static void n2fv_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 = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(24, is), MAKE_VOLATILE_STRIDE(24, os)) {
V T5, Ta, TG, TF, TB, Tt, Ti, Tm, TJ, TI, TA, Tp;
{
V T1, T6, T4, Tr, T9, Ts;
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);
Tr = VSUB(T3, T2);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Ts = VSUB(T8, T7);
}
T5 = VFNMS(LDK(KP500000000), T4, T1);
Ta = VFNMS(LDK(KP500000000), T9, T6);
TG = VADD(T6, T9);
TF = VADD(T1, T4);
TB = VADD(Tr, Ts);
Tt = VSUB(Tr, Ts);
}
{
V Tk, Tn, Te, Tl, Th, To;
Tk = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tn = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Tc, Td, Tf, Tg;
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = VSUB(Tc, Td);
Tl = VADD(Td, Tc);
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);
To = VADD(Tf, Tg);
}
Ti = VADD(Te, Th);
Tm = VFNMS(LDK(KP500000000), Tl, Tk);
TJ = VADD(Tn, To);
TI = VADD(Tk, Tl);
TA = VSUB(Te, Th);
Tp = VFNMS(LDK(KP500000000), To, Tn);
}
{
V TN, TO, TP, TQ, TT, TU;
{
V TH, TK, TL, TM;
TH = VSUB(TF, TG);
TK = VSUB(TI, TJ);
TN = VFNMSI(TK, TH);
STM2(&(xo[18]), TN, ovs, &(xo[2]));
TO = VFMAI(TK, TH);
STM2(&(xo[6]), TO, ovs, &(xo[2]));
TL = VADD(TF, TG);
TM = VADD(TI, TJ);
TP = VSUB(TL, TM);
STM2(&(xo[12]), TP, ovs, &(xo[0]));
TQ = VADD(TL, TM);
STM2(&(xo[0]), TQ, ovs, &(xo[0]));
}
{
V Tj, Tv, Tu, Tw, Tb, Tq, TR, TS;
Tb = VSUB(T5, Ta);
Tj = VFMA(LDK(KP866025403), Ti, Tb);
Tv = VFNMS(LDK(KP866025403), Ti, Tb);
Tq = VSUB(Tm, Tp);
Tu = VFNMS(LDK(KP866025403), Tt, Tq);
Tw = VFMA(LDK(KP866025403), Tt, Tq);
TR = VFNMSI(Tu, Tj);
STM2(&(xo[2]), TR, ovs, &(xo[2]));
STN2(&(xo[0]), TQ, TR, ovs);
TS = VFMAI(Tw, Tv);
STM2(&(xo[14]), TS, ovs, &(xo[2]));
STN2(&(xo[12]), TP, TS, ovs);
TT = VFMAI(Tu, Tj);
STM2(&(xo[22]), TT, ovs, &(xo[2]));
TU = VFNMSI(Tw, Tv);
STM2(&(xo[10]), TU, ovs, &(xo[2]));
}
{
V TC, TE, Tz, TD, Tx, Ty;
TC = VMUL(LDK(KP866025403), VSUB(TA, TB));
TE = VMUL(LDK(KP866025403), VADD(TB, TA));
Tx = VADD(T5, Ta);
Ty = VADD(Tm, Tp);
Tz = VSUB(Tx, Ty);
TD = VADD(Tx, Ty);
{
V TV, TW, TX, TY;
TV = VFMAI(TC, Tz);
STM2(&(xo[4]), TV, ovs, &(xo[0]));
STN2(&(xo[4]), TV, TO, ovs);
TW = VFNMSI(TE, TD);
STM2(&(xo[16]), TW, ovs, &(xo[0]));
STN2(&(xo[16]), TW, TN, ovs);
TX = VFNMSI(TC, Tz);
STM2(&(xo[20]), TX, ovs, &(xo[0]));
STN2(&(xo[20]), TX, TT, ovs);
TY = VFMAI(TE, TD);
STM2(&(xo[8]), TY, ovs, &(xo[0]));
STN2(&(xo[8]), TY, TU, ovs);
}
}
}
}
}
VLEAVE();
}
static void t2bv_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 = ii;
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(rs)) {
V Tu, TH, Tg, Tl, Tp, TD, TE, TJ, T5, Ta, To, TA, TB, TI, Tr;
V Tt, Ts;
Tr = LD(&(x[0]), ms, &(x[0]));
Ts = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tt = BYTW(&(W[TWVL * 8]), Ts);
Tu = VSUB(Tr, Tt);
TH = VADD(Tr, Tt);
{
V Td, Tk, Tf, Ti;
{
V Tc, Tj, Te, Th;
Tc = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Td = BYTW(&(W[TWVL * 6]), Tc);
Tj = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tk = BYTW(&(W[0]), Tj);
Te = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tf = BYTW(&(W[TWVL * 16]), Te);
Th = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Ti = BYTW(&(W[TWVL * 10]), Th);
}
Tg = VSUB(Td, Tf);
Tl = VSUB(Ti, Tk);
Tp = VADD(Tg, Tl);
TD = VADD(Td, Tf);
TE = VADD(Ti, Tk);
TJ = VADD(TD, TE);
}
{
V T2, T9, T4, T7;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T2 = BYTW(&(W[TWVL * 2]), T1);
T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = BYTW(&(W[TWVL * 4]), T8);
T3 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T4 = BYTW(&(W[TWVL * 12]), T3);
T6 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T7 = BYTW(&(W[TWVL * 14]), T6);
}
T5 = VSUB(T2, T4);
Ta = VSUB(T7, T9);
To = VADD(T5, Ta);
TA = VADD(T2, T4);
TB = VADD(T7, T9);
TI = VADD(TA, TB);
}
{
V Tq, Tv, Tw, Tn, Tz, Tb, Tm, Ty, Tx;
Tq = VMUL(LDK(KP559016994), VSUB(To, Tp));
Tv = VADD(To, Tp);
Tw = VFNMS(LDK(KP250000000), Tv, Tu);
Tb = VSUB(T5, Ta);
Tm = VSUB(Tg, Tl);
Tn = VBYI(VFMA(LDK(KP951056516), Tb, VMUL(LDK(KP587785252), Tm)));
Tz = VBYI(VFNMS(LDK(KP951056516), Tm, VMUL(LDK(KP587785252), Tb)));
ST(&(x[WS(rs, 5)]), VADD(Tu, Tv), ms, &(x[WS(rs, 1)]));
Ty = VSUB(Tw, Tq);
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)]));
Tx = VADD(Tq, Tw);
ST(&(x[WS(rs, 1)]), VADD(Tn, Tx), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VSUB(Tx, Tn), ms, &(x[WS(rs, 1)]));
}
{
V TM, TK, TL, TG, TP, TC, TF, TO, 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(KP951056516), TF, VMUL(LDK(KP587785252), TC)));
TP = VBYI(VFMA(LDK(KP951056516), TC, VMUL(LDK(KP587785252), TF)));
ST(&(x[0]), VADD(TH, TK), ms, &(x[0]));
TO = VADD(TM, TL);
ST(&(x[WS(rs, 4)]), VSUB(TO, TP), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VADD(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]));
}
}
}
static const R *t1fv_15(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
DVK(KP216506350, +0.216506350946109661690930792688234045867850657);
DVK(KP484122918, +0.484122918275927110647408174972799951354115213);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP509036960, +0.509036960455127183450980863393907648510733164);
DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 28), MAKE_VOLATILE_STRIDE(ios)) {
V T1e, T7, TP, T12, T15, Tf, Tn, To, T1b, T1c, T1f, TQ, TR, TS, Tw;
V TE, TF, TT, TU, TV;
{
V T1, T5, T3, T4, T2, T6;
T1 = LD(&(x[0]), dist, &(x[0]));
T4 = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 18]), T4);
T2 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[TWVL * 8]), T2);
T1e = VSUB(T5, T3);
T6 = VADD(T3, T5);
T7 = VADD(T1, T6);
TP = VFNMS(LDK(KP500000000), T6, T1);
}
{
V T9, Tq, Ty, Th, Te, T13, Tv, T10, TD, T11, Tm, T14;
{
V T8, Tp, Tx, Tg;
T8 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
T9 = BYTWJ(&(W[TWVL * 4]), T8);
Tp = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
Tq = BYTWJ(&(W[TWVL * 10]), Tp);
Tx = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
Ty = BYTWJ(&(W[TWVL * 16]), Tx);
Tg = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
Th = BYTWJ(&(W[TWVL * 22]), Tg);
}
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
Tb = BYTWJ(&(W[TWVL * 14]), Ta);
Tc = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
Td = BYTWJ(&(W[TWVL * 24]), Tc);
Te = VADD(Tb, Td);
T13 = VSUB(Td, Tb);
}
{
V Ts, Tu, Tr, Tt;
Tr = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
Ts = BYTWJ(&(W[TWVL * 20]), Tr);
Tt = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Tu = BYTWJ(&(W[0]), Tt);
Tv = VADD(Ts, Tu);
T10 = VSUB(Tu, Ts);
}
{
V TA, TC, Tz, TB;
Tz = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
TA = BYTWJ(&(W[TWVL * 26]), Tz);
TB = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
TC = BYTWJ(&(W[TWVL * 6]), TB);
TD = VADD(TA, TC);
T11 = VSUB(TC, TA);
}
{
V Tj, Tl, Ti, Tk;
Ti = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Tj = BYTWJ(&(W[TWVL * 2]), Ti);
Tk = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
Tl = BYTWJ(&(W[TWVL * 12]), Tk);
Tm = VADD(Tj, Tl);
T14 = VSUB(Tl, Tj);
}
T12 = VSUB(T10, T11);
T15 = VSUB(T13, T14);
Tf = VADD(T9, Te);
Tn = VADD(Th, Tm);
To = VADD(Tf, Tn);
T1b = VADD(T13, T14);
T1c = VADD(T10, T11);
T1f = VADD(T1b, T1c);
TQ = VFNMS(LDK(KP500000000), Te, T9);
TR = VFNMS(LDK(KP500000000), Tm, Th);
TS = VADD(TQ, TR);
Tw = VADD(Tq, Tv);
TE = VADD(Ty, TD);
TF = VADD(Tw, TE);
TT = VFNMS(LDK(KP500000000), Tv, Tq);
TU = VFNMS(LDK(KP500000000), TD, Ty);
TV = VADD(TT, TU);
}
{
V TI, TG, TH, TM, TO, TK, TL, TN, TJ;
TI = VMUL(LDK(KP559016994), VSUB(To, TF));
TG = VADD(To, TF);
TH = VFNMS(LDK(KP250000000), TG, T7);
TK = VSUB(Tw, TE);
TL = VSUB(Tf, Tn);
TM = VBYI(VFNMS(LDK(KP587785252), TL, VMUL(LDK(KP951056516), TK)));
TO = VBYI(VFMA(LDK(KP951056516), TL, VMUL(LDK(KP587785252), TK)));
ST(&(x[0]), VADD(T7, TG), dist, &(x[0]));
TN = VADD(TI, TH);
ST(&(x[WS(ios, 6)]), VSUB(TN, TO), dist, &(x[0]));
ST(&(x[WS(ios, 9)]), VADD(TO, TN), dist, &(x[WS(ios, 1)]));
TJ = VSUB(TH, TI);
ST(&(x[WS(ios, 3)]), VSUB(TJ, TM), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 12)]), VADD(TM, TJ), dist, &(x[0]));
}
{
V T16, T1m, T1u, T1h, T1o, T1a, T1p, TZ, T1t, T1l, T1d, T1g;
T16 = VFNMS(LDK(KP509036960), T15, VMUL(LDK(KP823639103), T12));
T1m = VFMA(LDK(KP823639103), T15, VMUL(LDK(KP509036960), T12));
T1u = VBYI(VMUL(LDK(KP866025403), VADD(T1e, T1f)));
T1d = VMUL(LDK(KP484122918), VSUB(T1b, T1c));
T1g = VFNMS(LDK(KP216506350), T1f, VMUL(LDK(KP866025403), T1e));
T1h = VSUB(T1d, T1g);
T1o = VADD(T1d, T1g);
{
V T18, T19, TY, TW, TX;
T18 = VSUB(TT, TU);
T19 = VSUB(TQ, TR);
T1a = VFNMS(LDK(KP587785252), T19, VMUL(LDK(KP951056516), T18));
T1p = VFMA(LDK(KP951056516), T19, VMUL(LDK(KP587785252), T18));
TY = VMUL(LDK(KP559016994), VSUB(TS, TV));
TW = VADD(TS, TV);
TX = VFNMS(LDK(KP250000000), TW, TP);
TZ = VSUB(TX, TY);
T1t = VADD(TP, TW);
T1l = VADD(TY, TX);
}
{
V T17, T1i, T1r, T1s;
ST(&(x[WS(ios, 5)]), VSUB(T1t, T1u), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 10)]), VADD(T1t, T1u), dist, &(x[0]));
T17 = VSUB(TZ, T16);
T1i = VBYI(VSUB(T1a, T1h));
ST(&(x[WS(ios, 8)]), VSUB(T17, T1i), dist, &(x[0]));
ST(&(x[WS(ios, 7)]), VADD(T17, T1i), dist, &(x[WS(ios, 1)]));
T1r = VSUB(T1l, T1m);
T1s = VBYI(VADD(T1p, T1o));
ST(&(x[WS(ios, 11)]), VSUB(T1r, T1s), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 4)]), VADD(T1r, T1s), dist, &(x[0]));
{
V T1n, T1q, T1j, T1k;
T1n = VADD(T1l, T1m);
T1q = VBYI(VSUB(T1o, T1p));
ST(&(x[WS(ios, 14)]), VSUB(T1n, T1q), dist, &(x[0]));
ST(&(x[WS(ios, 1)]), VADD(T1n, T1q), dist, &(x[WS(ios, 1)]));
T1j = VADD(TZ, T16);
T1k = VBYI(VADD(T1a, T1h));
ST(&(x[WS(ios, 13)]), VSUB(T1j, T1k), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 2)]), VADD(T1j, T1k), dist, &(x[0]));
}
}
}
}
return W;
}
static void n2bv_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(KP414213562, +0.414213562373095048801688724209698078569671875);
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 T7, Tu, TF, TB, T13, TL, TO, TX, TC, Te, TP, Th, TQ, Tk, TW;
V T16;
{
V TH, TU, Tz, Tf, TK, TV, TA, TM, Ta, TN, Td, Tg, Ti, Tj;
{
V T1, T2, T4, T5, To, Tp, Tr, Ts;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tp = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tr = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Ts = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T8, TI, Tq, TJ, Tt, T9, Tb, Tc, T3, T6;
T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
TH = VSUB(T1, T2);
T3 = VADD(T1, T2);
TU = VSUB(T4, T5);
T6 = VADD(T4, T5);
TI = VSUB(To, Tp);
Tq = VADD(To, Tp);
TJ = VSUB(Tr, Ts);
Tt = VADD(Tr, Ts);
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tb = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
T7 = VSUB(T3, T6);
Tz = VADD(T3, T6);
Tf = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
TK = VADD(TI, TJ);
TV = VSUB(TI, TJ);
TA = VADD(Tq, Tt);
Tu = VSUB(Tq, Tt);
TM = VSUB(T8, T9);
Ta = VADD(T8, T9);
TN = VSUB(Tb, Tc);
Td = VADD(Tb, Tc);
Tg = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Ti = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tj = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
}
}
TF = VADD(Tz, TA);
TB = VSUB(Tz, TA);
T13 = VFNMS(LDK(KP707106781), TK, TH);
TL = VFMA(LDK(KP707106781), TK, TH);
TO = VFNMS(LDK(KP414213562), TN, TM);
TX = VFMA(LDK(KP414213562), TM, TN);
TC = VADD(Ta, Td);
Te = VSUB(Ta, Td);
TP = VSUB(Tf, Tg);
Th = VADD(Tf, Tg);
TQ = VSUB(Tj, Ti);
Tk = VADD(Ti, Tj);
TW = VFMA(LDK(KP707106781), TV, TU);
T16 = VFNMS(LDK(KP707106781), TV, TU);
}
{
V TY, TR, Tl, TD;
TY = VFMA(LDK(KP414213562), TP, TQ);
TR = VFNMS(LDK(KP414213562), TQ, TP);
Tl = VSUB(Th, Tk);
TD = VADD(Th, Tk);
{
V TS, T17, TZ, T14;
TS = VADD(TO, TR);
T17 = VSUB(TO, TR);
TZ = VSUB(TX, TY);
T14 = VADD(TX, TY);
{
V TE, TG, Tm, Tv;
TE = VSUB(TC, TD);
TG = VADD(TC, TD);
Tm = VADD(Te, Tl);
Tv = VSUB(Te, Tl);
{
V T18, T1a, TT, T11;
T18 = VFMA(LDK(KP923879532), T17, T16);
T1a = VFNMS(LDK(KP923879532), T17, T16);
TT = VFNMS(LDK(KP923879532), TS, TL);
T11 = VFMA(LDK(KP923879532), TS, TL);
{
V T15, T19, T10, T12;
T15 = VFNMS(LDK(KP923879532), T14, T13);
T19 = VFMA(LDK(KP923879532), T14, T13);
T10 = VFNMS(LDK(KP923879532), TZ, TW);
T12 = VFMA(LDK(KP923879532), TZ, TW);
{
V T1b, T1c, T1d, T1e;
T1b = VADD(TF, TG);
STM2(&(xo[0]), T1b, ovs, &(xo[0]));
T1c = VSUB(TF, TG);
STM2(&(xo[16]), T1c, ovs, &(xo[0]));
T1d = VFMAI(TE, TB);
STM2(&(xo[8]), T1d, ovs, &(xo[0]));
T1e = VFNMSI(TE, TB);
STM2(&(xo[24]), T1e, ovs, &(xo[0]));
{
V Tw, Ty, Tn, Tx;
Tw = VFNMS(LDK(KP707106781), Tv, Tu);
Ty = VFMA(LDK(KP707106781), Tv, Tu);
Tn = VFNMS(LDK(KP707106781), Tm, T7);
Tx = VFMA(LDK(KP707106781), Tm, T7);
{
V T1f, T1g, T1h, T1i;
T1f = VFNMSI(T1a, T19);
STM2(&(xo[6]), T1f, ovs, &(xo[2]));
T1g = VFMAI(T1a, T19);
STM2(&(xo[26]), T1g, ovs, &(xo[2]));
STN2(&(xo[24]), T1e, T1g, ovs);
T1h = VFNMSI(T18, T15);
STM2(&(xo[22]), T1h, ovs, &(xo[2]));
T1i = VFMAI(T18, T15);
STM2(&(xo[10]), T1i, ovs, &(xo[2]));
STN2(&(xo[8]), T1d, T1i, ovs);
{
V T1j, T1k, T1l, T1m;
T1j = VFNMSI(T12, T11);
STM2(&(xo[30]), T1j, ovs, &(xo[2]));
T1k = VFMAI(T12, T11);
STM2(&(xo[2]), T1k, ovs, &(xo[2]));
STN2(&(xo[0]), T1b, T1k, ovs);
T1l = VFMAI(T10, TT);
STM2(&(xo[18]), T1l, ovs, &(xo[2]));
STN2(&(xo[16]), T1c, T1l, ovs);
T1m = VFNMSI(T10, TT);
STM2(&(xo[14]), T1m, ovs, &(xo[2]));
{
V T1n, T1o, T1p, T1q;
T1n = VFMAI(Ty, Tx);
STM2(&(xo[4]), T1n, ovs, &(xo[0]));
STN2(&(xo[4]), T1n, T1f, ovs);
T1o = VFNMSI(Ty, Tx);
STM2(&(xo[28]), T1o, ovs, &(xo[0]));
STN2(&(xo[28]), T1o, T1j, ovs);
T1p = VFMAI(Tw, Tn);
STM2(&(xo[20]), T1p, ovs, &(xo[0]));
STN2(&(xo[20]), T1p, T1h, ovs);
T1q = VFNMSI(Tw, Tn);
STM2(&(xo[12]), T1q, ovs, &(xo[0]));
STN2(&(xo[12]), T1q, T1m, ovs);
}
}
}
}
}
}
}
}
}
}
}
}
static void n2bv_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(28, is), MAKE_VOLATILE_STRIDE(28, 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);
}
{
V TR, TS, TU, TV;
TR = VADD(Tp, VADD(Ts, VADD(Tq, Tr)));
STM2(&(xo[14]), TR, ovs, &(xo[2]));
TS = VADD(Ty, VADD(TB, VADD(TE, TH)));
STM2(&(xo[0]), TS, ovs, &(xo[0]));
{
V TT, Tm, Tt, TQ, TP, TW;
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)));
TT = VADD(Tm, Tt);
STM2(&(xo[6]), TT, ovs, &(xo[2]));
TU = VSUB(Tt, Tm);
STM2(&(xo[22]), TU, ovs, &(xo[2]));
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)));
TV = VSUB(TP, TQ);
STM2(&(xo[24]), TV, ovs, &(xo[0]));
TW = VADD(TP, TQ);
STM2(&(xo[4]), TW, ovs, &(xo[0]));
STN2(&(xo[4]), TW, TT, ovs);
}
{
V T10, TM, TI, TZ;
{
V Tu, Tv, TX, TY;
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)));
TX = VADD(Tu, Tv);
STM2(&(xo[2]), TX, ovs, &(xo[2]));
STN2(&(xo[0]), TS, TX, ovs);
TY = VSUB(Tv, Tu);
STM2(&(xo[26]), TY, ovs, &(xo[2]));
STN2(&(xo[24]), TV, TY, ovs);
}
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)));
TZ = VSUB(TI, TM);
STM2(&(xo[12]), TZ, ovs, &(xo[0]));
STN2(&(xo[12]), TZ, TR, ovs);
T10 = VADD(TI, TM);
STM2(&(xo[16]), T10, ovs, &(xo[0]));
{
V T11, TO, TN, T12;
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)));
T11 = VSUB(TN, TO);
STM2(&(xo[8]), T11, ovs, &(xo[0]));
T12 = VADD(TN, TO);
STM2(&(xo[20]), T12, ovs, &(xo[0]));
STN2(&(xo[20]), T12, TU, ovs);
{
V Tx, Tw, T13, T14;
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)));
T13 = VSUB(Tw, Tx);
STM2(&(xo[10]), T13, ovs, &(xo[2]));
STN2(&(xo[8]), T11, T13, ovs);
T14 = VADD(Tx, Tw);
STM2(&(xo[18]), T14, ovs, &(xo[2]));
STN2(&(xo[16]), T10, T14, ovs);
}
}
}
}
}
}
VLEAVE();
}
static void t2sv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + (mb * 6); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 6), MAKE_VOLATILE_STRIDE(rs)) {
V T2, T5, T3, T6, T8, Tc, Tg, Ti, Tl, Tm, Tn, Tz, Tp, Tx;
{
V T4, Tb, T7, Ta;
T2 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 1]));
T3 = LDW(&(W[TWVL * 2]));
T6 = LDW(&(W[TWVL * 3]));
T4 = VMUL(T2, T3);
Tb = VMUL(T5, T3);
T7 = VMUL(T5, T6);
Ta = VMUL(T2, T6);
T8 = VSUB(T4, T7);
Tc = VADD(Ta, Tb);
Tg = VADD(T4, T7);
Ti = VSUB(Ta, Tb);
Tl = LDW(&(W[TWVL * 4]));
Tm = LDW(&(W[TWVL * 5]));
Tn = VFMA(T2, Tl, VMUL(T5, Tm));
Tz = VFNMS(Ti, Tl, VMUL(Tg, Tm));
Tp = VFNMS(T5, Tl, VMUL(T2, Tm));
Tx = VFMA(Tg, Tl, VMUL(Ti, Tm));
}
{
V Tf, T1i, TL, T1d, TJ, T17, TV, TY, Ts, T1j, TO, T1a, TC, T16, TQ;
V TT;
{
V T1, T1c, Te, T1b, T9, Td;
T1 = LD(&(ri[0]), ms, &(ri[0]));
T1c = LD(&(ii[0]), ms, &(ii[0]));
T9 = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
Td = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
Te = VFMA(T8, T9, VMUL(Tc, Td));
T1b = VFNMS(Tc, T9, VMUL(T8, Td));
Tf = VADD(T1, Te);
T1i = VSUB(T1c, T1b);
TL = VSUB(T1, Te);
T1d = VADD(T1b, T1c);
}
{
V TF, TW, TI, TX;
{
V TD, TE, TG, TH;
TD = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
TE = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
TF = VFMA(Tl, TD, VMUL(Tm, TE));
TW = VFNMS(Tm, TD, VMUL(Tl, TE));
TG = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
TH = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
TI = VFMA(T3, TG, VMUL(T6, TH));
TX = VFNMS(T6, TG, VMUL(T3, TH));
}
TJ = VADD(TF, TI);
T17 = VADD(TW, TX);
TV = VSUB(TF, TI);
TY = VSUB(TW, TX);
}
{
V Tk, TM, Tr, TN;
{
V Th, Tj, To, Tq;
Th = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
Tj = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
Tk = VFMA(Tg, Th, VMUL(Ti, Tj));
TM = VFNMS(Ti, Th, VMUL(Tg, Tj));
To = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
Tq = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
Tr = VFMA(Tn, To, VMUL(Tp, Tq));
TN = VFNMS(Tp, To, VMUL(Tn, Tq));
}
Ts = VADD(Tk, Tr);
T1j = VSUB(Tk, Tr);
TO = VSUB(TM, TN);
T1a = VADD(TM, TN);
}
{
V Tw, TR, TB, TS;
{
V Tu, Tv, Ty, TA;
Tu = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
Tv = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
Tw = VFMA(T2, Tu, VMUL(T5, Tv));
TR = VFNMS(T5, Tu, VMUL(T2, Tv));
Ty = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
TA = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
TB = VFMA(Tx, Ty, VMUL(Tz, TA));
TS = VFNMS(Tz, Ty, VMUL(Tx, TA));
}
TC = VADD(Tw, TB);
T16 = VADD(TR, TS);
TQ = VSUB(Tw, TB);
TT = VSUB(TR, TS);
}
{
V Tt, TK, T1f, T1g;
Tt = VADD(Tf, Ts);
TK = VADD(TC, TJ);
ST(&(ri[WS(rs, 4)]), VSUB(Tt, TK), ms, &(ri[0]));
ST(&(ri[0]), VADD(Tt, TK), ms, &(ri[0]));
{
V T19, T1e, T15, T18;
T19 = VADD(T16, T17);
T1e = VADD(T1a, T1d);
ST(&(ii[0]), VADD(T19, T1e), ms, &(ii[0]));
ST(&(ii[WS(rs, 4)]), VSUB(T1e, T19), ms, &(ii[0]));
T15 = VSUB(Tf, Ts);
T18 = VSUB(T16, T17);
ST(&(ri[WS(rs, 6)]), VSUB(T15, T18), ms, &(ri[0]));
ST(&(ri[WS(rs, 2)]), VADD(T15, T18), ms, &(ri[0]));
}
T1f = VSUB(TJ, TC);
T1g = VSUB(T1d, T1a);
ST(&(ii[WS(rs, 2)]), VADD(T1f, T1g), ms, &(ii[0]));
ST(&(ii[WS(rs, 6)]), VSUB(T1g, T1f), ms, &(ii[0]));
{
V T11, T1k, T14, T1h, T12, T13;
T11 = VSUB(TL, TO);
T1k = VSUB(T1i, T1j);
T12 = VSUB(TT, TQ);
T13 = VADD(TV, TY);
T14 = VMUL(LDK(KP707106781), VSUB(T12, T13));
T1h = VMUL(LDK(KP707106781), VADD(T12, T13));
ST(&(ri[WS(rs, 7)]), VSUB(T11, T14), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 5)]), VSUB(T1k, T1h), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 3)]), VADD(T11, T14), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 1)]), VADD(T1h, T1k), ms, &(ii[WS(rs, 1)]));
}
{
V TP, T1m, T10, T1l, TU, TZ;
TP = VADD(TL, TO);
T1m = VADD(T1j, T1i);
TU = VADD(TQ, TT);
TZ = VSUB(TV, TY);
T10 = VMUL(LDK(KP707106781), VADD(TU, TZ));
T1l = VMUL(LDK(KP707106781), VSUB(TZ, TU));
ST(&(ri[WS(rs, 5)]), VSUB(TP, T10), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 7)]), VSUB(T1m, T1l), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 1)]), VADD(TP, T10), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 3)]), VADD(T1l, T1m), ms, &(ii[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
static void t1bv_15(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP216506350, +0.216506350946109661690930792688234045867850657);
DVK(KP484122918, +0.484122918275927110647408174972799951354115213);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP509036960, +0.509036960455127183450980863393907648510733164);
DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 28)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 28), MAKE_VOLATILE_STRIDE(rs)) {
V Ts, TV, T1f, TZ, T10, Tb, Tm, Tt, T1j, T1k, T1l, TI, TM, TR, Tz;
V TD, TQ, T1g, T1h, T1i;
{
V TT, Tr, Tp, Tq, To, TU;
TT = LD(&(x[0]), ms, &(x[0]));
Tq = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
Tr = BYTW(&(W[TWVL * 18]), Tq);
To = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tp = BYTW(&(W[TWVL * 8]), To);
Ts = VSUB(Tp, Tr);
TU = VADD(Tp, Tr);
TV = VFNMS(LDK(KP500000000), TU, TT);
T1f = VADD(TT, TU);
}
{
V Tx, TG, TK, TB, T5, Ty, Tg, TH, Tl, TL, Ta, TC;
{
V Tw, TF, TJ, TA;
Tw = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tx = BYTW(&(W[TWVL * 4]), Tw);
TF = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
TG = BYTW(&(W[TWVL * 10]), TF);
TJ = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
TK = BYTW(&(W[TWVL * 16]), TJ);
TA = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TB = BYTW(&(W[TWVL * 22]), TA);
}
{
V T2, T4, T1, T3;
T1 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T2 = BYTW(&(W[TWVL * 14]), T1);
T3 = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
T4 = BYTW(&(W[TWVL * 24]), T3);
T5 = VSUB(T2, T4);
Ty = VADD(T2, T4);
}
{
V Td, Tf, Tc, Te;
Tc = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Td = BYTW(&(W[TWVL * 20]), Tc);
Te = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tf = BYTW(&(W[0]), Te);
Tg = VSUB(Td, Tf);
TH = VADD(Td, Tf);
}
{
V Ti, Tk, Th, Tj;
Th = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
Ti = BYTW(&(W[TWVL * 26]), Th);
Tj = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tk = BYTW(&(W[TWVL * 6]), Tj);
Tl = VSUB(Ti, Tk);
TL = VADD(Ti, Tk);
}
{
V T7, T9, T6, T8;
T6 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T7 = BYTW(&(W[TWVL * 2]), T6);
T8 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T9 = BYTW(&(W[TWVL * 12]), T8);
Ta = VSUB(T7, T9);
TC = VADD(T7, T9);
}
TZ = VSUB(T5, Ta);
T10 = VSUB(Tg, Tl);
Tb = VADD(T5, Ta);
Tm = VADD(Tg, Tl);
Tt = VADD(Tb, Tm);
T1j = VADD(TG, TH);
T1k = VADD(TK, TL);
T1l = VADD(T1j, T1k);
TI = VFNMS(LDK(KP500000000), TH, TG);
TM = VFNMS(LDK(KP500000000), TL, TK);
TR = VADD(TI, TM);
Tz = VFNMS(LDK(KP500000000), Ty, Tx);
TD = VFNMS(LDK(KP500000000), TC, TB);
TQ = VADD(Tz, TD);
T1g = VADD(Tx, Ty);
T1h = VADD(TB, TC);
T1i = VADD(T1g, T1h);
}
{
V T1o, T1m, T1n, T1s, T1t, T1q, T1r, T1u, T1p;
T1o = VMUL(LDK(KP559016994), VSUB(T1i, T1l));
T1m = VADD(T1i, T1l);
T1n = VFNMS(LDK(KP250000000), T1m, T1f);
T1q = VSUB(T1g, T1h);
T1r = VSUB(T1j, T1k);
T1s = VBYI(VFNMS(LDK(KP951056516), T1r, VMUL(LDK(KP587785252), T1q)));
T1t = VBYI(VFMA(LDK(KP951056516), T1q, VMUL(LDK(KP587785252), T1r)));
ST(&(x[0]), VADD(T1f, T1m), ms, &(x[0]));
T1u = VADD(T1o, T1n);
ST(&(x[WS(rs, 6)]), VADD(T1t, T1u), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VSUB(T1u, T1t), ms, &(x[WS(rs, 1)]));
T1p = VSUB(T1n, T1o);
ST(&(x[WS(rs, 3)]), VSUB(T1p, T1s), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 12)]), VADD(T1s, T1p), ms, &(x[0]));
}
{
V T11, T18, T1e, TO, T16, Tv, T15, TY, T1d, T19, TE, TN;
T11 = VFMA(LDK(KP823639103), TZ, VMUL(LDK(KP509036960), T10));
T18 = VFNMS(LDK(KP823639103), T10, VMUL(LDK(KP509036960), TZ));
T1e = VBYI(VMUL(LDK(KP866025403), VADD(Ts, Tt)));
TE = VSUB(Tz, TD);
TN = VSUB(TI, TM);
TO = VFMA(LDK(KP951056516), TE, VMUL(LDK(KP587785252), TN));
T16 = VFNMS(LDK(KP951056516), TN, VMUL(LDK(KP587785252), TE));
{
V Tn, Tu, TS, TW, TX;
Tn = VMUL(LDK(KP484122918), VSUB(Tb, Tm));
Tu = VFNMS(LDK(KP216506350), Tt, VMUL(LDK(KP866025403), Ts));
Tv = VADD(Tn, Tu);
T15 = VSUB(Tn, Tu);
TS = VMUL(LDK(KP559016994), VSUB(TQ, TR));
TW = VADD(TQ, TR);
TX = VFNMS(LDK(KP250000000), TW, TV);
TY = VADD(TS, TX);
T1d = VADD(TV, TW);
T19 = VSUB(TX, TS);
}
{
V TP, T12, T1b, T1c;
ST(&(x[WS(rs, 5)]), VSUB(T1d, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 10)]), VADD(T1e, T1d), ms, &(x[0]));
TP = VBYI(VADD(Tv, TO));
T12 = VSUB(TY, T11);
ST(&(x[WS(rs, 1)]), VADD(TP, T12), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 14)]), VSUB(T12, TP), ms, &(x[0]));
T1b = VBYI(VSUB(T16, T15));
T1c = VSUB(T19, T18);
ST(&(x[WS(rs, 7)]), VADD(T1b, T1c), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 8)]), VSUB(T1c, T1b), ms, &(x[0]));
{
V T17, T1a, T13, T14;
T17 = VBYI(VADD(T15, T16));
T1a = VADD(T18, T19);
ST(&(x[WS(rs, 2)]), VADD(T17, T1a), ms, &(x[0]));
ST(&(x[WS(rs, 13)]), VSUB(T1a, T17), ms, &(x[WS(rs, 1)]));
T13 = VBYI(VSUB(Tv, TO));
T14 = VADD(T11, TY);
ST(&(x[WS(rs, 4)]), VADD(T13, T14), ms, &(x[0]));
ST(&(x[WS(rs, 11)]), VSUB(T14, T13), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
void SHIP_cell_pos(struct actor_t* a, int i, int j, vec3f pos)
{
random_ship_t* s = (random_ship_t*)a->child;
VMOV3(pos, 1.0f*(i - (s->w >> 1)) , 1.0f*j, 0.0f);
VADD(pos, a->pos, pos);
}
static const R *t1bv_32(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
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);
int i;
R *x;
x = ii;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 62)) {
V T4, T1D, T2P, T3h, Tf, T1y, T2K, T3i, TC, T1w, T2G, T3e, Tr, T1v, T2D;
V T3d, T1k, T20, T2y, T3a, T1r, T21, T2v, T39, TV, T1X, T2r, T37, T12, T1Y;
V T2o, T36;
{
V T1, T1C, T3, T1A, T1B, T2, T1z, T2N, T2O;
T1 = LD(&(x[0]), dist, &(x[0]));
T1B = LD(&(x[WS(ios, 24)]), dist, &(x[0]));
T1C = BYTW(&(W[TWVL * 46]), T1B);
T2 = LD(&(x[WS(ios, 16)]), dist, &(x[0]));
T3 = BYTW(&(W[TWVL * 30]), T2);
T1z = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
T1A = BYTW(&(W[TWVL * 14]), T1z);
T4 = VSUB(T1, T3);
T1D = VSUB(T1A, T1C);
T2N = VADD(T1, T3);
T2O = VADD(T1A, T1C);
T2P = VSUB(T2N, T2O);
T3h = VADD(T2N, T2O);
}
{
V T6, Td, T8, Tb;
{
V T5, Tc, T7, Ta;
T5 = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
T6 = BYTW(&(W[TWVL * 6]), T5);
Tc = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
Td = BYTW(&(W[TWVL * 22]), Tc);
T7 = LD(&(x[WS(ios, 20)]), dist, &(x[0]));
T8 = BYTW(&(W[TWVL * 38]), T7);
Ta = LD(&(x[WS(ios, 28)]), dist, &(x[0]));
Tb = BYTW(&(W[TWVL * 54]), Ta);
}
{
V T9, Te, T2I, T2J;
T9 = VSUB(T6, T8);
Te = VSUB(Tb, Td);
Tf = VMUL(LDK(KP707106781), VADD(T9, Te));
T1y = VMUL(LDK(KP707106781), VSUB(T9, Te));
T2I = VADD(T6, T8);
T2J = VADD(Tb, Td);
T2K = VSUB(T2I, T2J);
T3i = VADD(T2I, T2J);
}
}
{
V Tt, TA, Tv, Ty;
{
V Ts, Tz, Tu, Tx;
Ts = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
Tt = BYTW(&(W[TWVL * 10]), Ts);
Tz = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
TA = BYTW(&(W[TWVL * 26]), Tz);
Tu = LD(&(x[WS(ios, 22)]), dist, &(x[0]));
Tv = BYTW(&(W[TWVL * 42]), Tu);
Tx = LD(&(x[WS(ios, 30)]), dist, &(x[0]));
Ty = BYTW(&(W[TWVL * 58]), Tx);
}
{
V Tw, TB, T2E, T2F;
Tw = VSUB(Tt, Tv);
TB = VSUB(Ty, TA);
TC = VFNMS(LDK(KP382683432), TB, VMUL(LDK(KP923879532), Tw));
T1w = VFMA(LDK(KP923879532), TB, VMUL(LDK(KP382683432), Tw));
T2E = VADD(Ty, TA);
T2F = VADD(Tt, Tv);
T2G = VSUB(T2E, T2F);
T3e = VADD(T2E, T2F);
}
}
{
V Ti, Tp, Tk, Tn;
{
V Th, To, Tj, Tm;
Th = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Ti = BYTW(&(W[TWVL * 2]), Th);
To = LD(&(x[WS(ios, 26)]), dist, &(x[0]));
Tp = BYTW(&(W[TWVL * 50]), To);
Tj = LD(&(x[WS(ios, 18)]), dist, &(x[0]));
Tk = BYTW(&(W[TWVL * 34]), Tj);
Tm = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
Tn = BYTW(&(W[TWVL * 18]), Tm);
}
{
V Tl, Tq, T2B, T2C;
Tl = VSUB(Ti, Tk);
Tq = VSUB(Tn, Tp);
Tr = VFMA(LDK(KP382683432), Tl, VMUL(LDK(KP923879532), Tq));
T1v = VFNMS(LDK(KP382683432), Tq, VMUL(LDK(KP923879532), Tl));
T2B = VADD(Ti, Tk);
T2C = VADD(Tn, Tp);
T2D = VSUB(T2B, T2C);
T3d = VADD(T2B, T2C);
}
}
{
V T1g, T1i, T1o, T1m, T1a, T1c, T1d, T15, T17, T18;
{
V T1f, T1h, T1n, T1l;
T1f = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
T1g = BYTW(&(W[TWVL * 12]), T1f);
T1h = LD(&(x[WS(ios, 23)]), dist, &(x[WS(ios, 1)]));
T1i = BYTW(&(W[TWVL * 44]), T1h);
T1n = LD(&(x[WS(ios, 15)]), dist, &(x[WS(ios, 1)]));
T1o = BYTW(&(W[TWVL * 28]), T1n);
T1l = LD(&(x[WS(ios, 31)]), dist, &(x[WS(ios, 1)]));
T1m = BYTW(&(W[TWVL * 60]), T1l);
{
V T19, T1b, T14, T16;
T19 = LD(&(x[WS(ios, 27)]), dist, &(x[WS(ios, 1)]));
T1a = BYTW(&(W[TWVL * 52]), T19);
T1b = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
T1c = BYTW(&(W[TWVL * 20]), T1b);
T1d = VSUB(T1a, T1c);
T14 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
T15 = BYTW(&(W[TWVL * 4]), T14);
T16 = LD(&(x[WS(ios, 19)]), dist, &(x[WS(ios, 1)]));
T17 = BYTW(&(W[TWVL * 36]), T16);
T18 = VSUB(T15, T17);
}
}
{
V T1e, T1j, T2w, T2x;
T1e = VMUL(LDK(KP707106781), VSUB(T18, T1d));
T1j = VSUB(T1g, T1i);
T1k = VSUB(T1e, T1j);
T20 = VADD(T1j, T1e);
T2w = VADD(T15, T17);
T2x = VADD(T1a, T1c);
T2y = VSUB(T2w, T2x);
T3a = VADD(T2w, T2x);
}
{
V T1p, T1q, T2t, T2u;
T1p = VSUB(T1m, T1o);
T1q = VMUL(LDK(KP707106781), VADD(T18, T1d));
T1r = VSUB(T1p, T1q);
T21 = VADD(T1p, T1q);
T2t = VADD(T1m, T1o);
T2u = VADD(T1g, T1i);
T2v = VSUB(T2t, T2u);
T39 = VADD(T2t, T2u);
}
}
{
V TR, TT, TZ, TX, TL, TN, TO, TG, TI, TJ;
{
V TQ, TS, TY, TW;
TQ = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
TR = BYTW(&(W[TWVL * 16]), TQ);
TS = LD(&(x[WS(ios, 25)]), dist, &(x[WS(ios, 1)]));
TT = BYTW(&(W[TWVL * 48]), TS);
TY = LD(&(x[WS(ios, 17)]), dist, &(x[WS(ios, 1)]));
TZ = BYTW(&(W[TWVL * 32]), TY);
TW = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
TX = BYTW(&(W[0]), TW);
{
V TK, TM, TF, TH;
TK = LD(&(x[WS(ios, 29)]), dist, &(x[WS(ios, 1)]));
TL = BYTW(&(W[TWVL * 56]), TK);
TM = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
TN = BYTW(&(W[TWVL * 24]), TM);
TO = VSUB(TL, TN);
TF = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
TG = BYTW(&(W[TWVL * 8]), TF);
TH = LD(&(x[WS(ios, 21)]), dist, &(x[WS(ios, 1)]));
TI = BYTW(&(W[TWVL * 40]), TH);
TJ = VSUB(TG, TI);
}
}
{
V TP, TU, T2p, T2q;
TP = VMUL(LDK(KP707106781), VSUB(TJ, TO));
TU = VSUB(TR, TT);
TV = VSUB(TP, TU);
T1X = VADD(TU, TP);
T2p = VADD(TG, TI);
T2q = VADD(TL, TN);
T2r = VSUB(T2p, T2q);
T37 = VADD(T2p, T2q);
}
{
V T10, T11, T2m, T2n;
T10 = VSUB(TX, TZ);
T11 = VMUL(LDK(KP707106781), VADD(TJ, TO));
T12 = VSUB(T10, T11);
T1Y = VADD(T10, T11);
T2m = VADD(TX, TZ);
T2n = VADD(TR, TT);
T2o = VSUB(T2m, T2n);
T36 = VADD(T2m, T2n);
}
}
{
V T3q, T3u, T3t, T3v;
{
V T3o, T3p, T3r, T3s;
T3o = VADD(T3h, T3i);
T3p = VADD(T3d, T3e);
T3q = VSUB(T3o, T3p);
T3u = VADD(T3o, T3p);
T3r = VADD(T36, T37);
T3s = VADD(T39, T3a);
T3t = VBYI(VSUB(T3r, T3s));
T3v = VADD(T3r, T3s);
}
ST(&(x[WS(ios, 24)]), VSUB(T3q, T3t), dist, &(x[0]));
ST(&(x[0]), VADD(T3u, T3v), dist, &(x[0]));
ST(&(x[WS(ios, 8)]), VADD(T3q, T3t), dist, &(x[0]));
ST(&(x[WS(ios, 16)]), VSUB(T3u, T3v), dist, &(x[0]));
}
{
V T3f, T3j, T3c, T3k, T38, T3b;
T3f = VSUB(T3d, T3e);
T3j = VSUB(T3h, T3i);
T38 = VSUB(T36, T37);
T3b = VSUB(T39, T3a);
T3c = VMUL(LDK(KP707106781), VSUB(T38, T3b));
T3k = VMUL(LDK(KP707106781), VADD(T38, T3b));
{
V T3g, T3l, T3m, T3n;
T3g = VBYI(VSUB(T3c, T3f));
T3l = VSUB(T3j, T3k);
ST(&(x[WS(ios, 12)]), VADD(T3g, T3l), dist, &(x[0]));
ST(&(x[WS(ios, 20)]), VSUB(T3l, T3g), dist, &(x[0]));
T3m = VBYI(VADD(T3f, T3c));
T3n = VADD(T3j, T3k);
ST(&(x[WS(ios, 4)]), VADD(T3m, T3n), dist, &(x[0]));
ST(&(x[WS(ios, 28)]), VSUB(T3n, T3m), dist, &(x[0]));
}
}
{
V T2L, T31, T2R, T2Y, T2A, T2Z, T2U, T32, T2H, T2Q;
T2H = VMUL(LDK(KP707106781), VSUB(T2D, T2G));
T2L = VSUB(T2H, T2K);
T31 = VADD(T2K, T2H);
T2Q = VMUL(LDK(KP707106781), VADD(T2D, T2G));
T2R = VSUB(T2P, T2Q);
T2Y = VADD(T2P, T2Q);
{
V T2s, T2z, T2S, T2T;
T2s = VFNMS(LDK(KP382683432), T2r, VMUL(LDK(KP923879532), T2o));
T2z = VFMA(LDK(KP923879532), T2v, VMUL(LDK(KP382683432), T2y));
T2A = VSUB(T2s, T2z);
T2Z = VADD(T2s, T2z);
T2S = VFMA(LDK(KP382683432), T2o, VMUL(LDK(KP923879532), T2r));
T2T = VFNMS(LDK(KP382683432), T2v, VMUL(LDK(KP923879532), T2y));
T2U = VSUB(T2S, T2T);
T32 = VADD(T2S, T2T);
}
{
V T2M, T2V, T34, T35;
T2M = VBYI(VSUB(T2A, T2L));
T2V = VSUB(T2R, T2U);
ST(&(x[WS(ios, 10)]), VADD(T2M, T2V), dist, &(x[0]));
ST(&(x[WS(ios, 22)]), VSUB(T2V, T2M), dist, &(x[0]));
T34 = VSUB(T2Y, T2Z);
T35 = VBYI(VSUB(T32, T31));
ST(&(x[WS(ios, 18)]), VSUB(T34, T35), dist, &(x[0]));
ST(&(x[WS(ios, 14)]), VADD(T34, T35), dist, &(x[0]));
}
{
V T2W, T2X, T30, T33;
T2W = VBYI(VADD(T2L, T2A));
T2X = VADD(T2R, T2U);
ST(&(x[WS(ios, 6)]), VADD(T2W, T2X), dist, &(x[0]));
ST(&(x[WS(ios, 26)]), VSUB(T2X, T2W), dist, &(x[0]));
T30 = VADD(T2Y, T2Z);
T33 = VBYI(VADD(T31, T32));
ST(&(x[WS(ios, 30)]), VSUB(T30, T33), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VADD(T30, T33), dist, &(x[0]));
}
}
{
V TE, T1P, T1I, T1Q, T1t, T1M, T1F, T1N;
{
V Tg, TD, T1G, T1H;
Tg = VSUB(T4, Tf);
TD = VSUB(Tr, TC);
TE = VSUB(Tg, TD);
T1P = VADD(Tg, TD);
T1G = VFNMS(LDK(KP555570233), TV, VMUL(LDK(KP831469612), T12));
T1H = VFMA(LDK(KP555570233), T1k, VMUL(LDK(KP831469612), T1r));
T1I = VSUB(T1G, T1H);
T1Q = VADD(T1G, T1H);
}
{
V T13, T1s, T1x, T1E;
T13 = VFMA(LDK(KP831469612), TV, VMUL(LDK(KP555570233), T12));
T1s = VFNMS(LDK(KP555570233), T1r, VMUL(LDK(KP831469612), T1k));
T1t = VSUB(T13, T1s);
T1M = VADD(T13, T1s);
T1x = VSUB(T1v, T1w);
T1E = VSUB(T1y, T1D);
T1F = VSUB(T1x, T1E);
T1N = VADD(T1E, T1x);
}
{
V T1u, T1J, T1S, T1T;
T1u = VADD(TE, T1t);
T1J = VBYI(VADD(T1F, T1I));
ST(&(x[WS(ios, 27)]), VSUB(T1u, T1J), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VADD(T1u, T1J), dist, &(x[WS(ios, 1)]));
T1S = VBYI(VADD(T1N, T1M));
T1T = VADD(T1P, T1Q);
ST(&(x[WS(ios, 3)]), VADD(T1S, T1T), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 29)]), VSUB(T1T, T1S), dist, &(x[WS(ios, 1)]));
}
{
V T1K, T1L, T1O, T1R;
T1K = VSUB(TE, T1t);
T1L = VBYI(VSUB(T1I, T1F));
ST(&(x[WS(ios, 21)]), VSUB(T1K, T1L), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 11)]), VADD(T1K, T1L), dist, &(x[WS(ios, 1)]));
T1O = VBYI(VSUB(T1M, T1N));
T1R = VSUB(T1P, T1Q);
ST(&(x[WS(ios, 13)]), VADD(T1O, T1R), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 19)]), VSUB(T1R, T1O), dist, &(x[WS(ios, 1)]));
}
}
{
V T1W, T2h, T2a, T2i, T23, T2e, T27, T2f;
{
V T1U, T1V, T28, T29;
T1U = VADD(T4, Tf);
T1V = VADD(T1v, T1w);
T1W = VSUB(T1U, T1V);
T2h = VADD(T1U, T1V);
T28 = VFNMS(LDK(KP195090322), T1X, VMUL(LDK(KP980785280), T1Y));
T29 = VFMA(LDK(KP195090322), T20, VMUL(LDK(KP980785280), T21));
T2a = VSUB(T28, T29);
T2i = VADD(T28, T29);
}
{
V T1Z, T22, T25, T26;
T1Z = VFMA(LDK(KP980785280), T1X, VMUL(LDK(KP195090322), T1Y));
T22 = VFNMS(LDK(KP195090322), T21, VMUL(LDK(KP980785280), T20));
T23 = VSUB(T1Z, T22);
T2e = VADD(T1Z, T22);
T25 = VADD(Tr, TC);
T26 = VADD(T1D, T1y);
T27 = VSUB(T25, T26);
T2f = VADD(T26, T25);
}
{
V T24, T2b, T2k, T2l;
T24 = VADD(T1W, T23);
T2b = VBYI(VADD(T27, T2a));
ST(&(x[WS(ios, 25)]), VSUB(T24, T2b), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VADD(T24, T2b), dist, &(x[WS(ios, 1)]));
T2k = VBYI(VADD(T2f, T2e));
T2l = VADD(T2h, T2i);
ST(&(x[WS(ios, 1)]), VADD(T2k, T2l), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 31)]), VSUB(T2l, T2k), dist, &(x[WS(ios, 1)]));
}
{
V T2c, T2d, T2g, T2j;
T2c = VSUB(T1W, T23);
T2d = VBYI(VSUB(T2a, T27));
ST(&(x[WS(ios, 23)]), VSUB(T2c, T2d), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 9)]), VADD(T2c, T2d), dist, &(x[WS(ios, 1)]));
T2g = VBYI(VSUB(T2e, T2f));
T2j = VSUB(T2h, T2i);
ST(&(x[WS(ios, 15)]), VADD(T2g, T2j), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 17)]), VSUB(T2j, T2g), dist, &(x[WS(ios, 1)]));
}
}
}
END_SIMD();
return W;
}
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(18, is), MAKE_VOLATILE_STRIDE(18, 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]));
}
}
}
}
}
}
}
VLEAVE();
}
static void n1bv_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(KP654860733, +0.654860733945285064056925072466293553183791199);
DVK(KP142314838, +0.142314838273285140443792668616369668791051361);
DVK(KP415415013, +0.415415013001886425529274149229623203524004910);
DVK(KP841253532, +0.841253532831181168861811648919367717513292498);
DVK(KP540640817, +0.540640817455597582107635954318691695431770608);
DVK(KP909631995, +0.909631995354518371411715383079028460060241051);
DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
DVK(KP755749574, +0.755749574354258283774035843972344420179717445);
DVK(KP281732556, +0.281732556841429697711417915346616899035777899);
int i;
const R *xi;
R *xo;
xi = ii;
xo = io;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V Th, T3, Tm, Tf, Ti, Tc, Tj, T9, Tk, T6, Tl, Ta, Tb, Ts, Tt;
Th = LD(&(xi[0]), ivs, &(xi[0]));
{
V T1, T2, Td, Te;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
Tm = VADD(T1, T2);
Td = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tf = VSUB(Td, Te);
Ti = VADD(Td, Te);
}
Ta = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tb = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tc = VSUB(Ta, Tb);
Tj = VADD(Ta, Tb);
{
V T7, T8, T4, T5;
T7 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
Tk = VADD(T7, T8);
T4 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
Tl = VADD(T4, T5);
}
ST(&(xo[0]), VADD(Th, VADD(Tm, VADD(Ti, VADD(Tl, VADD(Tj, Tk))))), ovs, &(xo[0]));
{
V Tg, Tn, Tu, Tv;
Tg = VBYI(VFMA(LDK(KP281732556), T3, VFMA(LDK(KP755749574), T6, VFNMS(LDK(KP909631995), Tc, VFNMS(LDK(KP540640817), Tf, VMUL(LDK(KP989821441), T9))))));
Tn = VFMA(LDK(KP841253532), Ti, VFMA(LDK(KP415415013), Tj, VFNMS(LDK(KP142314838), Tk, VFNMS(LDK(KP654860733), Tl, VFNMS(LDK(KP959492973), Tm, Th)))));
ST(&(xo[WS(os, 5)]), VADD(Tg, Tn), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VSUB(Tn, Tg), ovs, &(xo[0]));
Tu = VBYI(VFMA(LDK(KP755749574), T3, VFMA(LDK(KP540640817), T6, VFNMS(LDK(KP909631995), T9, VFNMS(LDK(KP989821441), Tf, VMUL(LDK(KP281732556), Tc))))));
Tv = VFMA(LDK(KP841253532), Tl, VFMA(LDK(KP415415013), Tk, VFNMS(LDK(KP959492973), Tj, VFNMS(LDK(KP142314838), Ti, VFNMS(LDK(KP654860733), Tm, Th)))));
ST(&(xo[WS(os, 4)]), VADD(Tu, Tv), ovs, &(xo[0]));
ST(&(xo[WS(os, 7)]), VSUB(Tv, Tu), ovs, &(xo[WS(os, 1)]));
}
Ts = VBYI(VFMA(LDK(KP909631995), T3, VFNMS(LDK(KP540640817), T9, VFNMS(LDK(KP989821441), Tc, VFNMS(LDK(KP281732556), T6, VMUL(LDK(KP755749574), Tf))))));
Tt = VFMA(LDK(KP415415013), Tm, VFMA(LDK(KP841253532), Tk, VFNMS(LDK(KP142314838), Tj, VFNMS(LDK(KP959492973), Tl, VFNMS(LDK(KP654860733), Ti, Th)))));
ST(&(xo[WS(os, 2)]), VADD(Ts, Tt), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VSUB(Tt, Ts), ovs, &(xo[WS(os, 1)]));
{
V Tq, Tr, To, Tp;
Tq = VBYI(VFMA(LDK(KP540640817), T3, VFMA(LDK(KP909631995), Tf, VFMA(LDK(KP989821441), T6, VFMA(LDK(KP755749574), Tc, VMUL(LDK(KP281732556), T9))))));
Tr = VFMA(LDK(KP841253532), Tm, VFMA(LDK(KP415415013), Ti, VFNMS(LDK(KP959492973), Tk, VFNMS(LDK(KP654860733), Tj, VFNMS(LDK(KP142314838), Tl, Th)))));
ST(&(xo[WS(os, 1)]), VADD(Tq, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 10)]), VSUB(Tr, Tq), ovs, &(xo[0]));
To = VBYI(VFMA(LDK(KP989821441), T3, VFMA(LDK(KP540640817), Tc, VFNMS(LDK(KP909631995), T6, VFNMS(LDK(KP281732556), Tf, VMUL(LDK(KP755749574), T9))))));
Tp = VFMA(LDK(KP415415013), Tl, VFMA(LDK(KP841253532), Tj, VFNMS(LDK(KP654860733), Tk, VFNMS(LDK(KP959492973), Ti, VFNMS(LDK(KP142314838), Tm, Th)))));
ST(&(xo[WS(os, 3)]), VADD(To, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VSUB(Tp, To), ovs, &(xo[0]));
}
}
END_SIMD();
}
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(18, is), MAKE_VOLATILE_STRIDE(18, 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)]));
}
}
}
}
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(32, is), MAKE_VOLATILE_STRIDE(32, os)) {
V T3, Tn, Ti, TC, T6, TB, Tl, To, Td, TN, Tz, TH, Ta, TM, Tu;
V TG;
{
V T1, T2, Tj, Tk;
T1 = LD(&(ri[0]), ivs, &(ri[0]));
T2 = LD(&(ri[WS(is, 4)]), ivs, &(ri[0]));
T3 = VADD(T1, T2);
Tn = VSUB(T1, T2);
{
V Tg, Th, T4, T5;
Tg = LD(&(ii[0]), ivs, &(ii[0]));
Th = LD(&(ii[WS(is, 4)]), ivs, &(ii[0]));
Ti = VADD(Tg, Th);
TC = VSUB(Tg, Th);
T4 = LD(&(ri[WS(is, 2)]), ivs, &(ri[0]));
T5 = LD(&(ri[WS(is, 6)]), ivs, &(ri[0]));
T6 = VADD(T4, T5);
TB = VSUB(T4, T5);
}
Tj = LD(&(ii[WS(is, 2)]), ivs, &(ii[0]));
Tk = LD(&(ii[WS(is, 6)]), ivs, &(ii[0]));
Tl = VADD(Tj, Tk);
To = VSUB(Tj, Tk);
{
V Tb, Tc, Tv, Tw, Tx, Ty;
Tb = LD(&(ri[WS(is, 7)]), ivs, &(ri[WS(is, 1)]));
Tc = LD(&(ri[WS(is, 3)]), ivs, &(ri[WS(is, 1)]));
Tv = VSUB(Tb, Tc);
Tw = LD(&(ii[WS(is, 7)]), ivs, &(ii[WS(is, 1)]));
Tx = LD(&(ii[WS(is, 3)]), ivs, &(ii[WS(is, 1)]));
Ty = VSUB(Tw, Tx);
Td = VADD(Tb, Tc);
TN = VADD(Tw, Tx);
Tz = VSUB(Tv, Ty);
TH = VADD(Tv, Ty);
}
{
V T8, T9, Tq, Tr, Ts, Tt;
T8 = LD(&(ri[WS(is, 1)]), ivs, &(ri[WS(is, 1)]));
T9 = LD(&(ri[WS(is, 5)]), ivs, &(ri[WS(is, 1)]));
Tq = VSUB(T8, T9);
Tr = LD(&(ii[WS(is, 1)]), ivs, &(ii[WS(is, 1)]));
Ts = LD(&(ii[WS(is, 5)]), ivs, &(ii[WS(is, 1)]));
Tt = VSUB(Tr, Ts);
Ta = VADD(T8, T9);
TM = VADD(Tr, Ts);
Tu = VADD(Tq, Tt);
TG = VSUB(Tt, Tq);
}
}
{
V TR, TS, TT, TU, TV, TW, TX, TY;
{
V T7, Te, TP, TQ;
T7 = VADD(T3, T6);
Te = VADD(Ta, Td);
TR = VSUB(T7, Te);
STM4(&(ro[4]), TR, ovs, &(ro[0]));
TS = VADD(T7, Te);
STM4(&(ro[0]), TS, ovs, &(ro[0]));
TP = VADD(Ti, Tl);
TQ = VADD(TM, TN);
TT = VSUB(TP, TQ);
STM4(&(io[4]), TT, ovs, &(io[0]));
TU = VADD(TP, TQ);
STM4(&(io[0]), TU, ovs, &(io[0]));
}
{
V Tf, Tm, TL, TO;
Tf = VSUB(Td, Ta);
Tm = VSUB(Ti, Tl);
TV = VADD(Tf, Tm);
STM4(&(io[2]), TV, ovs, &(io[0]));
TW = VSUB(Tm, Tf);
STM4(&(io[6]), TW, ovs, &(io[0]));
TL = VSUB(T3, T6);
TO = VSUB(TM, TN);
TX = VSUB(TL, TO);
STM4(&(ro[6]), TX, ovs, &(ro[0]));
TY = VADD(TL, TO);
STM4(&(ro[2]), TY, ovs, &(ro[0]));
}
{
V TZ, T10, T11, T12;
{
V Tp, TA, TJ, TK;
Tp = VADD(Tn, To);
TA = VADD(Tu, Tz);
TZ = VFNMS(LDK(KP707106781), TA, Tp);
STM4(&(ro[5]), TZ, ovs, &(ro[1]));
T10 = VFMA(LDK(KP707106781), TA, Tp);
STM4(&(ro[1]), T10, ovs, &(ro[1]));
TJ = VSUB(TC, TB);
TK = VADD(TG, TH);
T11 = VFNMS(LDK(KP707106781), TK, TJ);
STM4(&(io[5]), T11, ovs, &(io[1]));
T12 = VFMA(LDK(KP707106781), TK, TJ);
STM4(&(io[1]), T12, ovs, &(io[1]));
}
{
V TD, TE, T13, T14;
TD = VADD(TB, TC);
TE = VSUB(Tz, Tu);
T13 = VFNMS(LDK(KP707106781), TE, TD);
STM4(&(io[7]), T13, ovs, &(io[1]));
STN4(&(io[4]), TT, T11, TW, T13, ovs);
T14 = VFMA(LDK(KP707106781), TE, TD);
STM4(&(io[3]), T14, ovs, &(io[1]));
STN4(&(io[0]), TU, T12, TV, T14, ovs);
}
{
V TF, TI, T15, T16;
TF = VSUB(Tn, To);
TI = VSUB(TG, TH);
T15 = VFNMS(LDK(KP707106781), TI, TF);
STM4(&(ro[7]), T15, ovs, &(ro[1]));
STN4(&(ro[4]), TR, TZ, TX, T15, ovs);
T16 = VFMA(LDK(KP707106781), TI, TF);
STM4(&(ro[3]), T16, ovs, &(ro[1]));
STN4(&(ro[0]), TS, T10, TY, T16, ovs);
}
}
}
}
}
VLEAVE();
}
static void t1fv_20(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 38)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 38), MAKE_VOLATILE_STRIDE(20, rs)) {
V T4, Tx, T1m, T1K, TZ, T16, T17, T10, Tf, Tq, Tr, T1O, T1P, T1Q, T1w;
V T1z, T1A, TI, TT, TU, T1L, T1M, T1N, T1p, T1s, T1t, Ts, TV;
{
V T1, Tw, T3, Tu, Tv, T2, Tt, T1k, T1l;
T1 = LD(&(x[0]), ms, &(x[0]));
Tv = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
Tw = BYTWJ(&(W[TWVL * 28]), Tv);
T2 = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
T3 = BYTWJ(&(W[TWVL * 18]), T2);
Tt = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tu = BYTWJ(&(W[TWVL * 8]), Tt);
T4 = VSUB(T1, T3);
Tx = VSUB(Tu, Tw);
T1k = VADD(T1, T3);
T1l = VADD(Tu, Tw);
T1m = VSUB(T1k, T1l);
T1K = VADD(T1k, T1l);
}
{
V T9, T1n, TN, T1v, TS, T1y, Te, T1q, Tk, T1u, TC, T1o, TH, T1r, Tp;
V T1x;
{
V T6, T8, T5, T7;
T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T6 = BYTWJ(&(W[TWVL * 6]), T5);
T7 = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
T8 = BYTWJ(&(W[TWVL * 26]), T7);
T9 = VSUB(T6, T8);
T1n = VADD(T6, T8);
}
{
V TK, TM, TJ, TL;
TJ = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
TK = BYTWJ(&(W[TWVL * 24]), TJ);
TL = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
TM = BYTWJ(&(W[TWVL * 4]), TL);
TN = VSUB(TK, TM);
T1v = VADD(TK, TM);
}
{
V TP, TR, TO, TQ;
TO = LD(&(x[WS(rs, 17)]), ms, &(x[WS(rs, 1)]));
TP = BYTWJ(&(W[TWVL * 32]), TO);
TQ = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
TR = BYTWJ(&(W[TWVL * 12]), TQ);
TS = VSUB(TP, TR);
T1y = VADD(TP, TR);
}
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(rs, 16)]), ms, &(x[0]));
Tb = BYTWJ(&(W[TWVL * 30]), Ta);
Tc = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Td = BYTWJ(&(W[TWVL * 10]), Tc);
Te = VSUB(Tb, Td);
T1q = VADD(Tb, Td);
}
{
V Th, Tj, Tg, Ti;
Tg = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Th = BYTWJ(&(W[TWVL * 14]), Tg);
Ti = LD(&(x[WS(rs, 18)]), ms, &(x[0]));
Tj = BYTWJ(&(W[TWVL * 34]), Ti);
Tk = VSUB(Th, Tj);
T1u = VADD(Th, Tj);
}
{
V Tz, TB, Ty, TA;
Ty = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tz = BYTWJ(&(W[TWVL * 16]), Ty);
TA = LD(&(x[WS(rs, 19)]), ms, &(x[WS(rs, 1)]));
TB = BYTWJ(&(W[TWVL * 36]), TA);
TC = VSUB(Tz, TB);
T1o = VADD(Tz, TB);
}
{
V TE, TG, TD, TF;
TD = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
TE = BYTWJ(&(W[0]), TD);
TF = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
TG = BYTWJ(&(W[TWVL * 20]), TF);
TH = VSUB(TE, TG);
T1r = VADD(TE, TG);
}
{
V Tm, To, Tl, Tn;
Tl = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
Tm = BYTWJ(&(W[TWVL * 22]), Tl);
Tn = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
To = BYTWJ(&(W[TWVL * 2]), Tn);
Tp = VSUB(Tm, To);
T1x = VADD(Tm, To);
}
TZ = VSUB(TH, TC);
T16 = VSUB(T9, Te);
T17 = VSUB(Tk, Tp);
T10 = VSUB(TS, TN);
Tf = VADD(T9, Te);
Tq = VADD(Tk, Tp);
Tr = VADD(Tf, Tq);
T1O = VADD(T1u, T1v);
T1P = VADD(T1x, T1y);
T1Q = VADD(T1O, T1P);
T1w = VSUB(T1u, T1v);
T1z = VSUB(T1x, T1y);
T1A = VADD(T1w, T1z);
TI = VADD(TC, TH);
TT = VADD(TN, TS);
TU = VADD(TI, TT);
T1L = VADD(T1n, T1o);
T1M = VADD(T1q, T1r);
T1N = VADD(T1L, T1M);
T1p = VSUB(T1n, T1o);
T1s = VSUB(T1q, T1r);
T1t = VADD(T1p, T1s);
}
Ts = VADD(T4, Tr);
TV = VADD(Tx, TU);
ST(&(x[WS(rs, 5)]), VFNMSI(TV, Ts), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 15)]), VFMAI(TV, Ts), ms, &(x[WS(rs, 1)]));
{
V T1T, T1R, T1S, T1X, T1Z, T1V, T1W, T1Y, T1U;
T1T = VSUB(T1N, T1Q);
T1R = VADD(T1N, T1Q);
T1S = VFNMS(LDK(KP250000000), T1R, T1K);
T1V = VSUB(T1L, T1M);
T1W = VSUB(T1O, T1P);
T1X = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1W, T1V));
T1Z = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1V, T1W));
ST(&(x[0]), VADD(T1K, T1R), ms, &(x[0]));
T1Y = VFNMS(LDK(KP559016994), T1T, T1S);
ST(&(x[WS(rs, 8)]), VFNMSI(T1Z, T1Y), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VFMAI(T1Z, T1Y), ms, &(x[0]));
T1U = VFMA(LDK(KP559016994), T1T, T1S);
ST(&(x[WS(rs, 4)]), VFMAI(T1X, T1U), ms, &(x[0]));
ST(&(x[WS(rs, 16)]), VFNMSI(T1X, T1U), ms, &(x[0]));
}
{
V T1D, T1B, T1C, T1H, T1J, T1F, T1G, T1I, T1E;
T1D = VSUB(T1t, T1A);
T1B = VADD(T1t, T1A);
T1C = VFNMS(LDK(KP250000000), T1B, T1m);
T1F = VSUB(T1w, T1z);
T1G = VSUB(T1p, T1s);
T1H = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1G, T1F));
T1J = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1F, T1G));
ST(&(x[WS(rs, 10)]), VADD(T1m, T1B), ms, &(x[0]));
T1I = VFMA(LDK(KP559016994), T1D, T1C);
ST(&(x[WS(rs, 6)]), VFNMSI(T1J, T1I), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VFMAI(T1J, T1I), ms, &(x[0]));
T1E = VFNMS(LDK(KP559016994), T1D, T1C);
ST(&(x[WS(rs, 2)]), VFMAI(T1H, T1E), ms, &(x[0]));
ST(&(x[WS(rs, 18)]), VFNMSI(T1H, T1E), ms, &(x[0]));
}
{
V T11, T18, T1g, T1d, T15, T1f, TY, T1c;
T11 = VFMA(LDK(KP618033988), T10, TZ);
T18 = VFMA(LDK(KP618033988), T17, T16);
T1g = VFNMS(LDK(KP618033988), T16, T17);
T1d = VFNMS(LDK(KP618033988), TZ, T10);
{
V T13, T14, TW, TX;
T13 = VFNMS(LDK(KP250000000), TU, Tx);
T14 = VSUB(TT, TI);
T15 = VFNMS(LDK(KP559016994), T14, T13);
T1f = VFMA(LDK(KP559016994), T14, T13);
TW = VFNMS(LDK(KP250000000), Tr, T4);
TX = VSUB(Tf, Tq);
TY = VFMA(LDK(KP559016994), TX, TW);
T1c = VFNMS(LDK(KP559016994), TX, TW);
}
{
V T12, T19, T1i, T1j;
T12 = VFMA(LDK(KP951056516), T11, TY);
T19 = VFMA(LDK(KP951056516), T18, T15);
ST(&(x[WS(rs, 1)]), VFNMSI(T19, T12), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 19)]), VFMAI(T19, T12), ms, &(x[WS(rs, 1)]));
T1i = VFMA(LDK(KP951056516), T1d, T1c);
T1j = VFMA(LDK(KP951056516), T1g, T1f);
ST(&(x[WS(rs, 13)]), VFNMSI(T1j, T1i), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(T1j, T1i), ms, &(x[WS(rs, 1)]));
}
{
V T1a, T1b, T1e, T1h;
T1a = VFNMS(LDK(KP951056516), T11, TY);
T1b = VFNMS(LDK(KP951056516), T18, T15);
ST(&(x[WS(rs, 9)]), VFNMSI(T1b, T1a), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFMAI(T1b, T1a), ms, &(x[WS(rs, 1)]));
T1e = VFNMS(LDK(KP951056516), T1d, T1c);
T1h = VFNMS(LDK(KP951056516), T1g, T1f);
ST(&(x[WS(rs, 17)]), VFNMSI(T1h, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(T1h, T1e), ms, &(x[WS(rs, 1)]));
}
}
}
}
VLEAVE();
}
static void t2sv_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;
for (m = mb, W = W + (mb * 8); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 8), MAKE_VOLATILE_STRIDE(32, rs)) {
V T34, T30, T2N, T2v, T2M, T2g, T3V, T3X, T32, T2U, T33, T2X, T2O, T2K, T3P;
V T3R;
{
V T2, Tf, TM, TO, T3, T6, T5, Th;
T2 = LDW(&(W[0]));
Tf = LDW(&(W[TWVL * 2]));
TM = LDW(&(W[TWVL * 6]));
TO = LDW(&(W[TWVL * 7]));
T3 = LDW(&(W[TWVL * 4]));
T6 = LDW(&(W[TWVL * 5]));
T5 = LDW(&(W[TWVL * 1]));
Th = LDW(&(W[TWVL * 3]));
{
V TW, TZ, Te, T1U, T3A, T3L, T2D, T1G, T3h, T2A, T2B, T1R, T3i, T2I, Tx;
V T3M, T1Z, T3w, TL, T26, T25, T37, T1l, T2q, T1d, T2o, T2l, T3c, T1r, T2s;
V TX, T10, TV, T2a;
{
V Tz, TP, TT, Tq, TF, Tu, TI, Tm, TC, T1j, T1p, T1m, T1f, T1O, T1M;
V T1K, T2F, Tj, Tn, T1Q, T2G, Tk, T1V, Tr, Tv;
{
V T1, Ti, Tb, T3z, T8, Tc, T1u, T1D, T1L, T1z, T9, T3x, T1v, T1w, T1A;
V T1E;
{
V T7, T1i, T1e, T1C, T1y;
T1 = LD(&(ri[0]), ms, &(ri[0]));
{
V Tg, TN, TS, Tp;
Tg = VMUL(T2, Tf);
TN = VMUL(T2, TM);
TS = VMUL(T2, TO);
Tp = VMUL(Tf, T3);
{
V T4, Tt, Ta, Tl;
T4 = VMUL(T2, T3);
Tt = VMUL(Tf, T6);
Ta = VMUL(T2, T6);
Tl = VMUL(T2, Th);
Ti = VFNMS(T5, Th, Tg);
Tz = VFMA(T5, Th, Tg);
TP = VFMA(T5, TO, TN);
TT = VFNMS(T5, TM, TS);
TW = VFMA(Th, T6, Tp);
Tq = VFNMS(Th, T6, Tp);
TF = VFNMS(T5, T6, T4);
T7 = VFMA(T5, T6, T4);
Tu = VFMA(Th, T3, Tt);
TZ = VFNMS(Th, T3, Tt);
TI = VFMA(T5, T3, Ta);
Tb = VFNMS(T5, T3, Ta);
Tm = VFMA(T5, Tf, Tl);
TC = VFNMS(T5, Tf, Tl);
T1i = VMUL(Ti, T6);
T1e = VMUL(Ti, T3);
T1C = VMUL(Tz, T6);
T1y = VMUL(Tz, T3);
T3z = LD(&(ii[0]), ms, &(ii[0]));
}
}
T8 = LD(&(ri[WS(rs, 8)]), ms, &(ri[0]));
Tc = LD(&(ii[WS(rs, 8)]), ms, &(ii[0]));
T1u = LD(&(ri[WS(rs, 15)]), ms, &(ri[WS(rs, 1)]));
T1j = VFNMS(Tm, T3, T1i);
T1p = VFMA(Tm, T3, T1i);
T1m = VFNMS(Tm, T6, T1e);
T1f = VFMA(Tm, T6, T1e);
T1D = VFNMS(TC, T3, T1C);
T1O = VFMA(TC, T3, T1C);
T1L = VFNMS(TC, T6, T1y);
T1z = VFMA(TC, T6, T1y);
T9 = VMUL(T7, T8);
T3x = VMUL(T7, Tc);
T1v = VMUL(TM, T1u);
T1w = LD(&(ii[WS(rs, 15)]), ms, &(ii[WS(rs, 1)]));
T1A = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
T1E = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
}
{
V T1x, T2x, T1F, T2z, T1N, T1P;
{
V T1H, T1J, T1I, T2E;
{
V Td, T3y, T2w, T1B, T2y;
T1H = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
T1J = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
Td = VFMA(Tb, Tc, T9);
T3y = VFNMS(Tb, T8, T3x);
T1M = LD(&(ri[WS(rs, 11)]), ms, &(ri[WS(rs, 1)]));
T1x = VFMA(TO, T1w, T1v);
T2w = VMUL(TM, T1w);
T1B = VMUL(T1z, T1A);
T2y = VMUL(T1z, T1E);
T1I = VMUL(Tf, T1H);
T2E = VMUL(Tf, T1J);
Te = VADD(T1, Td);
T1U = VSUB(T1, Td);
T3A = VADD(T3y, T3z);
T3L = VSUB(T3z, T3y);
T2x = VFNMS(TO, T1u, T2w);
T1F = VFMA(T1D, T1E, T1B);
T2z = VFNMS(T1D, T1A, T2y);
T1N = VMUL(T1L, T1M);
T1P = LD(&(ii[WS(rs, 11)]), ms, &(ii[WS(rs, 1)]));
}
T1K = VFMA(Th, T1J, T1I);
T2F = VFNMS(Th, T1H, T2E);
}
Tj = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
Tn = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
T2D = VSUB(T1x, T1F);
T1G = VADD(T1x, T1F);
T3h = VADD(T2x, T2z);
T2A = VSUB(T2x, T2z);
T1Q = VFMA(T1O, T1P, T1N);
T2G = VMUL(T1L, T1P);
Tk = VMUL(Ti, Tj);
T1V = VMUL(Ti, Tn);
Tr = LD(&(ri[WS(rs, 12)]), ms, &(ri[0]));
Tv = LD(&(ii[WS(rs, 12)]), ms, &(ii[0]));
}
}
{
V TE, T22, T15, T17, TK, T16, T2h, T24, T19, T1b;
{
V To, T1W, TG, TJ, Tw, T1Y, TH, T23;
{
V TA, TD, TB, T21, T2H, Ts, T1X;
TA = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
TD = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
T2B = VSUB(T1K, T1Q);
T1R = VADD(T1K, T1Q);
T2H = VFNMS(T1O, T1M, T2G);
To = VFMA(Tm, Tn, Tk);
T1W = VFNMS(Tm, Tj, T1V);
Ts = VMUL(Tq, Tr);
T1X = VMUL(Tq, Tv);
TB = VMUL(Tz, TA);
T21 = VMUL(Tz, TD);
TG = LD(&(ri[WS(rs, 10)]), ms, &(ri[0]));
T3i = VADD(T2F, T2H);
T2I = VSUB(T2F, T2H);
TJ = LD(&(ii[WS(rs, 10)]), ms, &(ii[0]));
Tw = VFMA(Tu, Tv, Ts);
T1Y = VFNMS(Tu, Tr, T1X);
TE = VFMA(TC, TD, TB);
T22 = VFNMS(TC, TA, T21);
TH = VMUL(TF, TG);
}
T15 = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
T17 = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
T23 = VMUL(TF, TJ);
Tx = VADD(To, Tw);
T3M = VSUB(To, Tw);
T1Z = VSUB(T1W, T1Y);
T3w = VADD(T1W, T1Y);
TK = VFMA(TI, TJ, TH);
T16 = VMUL(T2, T15);
T2h = VMUL(T2, T17);
T24 = VFNMS(TI, TG, T23);
T19 = LD(&(ri[WS(rs, 9)]), ms, &(ri[WS(rs, 1)]));
T1b = LD(&(ii[WS(rs, 9)]), ms, &(ii[WS(rs, 1)]));
}
{
V T1g, T1k, T18, T2i, T1a, T2j, T1h, T2p, T1n, T1q;
T1g = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
T1k = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
TL = VADD(TE, TK);
T26 = VSUB(TE, TK);
T18 = VFMA(T5, T17, T16);
T2i = VFNMS(T5, T15, T2h);
T25 = VSUB(T22, T24);
T37 = VADD(T22, T24);
T1a = VMUL(T3, T19);
T2j = VMUL(T3, T1b);
T1h = VMUL(T1f, T1g);
T2p = VMUL(T1f, T1k);
T1n = LD(&(ri[WS(rs, 13)]), ms, &(ri[WS(rs, 1)]));
T1q = LD(&(ii[WS(rs, 13)]), ms, &(ii[WS(rs, 1)]));
{
V TQ, TU, TR, T29;
{
V T1c, T2k, T1o, T2r;
TQ = LD(&(ri[WS(rs, 14)]), ms, &(ri[0]));
TU = LD(&(ii[WS(rs, 14)]), ms, &(ii[0]));
T1c = VFMA(T6, T1b, T1a);
T2k = VFNMS(T6, T19, T2j);
T1l = VFMA(T1j, T1k, T1h);
T2q = VFNMS(T1j, T1g, T2p);
T1o = VMUL(T1m, T1n);
T2r = VMUL(T1m, T1q);
TR = VMUL(TP, TQ);
T29 = VMUL(TP, TU);
T1d = VADD(T18, T1c);
T2o = VSUB(T18, T1c);
T2l = VSUB(T2i, T2k);
T3c = VADD(T2i, T2k);
T1r = VFMA(T1p, T1q, T1o);
T2s = VFNMS(T1p, T1n, T2r);
TX = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
T10 = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
}
TV = VFMA(TT, TU, TR);
T2a = VFNMS(TT, TQ, T29);
}
}
}
}
{
V T36, Ty, T3B, T3G, T1s, T2m, T2t, T3d, TY, T2b, T3g, T1S, T3s, T3j;
T36 = VSUB(Te, Tx);
Ty = VADD(Te, Tx);
T3B = VADD(T3w, T3A);
T3G = VSUB(T3A, T3w);
T1s = VADD(T1l, T1r);
T2m = VSUB(T1l, T1r);
T2t = VSUB(T2q, T2s);
T3d = VADD(T2q, T2s);
TY = VMUL(TW, TX);
T2b = VMUL(TW, T10);
T3g = VSUB(T1G, T1R);
T1S = VADD(T1G, T1R);
T3s = VADD(T3h, T3i);
T3j = VSUB(T3h, T3i);
{
V T3D, T1T, T3u, T3t, T28, T12, T38, T2d, T3n, T3f;
{
V T1t, T3b, T3e, T3r, T11, T2c;
T1t = VADD(T1d, T1s);
T3b = VSUB(T1d, T1s);
T3e = VSUB(T3c, T3d);
T3r = VADD(T3c, T3d);
T11 = VFMA(TZ, T10, TY);
T2c = VFNMS(TZ, TX, T2b);
T3D = VSUB(T1S, T1t);
T1T = VADD(T1t, T1S);
T3u = VADD(T3r, T3s);
T3t = VSUB(T3r, T3s);
T28 = VSUB(TV, T11);
T12 = VADD(TV, T11);
T38 = VADD(T2a, T2c);
T2d = VSUB(T2a, T2c);
T3n = VSUB(T3e, T3b);
T3f = VADD(T3b, T3e);
}
{
V T2Q, T20, T3N, T3T, T2J, T2C, T2W, T2V, T3O, T2f, T3U, T2T;
{
V T2R, T27, T2e, T2S, T13, T3F;
T2Q = VADD(T1U, T1Z);
T20 = VSUB(T1U, T1Z);
T3N = VSUB(T3L, T3M);
T3T = VADD(T3M, T3L);
T13 = VADD(TL, T12);
T3F = VSUB(T12, TL);
{
V T3v, T39, T3o, T3k;
T3v = VADD(T37, T38);
T39 = VSUB(T37, T38);
T3o = VADD(T3g, T3j);
T3k = VSUB(T3g, T3j);
{
V T3H, T3J, T14, T3q;
T3H = VADD(T3F, T3G);
T3J = VSUB(T3G, T3F);
T14 = VADD(Ty, T13);
T3q = VSUB(Ty, T13);
{
V T3a, T3m, T3C, T3E;
T3a = VADD(T36, T39);
T3m = VSUB(T36, T39);
T3C = VADD(T3v, T3B);
T3E = VSUB(T3B, T3v);
{
V T3I, T3p, T3l, T3K;
T3I = VADD(T3n, T3o);
T3p = VSUB(T3n, T3o);
T3l = VADD(T3f, T3k);
T3K = VSUB(T3k, T3f);
ST(&(ri[WS(rs, 4)]), VADD(T3q, T3t), ms, &(ri[0]));
ST(&(ri[WS(rs, 12)]), VSUB(T3q, T3t), ms, &(ri[0]));
ST(&(ri[0]), VADD(T14, T1T), ms, &(ri[0]));
ST(&(ri[WS(rs, 8)]), VSUB(T14, T1T), ms, &(ri[0]));
ST(&(ii[WS(rs, 4)]), VADD(T3D, T3E), ms, &(ii[0]));
ST(&(ii[WS(rs, 12)]), VSUB(T3E, T3D), ms, &(ii[0]));
ST(&(ii[0]), VADD(T3u, T3C), ms, &(ii[0]));
ST(&(ii[WS(rs, 8)]), VSUB(T3C, T3u), ms, &(ii[0]));
ST(&(ri[WS(rs, 6)]), VFMA(LDK(KP707106781), T3p, T3m), ms, &(ri[0]));
ST(&(ri[WS(rs, 14)]), VFNMS(LDK(KP707106781), T3p, T3m), ms, &(ri[0]));
ST(&(ii[WS(rs, 10)]), VFNMS(LDK(KP707106781), T3I, T3H), ms, &(ii[0]));
ST(&(ii[WS(rs, 2)]), VFMA(LDK(KP707106781), T3I, T3H), ms, &(ii[0]));
ST(&(ii[WS(rs, 14)]), VFNMS(LDK(KP707106781), T3K, T3J), ms, &(ii[0]));
ST(&(ii[WS(rs, 6)]), VFMA(LDK(KP707106781), T3K, T3J), ms, &(ii[0]));
ST(&(ri[WS(rs, 2)]), VFMA(LDK(KP707106781), T3l, T3a), ms, &(ri[0]));
ST(&(ri[WS(rs, 10)]), VFNMS(LDK(KP707106781), T3l, T3a), ms, &(ri[0]));
T2R = VADD(T26, T25);
T27 = VSUB(T25, T26);
T2e = VADD(T28, T2d);
T2S = VSUB(T28, T2d);
}
}
}
}
{
V T2Y, T2Z, T2n, T2u;
T2J = VSUB(T2D, T2I);
T2Y = VADD(T2D, T2I);
T2Z = VSUB(T2A, T2B);
T2C = VADD(T2A, T2B);
T2W = VSUB(T2l, T2m);
T2n = VADD(T2l, T2m);
T2u = VSUB(T2o, T2t);
T2V = VADD(T2o, T2t);
T3O = VADD(T27, T2e);
T2f = VSUB(T27, T2e);
T34 = VFMA(LDK(KP414213562), T2Y, T2Z);
T30 = VFNMS(LDK(KP414213562), T2Z, T2Y);
T3U = VSUB(T2S, T2R);
T2T = VADD(T2R, T2S);
T2N = VFNMS(LDK(KP414213562), T2n, T2u);
T2v = VFMA(LDK(KP414213562), T2u, T2n);
}
}
T2M = VFNMS(LDK(KP707106781), T2f, T20);
T2g = VFMA(LDK(KP707106781), T2f, T20);
T3V = VFMA(LDK(KP707106781), T3U, T3T);
T3X = VFNMS(LDK(KP707106781), T3U, T3T);
T32 = VFNMS(LDK(KP707106781), T2T, T2Q);
T2U = VFMA(LDK(KP707106781), T2T, T2Q);
T33 = VFNMS(LDK(KP414213562), T2V, T2W);
T2X = VFMA(LDK(KP414213562), T2W, T2V);
T2O = VFMA(LDK(KP414213562), T2C, T2J);
T2K = VFNMS(LDK(KP414213562), T2J, T2C);
T3P = VFMA(LDK(KP707106781), T3O, T3N);
T3R = VFNMS(LDK(KP707106781), T3O, T3N);
}
}
}
}
}
{
V T3Q, T35, T31, T3S;
T3Q = VADD(T33, T34);
T35 = VSUB(T33, T34);
T31 = VADD(T2X, T30);
T3S = VSUB(T30, T2X);
{
V T3W, T2P, T2L, T3Y;
T3W = VSUB(T2O, T2N);
T2P = VADD(T2N, T2O);
T2L = VSUB(T2v, T2K);
T3Y = VADD(T2v, T2K);
ST(&(ri[WS(rs, 5)]), VFMA(LDK(KP923879532), T35, T32), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 13)]), VFNMS(LDK(KP923879532), T35, T32), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 9)]), VFNMS(LDK(KP923879532), T3Q, T3P), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 1)]), VFMA(LDK(KP923879532), T3Q, T3P), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 13)]), VFNMS(LDK(KP923879532), T3S, T3R), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 5)]), VFMA(LDK(KP923879532), T3S, T3R), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 1)]), VFMA(LDK(KP923879532), T31, T2U), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 9)]), VFNMS(LDK(KP923879532), T31, T2U), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 15)]), VFMA(LDK(KP923879532), T2P, T2M), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 7)]), VFNMS(LDK(KP923879532), T2P, T2M), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 11)]), VFNMS(LDK(KP923879532), T3W, T3V), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 3)]), VFMA(LDK(KP923879532), T3W, T3V), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 15)]), VFMA(LDK(KP923879532), T3Y, T3X), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 7)]), VFNMS(LDK(KP923879532), T3Y, T3X), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 3)]), VFMA(LDK(KP923879532), T2L, T2g), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 11)]), VFNMS(LDK(KP923879532), T2L, T2g), ms, &(ri[WS(rs, 1)]));
}
}
}
}
VLEAVE();
}
static void n2sv_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;
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 T2p, T2q, T2r, T2s, T2x, T2y, T2z, T2A, T1M, T1N, T1L, T1P, T2F, T2G, T2H;
V T2I, T1O, T1Q;
{
V T1l, T1H, T1R, T7, T1x, TN, TC, T25, T1E, T1b, T1Z, Tt, T2h, T22, T1D;
V T1g, T1n, TQ, T11, Ti, Te, T26, T1m, TT, T1S, TJ, TZ, T1V, TW, Tl;
V T12, T13;
{
V Tq, T1c, Tp, T20, T1a, Tr, T1d, T1e;
{
V T1, T2, Tw, Tx, T4, T5, Tz, TA;
T1 = LD(&(ri[0]), ivs, &(ri[0]));
T2 = LD(&(ri[WS(is, 8)]), ivs, &(ri[0]));
Tw = LD(&(ii[0]), ivs, &(ii[0]));
Tx = LD(&(ii[WS(is, 8)]), ivs, &(ii[0]));
T4 = LD(&(ri[WS(is, 4)]), ivs, &(ri[0]));
T5 = LD(&(ri[WS(is, 12)]), ivs, &(ri[0]));
Tz = LD(&(ii[WS(is, 4)]), ivs, &(ii[0]));
TA = LD(&(ii[WS(is, 12)]), ivs, &(ii[0]));
{
V Tn, TL, T3, T1k, Ty, T1j, T6, TM, TB, To, T18, T19;
Tn = LD(&(ri[WS(is, 15)]), ivs, &(ri[WS(is, 1)]));
TL = VSUB(T1, T2);
T3 = VADD(T1, T2);
T1k = VSUB(Tw, Tx);
Ty = VADD(Tw, Tx);
T1j = VSUB(T4, T5);
T6 = VADD(T4, T5);
TM = VSUB(Tz, TA);
TB = VADD(Tz, TA);
To = LD(&(ri[WS(is, 7)]), ivs, &(ri[WS(is, 1)]));
T18 = LD(&(ii[WS(is, 15)]), ivs, &(ii[WS(is, 1)]));
T19 = LD(&(ii[WS(is, 7)]), ivs, &(ii[WS(is, 1)]));
Tq = LD(&(ri[WS(is, 3)]), ivs, &(ri[WS(is, 1)]));
T1l = VADD(T1j, T1k);
T1H = VSUB(T1k, T1j);
T1R = VSUB(T3, T6);
T7 = VADD(T3, T6);
T1x = VADD(TL, TM);
TN = VSUB(TL, TM);
TC = VADD(Ty, TB);
T25 = VSUB(Ty, TB);
T1c = VSUB(Tn, To);
Tp = VADD(Tn, To);
T20 = VADD(T18, T19);
T1a = VSUB(T18, T19);
Tr = LD(&(ri[WS(is, 11)]), ivs, &(ri[WS(is, 1)]));
T1d = LD(&(ii[WS(is, 3)]), ivs, &(ii[WS(is, 1)]));
T1e = LD(&(ii[WS(is, 11)]), ivs, &(ii[WS(is, 1)]));
}
}
{
V Tb, Ta, TF, Tc, TG, TH, TP, TO;
{
V T8, T9, TD, TE;
T8 = LD(&(ri[WS(is, 2)]), ivs, &(ri[0]));
T9 = LD(&(ri[WS(is, 10)]), ivs, &(ri[0]));
TD = LD(&(ii[WS(is, 2)]), ivs, &(ii[0]));
TE = LD(&(ii[WS(is, 10)]), ivs, &(ii[0]));
Tb = LD(&(ri[WS(is, 14)]), ivs, &(ri[0]));
{
V T17, Ts, T21, T1f;
T17 = VSUB(Tq, Tr);
Ts = VADD(Tq, Tr);
T21 = VADD(T1d, T1e);
T1f = VSUB(T1d, T1e);
TP = VSUB(T8, T9);
Ta = VADD(T8, T9);
TO = VSUB(TD, TE);
TF = VADD(TD, TE);
T1E = VSUB(T1a, T17);
T1b = VADD(T17, T1a);
T1Z = VSUB(Tp, Ts);
Tt = VADD(Tp, Ts);
T2h = VADD(T20, T21);
T22 = VSUB(T20, T21);
T1D = VADD(T1c, T1f);
T1g = VSUB(T1c, T1f);
Tc = LD(&(ri[WS(is, 6)]), ivs, &(ri[0]));
}
TG = LD(&(ii[WS(is, 14)]), ivs, &(ii[0]));
TH = LD(&(ii[WS(is, 6)]), ivs, &(ii[0]));
}
T1n = VADD(TP, TO);
TQ = VSUB(TO, TP);
{
V Tg, Th, TX, TR, Td, TS, TI, TY, Tj, Tk;
Tg = LD(&(ri[WS(is, 1)]), ivs, &(ri[WS(is, 1)]));
Th = LD(&(ri[WS(is, 9)]), ivs, &(ri[WS(is, 1)]));
TX = LD(&(ii[WS(is, 1)]), ivs, &(ii[WS(is, 1)]));
TR = VSUB(Tb, Tc);
Td = VADD(Tb, Tc);
TS = VSUB(TG, TH);
TI = VADD(TG, TH);
TY = LD(&(ii[WS(is, 9)]), ivs, &(ii[WS(is, 1)]));
Tj = LD(&(ri[WS(is, 5)]), ivs, &(ri[WS(is, 1)]));
T11 = VSUB(Tg, Th);
Ti = VADD(Tg, Th);
Tk = LD(&(ri[WS(is, 13)]), ivs, &(ri[WS(is, 1)]));
Te = VADD(Ta, Td);
T26 = VSUB(Td, Ta);
T1m = VSUB(TR, TS);
TT = VADD(TR, TS);
T1S = VSUB(TF, TI);
TJ = VADD(TF, TI);
TZ = VSUB(TX, TY);
T1V = VADD(TX, TY);
TW = VSUB(Tj, Tk);
Tl = VADD(Tj, Tk);
T12 = LD(&(ii[WS(is, 5)]), ivs, &(ii[WS(is, 1)]));
T13 = LD(&(ii[WS(is, 13)]), ivs, &(ii[WS(is, 1)]));
}
}
}
{
V T2f, Tf, T2j, TK, Tm, T1U, T10, T1B, T14, T1W;
T2f = VSUB(T7, Te);
Tf = VADD(T7, Te);
T2j = VADD(TC, TJ);
TK = VSUB(TC, TJ);
Tm = VADD(Ti, Tl);
T1U = VSUB(Ti, Tl);
T10 = VADD(TW, TZ);
T1B = VSUB(TZ, TW);
T14 = VSUB(T12, T13);
T1W = VADD(T12, T13);
{
V T29, T1T, T27, T2d, T2b, T23, T15, T1A, T2l, T2m, T2n, T2o, T2i, T2k, T1Y;
V T2a;
{
V Tv, Tu, T1X, T2g;
T29 = VSUB(T1R, T1S);
T1T = VADD(T1R, T1S);
T27 = VSUB(T25, T26);
T2d = VADD(T26, T25);
T2b = VADD(T1Z, T22);
T23 = VSUB(T1Z, T22);
Tv = VSUB(Tt, Tm);
Tu = VADD(Tm, Tt);
T1X = VSUB(T1V, T1W);
T2g = VADD(T1V, T1W);
T15 = VSUB(T11, T14);
T1A = VADD(T11, T14);
T2l = VSUB(TK, Tv);
STM4(&(io[12]), T2l, ovs, &(io[0]));
T2m = VADD(Tv, TK);
STM4(&(io[4]), T2m, ovs, &(io[0]));
T2n = VADD(Tf, Tu);
STM4(&(ro[0]), T2n, ovs, &(ro[0]));
T2o = VSUB(Tf, Tu);
STM4(&(ro[8]), T2o, ovs, &(ro[0]));
T2i = VSUB(T2g, T2h);
T2k = VADD(T2g, T2h);
T1Y = VADD(T1U, T1X);
T2a = VSUB(T1X, T1U);
}
{
V T1I, T1y, T1t, T16, T1v, TV, T1r, T1p, T2t, T2u, T2v, T2w, T1h, T1s, TU;
V T1o;
T1I = VADD(TQ, TT);
TU = VSUB(TQ, TT);
T1o = VSUB(T1m, T1n);
T1y = VADD(T1n, T1m);
T1t = VFNMS(LDK(KP414213562), T10, T15);
T16 = VFMA(LDK(KP414213562), T15, T10);
T2p = VADD(T2f, T2i);
STM4(&(ro[4]), T2p, ovs, &(ro[0]));
T2q = VSUB(T2f, T2i);
STM4(&(ro[12]), T2q, ovs, &(ro[0]));
T2r = VADD(T2j, T2k);
STM4(&(io[0]), T2r, ovs, &(io[0]));
T2s = VSUB(T2j, T2k);
STM4(&(io[8]), T2s, ovs, &(io[0]));
{
V T28, T24, T2e, T2c;
T28 = VSUB(T23, T1Y);
T24 = VADD(T1Y, T23);
T2e = VADD(T2a, T2b);
T2c = VSUB(T2a, T2b);
T1v = VFNMS(LDK(KP707106781), TU, TN);
TV = VFMA(LDK(KP707106781), TU, TN);
T1r = VFMA(LDK(KP707106781), T1o, T1l);
T1p = VFNMS(LDK(KP707106781), T1o, T1l);
T2t = VFNMS(LDK(KP707106781), T28, T27);
STM4(&(io[14]), T2t, ovs, &(io[0]));
T2u = VFMA(LDK(KP707106781), T28, T27);
STM4(&(io[6]), T2u, ovs, &(io[0]));
T2v = VFMA(LDK(KP707106781), T24, T1T);
STM4(&(ro[2]), T2v, ovs, &(ro[0]));
T2w = VFNMS(LDK(KP707106781), T24, T1T);
STM4(&(ro[10]), T2w, ovs, &(ro[0]));
T2x = VFNMS(LDK(KP707106781), T2e, T2d);
STM4(&(io[10]), T2x, ovs, &(io[0]));
T2y = VFMA(LDK(KP707106781), T2e, T2d);
STM4(&(io[2]), T2y, ovs, &(io[0]));
T2z = VFMA(LDK(KP707106781), T2c, T29);
STM4(&(ro[6]), T2z, ovs, &(ro[0]));
T2A = VFNMS(LDK(KP707106781), T2c, T29);
STM4(&(ro[14]), T2A, ovs, &(ro[0]));
T1h = VFNMS(LDK(KP414213562), T1g, T1b);
T1s = VFMA(LDK(KP414213562), T1b, T1g);
}
{
V T1z, T1J, T1K, T1G, T2B, T2C, T2D, T2E, T1C, T1F;
T1M = VFNMS(LDK(KP414213562), T1A, T1B);
T1C = VFMA(LDK(KP414213562), T1B, T1A);
T1F = VFNMS(LDK(KP414213562), T1E, T1D);
T1N = VFMA(LDK(KP414213562), T1D, T1E);
{
V T1q, T1i, T1w, T1u;
T1q = VADD(T16, T1h);
T1i = VSUB(T16, T1h);
T1w = VADD(T1t, T1s);
T1u = VSUB(T1s, T1t);
T1L = VFNMS(LDK(KP707106781), T1y, T1x);
T1z = VFMA(LDK(KP707106781), T1y, T1x);
T1P = VFMA(LDK(KP707106781), T1I, T1H);
T1J = VFNMS(LDK(KP707106781), T1I, T1H);
T1K = VSUB(T1F, T1C);
T1G = VADD(T1C, T1F);
T2B = VFMA(LDK(KP923879532), T1q, T1p);
STM4(&(io[15]), T2B, ovs, &(io[1]));
T2C = VFNMS(LDK(KP923879532), T1q, T1p);
STM4(&(io[7]), T2C, ovs, &(io[1]));
T2D = VFMA(LDK(KP923879532), T1i, TV);
STM4(&(ro[3]), T2D, ovs, &(ro[1]));
T2E = VFNMS(LDK(KP923879532), T1i, TV);
STM4(&(ro[11]), T2E, ovs, &(ro[1]));
T2F = VFMA(LDK(KP923879532), T1w, T1v);
STM4(&(ro[15]), T2F, ovs, &(ro[1]));
T2G = VFNMS(LDK(KP923879532), T1w, T1v);
STM4(&(ro[7]), T2G, ovs, &(ro[1]));
T2H = VFMA(LDK(KP923879532), T1u, T1r);
STM4(&(io[3]), T2H, ovs, &(io[1]));
T2I = VFNMS(LDK(KP923879532), T1u, T1r);
STM4(&(io[11]), T2I, ovs, &(io[1]));
}
{
V T2J, T2K, T2L, T2M;
T2J = VFNMS(LDK(KP923879532), T1G, T1z);
STM4(&(ro[9]), T2J, ovs, &(ro[1]));
STN4(&(ro[8]), T2o, T2J, T2w, T2E, ovs);
T2K = VFMA(LDK(KP923879532), T1G, T1z);
STM4(&(ro[1]), T2K, ovs, &(ro[1]));
STN4(&(ro[0]), T2n, T2K, T2v, T2D, ovs);
T2L = VFNMS(LDK(KP923879532), T1K, T1J);
STM4(&(io[13]), T2L, ovs, &(io[1]));
STN4(&(io[12]), T2l, T2L, T2t, T2B, ovs);
T2M = VFMA(LDK(KP923879532), T1K, T1J);
STM4(&(io[5]), T2M, ovs, &(io[1]));
STN4(&(io[4]), T2m, T2M, T2u, T2C, ovs);
}
}
}
}
}
}
T1O = VSUB(T1M, T1N);
T1Q = VADD(T1M, T1N);
{
V T2N, T2O, T2P, T2Q;
T2N = VFMA(LDK(KP923879532), T1Q, T1P);
STM4(&(io[1]), T2N, ovs, &(io[1]));
STN4(&(io[0]), T2r, T2N, T2y, T2H, ovs);
T2O = VFNMS(LDK(KP923879532), T1Q, T1P);
STM4(&(io[9]), T2O, ovs, &(io[1]));
STN4(&(io[8]), T2s, T2O, T2x, T2I, ovs);
T2P = VFMA(LDK(KP923879532), T1O, T1L);
STM4(&(ro[5]), T2P, ovs, &(ro[1]));
STN4(&(ro[4]), T2p, T2P, T2z, T2G, ovs);
T2Q = VFNMS(LDK(KP923879532), T1O, T1L);
STM4(&(ro[13]), T2Q, ovs, &(ro[1]));
STN4(&(ro[12]), T2q, T2Q, T2A, T2F, ovs);
}
}
}
static void t2sv_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;
for (m = mb, W = W + (mb * 8); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 8), MAKE_VOLATILE_STRIDE(32, rs)) {
V T2, T5, Tg, Ti, Tk, To, TE, TC, T6, T3, T8, TW, TJ, Tt, TU;
V Tc, Tx, TH, TN, TO, TP, TR, T1f, T1k, T1b, T1i, T1y, T1H, T1u, T1F;
{
V T7, Tv, Ta, Ts, T4, Tw, Tb, Tr;
{
V Th, Tn, Tj, Tm;
T2 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 1]));
Tg = LDW(&(W[TWVL * 2]));
Ti = LDW(&(W[TWVL * 3]));
Th = VMUL(T2, Tg);
Tn = VMUL(T5, Tg);
Tj = VMUL(T5, Ti);
Tm = VMUL(T2, Ti);
Tk = VSUB(Th, Tj);
To = VADD(Tm, Tn);
TE = VSUB(Tm, Tn);
TC = VADD(Th, Tj);
T6 = LDW(&(W[TWVL * 5]));
T7 = VMUL(T5, T6);
Tv = VMUL(Tg, T6);
Ta = VMUL(T2, T6);
Ts = VMUL(Ti, T6);
T3 = LDW(&(W[TWVL * 4]));
T4 = VMUL(T2, T3);
Tw = VMUL(Ti, T3);
Tb = VMUL(T5, T3);
Tr = VMUL(Tg, T3);
}
T8 = VADD(T4, T7);
TW = VSUB(Tv, Tw);
TJ = VADD(Ta, Tb);
Tt = VSUB(Tr, Ts);
TU = VADD(Tr, Ts);
Tc = VSUB(Ta, Tb);
Tx = VADD(Tv, Tw);
TH = VSUB(T4, T7);
TN = LDW(&(W[TWVL * 6]));
TO = LDW(&(W[TWVL * 7]));
TP = VFMA(T2, TN, VMUL(T5, TO));
TR = VFNMS(T5, TN, VMUL(T2, TO));
{
V T1d, T1e, T19, T1a;
T1d = VMUL(Tk, T6);
T1e = VMUL(To, T3);
T1f = VSUB(T1d, T1e);
T1k = VADD(T1d, T1e);
T19 = VMUL(Tk, T3);
T1a = VMUL(To, T6);
T1b = VADD(T19, T1a);
T1i = VSUB(T19, T1a);
}
{
V T1w, T1x, T1s, T1t;
T1w = VMUL(TC, T6);
T1x = VMUL(TE, T3);
T1y = VSUB(T1w, T1x);
T1H = VADD(T1w, T1x);
T1s = VMUL(TC, T3);
T1t = VMUL(TE, T6);
T1u = VADD(T1s, T1t);
T1F = VSUB(T1s, T1t);
}
}
{
V Tf, T3r, T1N, T3e, TA, T3s, T1Q, T3b, TM, T2M, T1W, T2w, TZ, T2N, T21;
V T2x, T1B, T1K, T2V, T2W, T2X, T2Y, T2j, T2D, T2o, T2E, T18, T1n, T2Q, T2R;
V T2S, T2T, T28, T2A, T2d, T2B;
{
V T1, T3d, Te, T3c, T9, Td;
T1 = LD(&(ri[0]), ms, &(ri[0]));
T3d = LD(&(ii[0]), ms, &(ii[0]));
T9 = LD(&(ri[WS(rs, 8)]), ms, &(ri[0]));
Td = LD(&(ii[WS(rs, 8)]), ms, &(ii[0]));
Te = VFMA(T8, T9, VMUL(Tc, Td));
T3c = VFNMS(Tc, T9, VMUL(T8, Td));
Tf = VADD(T1, Te);
T3r = VSUB(T3d, T3c);
T1N = VSUB(T1, Te);
T3e = VADD(T3c, T3d);
}
{
V Tq, T1O, Tz, T1P;
{
V Tl, Tp, Tu, Ty;
Tl = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
Tp = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
Tq = VFMA(Tk, Tl, VMUL(To, Tp));
T1O = VFNMS(To, Tl, VMUL(Tk, Tp));
Tu = LD(&(ri[WS(rs, 12)]), ms, &(ri[0]));
Ty = LD(&(ii[WS(rs, 12)]), ms, &(ii[0]));
Tz = VFMA(Tt, Tu, VMUL(Tx, Ty));
T1P = VFNMS(Tx, Tu, VMUL(Tt, Ty));
}
TA = VADD(Tq, Tz);
T3s = VSUB(Tq, Tz);
T1Q = VSUB(T1O, T1P);
T3b = VADD(T1O, T1P);
}
{
V TG, T1S, TL, T1T, T1U, T1V;
{
V TD, TF, TI, TK;
TD = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
TF = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
TG = VFMA(TC, TD, VMUL(TE, TF));
T1S = VFNMS(TE, TD, VMUL(TC, TF));
TI = LD(&(ri[WS(rs, 10)]), ms, &(ri[0]));
TK = LD(&(ii[WS(rs, 10)]), ms, &(ii[0]));
TL = VFMA(TH, TI, VMUL(TJ, TK));
T1T = VFNMS(TJ, TI, VMUL(TH, TK));
}
TM = VADD(TG, TL);
T2M = VADD(T1S, T1T);
T1U = VSUB(T1S, T1T);
T1V = VSUB(TG, TL);
T1W = VSUB(T1U, T1V);
T2w = VADD(T1V, T1U);
}
{
V TT, T1Y, TY, T1Z, T1X, T20;
{
V TQ, TS, TV, TX;
TQ = LD(&(ri[WS(rs, 14)]), ms, &(ri[0]));
TS = LD(&(ii[WS(rs, 14)]), ms, &(ii[0]));
TT = VFMA(TP, TQ, VMUL(TR, TS));
T1Y = VFNMS(TR, TQ, VMUL(TP, TS));
TV = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
TX = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
TY = VFMA(TU, TV, VMUL(TW, TX));
T1Z = VFNMS(TW, TV, VMUL(TU, TX));
}
TZ = VADD(TT, TY);
T2N = VADD(T1Y, T1Z);
T1X = VSUB(TT, TY);
T20 = VSUB(T1Y, T1Z);
T21 = VADD(T1X, T20);
T2x = VSUB(T1X, T20);
}
{
V T1r, T2k, T1J, T2h, T1A, T2l, T1E, T2g;
{
V T1p, T1q, T1G, T1I;
T1p = LD(&(ri[WS(rs, 15)]), ms, &(ri[WS(rs, 1)]));
T1q = LD(&(ii[WS(rs, 15)]), ms, &(ii[WS(rs, 1)]));
T1r = VFMA(TN, T1p, VMUL(TO, T1q));
T2k = VFNMS(TO, T1p, VMUL(TN, T1q));
T1G = LD(&(ri[WS(rs, 11)]), ms, &(ri[WS(rs, 1)]));
T1I = LD(&(ii[WS(rs, 11)]), ms, &(ii[WS(rs, 1)]));
T1J = VFMA(T1F, T1G, VMUL(T1H, T1I));
T2h = VFNMS(T1H, T1G, VMUL(T1F, T1I));
}
{
V T1v, T1z, T1C, T1D;
T1v = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
T1z = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
T1A = VFMA(T1u, T1v, VMUL(T1y, T1z));
T2l = VFNMS(T1y, T1v, VMUL(T1u, T1z));
T1C = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
T1D = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
T1E = VFMA(Tg, T1C, VMUL(Ti, T1D));
T2g = VFNMS(Ti, T1C, VMUL(Tg, T1D));
}
T1B = VADD(T1r, T1A);
T1K = VADD(T1E, T1J);
T2V = VSUB(T1B, T1K);
T2W = VADD(T2k, T2l);
T2X = VADD(T2g, T2h);
T2Y = VSUB(T2W, T2X);
{
V T2f, T2i, T2m, T2n;
T2f = VSUB(T1r, T1A);
T2i = VSUB(T2g, T2h);
T2j = VSUB(T2f, T2i);
T2D = VADD(T2f, T2i);
T2m = VSUB(T2k, T2l);
T2n = VSUB(T1E, T1J);
T2o = VADD(T2m, T2n);
T2E = VSUB(T2m, T2n);
}
}
{
V T14, T24, T1m, T2b, T17, T25, T1h, T2a;
{
V T12, T13, T1j, T1l;
T12 = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
T13 = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
T14 = VFMA(T2, T12, VMUL(T5, T13));
T24 = VFNMS(T5, T12, VMUL(T2, T13));
T1j = LD(&(ri[WS(rs, 13)]), ms, &(ri[WS(rs, 1)]));
T1l = LD(&(ii[WS(rs, 13)]), ms, &(ii[WS(rs, 1)]));
T1m = VFMA(T1i, T1j, VMUL(T1k, T1l));
T2b = VFNMS(T1k, T1j, VMUL(T1i, T1l));
}
{
V T15, T16, T1c, T1g;
T15 = LD(&(ri[WS(rs, 9)]), ms, &(ri[WS(rs, 1)]));
T16 = LD(&(ii[WS(rs, 9)]), ms, &(ii[WS(rs, 1)]));
T17 = VFMA(T3, T15, VMUL(T6, T16));
T25 = VFNMS(T6, T15, VMUL(T3, T16));
T1c = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
T1g = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
T1h = VFMA(T1b, T1c, VMUL(T1f, T1g));
T2a = VFNMS(T1f, T1c, VMUL(T1b, T1g));
}
T18 = VADD(T14, T17);
T1n = VADD(T1h, T1m);
T2Q = VSUB(T18, T1n);
T2R = VADD(T24, T25);
T2S = VADD(T2a, T2b);
T2T = VSUB(T2R, T2S);
{
V T26, T27, T29, T2c;
T26 = VSUB(T24, T25);
T27 = VSUB(T1h, T1m);
T28 = VADD(T26, T27);
T2A = VSUB(T26, T27);
T29 = VSUB(T14, T17);
T2c = VSUB(T2a, T2b);
T2d = VSUB(T29, T2c);
T2B = VADD(T29, T2c);
}
}
{
V T23, T2r, T3A, T3C, T2q, T3B, T2u, T3x;
{
V T1R, T22, T3y, T3z;
T1R = VSUB(T1N, T1Q);
T22 = VMUL(LDK(KP707106781), VSUB(T1W, T21));
T23 = VADD(T1R, T22);
T2r = VSUB(T1R, T22);
T3y = VMUL(LDK(KP707106781), VSUB(T2x, T2w));
T3z = VADD(T3s, T3r);
T3A = VADD(T3y, T3z);
T3C = VSUB(T3z, T3y);
}
{
V T2e, T2p, T2s, T2t;
T2e = VFMA(LDK(KP923879532), T28, VMUL(LDK(KP382683432), T2d));
T2p = VFNMS(LDK(KP923879532), T2o, VMUL(LDK(KP382683432), T2j));
T2q = VADD(T2e, T2p);
T3B = VSUB(T2p, T2e);
T2s = VFNMS(LDK(KP923879532), T2d, VMUL(LDK(KP382683432), T28));
T2t = VFMA(LDK(KP382683432), T2o, VMUL(LDK(KP923879532), T2j));
T2u = VSUB(T2s, T2t);
T3x = VADD(T2s, T2t);
}
ST(&(ri[WS(rs, 11)]), VSUB(T23, T2q), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 11)]), VSUB(T3A, T3x), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 3)]), VADD(T23, T2q), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 3)]), VADD(T3x, T3A), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 15)]), VSUB(T2r, T2u), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 15)]), VSUB(T3C, T3B), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 7)]), VADD(T2r, T2u), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 7)]), VADD(T3B, T3C), ms, &(ii[WS(rs, 1)]));
}
{
V T2P, T31, T3m, T3o, T30, T3n, T34, T3j;
{
V T2L, T2O, T3k, T3l;
T2L = VSUB(Tf, TA);
T2O = VSUB(T2M, T2N);
T2P = VADD(T2L, T2O);
T31 = VSUB(T2L, T2O);
T3k = VSUB(TZ, TM);
T3l = VSUB(T3e, T3b);
T3m = VADD(T3k, T3l);
T3o = VSUB(T3l, T3k);
}
{
V T2U, T2Z, T32, T33;
T2U = VADD(T2Q, T2T);
T2Z = VSUB(T2V, T2Y);
T30 = VMUL(LDK(KP707106781), VADD(T2U, T2Z));
T3n = VMUL(LDK(KP707106781), VSUB(T2Z, T2U));
T32 = VSUB(T2T, T2Q);
T33 = VADD(T2V, T2Y);
T34 = VMUL(LDK(KP707106781), VSUB(T32, T33));
T3j = VMUL(LDK(KP707106781), VADD(T32, T33));
}
ST(&(ri[WS(rs, 10)]), VSUB(T2P, T30), ms, &(ri[0]));
ST(&(ii[WS(rs, 10)]), VSUB(T3m, T3j), ms, &(ii[0]));
ST(&(ri[WS(rs, 2)]), VADD(T2P, T30), ms, &(ri[0]));
ST(&(ii[WS(rs, 2)]), VADD(T3j, T3m), ms, &(ii[0]));
ST(&(ri[WS(rs, 14)]), VSUB(T31, T34), ms, &(ri[0]));
ST(&(ii[WS(rs, 14)]), VSUB(T3o, T3n), ms, &(ii[0]));
ST(&(ri[WS(rs, 6)]), VADD(T31, T34), ms, &(ri[0]));
ST(&(ii[WS(rs, 6)]), VADD(T3n, T3o), ms, &(ii[0]));
}
{
V T2z, T2H, T3u, T3w, T2G, T3v, T2K, T3p;
{
V T2v, T2y, T3q, T3t;
T2v = VADD(T1N, T1Q);
T2y = VMUL(LDK(KP707106781), VADD(T2w, T2x));
T2z = VADD(T2v, T2y);
T2H = VSUB(T2v, T2y);
T3q = VMUL(LDK(KP707106781), VADD(T1W, T21));
T3t = VSUB(T3r, T3s);
T3u = VADD(T3q, T3t);
T3w = VSUB(T3t, T3q);
}
{
V T2C, T2F, T2I, T2J;
T2C = VFMA(LDK(KP382683432), T2A, VMUL(LDK(KP923879532), T2B));
T2F = VFNMS(LDK(KP382683432), T2E, VMUL(LDK(KP923879532), T2D));
T2G = VADD(T2C, T2F);
T3v = VSUB(T2F, T2C);
T2I = VFNMS(LDK(KP382683432), T2B, VMUL(LDK(KP923879532), T2A));
T2J = VFMA(LDK(KP923879532), T2E, VMUL(LDK(KP382683432), T2D));
T2K = VSUB(T2I, T2J);
T3p = VADD(T2I, T2J);
}
ST(&(ri[WS(rs, 9)]), VSUB(T2z, T2G), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 9)]), VSUB(T3u, T3p), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 1)]), VADD(T2z, T2G), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 1)]), VADD(T3p, T3u), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 13)]), VSUB(T2H, T2K), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 13)]), VSUB(T3w, T3v), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 5)]), VADD(T2H, T2K), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 5)]), VADD(T3v, T3w), ms, &(ii[WS(rs, 1)]));
}
{
V T11, T35, T3g, T3i, T1M, T3h, T38, T39;
{
V TB, T10, T3a, T3f;
TB = VADD(Tf, TA);
T10 = VADD(TM, TZ);
T11 = VADD(TB, T10);
T35 = VSUB(TB, T10);
T3a = VADD(T2M, T2N);
T3f = VADD(T3b, T3e);
T3g = VADD(T3a, T3f);
T3i = VSUB(T3f, T3a);
}
{
V T1o, T1L, T36, T37;
T1o = VADD(T18, T1n);
T1L = VADD(T1B, T1K);
T1M = VADD(T1o, T1L);
T3h = VSUB(T1L, T1o);
T36 = VADD(T2R, T2S);
T37 = VADD(T2W, T2X);
T38 = VSUB(T36, T37);
T39 = VADD(T36, T37);
}
ST(&(ri[WS(rs, 8)]), VSUB(T11, T1M), ms, &(ri[0]));
ST(&(ii[WS(rs, 8)]), VSUB(T3g, T39), ms, &(ii[0]));
ST(&(ri[0]), VADD(T11, T1M), ms, &(ri[0]));
ST(&(ii[0]), VADD(T39, T3g), ms, &(ii[0]));
ST(&(ri[WS(rs, 12)]), VSUB(T35, T38), ms, &(ri[0]));
ST(&(ii[WS(rs, 12)]), VSUB(T3i, T3h), ms, &(ii[0]));
ST(&(ri[WS(rs, 4)]), VADD(T35, T38), ms, &(ri[0]));
ST(&(ii[WS(rs, 4)]), VADD(T3h, T3i), ms, &(ii[0]));
}
}
}
}
VLEAVE();
}
static void n2fv_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 = 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, 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 = VSUB(Tj, Tk);
Tl = VADD(Tj, Tk);
{
V Tq, To, Ta, Tf;
Tq = VSUB(Tn, Tm);
To = VADD(Tm, Tn);
Ta = VADD(T6, T9);
Tf = VSUB(T9, T6);
{
V Tr, Ts, Tt, Tu, Tg, Ti, Tb, Th;
Tr = VADD(Tl, To);
STM2(&(xo[0]), Tr, ovs, &(xo[0]));
Ts = VSUB(Tl, To);
STM2(&(xo[8]), Ts, ovs, &(xo[0]));
Tt = VFMAI(Tq, Tp);
STM2(&(xo[4]), Tt, ovs, &(xo[0]));
Tu = VFNMSI(Tq, Tp);
STM2(&(xo[12]), Tu, ovs, &(xo[0]));
Tg = VFNMS(LDK(KP707106781), Tf, Te);
Ti = VFMA(LDK(KP707106781), Tf, Te);
Tb = VFMA(LDK(KP707106781), Ta, T3);
Th = VFNMS(LDK(KP707106781), Ta, T3);
{
V Tv, Tw, Tx, Ty;
Tv = VFMAI(Ti, Th);
STM2(&(xo[6]), Tv, ovs, &(xo[2]));
STN2(&(xo[4]), Tt, Tv, ovs);
Tw = VFNMSI(Ti, Th);
STM2(&(xo[10]), Tw, ovs, &(xo[2]));
STN2(&(xo[8]), Ts, Tw, ovs);
Tx = VFMAI(Tg, Tb);
STM2(&(xo[14]), Tx, ovs, &(xo[2]));
STN2(&(xo[12]), Tu, Tx, ovs);
Ty = VFNMSI(Tg, Tb);
STM2(&(xo[2]), Ty, ovs, &(xo[2]));
STN2(&(xo[0]), Tr, Ty, ovs);
}
}
}
}
}
}
static void t2sv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + (mb * 6); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 6), MAKE_VOLATILE_STRIDE(rs)) {
V T1m, T1l, T1k, T1u, T1n, T1o;
{
V T2, T3, Tl, Tn, T5, T6;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
Tl = LDW(&(W[TWVL * 4]));
Tn = LDW(&(W[TWVL * 5]));
T5 = LDW(&(W[TWVL * 1]));
T6 = LDW(&(W[TWVL * 3]));
{
V T1, T1s, TK, T1r, Td, Tk, TG, TC, TY, Tu, TW, TL, TM, TO, TQ;
V Tx, Tz, TD, TH;
{
V T8, T4, Tm, Tr, Tc, Ta;
T1 = LD(&(ri[0]), ms, &(ri[0]));
T1s = LD(&(ii[0]), ms, &(ii[0]));
T8 = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
T4 = VMUL(T2, T3);
Tm = VMUL(T2, Tl);
Tr = VMUL(T2, Tn);
Tc = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
Ta = VMUL(T2, T6);
{
V Tp, Tt, Tg, T7, Tf, To, Ts, Ti, Tb, Tj;
Tp = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
Tt = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
Tg = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
T7 = VFNMS(T5, T6, T4);
Tf = VFMA(T5, T6, T4);
To = VFMA(T5, Tn, Tm);
Ts = VFNMS(T5, Tl, Tr);
Ti = VFNMS(T5, T3, Ta);
Tb = VFMA(T5, T3, Ta);
Tj = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
TK = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
{
V T1q, T9, Th, TF;
T1q = VMUL(T7, Tc);
T9 = VMUL(T7, T8);
Th = VMUL(Tf, Tg);
TF = VMUL(Tf, Tn);
{
V TB, TX, Tq, TV;
TB = VMUL(Tf, Tl);
TX = VMUL(To, Tt);
Tq = VMUL(To, Tp);
TV = VMUL(Tf, Tj);
T1r = VFNMS(Tb, T8, T1q);
Td = VFMA(Tb, Tc, T9);
Tk = VFMA(Ti, Tj, Th);
TG = VFNMS(Ti, Tl, TF);
TC = VFMA(Ti, Tn, TB);
TY = VFNMS(Ts, Tp, TX);
Tu = VFMA(Ts, Tt, Tq);
TW = VFNMS(Ti, Tg, TV);
TL = VMUL(Tl, TK);
}
}
TM = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
TO = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
TQ = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
Tx = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
Tz = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
TD = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
TH = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
}
}
{
V Te, T1p, T1g, T10, TS, T18, T1d, T1t, T1x, T1y, Tv, TJ, T11, T16;
{
V TN, T1a, TR, T1c, TA, T13, TI, T15;
{
V TU, T19, TP, T1b, Ty, T12, TE, T14, TZ;
TU = VSUB(T1, Td);
Te = VADD(T1, Td);
TN = VFMA(Tn, TM, TL);
T19 = VMUL(Tl, TM);
TP = VMUL(T3, TO);
T1b = VMUL(T3, TQ);
Ty = VMUL(T2, Tx);
T12 = VMUL(T2, Tz);
TE = VMUL(TC, TD);
T14 = VMUL(TC, TH);
T1p = VADD(TW, TY);
TZ = VSUB(TW, TY);
T1a = VFNMS(Tn, TK, T19);
TR = VFMA(T6, TQ, TP);
T1c = VFNMS(T6, TO, T1b);
TA = VFMA(T5, Tz, Ty);
T13 = VFNMS(T5, Tx, T12);
TI = VFMA(TG, TH, TE);
T15 = VFNMS(TG, TD, T14);
T1g = VSUB(TU, TZ);
T10 = VADD(TU, TZ);
}
TS = VADD(TN, TR);
T18 = VSUB(TN, TR);
T1d = VSUB(T1a, T1c);
T1m = VADD(T1a, T1c);
T1t = VADD(T1r, T1s);
T1x = VSUB(T1s, T1r);
T1y = VSUB(Tk, Tu);
Tv = VADD(Tk, Tu);
TJ = VADD(TA, TI);
T11 = VSUB(TA, TI);
T16 = VSUB(T13, T15);
T1l = VADD(T13, T15);
}
{
V Tw, T1w, T1v, TT;
{
V T1i, T1e, T1B, T1z, T1h, T17;
T1i = VADD(T18, T1d);
T1e = VSUB(T18, T1d);
T1B = VADD(T1y, T1x);
T1z = VSUB(T1x, T1y);
T1h = VSUB(T16, T11);
T17 = VADD(T11, T16);
T1k = VSUB(Te, Tv);
Tw = VADD(Te, Tv);
{
V T1A, T1j, T1C, T1f;
T1A = VADD(T1h, T1i);
T1j = VSUB(T1h, T1i);
T1C = VSUB(T1e, T17);
T1f = VADD(T17, T1e);
T1w = VSUB(T1t, T1p);
T1u = VADD(T1p, T1t);
T1v = VSUB(TS, TJ);
TT = VADD(TJ, TS);
ST(&(ii[WS(rs, 1)]), VFMA(LDK(KP707106781), T1A, T1z), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 5)]), VFNMS(LDK(KP707106781), T1A, T1z), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 3)]), VFMA(LDK(KP707106781), T1j, T1g), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 7)]), VFNMS(LDK(KP707106781), T1j, T1g), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 3)]), VFMA(LDK(KP707106781), T1C, T1B), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 7)]), VFNMS(LDK(KP707106781), T1C, T1B), ms, &(ii[WS(rs, 1)]));
ST(&(ri[WS(rs, 1)]), VFMA(LDK(KP707106781), T1f, T10), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 5)]), VFNMS(LDK(KP707106781), T1f, T10), ms, &(ri[WS(rs, 1)]));
}
}
ST(&(ri[WS(rs, 4)]), VSUB(Tw, TT), ms, &(ri[0]));
ST(&(ri[0]), VADD(Tw, TT), ms, &(ri[0]));
ST(&(ii[WS(rs, 6)]), VSUB(T1w, T1v), ms, &(ii[0]));
ST(&(ii[WS(rs, 2)]), VADD(T1v, T1w), ms, &(ii[0]));
}
}
}
}
T1n = VSUB(T1l, T1m);
T1o = VADD(T1l, T1m);
ST(&(ii[0]), VADD(T1o, T1u), ms, &(ii[0]));
ST(&(ii[WS(rs, 4)]), VSUB(T1u, T1o), ms, &(ii[0]));
ST(&(ri[WS(rs, 2)]), VADD(T1k, T1n), ms, &(ri[0]));
ST(&(ri[WS(rs, 6)]), VSUB(T1k, T1n), ms, &(ri[0]));
}
}
VLEAVE();
}
static const R *t1fv_8(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
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 * 14), MAKE_VOLATILE_STRIDE(ios)) {
V T4, Tq, Tm, Tr, T9, Tt, Te, Tu, T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 4)]), dist, &(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(ios, 2)]), dist, &(x[0]));
Tj = BYTWJ(&(W[TWVL * 2]), Ti);
Tk = LD(&(x[WS(ios, 6)]), dist, &(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(ios, 1)]), dist, &(x[WS(ios, 1)]));
T6 = BYTWJ(&(W[0]), T5);
T7 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
T8 = BYTWJ(&(W[TWVL * 8]), T7);
T9 = VSUB(T6, T8);
Tt = VADD(T6, T8);
}
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
Tb = BYTWJ(&(W[TWVL * 12]), Ta);
Tc = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 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(ios, 4)]), VSUB(Ts, Tv), dist, &(x[0]));
ST(&(x[0]), VADD(Ts, Tv), dist, &(x[0]));
Tw = VSUB(Tq, Tr);
Tx = VBYI(VSUB(Tu, Tt));
ST(&(x[WS(ios, 6)]), VSUB(Tw, Tx), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VADD(Tw, Tx), dist, &(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(ios, 7)]), VSUB(Tg, Tn), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 3)]), VADD(To, Tp), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(Tg, Tn), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VSUB(To, Tp), dist, &(x[WS(ios, 1)]));
}
}
}
return W;
}
static void n2bv_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
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 Tp, T13, Tu, TY, Tm, T14, Tv, TU, T7, T16, Tx, TN, Te, T17, Ty;
V TQ;
{
V Tn, To, TX, Ts, Tt, TW;
Tn = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
TX = VADD(Tn, To);
Ts = LD(&(xi[0]), ivs, &(xi[0]));
Tt = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
TW = VADD(Ts, Tt);
Tp = VSUB(Tn, To);
T13 = VADD(TW, TX);
Tu = VSUB(Ts, Tt);
TY = VSUB(TW, TX);
}
{
V Ti, TS, Tl, TT;
{
V Tg, Th, Tj, Tk;
Tg = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Ti = VSUB(Tg, Th);
TS = VADD(Tg, Th);
Tj = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tl = VSUB(Tj, Tk);
TT = VADD(Tj, Tk);
}
Tm = VMUL(LDK(KP707106781), VSUB(Ti, Tl));
T14 = VADD(TS, TT);
Tv = VMUL(LDK(KP707106781), VADD(Ti, Tl));
TU = VSUB(TS, TT);
}
{
V T3, TL, T6, TM;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
TL = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TM = VADD(T4, T5);
}
T7 = VFNMS(LDK(KP382683432), T6, VMUL(LDK(KP923879532), T3));
T16 = VADD(TL, TM);
Tx = VFMA(LDK(KP382683432), T3, VMUL(LDK(KP923879532), T6));
TN = VSUB(TL, TM);
}
{
V Ta, TO, Td, TP;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
TO = VADD(T8, T9);
Tb = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TP = VADD(Tb, Tc);
}
Te = VFMA(LDK(KP923879532), Ta, VMUL(LDK(KP382683432), Td));
T17 = VADD(TO, TP);
Ty = VFNMS(LDK(KP382683432), Ta, VMUL(LDK(KP923879532), Td));
TQ = VSUB(TO, TP);
}
{
V T1b, T1c, T1d, T1e;
{
V T15, T18, T19, T1a;
T15 = VSUB(T13, T14);
T18 = VBYI(VSUB(T16, T17));
T1b = VSUB(T15, T18);
STM2(&(xo[24]), T1b, ovs, &(xo[0]));
T1c = VADD(T15, T18);
STM2(&(xo[8]), T1c, ovs, &(xo[0]));
T19 = VADD(T13, T14);
T1a = VADD(T16, T17);
T1d = VSUB(T19, T1a);
STM2(&(xo[16]), T1d, ovs, &(xo[0]));
T1e = VADD(T19, T1a);
STM2(&(xo[0]), T1e, ovs, &(xo[0]));
}
{
V T1f, T1g, T1h, T1i;
{
V TV, T11, T10, T12, TR, TZ;
TR = VMUL(LDK(KP707106781), VSUB(TN, TQ));
TV = VBYI(VSUB(TR, TU));
T11 = VBYI(VADD(TU, TR));
TZ = VMUL(LDK(KP707106781), VADD(TN, TQ));
T10 = VSUB(TY, TZ);
T12 = VADD(TY, TZ);
T1f = VADD(TV, T10);
STM2(&(xo[12]), T1f, ovs, &(xo[0]));
T1g = VSUB(T12, T11);
STM2(&(xo[28]), T1g, ovs, &(xo[0]));
T1h = VSUB(T10, TV);
STM2(&(xo[20]), T1h, ovs, &(xo[0]));
T1i = VADD(T11, T12);
STM2(&(xo[4]), T1i, 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 = VSUB(Tu, Tv);
Tz = VSUB(Tx, Ty);
TA = VSUB(Tw, Tz);
TC = VADD(Tw, Tz);
}
{
V T1j, T1k, T1l, T1m;
T1j = VADD(Tr, TA);
STM2(&(xo[10]), T1j, ovs, &(xo[2]));
STN2(&(xo[8]), T1c, T1j, ovs);
T1k = VSUB(TC, TB);
STM2(&(xo[26]), T1k, ovs, &(xo[2]));
STN2(&(xo[24]), T1b, T1k, ovs);
T1l = VSUB(TA, Tr);
STM2(&(xo[22]), T1l, ovs, &(xo[2]));
STN2(&(xo[20]), T1h, T1l, ovs);
T1m = VADD(TB, TC);
STM2(&(xo[6]), T1m, ovs, &(xo[2]));
STN2(&(xo[4]), T1i, T1m, ovs);
}
}
{
V TF, TJ, TI, TK;
{
V TD, TE, TG, TH;
TD = VADD(Tu, Tv);
TE = VADD(T7, Te);
TF = VADD(TD, TE);
TJ = VSUB(TD, TE);
TG = VADD(Tp, Tm);
TH = VADD(Tx, Ty);
TI = VBYI(VADD(TG, TH));
TK = VBYI(VSUB(TH, TG));
}
{
V T1n, T1o, T1p, T1q;
T1n = VSUB(TF, TI);
STM2(&(xo[30]), T1n, ovs, &(xo[2]));
STN2(&(xo[28]), T1g, T1n, ovs);
T1o = VADD(TJ, TK);
STM2(&(xo[14]), T1o, ovs, &(xo[2]));
STN2(&(xo[12]), T1f, T1o, ovs);
T1p = VADD(TF, TI);
STM2(&(xo[2]), T1p, ovs, &(xo[2]));
STN2(&(xo[0]), T1e, T1p, ovs);
T1q = VSUB(TJ, TK);
STM2(&(xo[18]), T1q, ovs, &(xo[2]));
STN2(&(xo[16]), T1d, T1q, ovs);
}
}
}
}
}
}
static void n1fv_7(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(KP692021471, +0.692021471630095869627814897002069140197260599);
DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
{
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(14, is), MAKE_VOLATILE_STRIDE(14, os)) {
V T1, T2, T3, T8, T9, T5, T6;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
{
V Te, T4, Tf, Ta, Tg, T7;
Te = VSUB(T3, T2);
T4 = VADD(T2, T3);
Tf = VSUB(T9, T8);
Ta = VADD(T8, T9);
Tg = VSUB(T6, T5);
T7 = VADD(T5, T6);
{
V Tm, Tb, Tr, Th, Tj, To;
Tm = VFMA(LDK(KP554958132), Tf, Te);
Tb = VFNMS(LDK(KP356895867), T4, Ta);
Tr = VFNMS(LDK(KP554958132), Te, Tg);
Th = VFMA(LDK(KP554958132), Tg, Tf);
ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, Ta))), ovs, &(xo[0]));
Tj = VFNMS(LDK(KP356895867), T7, T4);
To = VFNMS(LDK(KP356895867), Ta, T7);
{
V Tn, Tc, Ts, Ti;
Tn = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), Tm, Tg));
Tc = VFNMS(LDK(KP692021471), Tb, T7);
Ts = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tr, Tf));
Ti = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Th, Te));
{
V Tk, Tp, Td, Tl, Tq;
Tk = VFNMS(LDK(KP692021471), Tj, Ta);
Tp = VFNMS(LDK(KP692021471), To, T4);
Td = VFNMS(LDK(KP900968867), Tc, T1);
Tl = VFNMS(LDK(KP900968867), Tk, T1);
Tq = VFNMS(LDK(KP900968867), Tp, T1);
ST(&(xo[WS(os, 2)]), VFMAI(Ti, Td), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VFNMSI(Ti, Td), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VFMAI(Tn, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VFNMSI(Tn, Tl), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(Ts, Tq), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VFNMSI(Ts, Tq), ovs, &(xo[0]));
}
}
}
}
}
}
VLEAVE();
}
static void n2fv_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 = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T3, Tj, Tf, Tk, Ta, Tn, Tc, Tm;
{
V T1, T2, Td, Te;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
Tj = VADD(T1, T2);
Td = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tf = VSUB(Td, Te);
Tk = VADD(Td, Te);
{
V T4, T5, T6, T7, T8, T9;
T4 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
T7 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T9 = VSUB(T7, T8);
Ta = VMUL(LDK(KP707106781), VADD(T6, T9));
Tn = VADD(T7, T8);
Tc = VMUL(LDK(KP707106781), VSUB(T9, T6));
Tm = VADD(T4, T5);
}
}
{
V Tb, Tg, Tp, Tq;
Tb = VADD(T3, Ta);
Tg = VBYI(VSUB(Tc, Tf));
ST(&(xo[14]), VSUB(Tb, Tg), ovs, &(xo[2]));
ST(&(xo[2]), VADD(Tb, Tg), ovs, &(xo[2]));
Tp = VSUB(Tj, Tk);
Tq = VBYI(VSUB(Tn, Tm));
ST(&(xo[12]), VSUB(Tp, Tq), ovs, &(xo[0]));
ST(&(xo[4]), VADD(Tp, Tq), ovs, &(xo[0]));
}
{
V Th, Ti, Tl, To;
Th = VSUB(T3, Ta);
Ti = VBYI(VADD(Tf, Tc));
ST(&(xo[10]), VSUB(Th, Ti), ovs, &(xo[2]));
ST(&(xo[6]), VADD(Th, Ti), ovs, &(xo[2]));
Tl = VADD(Tj, Tk);
To = VADD(Tm, Tn);
ST(&(xo[8]), VSUB(Tl, To), ovs, &(xo[0]));
ST(&(xo[0]), VADD(Tl, To), ovs, &(xo[0]));
}
}
END_SIMD();
}
static void t3bv_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) * 8)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(16, rs)) {
V T1, T8, T9, Tl, Ti, TE, T4, Ta, TO, TV, Td, Tm, TA, TH, Ts;
T1 = LDW(&(W[0]));
T8 = LDW(&(W[TWVL * 2]));
T9 = VZMUL(T1, T8);
Tl = VZMULJ(T1, T8);
Ti = LDW(&(W[TWVL * 6]));
TE = VZMULJ(T1, Ti);
T4 = LDW(&(W[TWVL * 4]));
Ta = VZMULJ(T9, T4);
TO = VZMUL(T8, T4);
TV = VZMULJ(T1, T4);
Td = VZMUL(T9, T4);
Tm = VZMULJ(Tl, T4);
TA = VZMUL(T1, T4);
TH = VZMULJ(T8, T4);
Ts = VZMUL(Tl, T4);
{
V TY, T1q, TR, T1r, T1m, T1n, TL, TZ, T1f, T1g, T1h, Th, T11, T1i, T1j;
V T1k, Tw, T12, TU, TX, TW;
TU = LD(&(x[0]), ms, &(x[0]));
TW = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
TX = VZMUL(TV, TW);
TY = VSUB(TU, TX);
T1q = VADD(TU, TX);
{
V TN, TQ, TM, TP;
TM = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
TN = VZMUL(T9, TM);
TP = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TQ = VZMUL(TO, TP);
TR = VSUB(TN, TQ);
T1r = VADD(TN, TQ);
}
{
V Tz, TJ, TC, TG, TD, TK;
{
V Ty, TI, TB, TF;
Ty = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tz = VZMUL(Tl, Ty);
TI = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
TJ = VZMUL(TH, TI);
TB = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
TC = VZMUL(TA, TB);
TF = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
TG = VZMUL(TE, TF);
}
T1m = VADD(Tz, TC);
T1n = VADD(TG, TJ);
TD = VSUB(Tz, TC);
TK = VSUB(TG, TJ);
TL = VMUL(LDK(KP707106781), VSUB(TD, TK));
TZ = VMUL(LDK(KP707106781), VADD(TD, TK));
}
{
V T3, Tf, T6, Tc, T7, Tg;
{
V T2, Te, T5, Tb;
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = VZMUL(T1, T2);
Te = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
Tf = VZMUL(Td, Te);
T5 = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
T6 = VZMUL(T4, T5);
Tb = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tc = VZMUL(Ta, Tb);
}
T1f = VADD(T3, T6);
T1g = VADD(Tc, Tf);
T1h = VSUB(T1f, T1g);
T7 = VSUB(T3, T6);
Tg = VSUB(Tc, Tf);
Th = VFNMS(LDK(KP382683432), Tg, VMUL(LDK(KP923879532), T7));
T11 = VFMA(LDK(KP382683432), T7, VMUL(LDK(KP923879532), Tg));
}
{
V Tk, Tu, To, Tr, Tp, Tv;
{
V Tj, Tt, Tn, Tq;
Tj = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
Tk = VZMUL(Ti, Tj);
Tt = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tu = VZMUL(Ts, Tt);
Tn = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
To = VZMUL(Tm, Tn);
Tq = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tr = VZMUL(T8, Tq);
}
T1i = VADD(Tk, To);
T1j = VADD(Tr, Tu);
T1k = VSUB(T1i, T1j);
Tp = VSUB(Tk, To);
Tv = VSUB(Tr, Tu);
Tw = VFMA(LDK(KP923879532), Tp, VMUL(LDK(KP382683432), Tv));
T12 = VFNMS(LDK(KP382683432), Tp, VMUL(LDK(KP923879532), Tv));
}
{
V T1p, T1v, T1u, T1w;
{
V T1l, T1o, T1s, T1t;
T1l = VMUL(LDK(KP707106781), VSUB(T1h, T1k));
T1o = VSUB(T1m, T1n);
T1p = VBYI(VSUB(T1l, T1o));
T1v = VBYI(VADD(T1o, T1l));
T1s = VSUB(T1q, T1r);
T1t = VMUL(LDK(KP707106781), VADD(T1h, T1k));
T1u = VSUB(T1s, T1t);
T1w = VADD(T1s, T1t);
}
ST(&(x[WS(rs, 6)]), VADD(T1p, T1u), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VSUB(T1w, T1v), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VSUB(T1u, T1p), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(T1v, T1w), ms, &(x[0]));
}
{
V T1z, T1D, T1C, T1E;
{
V T1x, T1y, T1A, T1B;
T1x = VADD(T1q, T1r);
T1y = VADD(T1m, T1n);
T1z = VSUB(T1x, T1y);
T1D = VADD(T1x, T1y);
T1A = VADD(T1f, T1g);
T1B = VADD(T1i, T1j);
T1C = VBYI(VSUB(T1A, T1B));
T1E = VADD(T1A, T1B);
}
ST(&(x[WS(rs, 12)]), VSUB(T1z, T1C), ms, &(x[0]));
ST(&(x[0]), VADD(T1D, T1E), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VADD(T1z, T1C), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(T1D, T1E), ms, &(x[0]));
}
{
V TT, T15, T14, T16;
{
V Tx, TS, T10, T13;
Tx = VSUB(Th, Tw);
TS = VSUB(TL, TR);
TT = VBYI(VSUB(Tx, TS));
T15 = VBYI(VADD(TS, Tx));
T10 = VSUB(TY, TZ);
T13 = VSUB(T11, T12);
T14 = VSUB(T10, T13);
T16 = VADD(T10, T13);
}
ST(&(x[WS(rs, 5)]), VADD(TT, T14), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VSUB(T16, T15), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VSUB(T14, TT), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(T15, T16), ms, &(x[WS(rs, 1)]));
}
{
V T19, T1d, T1c, T1e;
{
V T17, T18, T1a, T1b;
T17 = VADD(TY, TZ);
T18 = VADD(Th, Tw);
T19 = VADD(T17, T18);
T1d = VSUB(T17, T18);
T1a = VADD(TR, TL);
T1b = VADD(T11, T12);
T1c = VBYI(VADD(T1a, T1b));
T1e = VBYI(VSUB(T1b, T1a));
}
ST(&(x[WS(rs, 15)]), VSUB(T19, T1c), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(T1d, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(T19, T1c), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VSUB(T1d, T1e), ms, &(x[WS(rs, 1)]));
}
}
}
}
VLEAVE();
}
static void n2bv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
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(28, is), MAKE_VOLATILE_STRIDE(28, os)) {
V T3, TH, Ts, TV, TW, Tt, Tu, TU, Ta, To, Th, Tp, TC, Tx, TK;
V TQ, TN, TR, T14, TZ, T1, T2;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
TH = VADD(T1, T2);
{
V T6, TI, T9, TJ, Tn, TP, Tk, TO, Tg, TM, Td, TL;
{
V T4, T5, Ti, Tj;
T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TI = VADD(T4, T5);
{
V T7, T8, Tl, Tm;
T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T9 = VSUB(T7, T8);
TJ = VADD(T7, T8);
Tl = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tm = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tn = VSUB(Tl, Tm);
TP = VADD(Tl, Tm);
}
Ti = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tj = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Tk = VSUB(Ti, Tj);
TO = VADD(Ti, Tj);
{
V Te, Tf, Tb, Tc;
Te = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tg = VSUB(Te, Tf);
TM = VADD(Te, Tf);
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TL = VADD(Tb, Tc);
}
}
Ts = VSUB(Tk, Tn);
TV = VSUB(TP, TO);
TW = VSUB(TM, TL);
Tt = VSUB(Td, Tg);
Tu = VSUB(T6, T9);
TU = VSUB(TI, TJ);
Ta = VADD(T6, T9);
To = VADD(Tk, Tn);
Th = VADD(Td, Tg);
Tp = VFNMS(LDK(KP356895867), To, Th);
TC = VFNMS(LDK(KP356895867), Th, Ta);
Tx = VFNMS(LDK(KP356895867), Ta, To);
TK = VADD(TI, TJ);
TQ = VADD(TO, TP);
TN = VADD(TL, TM);
TR = VFNMS(LDK(KP356895867), TK, TQ);
T14 = VFNMS(LDK(KP356895867), TQ, TN);
TZ = VFNMS(LDK(KP356895867), TN, TK);
}
{
V T19, T1a, T1b, T1e, T1c, T1g, T1h;
T19 = VADD(T3, VADD(Ta, VADD(Th, To)));
STM2(&(xo[14]), T19, ovs, &(xo[2]));
T1a = VADD(TH, VADD(TK, VADD(TN, TQ)));
STM2(&(xo[0]), T1a, ovs, &(xo[0]));
{
V Tr, Tw, Tq, Tv;
Tq = VFNMS(LDK(KP692021471), Tp, Ta);
Tr = VFNMS(LDK(KP900968867), Tq, T3);
Tv = VFNMS(LDK(KP554958132), Tu, Tt);
Tw = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tv, Ts));
T1b = VFMAI(Tw, Tr);
STM2(&(xo[6]), T1b, ovs, &(xo[2]));
T1c = VFNMSI(Tw, Tr);
STM2(&(xo[22]), T1c, ovs, &(xo[2]));
}
{
V T16, T18, T15, T17, T1d;
T15 = VFNMS(LDK(KP692021471), T14, TK);
T16 = VFNMS(LDK(KP900968867), T15, TH);
T17 = VFMA(LDK(KP554958132), TU, TW);
T18 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T17, TV));
T1d = VFNMSI(T18, T16);
STM2(&(xo[20]), T1d, ovs, &(xo[0]));
STN2(&(xo[20]), T1d, T1c, ovs);
T1e = VFMAI(T18, T16);
STM2(&(xo[8]), T1e, ovs, &(xo[0]));
}
{
V Tz, TB, Ty, TA, T1f;
Ty = VFNMS(LDK(KP692021471), Tx, Th);
Tz = VFNMS(LDK(KP900968867), Ty, T3);
TA = VFMA(LDK(KP554958132), Tt, Ts);
TB = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TA, Tu));
T1f = VFNMSI(TB, Tz);
STM2(&(xo[10]), T1f, ovs, &(xo[2]));
STN2(&(xo[8]), T1e, T1f, ovs);
T1g = VFMAI(TB, Tz);
STM2(&(xo[18]), T1g, ovs, &(xo[2]));
}
{
V TT, TY, TS, TX, T1i;
TS = VFNMS(LDK(KP692021471), TR, TN);
TT = VFNMS(LDK(KP900968867), TS, TH);
TX = VFMA(LDK(KP554958132), TW, TV);
TY = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TX, TU));
T1h = VFNMSI(TY, TT);
STM2(&(xo[24]), T1h, ovs, &(xo[0]));
T1i = VFMAI(TY, TT);
STM2(&(xo[4]), T1i, ovs, &(xo[0]));
STN2(&(xo[4]), T1i, T1b, ovs);
}
{
V T11, T13, T10, T12, T1j, T1k;
T10 = VFNMS(LDK(KP692021471), TZ, TQ);
T11 = VFNMS(LDK(KP900968867), T10, TH);
T12 = VFNMS(LDK(KP554958132), TV, TU);
T13 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T12, TW));
T1j = VFNMSI(T13, T11);
STM2(&(xo[16]), T1j, ovs, &(xo[0]));
STN2(&(xo[16]), T1j, T1g, ovs);
T1k = VFMAI(T13, T11);
STM2(&(xo[12]), T1k, ovs, &(xo[0]));
STN2(&(xo[12]), T1k, T19, ovs);
}
{
V TE, TG, TD, TF, T1l, T1m;
TD = VFNMS(LDK(KP692021471), TC, To);
TE = VFNMS(LDK(KP900968867), TD, T3);
TF = VFMA(LDK(KP554958132), Ts, Tu);
TG = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TF, Tt));
T1l = VFMAI(TG, TE);
STM2(&(xo[2]), T1l, ovs, &(xo[2]));
STN2(&(xo[0]), T1a, T1l, ovs);
T1m = VFNMSI(TG, TE);
STM2(&(xo[26]), T1m, ovs, &(xo[2]));
STN2(&(xo[24]), T1h, T1m, ovs);
}
}
}
}
VLEAVE();
}
static void t3bv_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) * 8)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(16, rs)) {
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, TQ, TX, T1g, Tl, To, Tw, TG, Tz, T1p;
V T1e, TC;
{
V T1, T5, Ta, Td;
T1 = LD(&(x[0]), ms, &(x[0]));
T5 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Td = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
{
V TR, TN, TM, TE, Tb, Tp, Tm, Te, T6, TW, TO, TS;
{
V TL, Tx, T9, TU, Tc, T4, TV;
TL = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tx = VZMULJ(T2, T8);
T9 = VZMUL(T2, T8);
TR = VZMULJ(T2, Tu);
TU = VZMULJ(T8, T3);
Tc = VZMUL(T8, T3);
T4 = VZMULJ(T2, T3);
TN = VZMUL(T2, T3);
TV = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
TM = VZMUL(Tx, TL);
Ty = VZMULJ(Tx, T3);
TE = VZMUL(Tx, T3);
Tb = VZMUL(T9, Ta);
Tp = VZMUL(T9, T3);
Tm = VZMULJ(T9, T3);
Te = VZMUL(Tc, Td);
T6 = VZMUL(T4, T5);
TW = VZMUL(TU, TV);
}
TO = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
TS = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
{
V TP, TT, Ti, Tk, Tn, Th, Tq, Tj;
Th = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tq = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
Tj = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
T1o = VSUB(Tb, Te);
Tf = VADD(Tb, Te);
T1b = VSUB(T1, T6);
T7 = VADD(T1, T6);
TP = VZMUL(TN, TO);
TT = VZMUL(TR, TS);
Ti = VZMUL(T2, Th);
Tr = VZMUL(Tp, Tq);
Tk = VZMUL(T3, Tj);
Tn = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V T1c, T1d, Tv, TF;
Tv = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
TF = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
T1c = VSUB(TM, TP);
TQ = VADD(TM, TP);
T1d = VSUB(TT, TW);
TX = VADD(TT, TW);
T1g = VSUB(Ti, Tk);
Tl = VADD(Ti, Tk);
To = VZMUL(Tm, Tn);
Tw = VZMUL(Tu, Tv);
TG = VZMUL(TE, TF);
Tz = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T1p = VSUB(T1c, T1d);
T1e = VADD(T1c, T1d);
TC = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
}
}
}
}
{
V T1h, Ts, TA, TD;
T13 = VADD(T7, Tf);
Tg = VSUB(T7, Tf);
T1h = VSUB(To, Tr);
Ts = VADD(To, Tr);
TY = VSUB(TQ, TX);
T14 = VADD(TQ, TX);
TA = VZMUL(Ty, Tz);
T1A = VFNMS(LDK(KP707106781), T1p, T1o);
T1q = VFMA(LDK(KP707106781), T1p, T1o);
T1f = VFMA(LDK(KP707106781), T1e, T1b);
T1x = VFNMS(LDK(KP707106781), T1e, T1b);
TD = VZMUL(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 = VSUB(T13, T14);
T19 = VADD(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(T1i, T1l);
T1m = VADD(T1i, T1l);
{
V T18, T1a, TJ, TZ;
T18 = VSUB(T16, T17);
T1a = VADD(T16, T17);
TJ = VADD(Tt, TI);
TZ = VSUB(Tt, TI);
{
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 = VFMA(LDK(KP923879532), T1B, T1A);
T1E = VFNMS(LDK(KP923879532), T1B, T1A);
ST(&(x[WS(rs, 8)]), VSUB(T19, T1a), ms, &(x[0]));
ST(&(x[0]), VADD(T19, T1a), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(T18, T15), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VFNMSI(T18, T15), ms, &(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(rs, 15)]), VFNMSI(T1w, T1v), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(T1w, T1v), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFMAI(T1u, T1n), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(T1u, T1n), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(T1E, T1D), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VFMAI(T1E, T1D), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFNMSI(T1C, T1z), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFMAI(T1C, T1z), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VFMAI(T12, T11), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VFNMSI(T12, T11), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VFMAI(T10, TK), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(T10, TK), ms, &(x[0]));
}
}
}
}
}
}
}
}
VLEAVE();
}
static void t2bv_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 = ii;
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(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 = BYTW(&(W[TWVL * 8]), T2);
Th = BYTW(&(W[TWVL * 6]), Tg);
To = BYTW(&(W[0]), Tn);
Tj = BYTW(&(W[TWVL * 16]), Ti);
Tm = BYTW(&(W[TWVL * 10]), Tl);
T6 = BYTW(&(W[TWVL * 2]), T5);
Td = BYTW(&(W[TWVL * 4]), Tc);
T8 = BYTW(&(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 = BYTW(&(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)]), VFMAI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFNMSI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFMAI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFNMSI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFNMSI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(Tz, Ty), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(Tz, Ty), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
static void t1fv_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 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 = BYTWJ(&(W[TWVL * 6]), T2);
Ti = BYTWJ(&(W[TWVL * 2]), Th);
Tk = BYTWJ(&(W[TWVL * 10]), Tj);
T6 = BYTWJ(&(W[0]), T5);
T8 = BYTWJ(&(W[TWVL * 8]), T7);
Tb = BYTWJ(&(W[TWVL * 12]), Ta);
Td = BYTWJ(&(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 = VSUB(Tq, Tr);
Ts = VADD(Tq, Tr);
{
V Tx, Tv, Tm, Tf;
Tx = VSUB(Tu, Tt);
Tv = VADD(Tt, Tu);
Tm = VSUB(Te, T9);
Tf = VADD(T9, Te);
{
V Tp, Tn, To, Tg;
ST(&(x[WS(rs, 2)]), VFMAI(Tx, Tw), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(Tx, Tw), ms, &(x[0]));
ST(&(x[0]), VADD(Ts, Tv), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VSUB(Ts, Tv), ms, &(x[0]));
Tp = VFMA(LDK(KP707106781), Tm, Tl);
Tn = VFNMS(LDK(KP707106781), Tm, Tl);
To = VFNMS(LDK(KP707106781), Tf, T4);
Tg = VFMA(LDK(KP707106781), Tf, T4);
ST(&(x[WS(rs, 5)]), VFNMSI(Tp, To), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(Tp, To), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(Tn, Tg), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(Tn, Tg), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
VLEAVE();
}
static const R *t1fv_15(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
DVK(KP910592997, +0.910592997310029334643087372129977886038870291);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT i;
R *x;
x = ri;
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 28), MAKE_VOLATILE_STRIDE(ios)) {
V Tq, Ty, Th, T1b, T10, Ts, TP, T7, Tu, TA, TC, Tj, Tk, TQ, Tf;
{
V T1, T4, T2, T9, Te;
T1 = LD(&(x[0]), dist, &(x[0]));
T4 = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
{
V T8, Tp, Tx, Tg;
T8 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Tp = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
Tx = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
Tg = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
{
V Tb, Td, Tr, T6, Tt, Tz, TB, Ti;
{
V T5, T3, Ta, Tc;
Ta = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
Tc = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
T5 = BYTWJ(&(W[TWVL * 18]), T4);
T3 = BYTWJ(&(W[TWVL * 8]), T2);
T9 = BYTWJ(&(W[TWVL * 4]), T8);
Tq = BYTWJ(&(W[TWVL * 10]), Tp);
Ty = BYTWJ(&(W[TWVL * 16]), Tx);
Th = BYTWJ(&(W[TWVL * 22]), Tg);
Tb = BYTWJ(&(W[TWVL * 14]), Ta);
Td = BYTWJ(&(W[TWVL * 24]), Tc);
Tr = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
T1b = VSUB(T5, T3);
T6 = VADD(T3, T5);
Tt = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
}
Tz = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
TB = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
Ti = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Te = VADD(Tb, Td);
T10 = VSUB(Td, Tb);
Ts = BYTWJ(&(W[TWVL * 20]), Tr);
TP = VFNMS(LDK(KP500000000), T6, T1);
T7 = VADD(T1, T6);
Tu = BYTWJ(&(W[0]), Tt);
TA = BYTWJ(&(W[TWVL * 26]), Tz);
TC = BYTWJ(&(W[TWVL * 6]), TB);
Tj = BYTWJ(&(W[TWVL * 2]), Ti);
Tk = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
}
}
TQ = VFNMS(LDK(KP500000000), Te, T9);
Tf = VADD(T9, Te);
}
{
V Tv, T13, TD, T14, Tl;
Tv = VADD(Ts, Tu);
T13 = VSUB(Tu, Ts);
TD = VADD(TA, TC);
T14 = VSUB(TC, TA);
Tl = BYTWJ(&(W[TWVL * 12]), Tk);
{
V TT, Tw, T1d, T15, TU, TE, T11, Tm;
TT = VFNMS(LDK(KP500000000), Tv, Tq);
Tw = VADD(Tq, Tv);
T1d = VADD(T13, T14);
T15 = VSUB(T13, T14);
TU = VFNMS(LDK(KP500000000), TD, Ty);
TE = VADD(Ty, TD);
T11 = VSUB(Tl, Tj);
Tm = VADD(Tj, Tl);
{
V T19, TV, TK, TF, T1c, T12, TR, Tn;
T19 = VSUB(TT, TU);
TV = VADD(TT, TU);
TK = VSUB(Tw, TE);
TF = VADD(Tw, TE);
T1c = VADD(T10, T11);
T12 = VSUB(T10, T11);
TR = VFNMS(LDK(KP500000000), Tm, Th);
Tn = VADD(Th, Tm);
{
V T1g, T1e, T1m, T16, T18, TS, TL, To, T1f, T1u;
T1g = VSUB(T1c, T1d);
T1e = VADD(T1c, T1d);
T1m = VFNMS(LDK(KP618033988), T12, T15);
T16 = VFMA(LDK(KP618033988), T15, T12);
T18 = VSUB(TQ, TR);
TS = VADD(TQ, TR);
TL = VSUB(Tf, Tn);
To = VADD(Tf, Tn);
T1f = VFNMS(LDK(KP250000000), T1e, T1b);
T1u = VMUL(LDK(KP866025403), VADD(T1b, T1e));
{
V T1o, T1a, TY, TO, TM, TG, TI, T1p, T1h, T1t, TX, TW;
T1o = VFNMS(LDK(KP618033988), T18, T19);
T1a = VFMA(LDK(KP618033988), T19, T18);
TW = VADD(TS, TV);
TY = VSUB(TS, TV);
TO = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TK, TL));
TM = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TL, TK));
TG = VADD(To, TF);
TI = VSUB(To, TF);
T1p = VFNMS(LDK(KP559016994), T1g, T1f);
T1h = VFMA(LDK(KP559016994), T1g, T1f);
T1t = VADD(TP, TW);
TX = VFNMS(LDK(KP250000000), TW, TP);
{
V T1q, T1s, T1k, T1i, T1l, TZ, TJ, TN, TH;
ST(&(x[0]), VADD(T7, TG), dist, &(x[0]));
TH = VFNMS(LDK(KP250000000), TG, T7);
T1q = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), T1p, T1o));
T1s = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), T1p, T1o));
T1k = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), T1h, T1a));
T1i = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), T1h, T1a));
ST(&(x[WS(ios, 10)]), VFMAI(T1u, T1t), dist, &(x[0]));
ST(&(x[WS(ios, 5)]), VFNMSI(T1u, T1t), dist, &(x[WS(ios, 1)]));
T1l = VFNMS(LDK(KP559016994), TY, TX);
TZ = VFMA(LDK(KP559016994), TY, TX);
TJ = VFNMS(LDK(KP559016994), TI, TH);
TN = VFMA(LDK(KP559016994), TI, TH);
{
V T1n, T1r, T1j, T17;
T1n = VFMA(LDK(KP823639103), T1m, T1l);
T1r = VFNMS(LDK(KP823639103), T1m, T1l);
T1j = VFNMS(LDK(KP823639103), T16, TZ);
T17 = VFMA(LDK(KP823639103), T16, TZ);
ST(&(x[WS(ios, 12)]), VFMAI(TM, TJ), dist, &(x[0]));
ST(&(x[WS(ios, 3)]), VFNMSI(TM, TJ), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 9)]), VFMAI(TO, TN), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 6)]), VFNMSI(TO, TN), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VFMAI(T1q, T1n), dist, &(x[0]));
ST(&(x[WS(ios, 13)]), VFNMSI(T1q, T1n), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VFMAI(T1s, T1r), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 8)]), VFNMSI(T1s, T1r), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VFMAI(T1k, T1j), dist, &(x[0]));
ST(&(x[WS(ios, 11)]), VFNMSI(T1k, T1j), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 14)]), VFMAI(T1i, T17), dist, &(x[0]));
ST(&(x[WS(ios, 1)]), VFNMSI(T1i, T17), dist, &(x[WS(ios, 1)]));
}
}
}
}
}
}
}
}
return W;
}
static void n1bv_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 = 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, Tb, T4, Tq, Tg, Tm, T7, Tp, Ta, To, Tc, T11;
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);
Tq = VSUB(T2, T3);
Tg = VADD(Te, Tf);
Tm = VSUB(Te, Tf);
T7 = VADD(T5, T6);
Tp = VSUB(T5, T6);
Ta = VADD(T8, T9);
To = VSUB(T8, T9);
Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
}
}
T11 = VFMA(LDK(KP521108558), Tm, Tq);
{
V TA, TS, TE, TW, Td, Tn, Ts, Tw, Tr, Tv, TT, TF;
Tr = VFNMS(LDK(KP521108558), Tq, Tp);
Tv = VFNMS(LDK(KP342584725), T7, Tg);
TA = VFMA(LDK(KP715370323), To, Tq);
TS = VFMA(LDK(KP521108558), To, Tm);
TE = VFNMS(LDK(KP342584725), T4, Ta);
TW = VFNMS(LDK(KP342584725), Ta, T7);
Td = VADD(Tb, Tc);
Tn = VSUB(Tb, Tc);
Ts = VFNMS(LDK(KP715370323), Tr, To);
Tw = VFNMS(LDK(KP634356270), Tv, T4);
TT = VFNMS(LDK(KP715370323), TS, Tp);
TF = VFNMS(LDK(KP634356270), TE, Tg);
{
V Tu, TV, TD, TL, T14, TP, TZ, Tj, Tz, TI, TB, TJ, TM;
TB = VFMA(LDK(KP372785597), Tn, TA);
TJ = VFNMS(LDK(KP521108558), Tp, Tn);
{
V T12, TN, TX, Th;
T12 = VFMA(LDK(KP715370323), T11, Tn);
ST(&(xo[0]), VADD(Tg, VADD(Td, VADD(Ta, VADD(T7, VADD(T4, T1))))), ovs, &(xo[0]));
TN = VFNMS(LDK(KP342584725), Td, T4);
TX = VFNMS(LDK(KP634356270), TW, Td);
Th = VFNMS(LDK(KP342584725), Tg, Td);
{
V Tt, Tx, TU, TG;
Tt = VFNMS(LDK(KP830830026), Ts, Tn);
Tx = VFNMS(LDK(KP778434453), Tw, Ta);
TU = VFMA(LDK(KP830830026), TT, Tq);
TG = VFNMS(LDK(KP778434453), TF, Td);
{
V TC, TK, T13, TO;
TC = VFNMS(LDK(KP830830026), TB, Tm);
TK = VFMA(LDK(KP715370323), TJ, Tm);
T13 = VFMA(LDK(KP830830026), T12, Tp);
TO = VFNMS(LDK(KP634356270), TN, T7);
{
V TY, Ti, Ty, TH;
TY = VFNMS(LDK(KP778434453), TX, T4);
Ti = VFNMS(LDK(KP634356270), Th, Ta);
Tu = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), Tt, Tm));
Ty = VFNMS(LDK(KP876768831), Tx, Td);
TV = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TU, Tn));
TH = VFNMS(LDK(KP876768831), TG, T7);
TD = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), TC, Tp));
TL = VFNMS(LDK(KP830830026), TK, To);
T14 = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), T13, To));
TP = VFNMS(LDK(KP778434453), TO, Tg);
TZ = VFNMS(LDK(KP876768831), TY, Tg);
Tj = VFNMS(LDK(KP778434453), Ti, T7);
Tz = VFNMS(LDK(KP959492973), Ty, T1);
TI = VFNMS(LDK(KP959492973), TH, T1);
}
}
}
}
TM = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TL, Tq));
{
V TQ, T10, Tk, TR, Tl;
TQ = VFNMS(LDK(KP876768831), TP, Ta);
T10 = VFNMS(LDK(KP959492973), TZ, T1);
Tk = VFNMS(LDK(KP876768831), Tj, T4);
ST(&(xo[WS(os, 7)]), VFMAI(TD, Tz), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VFNMSI(TD, Tz), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VFNMSI(TM, TI), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(TM, TI), ovs, &(xo[WS(os, 1)]));
TR = VFNMS(LDK(KP959492973), TQ, T1);
ST(&(xo[WS(os, 10)]), VFNMSI(T14, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VFMAI(T14, T10), ovs, &(xo[WS(os, 1)]));
Tl = VFNMS(LDK(KP959492973), Tk, T1);
ST(&(xo[WS(os, 9)]), VFMAI(TV, TR), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VFNMSI(TV, TR), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VFNMSI(Tu, Tl), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VFMAI(Tu, Tl), ovs, &(xo[WS(os, 1)]));
}
}
}
}
}
VLEAVE();
}
