static void n2sv_4(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
{
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(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
V T3, Tb, T9, Tf, T6, Ta, Te, Tg;
{
V T1, T2, T7, T8;
T1 = LD(&(ri[0]), ivs, &(ri[0]));
T2 = LD(&(ri[WS(is, 2)]), ivs, &(ri[0]));
T3 = VADD(T1, T2);
Tb = VSUB(T1, T2);
T7 = LD(&(ii[0]), ivs, &(ii[0]));
T8 = LD(&(ii[WS(is, 2)]), ivs, &(ii[0]));
T9 = VSUB(T7, T8);
Tf = VADD(T7, T8);
}
{
V T4, T5, Tc, Td;
T4 = LD(&(ri[WS(is, 1)]), ivs, &(ri[WS(is, 1)]));
T5 = LD(&(ri[WS(is, 3)]), ivs, &(ri[WS(is, 1)]));
T6 = VADD(T4, T5);
Ta = VSUB(T4, T5);
Tc = LD(&(ii[WS(is, 1)]), ivs, &(ii[WS(is, 1)]));
Td = LD(&(ii[WS(is, 3)]), ivs, &(ii[WS(is, 1)]));
Te = VSUB(Tc, Td);
Tg = VADD(Tc, Td);
}
{
V Th, Ti, Tj, Tk;
Th = VSUB(T3, T6);
STM4(&(ro[2]), Th, ovs, &(ro[0]));
Ti = VSUB(Tf, Tg);
STM4(&(io[2]), Ti, ovs, &(io[0]));
Tj = VADD(T3, T6);
STM4(&(ro[0]), Tj, ovs, &(ro[0]));
Tk = VADD(Tf, Tg);
STM4(&(io[0]), Tk, ovs, &(io[0]));
{
V Tl, Tm, Tn, To;
Tl = VSUB(T9, Ta);
STM4(&(io[1]), Tl, ovs, &(io[1]));
Tm = VADD(Tb, Te);
STM4(&(ro[1]), Tm, ovs, &(ro[1]));
Tn = VADD(Ta, T9);
STM4(&(io[3]), Tn, ovs, &(io[1]));
STN4(&(io[0]), Tk, Tl, Ti, Tn, ovs);
To = VSUB(Tb, Te);
STM4(&(ro[3]), To, ovs, &(ro[1]));
STN4(&(ro[0]), Tj, Tm, Th, To, ovs);
}
}
}
}
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 n2sv_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
for (i = v; i > 0; i = i - (2 * VL), ri = ri + ((2 * VL) * ivs), ii = ii + ((2 * VL) * ivs), ro = ro + ((2 * VL) * ovs), io = io + ((2 * VL) * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V TF, TJ, TD, TR, TS, TT, TU, TV, TW, TE, TX, TY, TK, TI, TZ;
V T10, T11, T12;
{
V Tb, Tn, T3, TC, Ti, TB, T6, To, Tl, Tc, Tw, Tx, T8, T9, Tr;
V Ts;
{
V T1, T2, Tg, Th, T4, T5, Tj, Tk;
T1 = LD(&(ri[0]), ivs, &(ri[0]));
T2 = LD(&(ri[WS(is, 4)]), ivs, &(ri[0]));
Tg = LD(&(ii[0]), ivs, &(ii[0]));
Th = LD(&(ii[WS(is, 4)]), ivs, &(ii[0]));
T4 = LD(&(ri[WS(is, 2)]), ivs, &(ri[0]));
T5 = LD(&(ri[WS(is, 6)]), ivs, &(ri[0]));
Tj = LD(&(ii[WS(is, 2)]), ivs, &(ii[0]));
Tk = LD(&(ii[WS(is, 6)]), ivs, &(ii[0]));
Tb = LD(&(ri[WS(is, 7)]), ivs, &(ri[WS(is, 1)]));
Tn = VSUB(T1, T2);
T3 = VADD(T1, T2);
TC = VSUB(Tg, Th);
Ti = VADD(Tg, Th);
TB = VSUB(T4, T5);
T6 = VADD(T4, T5);
To = VSUB(Tj, Tk);
Tl = VADD(Tj, Tk);
Tc = LD(&(ri[WS(is, 3)]), ivs, &(ri[WS(is, 1)]));
Tw = LD(&(ii[WS(is, 7)]), ivs, &(ii[WS(is, 1)]));
Tx = LD(&(ii[WS(is, 3)]), ivs, &(ii[WS(is, 1)]));
T8 = LD(&(ri[WS(is, 1)]), ivs, &(ri[WS(is, 1)]));
T9 = LD(&(ri[WS(is, 5)]), ivs, &(ri[WS(is, 1)]));
Tr = LD(&(ii[WS(is, 1)]), ivs, &(ii[WS(is, 1)]));
Ts = LD(&(ii[WS(is, 5)]), ivs, &(ii[WS(is, 1)]));
}
{
V TL, T7, TP, Tm, Tz, TH, Te, Tf, TO, TQ, TG, Tu, Tp, TA;
{
V Td, Tv, TN, Ty, Ta, Tq, TM, Tt;
TL = VSUB(T3, T6);
T7 = VADD(T3, T6);
Td = VADD(Tb, Tc);
Tv = VSUB(Tb, Tc);
TN = VADD(Tw, Tx);
Ty = VSUB(Tw, Tx);
Ta = VADD(T8, T9);
Tq = VSUB(T8, T9);
TM = VADD(Tr, Ts);
Tt = VSUB(Tr, Ts);
TP = VADD(Ti, Tl);
Tm = VSUB(Ti, Tl);
Tz = VSUB(Tv, Ty);
TH = VADD(Tv, Ty);
Te = VADD(Ta, Td);
Tf = VSUB(Td, Ta);
TO = VSUB(TM, TN);
TQ = VADD(TM, TN);
TG = VSUB(Tt, Tq);
Tu = VADD(Tq, Tt);
}
TF = VSUB(Tn, To);
Tp = VADD(Tn, To);
TJ = VSUB(TC, TB);
TD = VADD(TB, TC);
TR = VSUB(Tm, Tf);
STM4(&(io[6]), TR, ovs, &(io[0]));
TS = VADD(Tf, Tm);
STM4(&(io[2]), TS, ovs, &(io[0]));
TT = VADD(T7, Te);
STM4(&(ro[0]), TT, ovs, &(ro[0]));
TU = VSUB(T7, Te);
STM4(&(ro[4]), TU, ovs, &(ro[0]));
TV = VADD(TP, TQ);
STM4(&(io[0]), TV, ovs, &(io[0]));
TW = VSUB(TP, TQ);
STM4(&(io[4]), TW, ovs, &(io[0]));
TE = VSUB(Tz, Tu);
TA = VADD(Tu, Tz);
TX = VADD(TL, TO);
STM4(&(ro[2]), TX, ovs, &(ro[0]));
TY = VSUB(TL, TO);
STM4(&(ro[6]), TY, ovs, &(ro[0]));
TK = VADD(TG, TH);
TI = VSUB(TG, TH);
TZ = VFMA(LDK(KP707106781), TA, Tp);
STM4(&(ro[1]), TZ, ovs, &(ro[1]));
T10 = VFNMS(LDK(KP707106781), TA, Tp);
STM4(&(ro[5]), T10, ovs, &(ro[1]));
}
}
T11 = VFMA(LDK(KP707106781), TK, TJ);
STM4(&(io[1]), T11, ovs, &(io[1]));
T12 = VFNMS(LDK(KP707106781), TK, TJ);
STM4(&(io[5]), T12, ovs, &(io[1]));
{
V T13, T14, T15, T16;
T13 = VFMA(LDK(KP707106781), TE, TD);
STM4(&(io[3]), T13, ovs, &(io[1]));
STN4(&(io[0]), TV, T11, TS, T13, ovs);
T14 = VFNMS(LDK(KP707106781), TE, TD);
STM4(&(io[7]), T14, ovs, &(io[1]));
STN4(&(io[4]), TW, T12, TR, T14, ovs);
T15 = VFMA(LDK(KP707106781), TI, TF);
STM4(&(ro[3]), T15, ovs, &(ro[1]));
STN4(&(ro[0]), TT, TZ, TX, T15, ovs);
T16 = VFNMS(LDK(KP707106781), TI, TF);
STM4(&(ro[7]), T16, ovs, &(ro[1]));
STN4(&(ro[4]), TU, T10, TY, T16, ovs);
}
}
}
VLEAVE();
}
