17, T18, T1a, T1b;
T17 = VSUB(TY, TZ);
T18 = VADD(Tw, Ti);
T19 = VADD(T17, T18);
T1d = VSUB(T17, T18);
T1a = VADD(TR, TL);
T1b = VSUB(T12, T11);
T1c = VBYI(VADD(T1a, T1b));
T1e = VBYI(VSUB(T1b, T1a));
}
ST(&(x[WS(ios, 13)]), VSUB(T19, T1c), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VADD(T1d, T1e), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 3)]), VADD(T19, T1c), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 11)]), VSUB(T1d, T1e), dist, &(x[WS(ios, 1)]));
}
}
}
return W;
}
static void t3fv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
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(rs)) {
V T1, T7, Th, Tx, Tr, Td, Tp, T6, Tv, Tc, Te, Ti, Tl, T2, T3;
V T5;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
T5 = LDW(&(W[TWVL * 4]));
T1 = LD(&(x[0]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V To, Tw, Tq, Tu, Ta, T4, Tt, Tk, Tb;
To = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tw = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tq = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tu = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Ta = VZMULJ(T2, T3);
T4 = VZMUL(T2, T3);
Th = VZMULJ(T2, T5);
Tt = VZMULJ(T3, T5);
Tb = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tx = VZMULJ(T2, Tw);
Tr = VZMULJ(T5, Tq);
Tk = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = VZMULJ(Ta, T5);
Tp = VZMULJ(T4, To);
T6 = VZMULJ(T4, T5);
Tv = VZMULJ(Tt, Tu);
Tc = VZMULJ(Ta, Tb);
Te = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tl = VZMULJ(T3, Tk);
}
{
V TN, Ts, T8, Ty, TO, Tf, Tj;
TN = VADD(Tp, Tr);
Ts = VSUB(Tp, Tr);
T8 = VZMULJ(T6, T7);
Ty = VSUB(Tv, Tx);
TO = VADD(Tv, Tx);
Tf = VZMULJ(Td, Te);
Tj = VZMULJ(Th, Ti);
{
V T9, TJ, TP, TU, Tz, TF, Tg, TK, Tm, TL;
T9 = VSUB(T1, T8);
TJ = VADD(T1, T8);
TP = VADD(TN, TO);
TU = VSUB(TN, TO);
Tz = VADD(Ts, Ty);
TF = VSUB(Ts, Ty);
Tg = VSUB(Tc, Tf);
TK = VADD(Tc, Tf);
Tm = VSUB(Tj, Tl);
TL = VADD(Tj, Tl);
{
V TM, TV, Tn, TE;
TM = VADD(TK, TL);
TV = VSUB(TK, TL);
Tn = VADD(Tg, Tm);
TE = VSUB(Tg, Tm);
{
V TW, TY, TS, TQ, TG, TI, TC, TA, TR, TB;
TW = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TV, TU));
TY = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TU, TV));
TS = VSUB(TM, TP);
TQ = VADD(TM, TP);
TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TF, TE));
TI = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TE, TF));
TC = VSUB(Tn, Tz);
TA = VADD(Tn, Tz);
ST(&(x[0]), VADD(TJ, TQ), ms, &(x[0]));
TR = VFNMS(LDK(KP250000000), TQ, TJ);
ST(&(x[WS(rs, 5)]), VADD(T9, TA), ms, &(x[WS(rs, 1)]));
TB = VFNMS(LDK(KP250000000), TA, T9);
{
V TX, TT, TH, TD;
TX = VFMA(LDK(KP559016994), TS, TR);
TT = VFNMS(LDK(KP559016994), TS, TR);
TH = VFNMS(LDK(KP559016994), TC, TB);
TD = VFMA(LDK(KP559016994), TC, TB);
ST(&(x[WS(rs, 8)]), VFNMSI(TW, TT), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TW, TT), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TY, TX), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TY, TX), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFMAI(TG, TD), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(TG, TD), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(TI, TH), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(TI, TH), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
}
static void m1fv_32_0(const R *xi, R *xo, stride is, stride os, int ivs, int ovs)
{
DVK(KP195090322, +0.195090322016128267848284868477022240927691618);
DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
DVK(KP555570233, +0.555570233019602224742830813948532874374937191);
DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
V T1T, T1W, T2K, T2x, T16, T1A, Tb, T1p, TT, T1v, TY, T1w, T27, T2a, T2b;
V T2H, T2O, TC, T1s, TH, T1t, T20, T23, T24, T2E, T2N, T2g, T2j, Tq, T1B;
V T19, T1q, T2A, T2L;
{
V T3, T1R, T15, T1S, T6, T1U, T9, T1V, T12, Ta;
{
V T1, T2, T13, T14;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
T1R = VADD(T1, T2);
T13 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T14 = LD(&(xi[WS(is, 24)]), ivs, &(xi[0]));
T15 = VSUB(T13, T14);
T1S = VADD(T13, T14);
}
{
V T4, T5, T7, T8;
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 20)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
T1U = VADD(T4, T5);
T7 = LD(&(xi[WS(is, 28)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
T1V = VADD(T7, T8);
}
T1T = VADD(T1R, T1S);
T1W = VADD(T1U, T1V);
T2K = VSUB(T1V, T1U);
T2x = VSUB(T1R, T1S);
T12 = VMUL(LDK(KP707106781), VSUB(T9, T6));
T16 = VSUB(T12, T15);
T1A = VADD(T15, T12);
Ta = VMUL(LDK(KP707106781), VADD(T6, T9));
Tb = VADD(T3, Ta);
T1p = VSUB(T3, Ta);
}
{
V TL, T25, TX, T26, TO, T28, TR, T29;
{
V TJ, TK, TV, TW;
TJ = LD(&(xi[WS(is, 31)]), ivs, &(xi[WS(is, 1)]));
TK = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
TL = VSUB(TJ, TK);
T25 = VADD(TJ, TK);
TV = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
TW = LD(&(xi[WS(is, 23)]), ivs, &(xi[WS(is, 1)]));
TX = VSUB(TV, TW);
T26 = VADD(TV, TW);
}
{
V TM, TN, TP, TQ;
TM = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
TN = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
TO = VSUB(TM, TN);
T28 = VADD(TM, TN);
TP = LD(&(xi[WS(is, 27)]), ivs, &(xi[WS(is, 1)]));
TQ = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
TR = VSUB(TP, TQ);
T29 = VADD(TP, TQ);
}
{
V TS, TU, T2F, T2G;
TS = VMUL(LDK(KP707106781), VADD(TO, TR));
TT = VADD(TL, TS);
T1v = VSUB(TL, TS);
TU = VMUL(LDK(KP707106781), VSUB(TR, TO));
TY = VSUB(TU, TX);
T1w = VADD(TX, TU);
T27 = VADD(T25, T26);
T2a = VADD(T28, T29);
T2b = VSUB(T27, T2a);
T2F = VSUB(T25, T26);
T2G = VSUB(T29, T28);
T2H = VFNMS(LDK(KP382683432), T2G, VMUL(LDK(KP923879532), T2F));
T2O = VFMA(LDK(KP382683432), T2F, VMUL(LDK(KP923879532), T2G));
}
}
{
V Tu, T1Y, TG, T1Z, Tx, T21, TA, T22;
{
V Ts, Tt, TE, TF;
Ts = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tt = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
Tu = VSUB(Ts, Tt);
T1Y = VADD(Ts, Tt);
TE = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
TF = LD(&(xi[WS(is, 25)]), ivs, &(xi[WS(is, 1)]));
TG = VSUB(TE, TF);
T1Z = VADD(TE, TF);
}
{
V Tv, Tw, Ty, Tz;
Tv = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tw = LD(&(xi[WS(is, 21)]), ivs, &(xi[WS(is, 1)]));
Tx = VSUB(Tv, Tw);
T21 = VADD(Tv, Tw);
Ty = LD(&(xi[WS(is, 29)]), ivs, &(xi[WS(is, 1)]));
Tz = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
TA = VSUB(Ty, Tz);
T22 = VADD(Ty, Tz);
}
{
V TB, TD, T2C, T2D;
TB = VMUL(LDK(KP707106781), VADD(Tx, TA));
TC = VADD(Tu, TB);
T1s = VSUB(Tu, TB);
TD = VMUL(LDK(KP707106781), VSUB(TA, Tx));
TH = VSUB(TD, TG);
T1t = VADD(TG, TD);
T20 = VADD(T1Y, T1Z);
T23 = VADD(T21, T22);
T24 = VSUB(T20, T23);
T2C = VSUB(T1Y, T1Z);
T2D = VSUB(T22, T21);
T2E = VFMA(LDK(KP923879532), T2C, VMUL(LDK(KP382683432), T2D));
T2N = VFNMS(LDK(KP382683432), T2C, VMUL(LDK(KP923879532), T2D));
}
}
{
V Te, T2h, To, T2f, Th, T2i, Tl, T2e, Ti, Tp;
{
V Tc, Td, Tm, Tn;
Tc = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Td = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
Te = VSUB(Tc, Td);
T2h = VADD(Tc, Td);
Tm = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tn = LD(&(xi[WS(is, 22)]), ivs, &(xi[0]));
To = VSUB(Tm, Tn);
T2f = VADD(Tm, Tn);
}
{
V Tf, Tg, Tj, Tk;
Tf = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tg = LD(&(xi[WS(is, 26)]), ivs, &(xi[0]));
Th = VSUB(Tf, Tg);
T2i = VADD(Tf, Tg);
Tj = LD(&(xi[WS(is, 30)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Tl = VSUB(Tj, Tk);
T2e = VADD(Tj, Tk);
}
T2g = VADD(T2e, T2f);
T2j = VADD(T2h, T2i);
Ti = VFNMS(LDK(KP382683432), Th, VMUL(LDK(KP923879532), Te));
Tp = VFMA(LDK(KP923879532), Tl, VMUL(LDK(KP382683432), To));
Tq = VADD(Ti, Tp);
T1B = VSUB(Tp, Ti);
{
V T17, T18, T2y, T2z;
T17 = VFNMS(LDK(KP923879532), To, VMUL(LDK(KP382683432), Tl));
T18 = VFMA(LDK(KP382683432), Te, VMUL(LDK(KP923879532), Th));
T19 = VSUB(T17, T18);
T1q = VADD(T18, T17);
T2y = VSUB(T2h, T2i);
T2z = VSUB(T2e, T2f);
T2A = VMUL(LDK(KP707106781), VADD(T2y, T2z));
T2L = VMUL(LDK(KP707106781), VSUB(T2z, T2y));
}
}
{
V T2d, T2n, T2m, T2o;
{
V T1X, T2c, T2k, T2l;
T1X = VSUB(T1T, T1W);
T2c = VMUL(LDK(KP707106781), VADD(T24, T2b));
T2d = VADD(T1X, T2c);
T2n = VSUB(T1X, T2c);
T2k = VSUB(T2g, T2j);
T2l = VMUL(LDK(KP707106781), VSUB(T2b, T24));
T2m = VBYI(VADD(T2k, T2l));
T2o = VBYI(VSUB(T2l, T2k));
}
ST(&(xo[WS(os, 28)]), VSUB(T2d, T2m), ovs, &(xo[0]));
ST(&(xo[WS(os, 12)]), VADD(T2n, T2o), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(T2d, T2m), ovs, &(xo[0]));
ST(&(xo[WS(os, 20)]), VSUB(T2n, T2o), ovs, &(xo[0]));
}
{
V T2r, T2v, T2u, T2w;
{
V T2p, T2q, T2s, T2t;
T2p = VADD(T1T, T1W);
T2q = VADD(T2j, T2g);
T2r = VADD(T2p, T2q);
T2v = VSUB(T2p, T2q);
T2s = VADD(T20, T23);
T2t = VADD(T27, T2a);
T2u = VADD(T2s, T2t);
T2w = VBYI(VSUB(T2t, T2s));
}
ST(&(xo[WS(os, 16)]), VSUB(T2r, T2u), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VADD(T2v, T2w), ovs, &(xo[0]));
ST(&(xo[0]), VADD(T2r, T2u), ovs, &(xo[0]));
ST(&(xo[WS(os, 24)]), VSUB(T2v, T2w), ovs, &(xo[0]));
}
{
V T2V, T2Z, T2Y, T30;
{
V T2T, T2U, T2W, T2X;
T2T = VSUB(T2H, T2E);
T2U = VSUB(T2L, T2K);
T2V = VBYI(VSUB(T2T, T2U));
T2Z = VBYI(VADD(T2U, T2T));
T2W = VSUB(T2x, T2A);
T2X = VSUB(T2O, T2N);
T2Y = VSUB(T2W, T2X);
T30 = VADD(T2W, T2X);
}
ST(&(xo[WS(os, 10)]), VADD(T2V, T2Y), ovs, &(xo[0]));
ST(&(xo[WS(os, 26)]), VSUB(T30, T2Z), ovs, &(xo[0]));
ST(&(xo[WS(os, 22)]), VSUB(T2Y, T2V), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VADD(T2Z, T30), ovs, &(xo[0]));
}
{
V T2J, T2R, T2Q, T2S;
{
V T2B, T2I, T2M, T2P;
T2B = VADD(T2x, T2A);
T2I = VADD(T2E, T2H);
T2J = VADD(T2B, T2I);
T2R = VSUB(T2B, T2I);
T2M = VADD(T2K, T2L);
T2P = VADD(T2N, T2O);
T2Q = VBYI(VADD(T2M, T2P));
T2S = VBYI(VSUB(T2P, T2M));
}
ST(&(xo[WS(os, 30)]), VSUB(T2J, T2Q), ovs, &(xo[0]));
ST(&(xo[WS(os, 14)]), VADD(T2R, T2S), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VADD(T2J, T2Q), ovs, &(xo[0]));
ST(&(xo[WS(os, 18)]), VSUB(T2R, T2S), ovs, &(xo[0]));
}
{
V T1r, T1C, T1M, T1K, T1F, T1N, T1y, T1J;
T1r = VADD(T1p, T1q);
T1C = VADD(T1A, T1B);
T1M = VSUB(T1p, T1q);
T1K = VSUB(T1B, T1A);
{
V T1D, T1E, T1u, T1x;
T1D = VFNMS(LDK(KP555570233), T1s, VMUL(LDK(KP831469612), T1t));
T1E = VFMA(LDK(KP555570233), T1v, VMUL(LDK(KP831469612), T1w));
T1F = VADD(T1D, T1E);
T1N = VSUB(T1E, T1D);
T1u = VFMA(LDK(KP831469612), T1s, VMUL(LDK(KP555570233), T1t));
T1x = VFNMS(LDK(KP555570233), T1w, VMUL(LDK(KP831469612), T1v));
T1y = VADD(T1u, T1x);
T1J = VSUB(T1x, T1u);
}
{
V T1z, T1G, T1P, T1Q;
T1z = VADD(T1r, T1y);
T1G = VBYI(VADD(T1C, T1F));
ST(&(xo[WS(os, 29)]), VSUB(T1z, T1G), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VADD(T1z, T1G), ovs, &(xo[WS(os, 1)]));
T1P = VBYI(VADD(T1K, T1J));
T1Q = VADD(T1M, T1N);
ST(&(xo[WS(os, 5)]), VADD(T1P, T1Q), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 27)]), VSUB(T1Q, T1P), ovs, &(xo[WS(os, 1)]));
}
{
V T1H, T1I, T1L, T1O;
T1H = VSUB(T1r, T1y);
T1I = VBYI(VSUB(T1F, T1C));
ST(&(xo[WS(os, 19)]), VSUB(T1H, T1I), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 13)]), VADD(T1H, T1I), ovs, &(xo[WS(os, 1)]));
T1L = VBYI(VSUB(T1J, T1K));
T1O = VSUB(T1M, T1N);
ST(&(xo[WS(os, 11)]), VADD(T1L, T1O), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 21)]), VSUB(T1O, T1L), ovs, &(xo[WS(os, 1)]));
}
}
{
V Tr, T1a, T1k, T1i, T1d, T1l, T10, T1h;
Tr = VADD(Tb, Tq);
T1a = VADD(T16, T19);
T1k = VSUB(Tb, Tq);
T1i = VSUB(T19, T16);
{
V T1b, T1c, TI, TZ;
T1b = VFNMS(LDK(KP195090322), TC, VMUL(LDK(KP980785280), TH));
T1c = VFMA(LDK(KP195090322), TT, VMUL(LDK(KP980785280), TY));
T1d = VADD(T1b, T1c);
T1l = VSUB(T1c, T1b);
TI = VFMA(LDK(KP980785280), TC, VMUL(LDK(KP195090322), TH));
TZ = VFNMS(LDK(KP195090322), TY, VMUL(LDK(KP980785280), TT));
T10 = VADD(TI, TZ);
T1h = VSUB(TZ, TI);
}
{
V T11, T1e, T1n, T1o;
T11 = VADD(Tr, T10);
T1e = VBYI(VADD(T1a, T1d));
ST(&(xo[WS(os, 31)]), VSUB(T11, T1e), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VADD(T11, T1e), ovs, &(xo[WS(os, 1)]));
T1n = VBYI(VADD(T1i, T1h));
T1o = VADD(T1k, T1l);
ST(&(xo[WS(os, 7)]), VADD(T1n, T1o), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 25)]), VSUB(T1o, T1n), ovs, &(xo[WS(os, 1)]));
}
{
V T1f, T1g, T1j, T1m;
T1f = VSUB(Tr, T10);
T1g = VBYI(VSUB(T1d, T1a));
ST(&(xo[WS(os, 17)]), VSUB(T1f, T1g), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 15)]), VADD(T1f, T1g), ovs, &(xo[WS(os, 1)]));
T1j = VBYI(VSUB(T1h, T1i));
T1m = VSUB(T1k, T1l);
ST(&(xo[WS(os, 9)]), VADD(T1j, T1m), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 23)]), VSUB(T1m, T1j), ovs, &(xo[WS(os, 1)]));
}
}
}
}
static void n2bv_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T1, T2, Tc, Td, T4, T5, T7, T8;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Td = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T4 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
{
V T3, Tj, Te, Tk, T6, Tm, T9, Tn, Tp, Tl;
T3 = VSUB(T1, T2);
Tj = VADD(T1, T2);
Te = VSUB(Tc, Td);
Tk = VADD(Tc, Td);
T6 = VSUB(T4, T5);
Tm = VADD(T4, T5);
T9 = VSUB(T7, T8);
Tn = VADD(T7, T8);
Tp = VADD(Tj, Tk);
Tl = VSUB(Tj, Tk);
{
V Tq, To, Ta, Tf;
Tq = VADD(Tm, Tn);
To = VSUB(Tm, Tn);
Ta = VADD(T6, T9);
Tf = VSUB(T6, T9);
{
V Tr, Ts, Tt, Tu, Tg, Ti, Tb, Th;
Tr = VFMAI(To, Tl);
STM2(&(xo[4]), Tr, ovs, &(xo[0]));
Ts = VFNMSI(To, Tl);
STM2(&(xo[12]), Ts, ovs, &(xo[0]));
Tt = VADD(Tp, Tq);
STM2(&(xo[0]), Tt, ovs, &(xo[0]));
Tu = VSUB(Tp, Tq);
STM2(&(xo[8]), Tu, ovs, &(xo[0]));
Tg = VFNMS(LDK(KP707106781), Tf, Te);
Ti = VFMA(LDK(KP707106781), Tf, Te);
Tb = VFNMS(LDK(KP707106781), Ta, T3);
Th = VFMA(LDK(KP707106781), Ta, T3);
{
V Tv, Tw, Tx, Ty;
Tv = VFNMSI(Ti, Th);
STM2(&(xo[14]), Tv, ovs, &(xo[2]));
STN2(&(xo[12]), Ts, Tv, ovs);
Tw = VFMAI(Ti, Th);
STM2(&(xo[2]), Tw, ovs, &(xo[2]));
STN2(&(xo[0]), Tt, Tw, ovs);
Tx = VFMAI(Tg, Tb);
STM2(&(xo[10]), Tx, ovs, &(xo[2]));
STN2(&(xo[8]), Tu, Tx, ovs);
Ty = VFNMSI(Tg, Tb);
STM2(&(xo[6]), Ty, ovs, &(xo[2]));
STN2(&(xo[4]), Tr, Ty, ovs);
}
}
}
}
}
}
static void t1fv_12(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 22)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(12, rs)) {
V Tq, Ti, T7, TQ, Tu, TA, TU, Tk, TR, Tf, TE, TM;
{
V T9, TC, Tj, TD, Te;
{
V T1, T4, T2, Tm, Tx, To;
T1 = LD(&(x[0]), ms, &(x[0]));
T4 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tm = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tx = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
To = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V T5, T3, Tn, Ty, Tp, Td, Tb, T8, Tc, Ta;
T8 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Ta = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 14]), T4);
T3 = BYTWJ(&(W[TWVL * 6]), T2);
Tn = BYTWJ(&(W[0]), Tm);
Ty = BYTWJ(&(W[TWVL * 16]), Tx);
Tp = BYTWJ(&(W[TWVL * 8]), To);
T9 = BYTWJ(&(W[TWVL * 10]), T8);
Td = BYTWJ(&(W[TWVL * 2]), Tc);
Tb = BYTWJ(&(W[TWVL * 18]), Ta);
{
V Th, T6, Tt, Tz;
Th = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
TC = VSUB(T5, T3);
T6 = VADD(T3, T5);
Tt = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tz = VADD(Tn, Tp);
Tq = VSUB(Tn, Tp);
Tj = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
TD = VSUB(Td, Tb);
Te = VADD(Tb, Td);
Ti = BYTWJ(&(W[TWVL * 20]), Th);
T7 = VFNMS(LDK(KP500000000), T6, T1);
TQ = VADD(T1, T6);
Tu = BYTWJ(&(W[TWVL * 4]), Tt);
TA = VFNMS(LDK(KP500000000), Tz, Ty);
TU = VADD(Ty, Tz);
}
}
}
Tk = BYTWJ(&(W[TWVL * 12]), Tj);
TR = VADD(T9, Te);
Tf = VFNMS(LDK(KP500000000), Te, T9);
TE = VSUB(TC, TD);
TM = VADD(TC, TD);
}
{
V Tv, Tl, TI, Tg, TW, TS;
Tv = VADD(Tk, Ti);
Tl = VSUB(Ti, Tk);
TI = VADD(T7, Tf);
Tg = VSUB(T7, Tf);
TW = VADD(TQ, TR);
TS = VSUB(TQ, TR);
{
V TT, Tw, TL, Tr;
TT = VADD(Tu, Tv);
Tw = VFNMS(LDK(KP500000000), Tv, Tu);
TL = VSUB(Tl, Tq);
Tr = VADD(Tl, Tq);
{
V TP, TN, TG, Ts, TO, TK, TH, TF;
{
V TX, TV, TJ, TB;
TX = VADD(TT, TU);
TV = VSUB(TT, TU);
TJ = VADD(Tw, TA);
TB = VSUB(Tw, TA);
TP = VMUL(LDK(KP866025403), VADD(TM, TL));
TN = VMUL(LDK(KP866025403), VSUB(TL, TM));
TG = VFNMS(LDK(KP866025403), Tr, Tg);
Ts = VFMA(LDK(KP866025403), Tr, Tg);
ST(&(x[WS(rs, 6)]), VSUB(TW, TX), ms, &(x[0]));
ST(&(x[0]), VADD(TW, TX), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VFMAI(TV, TS), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFNMSI(TV, TS), ms, &(x[WS(rs, 1)]));
TO = VADD(TI, TJ);
TK = VSUB(TI, TJ);
TH = VFMA(LDK(KP866025403), TE, TB);
TF = VFNMS(LDK(KP866025403), TE, TB);
}
ST(&(x[WS(rs, 4)]), VFMAI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VFNMSI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VFNMSI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VFNMSI(TH, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(TH, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFMAI(TF, Ts), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(TF, Ts), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
static void hc2cbdftv_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 30)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(64, rs)) {
V T8, Tv, TE, T1t, TP, T1w, T10, T1p, Tn, Tw, T13, T1q, TL, T1x, TS;
V T1u;
{
V T4, TA, Tu, TC, T7, TN, Tr, TB, T2, T3, Ts, Tt, T5, T6, Tp;
V Tq, TD, TO, TY, TZ, Tb, TF, Tl, TJ, Te, TG, Ti, TI, T9, Ta;
V Tj, Tk, Tc, Td, Tg, Th, Tf, Tm, T11, T12, TH, TK, TQ, TR;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 7)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VFMACONJ(T3, T2);
TA = VFNMSCONJ(T3, T2);
Ts = LD(&(Rp[WS(rs, 6)]), ms, &(Rp[0]));
Tt = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tu = VFMACONJ(Tt, Ts);
TC = VFMSCONJ(Tt, Ts);
T5 = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
T7 = VFMACONJ(T6, T5);
TN = VFNMSCONJ(T6, T5);
Tp = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Tq = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
Tr = VFMACONJ(Tq, Tp);
TB = VFNMSCONJ(Tq, Tp);
T8 = VSUB(T4, T7);
Tv = VSUB(Tr, Tu);
TD = VADD(TB, TC);
TE = VFMA(LDK(KP707106781), TD, TA);
T1t = VFNMS(LDK(KP707106781), TD, TA);
TO = VSUB(TB, TC);
TP = VFMA(LDK(KP707106781), TO, TN);
T1w = VFNMS(LDK(KP707106781), TO, TN);
TY = VADD(T4, T7);
TZ = VADD(Tr, Tu);
T10 = VADD(TY, TZ);
T1p = VSUB(TY, TZ);
T9 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Ta = LD(&(Rm[WS(rs, 6)]), -ms, &(Rm[0]));
Tb = VFMACONJ(Ta, T9);
TF = VFNMSCONJ(Ta, T9);
Tj = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tk = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tl = VFMACONJ(Tk, Tj);
TJ = VFNMSCONJ(Tk, Tj);
Tc = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Td = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Te = VFMACONJ(Td, Tc);
TG = VFNMSCONJ(Td, Tc);
Tg = LD(&(Rp[WS(rs, 7)]), ms, &(Rp[WS(rs, 1)]));
Th = LD(&(Rm[0]), -ms, &(Rm[0]));
Ti = VFMACONJ(Th, Tg);
TI = VFMSCONJ(Th, Tg);
Tf = VSUB(Tb, Te);
Tm = VSUB(Ti, Tl);
Tn = VADD(Tf, Tm);
Tw = VSUB(Tf, Tm);
T11 = VADD(Tb, Te);
T12 = VADD(Ti, Tl);
T13 = VADD(T11, T12);
T1q = VSUB(T11, T12);
TH = VFNMS(LDK(KP414213562), TG, TF);
TK = VFMA(LDK(KP414213562), TJ, TI);
TL = VADD(TH, TK);
T1x = VSUB(TH, TK);
TQ = VFMA(LDK(KP414213562), TF, TG);
TR = VFNMS(LDK(KP414213562), TI, TJ);
TS = VADD(TQ, TR);
T1u = VSUB(TQ, TR);
}
{
V T1j, T1R, T1c, T1J, T1g, T1l, T1N, T1T, T1Q, T1a, T1b, T19, T1I, T1e, T1f;
V T1d, T1k, T1L, T1M, T1K, T1S, T1h, T1U, T1V, T1i, T1m, T1O, T1P, T1n, T14;
V T1r, Ty, T1D, TU, T16, T1z, T1F, TX, T1o, To, Tx, T1, T1C, TM, TT;
V Tz, T15, T1v, T1y, T1s, T1E, TV, T1G, T1H, TW, T17, T1A, T1B, T18;
T1j = VADD(T10, T13);
T1Q = LDW(&(W[TWVL * 22]));
T1R = VZMUL(T1Q, VFNMSI(T1q, T1p));
T1a = VFMA(LDK(KP707106781), Tn, T8);
T1b = VFMA(LDK(KP707106781), Tw, Tv);
T19 = LDW(&(W[TWVL * 26]));
T1c = VZMUL(T19, VFNMSI(T1b, T1a));
T1I = LDW(&(W[TWVL * 2]));
T1J = VZMUL(T1I, VFMAI(T1b, T1a));
T1e = VFMA(LDK(KP923879532), TL, TE);
T1f = VFMA(LDK(KP923879532), TS, TP);
T1d = LDW(&(W[TWVL * 28]));
T1g = VZMULI(T1d, VFNMSI(T1f, T1e));
T1k = LDW(&(W[0]));
T1l = VZMULI(T1k, VFMAI(T1f, T1e));
T1L = VFMA(LDK(KP923879532), T1u, T1t);
T1M = VFNMS(LDK(KP923879532), T1x, T1w);
T1K = LDW(&(W[TWVL * 4]));
T1N = VZMULI(T1K, VFNMSI(T1M, T1L));
T1S = LDW(&(W[TWVL * 24]));
T1T = VZMULI(T1S, VFMAI(T1M, T1L));
T1h = VCONJ(VSUB(T1c, T1g));
ST(&(Rm[WS(rs, 7)]), T1h, -ms, &(Rm[WS(rs, 1)]));
T1U = VCONJ(VSUB(T1R, T1T));
ST(&(Rm[WS(rs, 6)]), T1U, -ms, &(Rm[0]));
T1V = VADD(T1R, T1T);
ST(&(Rp[WS(rs, 6)]), T1V, ms, &(Rp[0]));
T1i = VADD(T1c, T1g);
ST(&(Rp[WS(rs, 7)]), T1i, ms, &(Rp[WS(rs, 1)]));
T1m = VCONJ(VSUB(T1j, T1l));
ST(&(Rm[0]), T1m, -ms, &(Rm[0]));
T1O = VCONJ(VSUB(T1J, T1N));
ST(&(Rm[WS(rs, 1)]), T1O, -ms, &(Rm[WS(rs, 1)]));
T1P = VADD(T1J, T1N);
ST(&(Rp[WS(rs, 1)]), T1P, ms, &(Rp[WS(rs, 1)]));
T1n = VADD(T1j, T1l);
ST(&(Rp[0]), T1n, ms, &(Rp[0]));
TX = LDW(&(W[TWVL * 14]));
T14 = VZMUL(TX, VSUB(T10, T13));
T1o = LDW(&(W[TWVL * 6]));
T1r = VZMUL(T1o, VFMAI(T1q, T1p));
To = VFNMS(LDK(KP707106781), Tn, T8);
Tx = VFNMS(LDK(KP707106781), Tw, Tv);
T1 = LDW(&(W[TWVL * 10]));
Ty = VZMUL(T1, VFNMSI(Tx, To));
T1C = LDW(&(W[TWVL * 18]));
T1D = VZMUL(T1C, VFMAI(Tx, To));
TM = VFNMS(LDK(KP923879532), TL, TE);
TT = VFNMS(LDK(KP923879532), TS, TP);
Tz = LDW(&(W[TWVL * 12]));
TU = VZMULI(Tz, VFNMSI(TT, TM));
T15 = LDW(&(W[TWVL * 16]));
T16 = VZMULI(T15, VFMAI(TT, TM));
T1v = VFNMS(LDK(KP923879532), T1u, T1t);
T1y = VFMA(LDK(KP923879532), T1x, T1w);
T1s = LDW(&(W[TWVL * 8]));
T1z = VZMULI(T1s, VFMAI(T1y, T1v));
T1E = LDW(&(W[TWVL * 20]));
T1F = VZMULI(T1E, VFNMSI(T1y, T1v));
TV = VCONJ(VSUB(Ty, TU));
ST(&(Rm[WS(rs, 3)]), TV, -ms, &(Rm[WS(rs, 1)]));
T1G = VCONJ(VSUB(T1D, T1F));
ST(&(Rm[WS(rs, 5)]), T1G, -ms, &(Rm[WS(rs, 1)]));
T1H = VADD(T1D, T1F);
ST(&(Rp[WS(rs, 5)]), T1H, ms, &(Rp[WS(rs, 1)]));
TW = VADD(Ty, TU);
ST(&(Rp[WS(rs, 3)]), TW, ms, &(Rp[WS(rs, 1)]));
T17 = VCONJ(VSUB(T14, T16));
ST(&(Rm[WS(rs, 4)]), T17, -ms, &(Rm[0]));
T1A = VCONJ(VSUB(T1r, T1z));
ST(&(Rm[WS(rs, 2)]), T1A, -ms, &(Rm[0]));
T1B = VADD(T1r, T1z);
ST(&(Rp[WS(rs, 2)]), T1B, ms, &(Rp[0]));
T18 = VADD(T14, T16);
ST(&(Rp[WS(rs, 4)]), T18, ms, &(Rp[0]));
}
}
}
VLEAVE();
}
static void n2sv_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
for (i = v; i > 0; i = i - (2 * VL), ri = ri + ((2 * VL) * ivs), ii = ii + ((2 * VL) * ivs), ro = ro + ((2 * VL) * ovs), io = io + ((2 * VL) * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V TF, TJ, TD, TR, TS, TT, TU, TV, TW, TE, TX, TY, TK, TI, TZ;
V T10, T11, T12;
{
V Tb, Tn, T3, TC, Ti, TB, T6, To, Tl, Tc, Tw, Tx, T8, T9, Tr;
V Ts;
{
V T1, T2, Tg, Th, T4, T5, Tj, Tk;
T1 = LD(&(ri[0]), ivs, &(ri[0]));
T2 = LD(&(ri[WS(is, 4)]), ivs, &(ri[0]));
Tg = LD(&(ii[0]), ivs, &(ii[0]));
Th = LD(&(ii[WS(is, 4)]), ivs, &(ii[0]));
T4 = LD(&(ri[WS(is, 2)]), ivs, &(ri[0]));
T5 = LD(&(ri[WS(is, 6)]), ivs, &(ri[0]));
Tj = LD(&(ii[WS(is, 2)]), ivs, &(ii[0]));
Tk = LD(&(ii[WS(is, 6)]), ivs, &(ii[0]));
Tb = LD(&(ri[WS(is, 7)]), ivs, &(ri[WS(is, 1)]));
Tn = VSUB(T1, T2);
T3 = VADD(T1, T2);
TC = VSUB(Tg, Th);
Ti = VADD(Tg, Th);
TB = VSUB(T4, T5);
T6 = VADD(T4, T5);
To = VSUB(Tj, Tk);
Tl = VADD(Tj, Tk);
Tc = LD(&(ri[WS(is, 3)]), ivs, &(ri[WS(is, 1)]));
Tw = LD(&(ii[WS(is, 7)]), ivs, &(ii[WS(is, 1)]));
Tx = LD(&(ii[WS(is, 3)]), ivs, &(ii[WS(is, 1)]));
T8 = LD(&(ri[WS(is, 1)]), ivs, &(ri[WS(is, 1)]));
T9 = LD(&(ri[WS(is, 5)]), ivs, &(ri[WS(is, 1)]));
Tr = LD(&(ii[WS(is, 1)]), ivs, &(ii[WS(is, 1)]));
Ts = LD(&(ii[WS(is, 5)]), ivs, &(ii[WS(is, 1)]));
}
{
V TL, T7, TP, Tm, Tz, TH, Te, Tf, TO, TQ, TG, Tu, Tp, TA;
{
V Td, Tv, TN, Ty, Ta, Tq, TM, Tt;
TL = VSUB(T3, T6);
T7 = VADD(T3, T6);
Td = VADD(Tb, Tc);
Tv = VSUB(Tb, Tc);
TN = VADD(Tw, Tx);
Ty = VSUB(Tw, Tx);
Ta = VADD(T8, T9);
Tq = VSUB(T8, T9);
TM = VADD(Tr, Ts);
Tt = VSUB(Tr, Ts);
TP = VADD(Ti, Tl);
Tm = VSUB(Ti, Tl);
Tz = VSUB(Tv, Ty);
TH = VADD(Tv, Ty);
Te = VADD(Ta, Td);
Tf = VSUB(Td, Ta);
TO = VSUB(TM, TN);
TQ = VADD(TM, TN);
TG = VSUB(Tt, Tq);
Tu = VADD(Tq, Tt);
}
TF = VSUB(Tn, To);
Tp = VADD(Tn, To);
TJ = VSUB(TC, TB);
TD = VADD(TB, TC);
TR = VSUB(Tm, Tf);
STM4(&(io[6]), TR, ovs, &(io[0]));
TS = VADD(Tf, Tm);
STM4(&(io[2]), TS, ovs, &(io[0]));
TT = VADD(T7, Te);
STM4(&(ro[0]), TT, ovs, &(ro[0]));
TU = VSUB(T7, Te);
STM4(&(ro[4]), TU, ovs, &(ro[0]));
TV = VADD(TP, TQ);
STM4(&(io[0]), TV, ovs, &(io[0]));
TW = VSUB(TP, TQ);
STM4(&(io[4]), TW, ovs, &(io[0]));
TE = VSUB(Tz, Tu);
TA = VADD(Tu, Tz);
TX = VADD(TL, TO);
STM4(&(ro[2]), TX, ovs, &(ro[0]));
TY = VSUB(TL, TO);
STM4(&(ro[6]), TY, ovs, &(ro[0]));
TK = VADD(TG, TH);
TI = VSUB(TG, TH);
TZ = VFMA(LDK(KP707106781), TA, Tp);
STM4(&(ro[1]), TZ, ovs, &(ro[1]));
T10 = VFNMS(LDK(KP707106781), TA, Tp);
STM4(&(ro[5]), T10, ovs, &(ro[1]));
}
}
T11 = VFMA(LDK(KP707106781), TK, TJ);
STM4(&(io[1]), T11, ovs, &(io[1]));
T12 = VFNMS(LDK(KP707106781), TK, TJ);
STM4(&(io[5]), T12, ovs, &(io[1]));
{
V T13, T14, T15, T16;
T13 = VFMA(LDK(KP707106781), TE, TD);
STM4(&(io[3]), T13, ovs, &(io[1]));
STN4(&(io[0]), TV, T11, TS, T13, ovs);
T14 = VFNMS(LDK(KP707106781), TE, TD);
STM4(&(io[7]), T14, ovs, &(io[1]));
STN4(&(io[4]), TW, T12, TR, T14, ovs);
T15 = VFMA(LDK(KP707106781), TI, TF);
STM4(&(ro[3]), T15, ovs, &(ro[1]));
STN4(&(ro[0]), TT, TZ, TX, T15, ovs);
T16 = VFNMS(LDK(KP707106781), TI, TF);
STM4(&(ro[7]), T16, ovs, &(ro[1]));
STN4(&(ro[4]), TU, T10, TY, T16, ovs);
}
}
}
VLEAVE();
}
static void n1fv_9(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP939692620, +0.939692620785908384054109277324731469936208134);
DVK(KP826351822, +0.826351822333069651148283373230685203999624323);
DVK(KP879385241, +0.879385241571816768108218554649462939872416269);
DVK(KP984807753, +0.984807753012208059366743024589523013670643252);
DVK(KP666666666, +0.666666666666666666666666666666666666666666667);
DVK(KP852868531, +0.852868531952443209628250963940074071936020296);
DVK(KP907603734, +0.907603734547952313649323976213898122064543220);
DVK(KP420276625, +0.420276625461206169731530603237061658838781920);
DVK(KP673648177, +0.673648177666930348851716626769314796000375677);
DVK(KP898197570, +0.898197570222573798468955502359086394667167570);
DVK(KP347296355, +0.347296355333860697703433253538629592000751354);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP439692620, +0.439692620785908384054109277324731469936208134);
DVK(KP203604859, +0.203604859554852403062088995281827210665664861);
DVK(KP152703644, +0.152703644666139302296566746461370407999248646);
DVK(KP586256827, +0.586256827714544512072145703099641959914944179);
DVK(KP968908795, +0.968908795874236621082202410917456709164223497);
DVK(KP726681596, +0.726681596905677465811651808188092531873167623);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T1, T2, T3, T6, Tb, T7, T8, Tc, Td, Tv, T4;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tb = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T7 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tv = VSUB(T3, T2);
T4 = VADD(T2, T3);
{
V Tl, T9, Tm, Te, Tj, T5;
Tl = VSUB(T7, T8);
T9 = VADD(T7, T8);
Tm = VSUB(Td, Tc);
Te = VADD(Tc, Td);
Tj = VFNMS(LDK(KP500000000), T4, T1);
T5 = VADD(T1, T4);
{
V Tn, Ta, Tk, Tf;
Tn = VFNMS(LDK(KP500000000), T9, T6);
Ta = VADD(T6, T9);
Tk = VFNMS(LDK(KP500000000), Te, Tb);
Tf = VADD(Tb, Te);
{
V Ty, TC, To, TB, Tx, Ts, Tg, Ti;
Ty = VFNMS(LDK(KP726681596), Tl, Tn);
TC = VFMA(LDK(KP968908795), Tn, Tl);
To = VFNMS(LDK(KP586256827), Tn, Tm);
TB = VFNMS(LDK(KP152703644), Tm, Tk);
Tx = VFMA(LDK(KP203604859), Tk, Tm);
Ts = VFNMS(LDK(KP439692620), Tl, Tk);
Tg = VADD(Ta, Tf);
Ti = VMUL(LDK(KP866025403), VSUB(Tf, Ta));
{
V Tz, TI, TF, TD, Tt, Th, Tq, Tp;
Tp = VFNMS(LDK(KP347296355), To, Tl);
Tz = VFMA(LDK(KP898197570), Ty, Tx);
TI = VFNMS(LDK(KP898197570), Ty, Tx);
TF = VFNMS(LDK(KP673648177), TC, TB);
TD = VFMA(LDK(KP673648177), TC, TB);
Tt = VFNMS(LDK(KP420276625), Ts, Tm);
ST(&(xo[0]), VADD(T5, Tg), ovs, &(xo[0]));
Th = VFNMS(LDK(KP500000000), Tg, T5);
Tq = VFNMS(LDK(KP907603734), Tp, Tk);
{
V TA, TJ, TE, TG, Tu, Tr, TK, TH, Tw;
TA = VFMA(LDK(KP852868531), Tz, Tj);
TJ = VFMA(LDK(KP666666666), TD, TI);
TE = VMUL(LDK(KP984807753), VFNMS(LDK(KP879385241), Tv, TD));
TG = VFNMS(LDK(KP500000000), Tz, TF);
Tu = VFNMS(LDK(KP826351822), Tt, Tn);
ST(&(xo[WS(os, 6)]), VFNMSI(Ti, Th), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(Ti, Th), ovs, &(xo[WS(os, 1)]));
Tr = VFNMS(LDK(KP939692620), Tq, Tj);
TK = VMUL(LDK(KP866025403), VFMA(LDK(KP852868531), TJ, Tv));
ST(&(xo[WS(os, 8)]), VFMAI(TE, TA), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VFNMSI(TE, TA), ovs, &(xo[WS(os, 1)]));
TH = VFMA(LDK(KP852868531), TG, Tj);
Tw = VMUL(LDK(KP984807753), VFMA(LDK(KP879385241), Tv, Tu));
ST(&(xo[WS(os, 4)]), VFMAI(TK, TH), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VFNMSI(TK, TH), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VFMAI(Tw, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VFNMSI(Tw, Tr), ovs, &(xo[0]));
}
}
}
}
}
}
}
static const R *t1bv_12(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 = ii;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 22)) {
V T1, Tt, T6, T7, TB, Tq, TC, TD, T9, Tu, Te, Tf, Tx, Tl, Ty;
V Tz;
{
V T5, T3, T4, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T4 = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
T5 = BYTW(&(W[TWVL * 14]), T4);
T2 = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
T3 = BYTW(&(W[TWVL * 6]), T2);
Tt = VSUB(T3, T5);
T6 = VADD(T3, T5);
T7 = VFNMS(LDK(KP500000000), T6, T1);
}
{
V Tn, Tp, Tm, TA, To;
Tm = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Tn = BYTW(&(W[0]), Tm);
TA = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
TB = BYTW(&(W[TWVL * 16]), TA);
To = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
Tp = BYTW(&(W[TWVL * 8]), To);
Tq = VSUB(Tn, Tp);
TC = VADD(Tn, Tp);
TD = VFNMS(LDK(KP500000000), TC, TB);
}
{
V Td, Tb, T8, Tc, Ta;
T8 = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
T9 = BYTW(&(W[TWVL * 10]), T8);
Tc = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Td = BYTW(&(W[TWVL * 2]), Tc);
Ta = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
Tb = BYTW(&(W[TWVL * 18]), Ta);
Tu = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Tf = VFNMS(LDK(KP500000000), Te, T9);
}
{
V Ti, Tk, Th, Tw, Tj;
Th = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
Ti = BYTW(&(W[TWVL * 12]), Th);
Tw = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Tx = BYTW(&(W[TWVL * 4]), Tw);
Tj = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
Tk = BYTW(&(W[TWVL * 20]), Tj);
Tl = VSUB(Ti, Tk);
Ty = VADD(Ti, Tk);
Tz = VFNMS(LDK(KP500000000), Ty, Tx);
}
{
V Ts, TG, TF, TH;
{
V Tg, Tr, Tv, TE;
Tg = VSUB(T7, Tf);
Tr = VMUL(LDK(KP866025403), VSUB(Tl, Tq));
Ts = VSUB(Tg, Tr);
TG = VADD(Tg, Tr);
Tv = VMUL(LDK(KP866025403), VSUB(Tt, Tu));
TE = VSUB(Tz, TD);
TF = VBYI(VADD(Tv, TE));
TH = VBYI(VSUB(TE, Tv));
}
ST(&(x[WS(ios, 11)]), VSUB(Ts, TF), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VADD(TG, TH), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(Ts, TF), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VSUB(TG, TH), dist, &(x[WS(ios, 1)]));
}
{
V TS, TW, TV, TX;
{
V TQ, TR, TT, TU;
TQ = VADD(T1, T6);
TR = VADD(T9, Te);
TS = VSUB(TQ, TR);
TW = VADD(TQ, TR);
TT = VADD(Tx, Ty);
TU = VADD(TB, TC);
TV = VBYI(VSUB(TT, TU));
TX = VADD(TT, TU);
}
ST(&(x[WS(ios, 3)]), VSUB(TS, TV), dist, &(x[WS(ios, 1)]));
ST(&(x[0]), VADD(TW, TX), dist, &(x[0]));
ST(&(x[WS(ios, 9)]), VADD(TS, TV), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 6)]), VSUB(TW, TX), dist, &(x[0]));
}
{
V TK, TO, TN, TP;
{
V TI, TJ, TL, TM;
TI = VADD(Tl, Tq);
TJ = VADD(Tt, Tu);
TK = VBYI(VMUL(LDK(KP866025403), VSUB(TI, TJ)));
TO = VBYI(VMUL(LDK(KP866025403), VADD(TJ, TI)));
TL = VADD(T7, Tf);
TM = VADD(Tz, TD);
TN = VSUB(TL, TM);
TP = VADD(TL, TM);
}
ST(&(x[WS(ios, 2)]), VADD(TK, TN), dist, &(x[0]));
ST(&(x[WS(ios, 8)]), VSUB(TP, TO), dist, &(x[0]));
ST(&(x[WS(ios, 10)]), VSUB(TN, TK), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(TO, TP), dist, &(x[0]));
}
}
END_SIMD();
return W;
}
static void n1fv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V Tb, Tr, T3, Ts, T6, Tw, Tg, Tt, T9, Tc, T1, T2;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
{
V T4, T5, Te, Tf, T7, T8;
T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tr = VADD(T1, T2);
T3 = VSUB(T1, T2);
Ts = VADD(T4, T5);
T6 = VSUB(T4, T5);
Tw = VADD(Te, Tf);
Tg = VSUB(Te, Tf);
Tt = VADD(T7, T8);
T9 = VSUB(T7, T8);
Tc = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
}
{
V TD, Tu, Tm, Ta, Td, Tv;
TD = VSUB(Ts, Tt);
Tu = VADD(Ts, Tt);
Tm = VSUB(T6, T9);
Ta = VADD(T6, T9);
Td = VSUB(Tb, Tc);
Tv = VADD(Tb, Tc);
{
V TC, Tx, Tn, Th;
TC = VSUB(Tv, Tw);
Tx = VADD(Tv, Tw);
Tn = VSUB(Td, Tg);
Th = VADD(Td, Tg);
{
V Ty, TA, TE, TG, Ti, Tk, To, Tq, Tz, Tj;
Ty = VADD(Tu, Tx);
TA = VSUB(Tu, Tx);
TE = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TD, TC));
TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TC, TD));
Ti = VADD(Ta, Th);
Tk = VSUB(Ta, Th);
To = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tn, Tm));
Tq = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tm, Tn));
Tz = VFNMS(LDK(KP250000000), Ty, Tr);
ST(&(xo[0]), VADD(Tr, Ty), ovs, &(xo[0]));
Tj = VFNMS(LDK(KP250000000), Ti, T3);
ST(&(xo[WS(os, 5)]), VADD(T3, Ti), ovs, &(xo[WS(os, 1)]));
{
V TB, TF, Tl, Tp;
TB = VFNMS(LDK(KP559016994), TA, Tz);
TF = VFMA(LDK(KP559016994), TA, Tz);
Tl = VFMA(LDK(KP559016994), Tk, Tj);
Tp = VFNMS(LDK(KP559016994), Tk, Tj);
ST(&(xo[WS(os, 4)]), VFMAI(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VFNMSI(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VFNMSI(TE, TB), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VFMAI(TE, TB), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFNMSI(Tq, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VFMAI(Tq, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VFMAI(To, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VFNMSI(To, Tl), ovs, &(xo[WS(os, 1)]));
}
}
}
}
}
}
static void n1fv_11(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP654860733, +0.654860733945285064056925072466293553183791199);
DVK(KP142314838, +0.142314838273285140443792668616369668791051361);
DVK(KP959492973, +0.959492973614497389890368057066327699062454848);
DVK(KP415415013, +0.415415013001886425529274149229623203524004910);
DVK(KP841253532, +0.841253532831181168861811648919367717513292498);
DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
DVK(KP909631995, +0.909631995354518371411715383079028460060241051);
DVK(KP281732556, +0.281732556841429697711417915346616899035777899);
DVK(KP540640817, +0.540640817455597582107635954318691695431770608);
DVK(KP755749574, +0.755749574354258283774035843972344420179717445);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(22, is), MAKE_VOLATILE_STRIDE(22, os)) {
V T1, T4, Ti, Tg, Tl, Td, Tk, Ta, Tj, T7, Tm, Tb, Tc, Tt, Ts;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
{
V T2, T3, Te, Tf;
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Ti = VSUB(T3, T2);
Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tf = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tg = VADD(Te, Tf);
Tl = VSUB(Tf, Te);
}
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Td = VADD(Tb, Tc);
Tk = VSUB(Tc, Tb);
{
V T8, T9, T5, T6;
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Ta = VADD(T8, T9);
Tj = VSUB(T9, T8);
T5 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T7 = VADD(T5, T6);
Tm = VSUB(T6, T5);
}
ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, VADD(Ta, VADD(Td, Tg))))), ovs, &(xo[0]));
{
V Tn, Th, Tv, Tu;
Tn = VBYI(VFMA(LDK(KP755749574), Ti, VFMA(LDK(KP540640817), Tj, VFNMS(LDK(KP909631995), Tl, VFNMS(LDK(KP989821441), Tm, VMUL(LDK(KP281732556), Tk))))));
Th = VFMA(LDK(KP841253532), Ta, VFMA(LDK(KP415415013), Tg, VFNMS(LDK(KP959492973), Td, VFNMS(LDK(KP142314838), T7, VFNMS(LDK(KP654860733), T4, T1)))));
ST(&(xo[WS(os, 7)]), VSUB(Th, Tn), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VADD(Th, Tn), ovs, &(xo[0]));
Tv = VBYI(VFMA(LDK(KP281732556), Ti, VFMA(LDK(KP755749574), Tj, VFNMS(LDK(KP909631995), Tk, VFNMS(LDK(KP540640817), Tm, VMUL(LDK(KP989821441), Tl))))));
Tu = VFMA(LDK(KP841253532), T7, VFMA(LDK(KP415415013), Td, VFNMS(LDK(KP142314838), Tg, VFNMS(LDK(KP654860733), Ta, VFNMS(LDK(KP959492973), T4, T1)))));
ST(&(xo[WS(os, 6)]), VSUB(Tu, Tv), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VADD(Tu, Tv), ovs, &(xo[WS(os, 1)]));
}
Tt = VBYI(VFMA(LDK(KP989821441), Ti, VFMA(LDK(KP540640817), Tk, VFNMS(LDK(KP909631995), Tj, VFNMS(LDK(KP281732556), Tm, VMUL(LDK(KP755749574), Tl))))));
Ts = VFMA(LDK(KP415415013), Ta, VFMA(LDK(KP841253532), Td, VFNMS(LDK(KP654860733), Tg, VFNMS(LDK(KP959492973), T7, VFNMS(LDK(KP142314838), T4, T1)))));
ST(&(xo[WS(os, 8)]), VSUB(Ts, Tt), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VADD(Ts, Tt), ovs, &(xo[WS(os, 1)]));
{
V Tr, Tq, Tp, To;
Tr = VBYI(VFMA(LDK(KP540640817), Ti, VFMA(LDK(KP909631995), Tm, VFMA(LDK(KP989821441), Tj, VFMA(LDK(KP755749574), Tk, VMUL(LDK(KP281732556), Tl))))));
Tq = VFMA(LDK(KP841253532), T4, VFMA(LDK(KP415415013), T7, VFNMS(LDK(KP959492973), Tg, VFNMS(LDK(KP654860733), Td, VFNMS(LDK(KP142314838), Ta, T1)))));
ST(&(xo[WS(os, 10)]), VSUB(Tq, Tr), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VADD(Tq, Tr), ovs, &(xo[WS(os, 1)]));
Tp = VBYI(VFMA(LDK(KP909631995), Ti, VFNMS(LDK(KP540640817), Tl, VFNMS(LDK(KP989821441), Tk, VFNMS(LDK(KP281732556), Tj, VMUL(LDK(KP755749574), Tm))))));
To = VFMA(LDK(KP415415013), T4, VFMA(LDK(KP841253532), Tg, VFNMS(LDK(KP142314838), Td, VFNMS(LDK(KP959492973), Ta, VFNMS(LDK(KP654860733), T7, T1)))));
ST(&(xo[WS(os, 9)]), VSUB(To, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VADD(To, Tp), ovs, &(xo[0]));
}
}
}
VLEAVE();
}
static void t1bv_16(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 30)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(rs)) {
V TO, Ta, TJ, TP, T14, Tq, T1i, T10, T1b, T1l, T13, T1c, TR, Tl, T15;
V Tv;
{
V Tc, TW, T4, T19, T9, TD, TI, Tj, TZ, T1a, Te, Th, Tn, Tr, Tu;
V Tp;
{
V T1, T2, T5, T7;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
{
V Tz, TG, TB, TE;
Tz = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
TG = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
TB = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
TE = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
{
V Ti, TX, TY, Td, Tg, Tm, Tt, To;
{
V T3, T6, T8, TA, TH, TC, TF, Tb;
Tb = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTW(&(W[TWVL * 14]), T2);
T6 = BYTW(&(W[TWVL * 6]), T5);
T8 = BYTW(&(W[TWVL * 22]), T7);
TA = BYTW(&(W[TWVL * 2]), Tz);
TH = BYTW(&(W[TWVL * 10]), TG);
TC = BYTW(&(W[TWVL * 18]), TB);
TF = BYTW(&(W[TWVL * 26]), TE);
Tc = BYTW(&(W[0]), Tb);
TW = VSUB(T1, T3);
T4 = VADD(T1, T3);
T19 = VSUB(T6, T8);
T9 = VADD(T6, T8);
Ti = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
TD = VADD(TA, TC);
TX = VSUB(TA, TC);
TI = VADD(TF, TH);
TY = VSUB(TF, TH);
}
Td = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tg = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tm = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
Tj = BYTW(&(W[TWVL * 24]), Ti);
Tt = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
To = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
TZ = VADD(TX, TY);
T1a = VSUB(TX, TY);
Te = BYTW(&(W[TWVL * 16]), Td);
Th = BYTW(&(W[TWVL * 8]), Tg);
Tn = BYTW(&(W[TWVL * 28]), Tm);
Tr = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tu = BYTW(&(W[TWVL * 20]), Tt);
Tp = BYTW(&(W[TWVL * 12]), To);
}
}
}
{
V Tf, T11, Tk, T12, Ts;
TO = VADD(T4, T9);
Ta = VSUB(T4, T9);
TJ = VSUB(TD, TI);
TP = VADD(TD, TI);
Tf = VADD(Tc, Te);
T11 = VSUB(Tc, Te);
Tk = VADD(Th, Tj);
T12 = VSUB(Th, Tj);
Ts = BYTW(&(W[TWVL * 4]), Tr);
T14 = VSUB(Tn, Tp);
Tq = VADD(Tn, Tp);
T1i = VFNMS(LDK(KP707106781), TZ, TW);
T10 = VFMA(LDK(KP707106781), TZ, TW);
T1b = VFMA(LDK(KP707106781), T1a, T19);
T1l = VFNMS(LDK(KP707106781), T1a, T19);
T13 = VFNMS(LDK(KP414213562), T12, T11);
T1c = VFMA(LDK(KP414213562), T11, T12);
TR = VADD(Tf, Tk);
Tl = VSUB(Tf, Tk);
T15 = VSUB(Tu, Ts);
Tv = VADD(Ts, Tu);
}
}
{
V T1d, T16, TS, Tw, TU, TQ;
T1d = VFMA(LDK(KP414213562), T14, T15);
T16 = VFNMS(LDK(KP414213562), T15, T14);
TS = VADD(Tq, Tv);
Tw = VSUB(Tq, Tv);
TU = VADD(TO, TP);
TQ = VSUB(TO, TP);
{
V T1e, T1j, T17, T1m;
T1e = VSUB(T1c, T1d);
T1j = VADD(T1c, T1d);
T17 = VADD(T13, T16);
T1m = VSUB(T13, T16);
{
V TV, TT, TK, Tx;
TV = VADD(TR, TS);
TT = VSUB(TR, TS);
TK = VSUB(Tl, Tw);
Tx = VADD(Tl, Tw);
{
V T1h, T1f, T1o, T1k;
T1h = VFMA(LDK(KP923879532), T1e, T1b);
T1f = VFNMS(LDK(KP923879532), T1e, T1b);
T1o = VFMA(LDK(KP923879532), T1j, T1i);
T1k = VFNMS(LDK(KP923879532), T1j, T1i);
{
V T1g, T18, T1p, T1n;
T1g = VFMA(LDK(KP923879532), T17, T10);
T18 = VFNMS(LDK(KP923879532), T17, T10);
T1p = VFNMS(LDK(KP923879532), T1m, T1l);
T1n = VFMA(LDK(KP923879532), T1m, T1l);
ST(&(x[WS(rs, 8)]), VSUB(TU, TV), ms, &(x[0]));
ST(&(x[0]), VADD(TU, TV), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TT, TQ), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VFNMSI(TT, TQ), ms, &(x[0]));
{
V TN, TL, TM, Ty;
TN = VFMA(LDK(KP707106781), TK, TJ);
TL = VFNMS(LDK(KP707106781), TK, TJ);
TM = VFMA(LDK(KP707106781), Tx, Ta);
Ty = VFNMS(LDK(KP707106781), Tx, Ta);
ST(&(x[WS(rs, 15)]), VFNMSI(T1h, T1g), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(T1h, T1g), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFMAI(T1f, T18), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(T1f, T18), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(T1p, T1o), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VFMAI(T1p, T1o), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFNMSI(T1n, T1k), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFMAI(T1n, T1k), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TM), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VFNMSI(TN, TM), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VFMAI(TL, Ty), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TL, Ty), ms, &(x[0]));
}
}
}
}
}
}
}
}
VLEAVE();
}
static void n1fv_11(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP959492973, +0.959492973614497389890368057066327699062454848);
DVK(KP876768831, +0.876768831002589333891339807079336796764054852);
DVK(KP918985947, +0.918985947228994779780736114132655398124909697);
DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
DVK(KP778434453, +0.778434453334651800608337670740821884709317477);
DVK(KP830830026, +0.830830026003772851058548298459246407048009821);
DVK(KP372785597, +0.372785597771792209609773152906148328659002598);
DVK(KP634356270, +0.634356270682424498893150776899916060542806975);
DVK(KP715370323, +0.715370323453429719112414662767260662417897278);
DVK(KP342584725, +0.342584725681637509502641509861112333758894680);
DVK(KP521108558, +0.521108558113202722944698153526659300680427422);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(22, is), MAKE_VOLATILE_STRIDE(22, os)) {
V T1, Tb, T4, Tp, Tg, Tq, T7, Tn, Ta, Tm, Tc, Tr;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
{
V T2, T3, Te, Tf;
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tf = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T5, T6, T8, T9;
T5 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Tp = VSUB(T3, T2);
Tg = VADD(Te, Tf);
Tq = VSUB(Tf, Te);
T7 = VADD(T5, T6);
Tn = VSUB(T6, T5);
Ta = VADD(T8, T9);
Tm = VSUB(T9, T8);
Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
}
}
Tr = VFMA(LDK(KP521108558), Tq, Tp);
{
V TS, TE, Th, Td, To, T12, TO, TB, T11, TN, TA, TF;
T11 = VFNMS(LDK(KP521108558), Tp, Tn);
TN = VFNMS(LDK(KP342584725), T7, Tg);
TA = VFMA(LDK(KP521108558), Tm, Tq);
TS = VFMA(LDK(KP715370323), Tm, Tp);
TE = VFNMS(LDK(KP342584725), T4, Ta);
Th = VFNMS(LDK(KP342584725), Ta, T7);
Td = VADD(Tb, Tc);
To = VSUB(Tc, Tb);
T12 = VFNMS(LDK(KP715370323), T11, Tm);
TO = VFNMS(LDK(KP634356270), TN, T4);
TB = VFNMS(LDK(KP715370323), TA, Tn);
TF = VFNMS(LDK(KP634356270), TE, Tg);
{
V T14, TD, TV, Tu, TY, Tx, Tk, TR, TI, TM, TJ, TT, Ts;
TJ = VFNMS(LDK(KP521108558), Tn, To);
TT = VFMA(LDK(KP372785597), To, TS);
Ts = VFMA(LDK(KP715370323), Tr, To);
ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, VADD(Ta, VADD(Td, Tg))))), ovs, &(xo[0]));
{
V TW, Tv, Ti, T13;
TW = VFNMS(LDK(KP342584725), Tg, Td);
Tv = VFNMS(LDK(KP342584725), Td, T4);
Ti = VFNMS(LDK(KP634356270), Th, Td);
T13 = VFNMS(LDK(KP830830026), T12, To);
{
V TP, TC, TG, TK;
TP = VFNMS(LDK(KP778434453), TO, Ta);
TC = VFMA(LDK(KP830830026), TB, Tp);
TG = VFNMS(LDK(KP778434453), TF, Td);
TK = VFMA(LDK(KP715370323), TJ, Tq);
{
V TU, Tt, TX, Tw;
TU = VFNMS(LDK(KP830830026), TT, Tq);
Tt = VFMA(LDK(KP830830026), Ts, Tn);
TX = VFNMS(LDK(KP634356270), TW, Ta);
Tw = VFNMS(LDK(KP634356270), Tv, T7);
{
V Tj, TQ, TH, TL;
Tj = VFNMS(LDK(KP778434453), Ti, T4);
T14 = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), T13, Tq));
TQ = VFNMS(LDK(KP876768831), TP, Td);
TD = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TC, To));
TH = VFNMS(LDK(KP876768831), TG, T7);
TL = VFNMS(LDK(KP830830026), TK, Tm);
TV = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), TU, Tn));
Tu = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), Tt, Tm));
TY = VFNMS(LDK(KP778434453), TX, T7);
Tx = VFNMS(LDK(KP778434453), Tw, Tg);
Tk = VFNMS(LDK(KP876768831), Tj, Tg);
TR = VFNMS(LDK(KP959492973), TQ, T1);
TI = VFNMS(LDK(KP959492973), TH, T1);
TM = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TL, Tp));
}
}
}
}
{
V TZ, Ty, Tl, T10, Tz;
TZ = VFNMS(LDK(KP876768831), TY, T4);
Ty = VFNMS(LDK(KP876768831), Tx, Ta);
Tl = VFNMS(LDK(KP959492973), Tk, T1);
ST(&(xo[WS(os, 7)]), VFMAI(TV, TR), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VFNMSI(TV, TR), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(TM, TI), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VFNMSI(TM, TI), ovs, &(xo[0]));
T10 = VFNMS(LDK(KP959492973), TZ, T1);
Tz = VFNMS(LDK(KP959492973), Ty, T1);
ST(&(xo[WS(os, 1)]), VFMAI(Tu, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 10)]), VFNMSI(Tu, Tl), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VFMAI(T14, T10), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VFNMSI(T14, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VFMAI(TD, Tz), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VFNMSI(TD, Tz), ovs, &(xo[0]));
}
}
}
}
}
VLEAVE();
}
static void n1fv_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(32, is), MAKE_VOLATILE_STRIDE(32, os)) {
V T7, TU, Tz, TH, Tu, TV, TA, TK, Te, TX, TC, TO, Tl, TY, TD;
V TR;
{
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T3 = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T6 = VADD(T4, T5);
T7 = VSUB(T3, T6);
TU = VSUB(T4, T5);
Tz = VADD(T3, T6);
TH = VSUB(T1, T2);
}
{
V Tq, TJ, Tt, TI;
{
V To, Tp, Tr, Ts;
To = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Tp = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tq = VADD(To, Tp);
TJ = VSUB(To, Tp);
Tr = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Ts = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tt = VADD(Tr, Ts);
TI = VSUB(Tr, Ts);
}
Tu = VSUB(Tq, Tt);
TV = VSUB(TJ, TI);
TA = VADD(Tt, Tq);
TK = VADD(TI, TJ);
}
{
V Ta, TM, Td, TN;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VADD(T8, T9);
TM = VSUB(T8, T9);
Tb = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Td = VADD(Tb, Tc);
TN = VSUB(Tb, Tc);
}
Te = VSUB(Ta, Td);
TX = VFMA(LDK(KP414213562), TM, TN);
TC = VADD(Ta, Td);
TO = VFNMS(LDK(KP414213562), TN, TM);
}
{
V Th, TP, Tk, TQ;
{
V Tf, Tg, Ti, Tj;
Tf = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
Tg = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Th = VADD(Tf, Tg);
TP = VSUB(Tf, Tg);
Ti = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tj = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tk = VADD(Ti, Tj);
TQ = VSUB(Tj, Ti);
}
Tl = VSUB(Th, Tk);
TY = VFMA(LDK(KP414213562), TP, TQ);
TD = VADD(Th, Tk);
TR = VFNMS(LDK(KP414213562), TQ, TP);
}
{
V TB, TE, TF, TG;
TB = VADD(Tz, TA);
TE = VADD(TC, TD);
ST(&(xo[WS(os, 8)]), VSUB(TB, TE), ovs, &(xo[0]));
ST(&(xo[0]), VADD(TB, TE), ovs, &(xo[0]));
TF = VSUB(Tz, TA);
TG = VSUB(TD, TC);
ST(&(xo[WS(os, 12)]), VFNMSI(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VFMAI(TG, TF), ovs, &(xo[0]));
}
{
V Tn, Tx, Tw, Ty, Tm, Tv;
Tm = VADD(Te, Tl);
Tn = VFNMS(LDK(KP707106781), Tm, T7);
Tx = VFMA(LDK(KP707106781), Tm, T7);
Tv = VSUB(Tl, Te);
Tw = VFNMS(LDK(KP707106781), Tv, Tu);
Ty = VFMA(LDK(KP707106781), Tv, Tu);
ST(&(xo[WS(os, 6)]), VFNMSI(Tw, Tn), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VFMAI(Ty, Tx), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VFMAI(Tw, Tn), ovs, &(xo[0]));
ST(&(xo[WS(os, 14)]), VFNMSI(Ty, Tx), ovs, &(xo[0]));
}
{
V TT, T11, T10, T12;
{
V TL, TS, TW, TZ;
TL = VFMA(LDK(KP707106781), TK, TH);
TS = VADD(TO, TR);
TT = VFNMS(LDK(KP923879532), TS, TL);
T11 = VFMA(LDK(KP923879532), TS, TL);
TW = VFNMS(LDK(KP707106781), TV, TU);
TZ = VSUB(TX, TY);
T10 = VFNMS(LDK(KP923879532), TZ, TW);
T12 = VFMA(LDK(KP923879532), TZ, TW);
}
ST(&(xo[WS(os, 9)]), VFNMSI(T10, TT), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 15)]), VFMAI(T12, T11), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VFMAI(T10, TT), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VFNMSI(T12, T11), ovs, &(xo[WS(os, 1)]));
}
{
V T15, T19, T18, T1a;
{
V T13, T14, T16, T17;
T13 = VFNMS(LDK(KP707106781), TK, TH);
T14 = VADD(TX, TY);
T15 = VFNMS(LDK(KP923879532), T14, T13);
T19 = VFMA(LDK(KP923879532), T14, T13);
T16 = VFMA(LDK(KP707106781), TV, TU);
T17 = VSUB(TR, TO);
T18 = VFNMS(LDK(KP923879532), T17, T16);
T1a = VFMA(LDK(KP923879532), T17, T16);
}
ST(&(xo[WS(os, 5)]), VFNMSI(T18, T15), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 13)]), VFNMSI(T1a, T19), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VFMAI(T18, T15), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VFMAI(T1a, T19), ovs, &(xo[WS(os, 1)]));
}
}
}
VLEAVE();
}
static void t3fv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
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(rs)) {
V T1, T2, T3, Ti, T6, T7, Tx, Tb, To;
T1 = LDW(&(W[0]));
T2 = LDW(&(W[TWVL * 2]));
T3 = VZMULJ(T1, T2);
Ti = VZMUL(T1, T2);
T6 = LDW(&(W[TWVL * 4]));
T7 = VZMULJ(T3, T6);
Tx = VZMULJ(Ti, T6);
Tb = VZMULJ(T1, T6);
To = VZMULJ(T2, T6);
{
V TA, TQ, Tn, Tt, Tu, TJ, TK, TS, Ta, Tg, Th, TM, TN, TR, Tw;
V Tz, Ty;
Tw = LD(&(x[0]), ms, &(x[0]));
Ty = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tz = VZMULJ(Tx, Ty);
TA = VSUB(Tw, Tz);
TQ = VADD(Tw, Tz);
{
V Tk, Ts, Tm, Tq;
{
V Tj, Tr, Tl, Tp;
Tj = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tk = VZMULJ(Ti, Tj);
Tr = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Ts = VZMULJ(T1, Tr);
Tl = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tm = VZMULJ(T6, Tl);
Tp = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tq = VZMULJ(To, Tp);
}
Tn = VSUB(Tk, Tm);
Tt = VSUB(Tq, Ts);
Tu = VADD(Tn, Tt);
TJ = VADD(Tk, Tm);
TK = VADD(Tq, Ts);
TS = VADD(TJ, TK);
}
{
V T5, Tf, T9, Td;
{
V T4, Te, T8, Tc;
T4 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T5 = VZMULJ(T3, T4);
Te = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tf = VZMULJ(T2, Te);
T8 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T9 = VZMULJ(T7, T8);
Tc = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Td = VZMULJ(Tb, Tc);
}
Ta = VSUB(T5, T9);
Tg = VSUB(Td, Tf);
Th = VADD(Ta, Tg);
TM = VADD(T5, T9);
TN = VADD(Td, Tf);
TR = VADD(TM, TN);
}
{
V Tv, TB, TC, TG, TI, TE, TF, TH, TD;
Tv = VMUL(LDK(KP559016994), VSUB(Th, Tu));
TB = VADD(Th, Tu);
TC = VFNMS(LDK(KP250000000), TB, TA);
TE = VSUB(Ta, Tg);
TF = VSUB(Tn, Tt);
TG = VBYI(VFMA(LDK(KP951056516), TE, VMUL(LDK(KP587785252), TF)));
TI = VBYI(VFNMS(LDK(KP587785252), TE, VMUL(LDK(KP951056516), TF)));
ST(&(x[WS(rs, 5)]), VADD(TA, TB), ms, &(x[WS(rs, 1)]));
TH = VSUB(TC, Tv);
ST(&(x[WS(rs, 3)]), VSUB(TH, TI), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(TI, TH), ms, &(x[WS(rs, 1)]));
TD = VADD(Tv, TC);
ST(&(x[WS(rs, 1)]), VSUB(TD, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(TG, TD), ms, &(x[WS(rs, 1)]));
}
{
V TV, TT, TU, TP, TX, TL, TO, TY, TW;
TV = VMUL(LDK(KP559016994), VSUB(TR, TS));
TT = VADD(TR, TS);
TU = VFNMS(LDK(KP250000000), TT, TQ);
TL = VSUB(TJ, TK);
TO = VSUB(TM, TN);
TP = VBYI(VFNMS(LDK(KP587785252), TO, VMUL(LDK(KP951056516), TL)));
TX = VBYI(VFMA(LDK(KP951056516), TO, VMUL(LDK(KP587785252), TL)));
ST(&(x[0]), VADD(TQ, TT), ms, &(x[0]));
TY = VADD(TV, TU);
ST(&(x[WS(rs, 4)]), VADD(TX, TY), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VSUB(TY, TX), ms, &(x[0]));
TW = VSUB(TU, TV);
ST(&(x[WS(rs, 2)]), VADD(TP, TW), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TW, TP), ms, &(x[0]));
}
}
}
}
static void n1bv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
{
INT i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V TH, T3, TP, Tn, Ta, Tu, TU, TK, TO, Tk, TM, Tg, TL, Td, T1;
V T2;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
{
V Ti, TI, T6, TJ, T9, Tj, Te, Tf, Tb, Tc;
{
V T4, T5, T7, T8, Tl, Tm;
T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tl = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tm = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Ti = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
TH = VADD(T1, T2);
T3 = VSUB(T1, T2);
TI = VADD(T4, T5);
T6 = VSUB(T4, T5);
TJ = VADD(T7, T8);
T9 = VSUB(T7, T8);
TP = VADD(Tl, Tm);
Tn = VSUB(Tl, Tm);
Tj = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
}
Ta = VADD(T6, T9);
Tu = VSUB(T6, T9);
TU = VSUB(TI, TJ);
TK = VADD(TI, TJ);
TO = VADD(Ti, Tj);
Tk = VSUB(Ti, Tj);
TM = VADD(Te, Tf);
Tg = VSUB(Te, Tf);
TL = VADD(Tb, Tc);
Td = VSUB(Tb, Tc);
}
{
V T13, TG, TY, T18, TB, Tw, TT, Tz, T11, T16, TE, Tr, TV, TQ;
TV = VSUB(TP, TO);
TQ = VADD(TO, TP);
{
V Ts, To, TW, TN;
Ts = VSUB(Tk, Tn);
To = VADD(Tk, Tn);
TW = VSUB(TM, TL);
TN = VADD(TL, TM);
{
V Tt, Th, TR, T12;
Tt = VSUB(Td, Tg);
Th = VADD(Td, Tg);
TR = VFNMS(LDK(KP356895867), TK, TQ);
T12 = VFNMS(LDK(KP554958132), TV, TU);
{
V Tx, TF, TZ, T14;
Tx = VFNMS(LDK(KP356895867), Ta, To);
TF = VFMA(LDK(KP554958132), Ts, Tu);
ST(&(xo[0]), VADD(TH, VADD(TK, VADD(TN, TQ))), ovs, &(xo[0]));
TZ = VFNMS(LDK(KP356895867), TN, TK);
T14 = VFNMS(LDK(KP356895867), TQ, TN);
{
V TX, T17, TC, Tp;
TX = VFMA(LDK(KP554958132), TW, TV);
T17 = VFMA(LDK(KP554958132), TU, TW);
ST(&(xo[WS(os, 7)]), VADD(T3, VADD(Ta, VADD(Th, To))), ovs, &(xo[WS(os, 1)]));
TC = VFNMS(LDK(KP356895867), Th, Ta);
Tp = VFNMS(LDK(KP356895867), To, Th);
{
V TA, Tv, TS, Ty;
TA = VFMA(LDK(KP554958132), Tt, Ts);
Tv = VFNMS(LDK(KP554958132), Tu, Tt);
TS = VFNMS(LDK(KP692021471), TR, TN);
T13 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T12, TW));
Ty = VFNMS(LDK(KP692021471), Tx, Th);
TG = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TF, Tt));
{
V T10, T15, TD, Tq;
T10 = VFNMS(LDK(KP692021471), TZ, TQ);
T15 = VFNMS(LDK(KP692021471), T14, TK);
TY = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TX, TU));
T18 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T17, TV));
TD = VFNMS(LDK(KP692021471), TC, To);
Tq = VFNMS(LDK(KP692021471), Tp, Ta);
TB = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TA, Tu));
Tw = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tv, Ts));
TT = VFNMS(LDK(KP900968867), TS, TH);
Tz = VFNMS(LDK(KP900968867), Ty, T3);
T11 = VFNMS(LDK(KP900968867), T10, TH);
T16 = VFNMS(LDK(KP900968867), T15, TH);
TE = VFNMS(LDK(KP900968867), TD, T3);
Tr = VFNMS(LDK(KP900968867), Tq, T3);
}
}
}
}
}
}
ST(&(xo[WS(os, 2)]), VFMAI(TY, TT), ovs, &(xo[0]));
ST(&(xo[WS(os, 12)]), VFNMSI(TY, TT), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VFMAI(TB, Tz), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VFNMSI(TB, Tz), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VFMAI(T13, T11), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VFNMSI(T13, T11), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VFMAI(T18, T16), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VFNMSI(T18, T16), ovs, &(xo[0]));
ST(&(xo[WS(os, 13)]), VFNMSI(TG, TE), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VFMAI(TG, TE), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VFNMSI(Tw, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VFMAI(Tw, Tr), ovs, &(xo[WS(os, 1)]));
}
}
}
VLEAVE();
}
static void t1fuv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
V Tr, TH, Tg, Tl, Tm, TA, TB, TJ, T5, Ta, Tb, TD, TE, TI, To;
V Tq, Tp;
To = LD(&(x[0]), ms, &(x[0]));
Tp = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tq = BYTWJ(&(W[TWVL * 8]), Tp);
Tr = VSUB(To, Tq);
TH = VADD(To, Tq);
{
V Td, Tk, Tf, Ti;
{
V Tc, Tj, Te, Th;
Tc = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Td = BYTWJ(&(W[TWVL * 6]), Tc);
Tj = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tk = BYTWJ(&(W[0]), Tj);
Te = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tf = BYTWJ(&(W[TWVL * 16]), Te);
Th = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Ti = BYTWJ(&(W[TWVL * 10]), Th);
}
Tg = VSUB(Td, Tf);
Tl = VSUB(Ti, Tk);
Tm = VADD(Tg, Tl);
TA = VADD(Td, Tf);
TB = VADD(Ti, Tk);
TJ = VADD(TA, TB);
}
{
V T2, T9, T4, T7;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T2 = BYTWJ(&(W[TWVL * 2]), T1);
T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = BYTWJ(&(W[TWVL * 4]), T8);
T3 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T4 = BYTWJ(&(W[TWVL * 12]), T3);
T6 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T7 = BYTWJ(&(W[TWVL * 14]), T6);
}
T5 = VSUB(T2, T4);
Ta = VSUB(T7, T9);
Tb = VADD(T5, Ta);
TD = VADD(T2, T4);
TE = VADD(T7, T9);
TI = VADD(TD, TE);
}
{
V Tn, Ts, Tt, Tx, Tz, Tv, Tw, Ty, Tu;
Tn = VMUL(LDK(KP559016994), VSUB(Tb, Tm));
Ts = VADD(Tb, Tm);
Tt = VFNMS(LDK(KP250000000), Ts, Tr);
Tv = VSUB(T5, Ta);
Tw = VSUB(Tg, Tl);
Tx = VBYI(VFMA(LDK(KP951056516), Tv, VMUL(LDK(KP587785252), Tw)));
Tz = VBYI(VFNMS(LDK(KP587785252), Tv, VMUL(LDK(KP951056516), Tw)));
ST(&(x[WS(rs, 5)]), VADD(Tr, Ts), ms, &(x[WS(rs, 1)]));
Ty = VSUB(Tt, Tn);
ST(&(x[WS(rs, 3)]), VSUB(Ty, Tz), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(Tz, Ty), ms, &(x[WS(rs, 1)]));
Tu = VADD(Tn, Tt);
ST(&(x[WS(rs, 1)]), VSUB(Tu, Tx), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(Tx, Tu), ms, &(x[WS(rs, 1)]));
}
{
V TM, TK, TL, TG, TO, TC, TF, TP, TN;
TM = VMUL(LDK(KP559016994), VSUB(TI, TJ));
TK = VADD(TI, TJ);
TL = VFNMS(LDK(KP250000000), TK, TH);
TC = VSUB(TA, TB);
TF = VSUB(TD, TE);
TG = VBYI(VFNMS(LDK(KP587785252), TF, VMUL(LDK(KP951056516), TC)));
TO = VBYI(VFMA(LDK(KP951056516), TF, VMUL(LDK(KP587785252), TC)));
ST(&(x[0]), VADD(TH, TK), ms, &(x[0]));
TP = VADD(TM, TL);
ST(&(x[WS(rs, 4)]), VADD(TO, TP), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VSUB(TP, TO), ms, &(x[0]));
TN = VSUB(TL, TM);
ST(&(x[WS(rs, 2)]), VADD(TG, TN), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TN, TG), ms, &(x[0]));
}
}
}
VLEAVE();
}