static void t1fv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
V Td, TA, T4, Ta, Tk, TE, Tp, TF, TB, T9, T1, T2, Tb;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V Tg, Tn, Ti, Tl;
Tg = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tn = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tl = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
{
V T6, T8, T5, Tc;
T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
{
V T3, Th, To, Tj, Tm, T7;
T7 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[TWVL * 8]), T2);
Th = BYTWJ(&(W[TWVL * 6]), Tg);
To = BYTWJ(&(W[0]), Tn);
Tj = BYTWJ(&(W[TWVL * 16]), Ti);
Tm = BYTWJ(&(W[TWVL * 10]), Tl);
T6 = BYTWJ(&(W[TWVL * 2]), T5);
Td = BYTWJ(&(W[TWVL * 4]), Tc);
T8 = BYTWJ(&(W[TWVL * 12]), T7);
TA = VADD(T1, T3);
T4 = VSUB(T1, T3);
Ta = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tk = VSUB(Th, Tj);
TE = VADD(Th, Tj);
Tp = VSUB(Tm, To);
TF = VADD(Tm, To);
}
TB = VADD(T6, T8);
T9 = VSUB(T6, T8);
}
}
Tb = BYTWJ(&(W[TWVL * 14]), Ta);
{
V TL, TG, Tw, Tq, TC, Te;
TL = VSUB(TE, TF);
TG = VADD(TE, TF);
Tw = VSUB(Tk, Tp);
Tq = VADD(Tk, Tp);
TC = VADD(Tb, Td);
Te = VSUB(Tb, Td);
{
V TM, TD, Tv, Tf;
TM = VSUB(TB, TC);
TD = VADD(TB, TC);
Tv = VSUB(T9, Te);
Tf = VADD(T9, Te);
{
V TP, TN, TH, TJ, Tz, Tx, Tr, Tt, TI, Ts;
TP = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TL, TM));
TN = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TM, TL));
TH = VADD(TD, TG);
TJ = VSUB(TD, TG);
Tz = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tv, Tw));
Tx = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tw, Tv));
Tr = VADD(Tf, Tq);
Tt = VSUB(Tf, Tq);
ST(&(x[0]), VADD(TA, TH), ms, &(x[0]));
TI = VFNMS(LDK(KP250000000), TH, TA);
ST(&(x[WS(rs, 5)]), VADD(T4, Tr), ms, &(x[WS(rs, 1)]));
Ts = VFNMS(LDK(KP250000000), Tr, T4);
{
V TK, TO, Tu, Ty;
TK = VFNMS(LDK(KP559016994), TJ, TI);
TO = VFMA(LDK(KP559016994), TJ, TI);
Tu = VFMA(LDK(KP559016994), Tt, Ts);
Ty = VFNMS(LDK(KP559016994), Tt, Ts);
ST(&(x[WS(rs, 8)]), VFNMSI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFMAI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(Tx, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(Tz, Ty), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(Tz, Ty), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
VLEAVE();
}
static void n2fv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V Tb, Tr, T3, Ts, T6, Tw, Tg, Tt, T9, Tc, T1, T2;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
{
V T4, T5, Te, Tf, T7, T8;
T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tr = VADD(T1, T2);
T3 = VSUB(T1, T2);
Ts = VADD(T4, T5);
T6 = VSUB(T4, T5);
Tw = VADD(Te, Tf);
Tg = VSUB(Te, Tf);
Tt = VADD(T7, T8);
T9 = VSUB(T7, T8);
Tc = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
}
{
V TD, Tu, Tm, Ta, Td, Tv;
TD = VSUB(Ts, Tt);
Tu = VADD(Ts, Tt);
Tm = VSUB(T6, T9);
Ta = VADD(T6, T9);
Td = VSUB(Tb, Tc);
Tv = VADD(Tb, Tc);
{
V TC, Tx, Tn, Th;
TC = VSUB(Tv, Tw);
Tx = VADD(Tv, Tw);
Tn = VSUB(Td, Tg);
Th = VADD(Td, Tg);
{
V Ty, TA, TE, TG, Ti, Tk, To, Tq;
Ty = VADD(Tu, Tx);
TA = VSUB(Tu, Tx);
TE = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TD, TC));
TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TC, TD));
Ti = VADD(Ta, Th);
Tk = VSUB(Ta, Th);
To = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tn, Tm));
Tq = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tm, Tn));
{
V Tz, TH, Tj, TI;
Tz = VFNMS(LDK(KP250000000), Ty, Tr);
TH = VADD(Tr, Ty);
STM2(&(xo[0]), TH, ovs, &(xo[0]));
Tj = VFNMS(LDK(KP250000000), Ti, T3);
TI = VADD(T3, Ti);
STM2(&(xo[10]), TI, ovs, &(xo[2]));
{
V TB, TF, Tl, Tp;
TB = VFNMS(LDK(KP559016994), TA, Tz);
TF = VFMA(LDK(KP559016994), TA, Tz);
Tl = VFMA(LDK(KP559016994), Tk, Tj);
Tp = VFNMS(LDK(KP559016994), Tk, Tj);
{
V TJ, TK, TL, TM;
TJ = VFMAI(TG, TF);
STM2(&(xo[8]), TJ, ovs, &(xo[0]));
STN2(&(xo[8]), TJ, TI, ovs);
TK = VFNMSI(TG, TF);
STM2(&(xo[12]), TK, ovs, &(xo[0]));
TL = VFNMSI(TE, TB);
STM2(&(xo[16]), TL, ovs, &(xo[0]));
TM = VFMAI(TE, TB);
STM2(&(xo[4]), TM, ovs, &(xo[0]));
{
V TN, TO, TP, TQ;
TN = VFNMSI(Tq, Tp);
STM2(&(xo[6]), TN, ovs, &(xo[2]));
STN2(&(xo[4]), TM, TN, ovs);
TO = VFMAI(Tq, Tp);
STM2(&(xo[14]), TO, ovs, &(xo[2]));
STN2(&(xo[12]), TK, TO, ovs);
TP = VFMAI(To, Tl);
STM2(&(xo[18]), TP, ovs, &(xo[2]));
STN2(&(xo[16]), TL, TP, ovs);
TQ = VFNMSI(To, Tl);
STM2(&(xo[2]), TQ, ovs, &(xo[2]));
STN2(&(xo[0]), TH, TQ, ovs);
}
}
}
}
}
}
}
}
}
static void t1bv_12(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 22)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(12, rs)) {
V TI, Ti, TA, T7, Tm, TE, Tw, Tk, Tf, TB, TU, TM;
{
V T9, TK, Tj, TL, Te;
{
V T1, T4, T2, Tp, Tt, Tr;
T1 = LD(&(x[0]), ms, &(x[0]));
T4 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tp = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tt = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tr = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
{
V T5, T3, Tq, Tu, Ts, Td, Tb, T8, Tc, Ta;
T8 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Ta = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
T5 = BYTW(&(W[TWVL * 14]), T4);
T3 = BYTW(&(W[TWVL * 6]), T2);
Tq = BYTW(&(W[TWVL * 16]), Tp);
Tu = BYTW(&(W[TWVL * 8]), Tt);
Ts = BYTW(&(W[0]), Tr);
T9 = BYTW(&(W[TWVL * 10]), T8);
Td = BYTW(&(W[TWVL * 2]), Tc);
Tb = BYTW(&(W[TWVL * 18]), Ta);
{
V Th, T6, Tl, Tv;
Th = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
TK = VSUB(T3, T5);
T6 = VADD(T3, T5);
Tl = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tv = VADD(Ts, Tu);
TI = VSUB(Tu, Ts);
Tj = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
TL = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Ti = BYTW(&(W[TWVL * 4]), Th);
TA = VFNMS(LDK(KP500000000), T6, T1);
T7 = VADD(T1, T6);
Tm = BYTW(&(W[TWVL * 20]), Tl);
TE = VFNMS(LDK(KP500000000), Tv, Tq);
Tw = VADD(Tq, Tv);
}
}
}
Tk = BYTW(&(W[TWVL * 12]), Tj);
Tf = VADD(T9, Te);
TB = VFNMS(LDK(KP500000000), Te, T9);
TU = VSUB(TK, TL);
TM = VADD(TK, TL);
}
{
V Tn, TH, TC, TQ, Ty, Tg;
Tn = VADD(Tk, Tm);
TH = VSUB(Tk, Tm);
TC = VADD(TA, TB);
TQ = VSUB(TA, TB);
Ty = VADD(T7, Tf);
Tg = VSUB(T7, Tf);
{
V To, TD, TJ, TR;
To = VADD(Ti, Tn);
TD = VFNMS(LDK(KP500000000), Tn, Ti);
TJ = VSUB(TH, TI);
TR = VADD(TH, TI);
{
V TP, TN, TW, TS, TO, TG, TX, TV;
{
V Tz, Tx, TF, TT;
Tz = VADD(To, Tw);
Tx = VSUB(To, Tw);
TF = VADD(TD, TE);
TT = VSUB(TD, TE);
TP = VMUL(LDK(KP866025403), VADD(TM, TJ));
TN = VMUL(LDK(KP866025403), VSUB(TJ, TM));
TW = VFMA(LDK(KP866025403), TR, TQ);
TS = VFNMS(LDK(KP866025403), TR, TQ);
ST(&(x[WS(rs, 6)]), VSUB(Ty, Tz), ms, &(x[0]));
ST(&(x[0]), VADD(Ty, Tz), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFMAI(Tx, Tg), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(Tx, Tg), ms, &(x[WS(rs, 1)]));
TO = VADD(TC, TF);
TG = VSUB(TC, TF);
TX = VFNMS(LDK(KP866025403), TU, TT);
TV = VFMA(LDK(KP866025403), TU, TT);
}
ST(&(x[WS(rs, 8)]), VFNMSI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TG), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VFNMSI(TN, TG), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VFMAI(TX, TW), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(TX, TW), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFNMSI(TV, TS), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(TV, TS), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
static void n1bv_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 = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V T1, T2, 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 Tg, T4, Te, Ta, Tf, T7;
Tg = VSUB(T2, T3);
T4 = VADD(T2, T3);
Te = VSUB(T8, T9);
Ta = VADD(T8, T9);
Tf = VSUB(T5, T6);
T7 = VADD(T5, T6);
{
V Tr, Tj, Tm, Th, To, Tb;
Tr = VFMA(LDK(KP554958132), Te, Tg);
Tj = VFNMS(LDK(KP356895867), T4, Ta);
Tm = VFMA(LDK(KP554958132), Tf, Te);
Th = VFNMS(LDK(KP554958132), Tg, Tf);
ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, Ta))), ovs, &(xo[0]));
To = VFNMS(LDK(KP356895867), T7, T4);
Tb = VFNMS(LDK(KP356895867), Ta, T7);
{
V Ts, Tk, Tn, Ti;
Ts = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), Tr, Tf));
Tk = VFNMS(LDK(KP692021471), Tj, T7);
Tn = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tm, Tg));
Ti = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Th, Te));
{
V Tp, Tc, Tl, Tq, Td;
Tp = VFNMS(LDK(KP692021471), To, Ta);
Tc = VFNMS(LDK(KP692021471), Tb, T4);
Tl = VFNMS(LDK(KP900968867), Tk, T1);
Tq = VFNMS(LDK(KP900968867), Tp, T1);
Td = VFNMS(LDK(KP900968867), Tc, T1);
ST(&(xo[WS(os, 5)]), VFNMSI(Tn, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VFMAI(Tn, Tl), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VFNMSI(Ts, Tq), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VFMAI(Ts, Tq), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VFNMSI(Ti, Td), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(Ti, Td), ovs, &(xo[WS(os, 1)]));
}
}
}
}
}
}
VLEAVE();
}
static void t1fv_9(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
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 m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 16)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 16), MAKE_VOLATILE_STRIDE(9, rs)) {
V T1, T3, T5, T9, Th, Tb, Td, Tj, Tl, TD, T6;
T1 = LD(&(x[0]), ms, &(x[0]));
{
V T2, T4, T8, Tg;
T2 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tg = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
{
V Ta, Tc, Ti, Tk;
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tk = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T3 = BYTWJ(&(W[TWVL * 4]), T2);
T5 = BYTWJ(&(W[TWVL * 10]), T4);
T9 = BYTWJ(&(W[0]), T8);
Th = BYTWJ(&(W[TWVL * 2]), Tg);
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Td = BYTWJ(&(W[TWVL * 12]), Tc);
Tj = BYTWJ(&(W[TWVL * 8]), Ti);
Tl = BYTWJ(&(W[TWVL * 14]), Tk);
}
}
TD = VSUB(T5, T3);
T6 = VADD(T3, T5);
{
V Tt, Te, Tu, Tm, Tr, T7;
Tt = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Tu = VSUB(Tl, Tj);
Tm = VADD(Tj, Tl);
Tr = VFNMS(LDK(KP500000000), T6, T1);
T7 = VADD(T1, T6);
{
V Tv, Tf, Ts, Tn;
Tv = VFNMS(LDK(KP500000000), Te, T9);
Tf = VADD(T9, Te);
Ts = VFNMS(LDK(KP500000000), Tm, Th);
Tn = VADD(Th, Tm);
{
V TG, TK, Tw, TJ, TF, TA, To, Tq;
TG = VFNMS(LDK(KP726681596), Tt, Tv);
TK = VFMA(LDK(KP968908795), Tv, Tt);
Tw = VFNMS(LDK(KP586256827), Tv, Tu);
TJ = VFNMS(LDK(KP152703644), Tu, Ts);
TF = VFMA(LDK(KP203604859), Ts, Tu);
TA = VFNMS(LDK(KP439692620), Tt, Ts);
To = VADD(Tf, Tn);
Tq = VMUL(LDK(KP866025403), VSUB(Tn, Tf));
{
V TQ, TH, TL, TN, TB, Tp, Ty, TI, Tx;
Tx = VFNMS(LDK(KP347296355), Tw, Tt);
TQ = VFNMS(LDK(KP898197570), TG, TF);
TH = VFMA(LDK(KP898197570), TG, TF);
TL = VFMA(LDK(KP673648177), TK, TJ);
TN = VFNMS(LDK(KP673648177), TK, TJ);
TB = VFNMS(LDK(KP420276625), TA, Tu);
ST(&(x[0]), VADD(T7, To), ms, &(x[0]));
Tp = VFNMS(LDK(KP500000000), To, T7);
Ty = VFNMS(LDK(KP907603734), Tx, Ts);
TI = VFMA(LDK(KP852868531), TH, Tr);
{
V TO, TR, TM, TC, Tz, TP, TS, TE;
TO = VFNMS(LDK(KP500000000), TH, TN);
TR = VFMA(LDK(KP666666666), TL, TQ);
TM = VMUL(LDK(KP984807753), VFNMS(LDK(KP879385241), TD, TL));
TC = VFNMS(LDK(KP826351822), TB, Tv);
ST(&(x[WS(rs, 6)]), VFNMSI(Tq, Tp), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VFMAI(Tq, Tp), ms, &(x[WS(rs, 1)]));
Tz = VFNMS(LDK(KP939692620), Ty, Tr);
TP = VFMA(LDK(KP852868531), TO, Tr);
TS = VMUL(LDK(KP866025403), VFMA(LDK(KP852868531), TR, TD));
ST(&(x[WS(rs, 8)]), VFMAI(TM, TI), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VFNMSI(TM, TI), ms, &(x[WS(rs, 1)]));
TE = VMUL(LDK(KP984807753), VFMA(LDK(KP879385241), TD, TC));
ST(&(x[WS(rs, 4)]), VFMAI(TS, TP), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VFNMSI(TS, TP), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(TE, Tz), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VFNMSI(TE, Tz), ms, &(x[0]));
}
}
}
}
}
}
}
VLEAVE();
}
static void m2fv_32_0(const R *xi, R *xo, stride is, int ivs, int ovs)
{
DVK(KP195090322, +0.195090322016128267848284868477022240927691618);
DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
DVK(KP555570233, +0.555570233019602224742830813948532874374937191);
DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
V T1T, T1W, T2K, T2x, T16, T1A, Tb, T1p, TT, T1v, TY, T1w, T27, T2a, T2b;
V T2H, T2O, TC, T1s, TH, T1t, T20, T23, T24, T2E, T2N, T2g, T2j, Tq, T1B;
V T19, T1q, T2A, T2L;
{
V T3, T1R, T15, T1S, T6, T1U, T9, T1V, T12, Ta;
{
V T1, T2, T13, T14;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
T1R = VADD(T1, T2);
T13 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T14 = LD(&(xi[WS(is, 24)]), ivs, &(xi[0]));
T15 = VSUB(T13, T14);
T1S = VADD(T13, T14);
}
{
V T4, T5, T7, T8;
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 20)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
T1U = VADD(T4, T5);
T7 = LD(&(xi[WS(is, 28)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
T1V = VADD(T7, T8);
}
T1T = VADD(T1R, T1S);
T1W = VADD(T1U, T1V);
T2K = VSUB(T1V, T1U);
T2x = VSUB(T1R, T1S);
T12 = VMUL(LDK(KP707106781), VSUB(T9, T6));
T16 = VSUB(T12, T15);
T1A = VADD(T15, T12);
Ta = VMUL(LDK(KP707106781), VADD(T6, T9));
Tb = VADD(T3, Ta);
T1p = VSUB(T3, Ta);
}
{
V TL, T25, TX, T26, TO, T28, TR, T29;
{
V TJ, TK, TV, TW;
TJ = LD(&(xi[WS(is, 31)]), ivs, &(xi[WS(is, 1)]));
TK = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
TL = VSUB(TJ, TK);
T25 = VADD(TJ, TK);
TV = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
TW = LD(&(xi[WS(is, 23)]), ivs, &(xi[WS(is, 1)]));
TX = VSUB(TV, TW);
T26 = VADD(TV, TW);
}
{
V TM, TN, TP, TQ;
TM = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
TN = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
TO = VSUB(TM, TN);
T28 = VADD(TM, TN);
TP = LD(&(xi[WS(is, 27)]), ivs, &(xi[WS(is, 1)]));
TQ = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
TR = VSUB(TP, TQ);
T29 = VADD(TP, TQ);
}
{
V TS, TU, T2F, T2G;
TS = VMUL(LDK(KP707106781), VADD(TO, TR));
TT = VADD(TL, TS);
T1v = VSUB(TL, TS);
TU = VMUL(LDK(KP707106781), VSUB(TR, TO));
TY = VSUB(TU, TX);
T1w = VADD(TX, TU);
T27 = VADD(T25, T26);
T2a = VADD(T28, T29);
T2b = VSUB(T27, T2a);
T2F = VSUB(T25, T26);
T2G = VSUB(T29, T28);
T2H = VFNMS(LDK(KP382683432), T2G, VMUL(LDK(KP923879532), T2F));
T2O = VFMA(LDK(KP382683432), T2F, VMUL(LDK(KP923879532), T2G));
}
}
{
V Tu, T1Y, TG, T1Z, Tx, T21, TA, T22;
{
V Ts, Tt, TE, TF;
Ts = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tt = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
Tu = VSUB(Ts, Tt);
T1Y = VADD(Ts, Tt);
TE = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
TF = LD(&(xi[WS(is, 25)]), ivs, &(xi[WS(is, 1)]));
TG = VSUB(TE, TF);
T1Z = VADD(TE, TF);
}
{
V Tv, Tw, Ty, Tz;
Tv = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tw = LD(&(xi[WS(is, 21)]), ivs, &(xi[WS(is, 1)]));
Tx = VSUB(Tv, Tw);
T21 = VADD(Tv, Tw);
Ty = LD(&(xi[WS(is, 29)]), ivs, &(xi[WS(is, 1)]));
Tz = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
TA = VSUB(Ty, Tz);
T22 = VADD(Ty, Tz);
}
{
V TB, TD, T2C, T2D;
TB = VMUL(LDK(KP707106781), VADD(Tx, TA));
TC = VADD(Tu, TB);
T1s = VSUB(Tu, TB);
TD = VMUL(LDK(KP707106781), VSUB(TA, Tx));
TH = VSUB(TD, TG);
T1t = VADD(TG, TD);
T20 = VADD(T1Y, T1Z);
T23 = VADD(T21, T22);
T24 = VSUB(T20, T23);
T2C = VSUB(T1Y, T1Z);
T2D = VSUB(T22, T21);
T2E = VFMA(LDK(KP923879532), T2C, VMUL(LDK(KP382683432), T2D));
T2N = VFNMS(LDK(KP382683432), T2C, VMUL(LDK(KP923879532), T2D));
}
}
{
V Te, T2h, To, T2f, Th, T2i, Tl, T2e, Ti, Tp;
{
V Tc, Td, Tm, Tn;
Tc = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Td = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
Te = VSUB(Tc, Td);
T2h = VADD(Tc, Td);
Tm = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tn = LD(&(xi[WS(is, 22)]), ivs, &(xi[0]));
To = VSUB(Tm, Tn);
T2f = VADD(Tm, Tn);
}
{
V Tf, Tg, Tj, Tk;
Tf = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tg = LD(&(xi[WS(is, 26)]), ivs, &(xi[0]));
Th = VSUB(Tf, Tg);
T2i = VADD(Tf, Tg);
Tj = LD(&(xi[WS(is, 30)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Tl = VSUB(Tj, Tk);
T2e = VADD(Tj, Tk);
}
T2g = VADD(T2e, T2f);
T2j = VADD(T2h, T2i);
Ti = VFNMS(LDK(KP382683432), Th, VMUL(LDK(KP923879532), Te));
Tp = VFMA(LDK(KP923879532), Tl, VMUL(LDK(KP382683432), To));
Tq = VADD(Ti, Tp);
T1B = VSUB(Tp, Ti);
{
V T17, T18, T2y, T2z;
T17 = VFNMS(LDK(KP923879532), To, VMUL(LDK(KP382683432), Tl));
T18 = VFMA(LDK(KP382683432), Te, VMUL(LDK(KP923879532), Th));
T19 = VSUB(T17, T18);
T1q = VADD(T18, T17);
T2y = VSUB(T2h, T2i);
T2z = VSUB(T2e, T2f);
T2A = VMUL(LDK(KP707106781), VADD(T2y, T2z));
T2L = VMUL(LDK(KP707106781), VSUB(T2z, T2y));
}
}
{
V T2d, T2n, T2m, T2o;
{
V T1X, T2c, T2k, T2l;
T1X = VSUB(T1T, T1W);
T2c = VMUL(LDK(KP707106781), VADD(T24, T2b));
T2d = VADD(T1X, T2c);
T2n = VSUB(T1X, T2c);
T2k = VSUB(T2g, T2j);
T2l = VMUL(LDK(KP707106781), VSUB(T2b, T24));
T2m = VBYI(VADD(T2k, T2l));
T2o = VBYI(VSUB(T2l, T2k));
}
ST(&(xo[56]), VSUB(T2d, T2m), ovs, &(xo[0]));
ST(&(xo[24]), VADD(T2n, T2o), ovs, &(xo[0]));
ST(&(xo[8]), VADD(T2d, T2m), ovs, &(xo[0]));
ST(&(xo[40]), VSUB(T2n, T2o), ovs, &(xo[0]));
}
{
V T2r, T2v, T2u, T2w;
{
V T2p, T2q, T2s, T2t;
T2p = VADD(T1T, T1W);
T2q = VADD(T2j, T2g);
T2r = VADD(T2p, T2q);
T2v = VSUB(T2p, T2q);
T2s = VADD(T20, T23);
T2t = VADD(T27, T2a);
T2u = VADD(T2s, T2t);
T2w = VBYI(VSUB(T2t, T2s));
}
ST(&(xo[32]), VSUB(T2r, T2u), ovs, &(xo[0]));
ST(&(xo[16]), VADD(T2v, T2w), ovs, &(xo[0]));
ST(&(xo[0]), VADD(T2r, T2u), ovs, &(xo[0]));
ST(&(xo[48]), VSUB(T2v, T2w), ovs, &(xo[0]));
}
{
V T2V, T2Z, T2Y, T30;
{
V T2T, T2U, T2W, T2X;
T2T = VSUB(T2H, T2E);
T2U = VSUB(T2L, T2K);
T2V = VBYI(VSUB(T2T, T2U));
T2Z = VBYI(VADD(T2U, T2T));
T2W = VSUB(T2x, T2A);
T2X = VSUB(T2O, T2N);
T2Y = VSUB(T2W, T2X);
T30 = VADD(T2W, T2X);
}
ST(&(xo[20]), VADD(T2V, T2Y), ovs, &(xo[0]));
ST(&(xo[52]), VSUB(T30, T2Z), ovs, &(xo[0]));
ST(&(xo[44]), VSUB(T2Y, T2V), ovs, &(xo[0]));
ST(&(xo[12]), VADD(T2Z, T30), ovs, &(xo[0]));
}
{
V T2J, T2R, T2Q, T2S;
{
V T2B, T2I, T2M, T2P;
T2B = VADD(T2x, T2A);
T2I = VADD(T2E, T2H);
T2J = VADD(T2B, T2I);
T2R = VSUB(T2B, T2I);
T2M = VADD(T2K, T2L);
T2P = VADD(T2N, T2O);
T2Q = VBYI(VADD(T2M, T2P));
T2S = VBYI(VSUB(T2P, T2M));
}
ST(&(xo[60]), VSUB(T2J, T2Q), ovs, &(xo[0]));
ST(&(xo[28]), VADD(T2R, T2S), ovs, &(xo[0]));
ST(&(xo[4]), VADD(T2J, T2Q), ovs, &(xo[0]));
ST(&(xo[36]), VSUB(T2R, T2S), ovs, &(xo[0]));
}
{
V T1r, T1C, T1M, T1K, T1F, T1N, T1y, T1J;
T1r = VADD(T1p, T1q);
T1C = VADD(T1A, T1B);
T1M = VSUB(T1p, T1q);
T1K = VSUB(T1B, T1A);
{
V T1D, T1E, T1u, T1x;
T1D = VFNMS(LDK(KP555570233), T1s, VMUL(LDK(KP831469612), T1t));
T1E = VFMA(LDK(KP555570233), T1v, VMUL(LDK(KP831469612), T1w));
T1F = VADD(T1D, T1E);
T1N = VSUB(T1E, T1D);
T1u = VFMA(LDK(KP831469612), T1s, VMUL(LDK(KP555570233), T1t));
T1x = VFNMS(LDK(KP555570233), T1w, VMUL(LDK(KP831469612), T1v));
T1y = VADD(T1u, T1x);
T1J = VSUB(T1x, T1u);
}
{
V T1z, T1G, T1P, T1Q;
T1z = VADD(T1r, T1y);
T1G = VBYI(VADD(T1C, T1F));
ST(&(xo[58]), VSUB(T1z, T1G), ovs, &(xo[2]));
ST(&(xo[6]), VADD(T1z, T1G), ovs, &(xo[2]));
T1P = VBYI(VADD(T1K, T1J));
T1Q = VADD(T1M, T1N);
ST(&(xo[10]), VADD(T1P, T1Q), ovs, &(xo[2]));
ST(&(xo[54]), VSUB(T1Q, T1P), ovs, &(xo[2]));
}
{
V T1H, T1I, T1L, T1O;
T1H = VSUB(T1r, T1y);
T1I = VBYI(VSUB(T1F, T1C));
ST(&(xo[38]), VSUB(T1H, T1I), ovs, &(xo[2]));
ST(&(xo[26]), VADD(T1H, T1I), ovs, &(xo[2]));
T1L = VBYI(VSUB(T1J, T1K));
T1O = VSUB(T1M, T1N);
ST(&(xo[22]), VADD(T1L, T1O), ovs, &(xo[2]));
ST(&(xo[42]), VSUB(T1O, T1L), ovs, &(xo[2]));
}
}
{
V Tr, T1a, T1k, T1i, T1d, T1l, T10, T1h;
Tr = VADD(Tb, Tq);
T1a = VADD(T16, T19);
T1k = VSUB(Tb, Tq);
T1i = VSUB(T19, T16);
{
V T1b, T1c, TI, TZ;
T1b = VFNMS(LDK(KP195090322), TC, VMUL(LDK(KP980785280), TH));
T1c = VFMA(LDK(KP195090322), TT, VMUL(LDK(KP980785280), TY));
T1d = VADD(T1b, T1c);
T1l = VSUB(T1c, T1b);
TI = VFMA(LDK(KP980785280), TC, VMUL(LDK(KP195090322), TH));
TZ = VFNMS(LDK(KP195090322), TY, VMUL(LDK(KP980785280), TT));
T10 = VADD(TI, TZ);
T1h = VSUB(TZ, TI);
}
{
V T11, T1e, T1n, T1o;
T11 = VADD(Tr, T10);
T1e = VBYI(VADD(T1a, T1d));
ST(&(xo[62]), VSUB(T11, T1e), ovs, &(xo[2]));
ST(&(xo[2]), VADD(T11, T1e), ovs, &(xo[2]));
T1n = VBYI(VADD(T1i, T1h));
T1o = VADD(T1k, T1l);
ST(&(xo[14]), VADD(T1n, T1o), ovs, &(xo[2]));
ST(&(xo[50]), VSUB(T1o, T1n), ovs, &(xo[2]));
}
{
V T1f, T1g, T1j, T1m;
T1f = VSUB(Tr, T10);
T1g = VBYI(VSUB(T1d, T1a));
ST(&(xo[34]), VSUB(T1f, T1g), ovs, &(xo[2]));
ST(&(xo[30]), VADD(T1f, T1g), ovs, &(xo[2]));
T1j = VBYI(VSUB(T1h, T1i));
T1m = VSUB(T1k, T1l);
ST(&(xo[18]), VADD(T1j, T1m), ovs, &(xo[2]));
ST(&(xo[46]), VSUB(T1m, T1j), ovs, &(xo[2]));
}
}
}
}
static void t1fuv_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 12)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 12), MAKE_VOLATILE_STRIDE(rs)) {
V T1, Tg, Tj, T6, Ti, Tb, Tk, Tp, To;
T1 = LD(&(x[0]), ms, &(x[0]));
{
V Td, Tf, Tc, Te;
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[TWVL * 4]), Tc);
Te = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tf = BYTWJ(&(W[TWVL * 6]), Te);
Tg = VADD(Td, Tf);
Tj = VSUB(Tf, Td);
}
{
V T3, T5, T2, T4;
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[0]), T2);
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 10]), T4);
T6 = VADD(T3, T5);
Ti = VSUB(T5, T3);
}
{
V T8, Ta, T7, T9;
T7 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T8 = BYTWJ(&(W[TWVL * 2]), T7);
T9 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Ta = BYTWJ(&(W[TWVL * 8]), T9);
Tb = VADD(T8, Ta);
Tk = VSUB(Ta, T8);
}
ST(&(x[0]), VADD(T1, VADD(T6, VADD(Tb, Tg))), ms, &(x[0]));
Tp = VBYI(VFMA(LDK(KP433883739), Ti, VFNMS(LDK(KP781831482), Tk, VMUL(LDK(KP974927912), Tj))));
To = VFMA(LDK(KP623489801), Tb, VFNMS(LDK(KP222520933), Tg, VFNMS(LDK(KP900968867), T6, T1)));
ST(&(x[WS(rs, 4)]), VSUB(To, Tp), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VADD(To, Tp), ms, &(x[WS(rs, 1)]));
{
V Tl, Th, Tn, Tm;
Tl = VBYI(VFNMS(LDK(KP781831482), Tj, VFNMS(LDK(KP433883739), Tk, VMUL(LDK(KP974927912), Ti))));
Th = VFMA(LDK(KP623489801), Tg, VFNMS(LDK(KP900968867), Tb, VFNMS(LDK(KP222520933), T6, T1)));
ST(&(x[WS(rs, 5)]), VSUB(Th, Tl), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VADD(Th, Tl), ms, &(x[0]));
Tn = VBYI(VFMA(LDK(KP781831482), Ti, VFMA(LDK(KP974927912), Tk, VMUL(LDK(KP433883739), Tj))));
Tm = VFMA(LDK(KP623489801), T6, VFNMS(LDK(KP900968867), Tg, VFNMS(LDK(KP222520933), Tb, T1)));
ST(&(x[WS(rs, 6)]), VSUB(Tm, Tn), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VADD(Tm, Tn), ms, &(x[WS(rs, 1)]));
}
}
}
VLEAVE();
}
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 T1, T6, Tk, Tn, Tc, Td, Tf, Tr, T4, Ts, T9, Tg, Te, Tl;
{
V T2, T3, T7, T8;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tn = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tr = VSUB(T3, T2);
T4 = VADD(T2, T3);
Ts = VSUB(T8, T7);
T9 = VADD(T7, T8);
Tg = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
}
Te = VSUB(Tc, Td);
Tl = VADD(Td, Tc);
{
V T5, TF, TB, Tt, Ta, TG, Th, To, Tm, TI;
T5 = VFNMS(LDK(KP500000000), T4, T1);
TF = VADD(T1, T4);
TB = VADD(Tr, Ts);
Tt = VSUB(Tr, Ts);
Ta = VFNMS(LDK(KP500000000), T9, T6);
TG = VADD(T6, T9);
Th = VSUB(Tf, Tg);
To = VADD(Tf, Tg);
Tm = VFNMS(LDK(KP500000000), Tl, Tk);
TI = VADD(Tk, Tl);
{
V TH, TL, Tb, Tx, TJ, Tp, Ti, TA;
TH = VSUB(TF, TG);
TL = VADD(TF, TG);
Tb = VSUB(T5, Ta);
Tx = VADD(T5, Ta);
TJ = VADD(Tn, To);
Tp = VFNMS(LDK(KP500000000), To, Tn);
Ti = VADD(Te, Th);
TA = VSUB(Te, Th);
{
V Tq, Ty, TK, TM;
Tq = VSUB(Tm, Tp);
Ty = VADD(Tm, Tp);
TK = VSUB(TI, TJ);
TM = VADD(TI, TJ);
{
V TC, TE, Tj, Tv;
TC = VMUL(LDK(KP866025403), VSUB(TA, TB));
TE = VMUL(LDK(KP866025403), VADD(TB, TA));
Tj = VFMA(LDK(KP866025403), Ti, Tb);
Tv = VFNMS(LDK(KP866025403), Ti, Tb);
{
V Tz, TD, Tu, Tw;
Tz = VSUB(Tx, Ty);
TD = VADD(Tx, Ty);
Tu = VFNMS(LDK(KP866025403), Tt, Tq);
Tw = VFMA(LDK(KP866025403), Tt, Tq);
{
V TN, TO, TP, TQ;
TN = VADD(TL, TM);
STM2(&(xo[0]), TN, ovs, &(xo[0]));
TO = VSUB(TL, TM);
STM2(&(xo[12]), TO, ovs, &(xo[0]));
TP = VFMAI(TK, TH);
STM2(&(xo[6]), TP, ovs, &(xo[2]));
TQ = VFNMSI(TK, TH);
STM2(&(xo[18]), TQ, ovs, &(xo[2]));
{
V TR, TS, TT, TU;
TR = VFMAI(TE, TD);
STM2(&(xo[8]), TR, ovs, &(xo[0]));
TS = VFNMSI(TE, TD);
STM2(&(xo[16]), TS, ovs, &(xo[0]));
STN2(&(xo[16]), TS, TQ, ovs);
TT = VFNMSI(TC, Tz);
STM2(&(xo[20]), TT, ovs, &(xo[0]));
TU = VFMAI(TC, Tz);
STM2(&(xo[4]), TU, ovs, &(xo[0]));
STN2(&(xo[4]), TU, TP, ovs);
{
V TV, TW, TX, TY;
TV = VFNMSI(Tw, Tv);
STM2(&(xo[10]), TV, ovs, &(xo[2]));
STN2(&(xo[8]), TR, TV, ovs);
TW = VFMAI(Tw, Tv);
STM2(&(xo[14]), TW, ovs, &(xo[2]));
STN2(&(xo[12]), TO, TW, ovs);
TX = VFMAI(Tu, Tj);
STM2(&(xo[22]), TX, ovs, &(xo[2]));
STN2(&(xo[20]), TT, TX, ovs);
TY = VFNMSI(Tu, Tj);
STM2(&(xo[2]), TY, ovs, &(xo[2]));
STN2(&(xo[0]), TN, TY, ovs);
}
}
}
}
}
}
}
}
}
}
VLEAVE();
}
static void n2bv_20(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
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 T3, T1y, TH, T1i, Ts, TL, TM, Tz, T13, T16, T1j, T1u, T1v, T1w, T1r;
V T1s, T1t, T1a, T1d, T1k, Ti, Tk, TE, TI;
{
V T1, T2, T1g, TF, TG, T1h;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T1g = VADD(T1, T2);
TF = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
TG = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
T1h = VADD(TF, TG);
T3 = VSUB(T1, T2);
T1y = VADD(T1g, T1h);
TH = VSUB(TF, TG);
T1i = VSUB(T1g, T1h);
}
{
V T6, T11, Tv, T19, Ty, T1c, T9, T14, Td, T18, To, T12, Tr, T15, Tg;
V T1b;
{
V T4, T5, Tt, Tu;
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
T11 = VADD(T4, T5);
Tt = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Tu = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tv = VSUB(Tt, Tu);
T19 = VADD(Tt, Tu);
}
{
V Tw, Tx, T7, T8;
Tw = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
Tx = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Ty = VSUB(Tw, Tx);
T1c = VADD(Tw, Tx);
T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
T14 = VADD(T7, T8);
}
{
V Tb, Tc, Tm, Tn;
Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
Td = VSUB(Tb, Tc);
T18 = VADD(Tb, Tc);
Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tn = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
To = VSUB(Tm, Tn);
T12 = VADD(Tm, Tn);
}
{
V Tp, Tq, Te, Tf;
Tp = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tq = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tr = VSUB(Tp, Tq);
T15 = VADD(Tp, Tq);
Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tg = VSUB(Te, Tf);
T1b = VADD(Te, Tf);
}
Ts = VSUB(To, Tr);
TL = VSUB(T6, T9);
TM = VSUB(Td, Tg);
Tz = VSUB(Tv, Ty);
T13 = VSUB(T11, T12);
T16 = VSUB(T14, T15);
T1j = VADD(T13, T16);
T1u = VADD(T18, T19);
T1v = VADD(T1b, T1c);
T1w = VADD(T1u, T1v);
T1r = VADD(T11, T12);
T1s = VADD(T14, T15);
T1t = VADD(T1r, T1s);
T1a = VSUB(T18, T19);
T1d = VSUB(T1b, T1c);
T1k = VADD(T1a, T1d);
{
V Ta, Th, TC, TD;
Ta = VADD(T6, T9);
Th = VADD(Td, Tg);
Ti = VADD(Ta, Th);
Tk = VMUL(LDK(KP559016994), VSUB(Ta, Th));
TC = VADD(To, Tr);
TD = VADD(Tv, Ty);
TE = VMUL(LDK(KP559016994), VSUB(TC, TD));
TI = VADD(TC, TD);
}
}
{
V T1H, T1J, T1K, T1L, T1N, T1I, TZ, T10;
TZ = VADD(T3, Ti);
T10 = VBYI(VADD(TH, TI));
T1H = VSUB(TZ, T10);
STM2(&(xo[30]), T1H, ovs, &(xo[2]));
T1I = VADD(TZ, T10);
STM2(&(xo[10]), T1I, ovs, &(xo[2]));
{
V T1x, T1z, T1A, T1E, T1G, T1C, T1D, T1F, T1B, T1M;
T1x = VMUL(LDK(KP559016994), VSUB(T1t, T1w));
T1z = VADD(T1t, T1w);
T1A = VFNMS(LDK(KP250000000), T1z, T1y);
T1C = VSUB(T1r, T1s);
T1D = VSUB(T1u, T1v);
T1E = VBYI(VFMA(LDK(KP951056516), T1C, VMUL(LDK(KP587785252), T1D)));
T1G = VBYI(VFNMS(LDK(KP951056516), T1D, VMUL(LDK(KP587785252), T1C)));
T1J = VADD(T1y, T1z);
STM2(&(xo[0]), T1J, ovs, &(xo[0]));
T1F = VSUB(T1A, T1x);
T1K = VSUB(T1F, T1G);
STM2(&(xo[16]), T1K, ovs, &(xo[0]));
T1L = VADD(T1G, T1F);
STM2(&(xo[24]), T1L, ovs, &(xo[0]));
T1B = VADD(T1x, T1A);
T1M = VSUB(T1B, T1E);
STM2(&(xo[8]), T1M, ovs, &(xo[0]));
STN2(&(xo[8]), T1M, T1I, ovs);
T1N = VADD(T1E, T1B);
STM2(&(xo[32]), T1N, ovs, &(xo[0]));
}
{
V T1O, T1P, T1R, T1S;
{
V T1n, T1l, T1m, T1f, T1p, T17, T1e, T1q, T1Q, T1o;
T1n = VMUL(LDK(KP559016994), VSUB(T1j, T1k));
T1l = VADD(T1j, T1k);
T1m = VFNMS(LDK(KP250000000), T1l, T1i);
T17 = VSUB(T13, T16);
T1e = VSUB(T1a, T1d);
T1f = VBYI(VFNMS(LDK(KP951056516), T1e, VMUL(LDK(KP587785252), T17)));
T1p = VBYI(VFMA(LDK(KP951056516), T17, VMUL(LDK(KP587785252), T1e)));
T1O = VADD(T1i, T1l);
STM2(&(xo[20]), T1O, ovs, &(xo[0]));
T1q = VADD(T1n, T1m);
T1P = VADD(T1p, T1q);
STM2(&(xo[12]), T1P, ovs, &(xo[0]));
T1Q = VSUB(T1q, T1p);
STM2(&(xo[28]), T1Q, ovs, &(xo[0]));
STN2(&(xo[28]), T1Q, T1H, ovs);
T1o = VSUB(T1m, T1n);
T1R = VADD(T1f, T1o);
STM2(&(xo[4]), T1R, ovs, &(xo[0]));
T1S = VSUB(T1o, T1f);
STM2(&(xo[36]), T1S, ovs, &(xo[0]));
}
{
V TA, TN, TU, TS, TK, TV, Tl, TR, TJ, Tj;
TA = VFNMS(LDK(KP951056516), Tz, VMUL(LDK(KP587785252), Ts));
TN = VFNMS(LDK(KP951056516), TM, VMUL(LDK(KP587785252), TL));
TU = VFMA(LDK(KP951056516), TL, VMUL(LDK(KP587785252), TM));
TS = VFMA(LDK(KP951056516), Ts, VMUL(LDK(KP587785252), Tz));
TJ = VFNMS(LDK(KP250000000), TI, TH);
TK = VSUB(TE, TJ);
TV = VADD(TE, TJ);
Tj = VFNMS(LDK(KP250000000), Ti, T3);
Tl = VSUB(Tj, Tk);
TR = VADD(Tk, Tj);
{
V TB, TO, T1T, T1U;
TB = VSUB(Tl, TA);
TO = VBYI(VSUB(TK, TN));
T1T = VSUB(TB, TO);
STM2(&(xo[34]), T1T, ovs, &(xo[2]));
STN2(&(xo[32]), T1N, T1T, ovs);
T1U = VADD(TB, TO);
STM2(&(xo[6]), T1U, ovs, &(xo[2]));
STN2(&(xo[4]), T1R, T1U, ovs);
}
{
V TX, TY, T1V, T1W;
TX = VADD(TR, TS);
TY = VBYI(VSUB(TV, TU));
T1V = VSUB(TX, TY);
STM2(&(xo[22]), T1V, ovs, &(xo[2]));
STN2(&(xo[20]), T1O, T1V, ovs);
T1W = VADD(TX, TY);
STM2(&(xo[18]), T1W, ovs, &(xo[2]));
STN2(&(xo[16]), T1K, T1W, ovs);
}
{
V TP, TQ, T1X, T1Y;
TP = VADD(Tl, TA);
TQ = VBYI(VADD(TN, TK));
T1X = VSUB(TP, TQ);
STM2(&(xo[26]), T1X, ovs, &(xo[2]));
STN2(&(xo[24]), T1L, T1X, ovs);
T1Y = VADD(TP, TQ);
STM2(&(xo[14]), T1Y, ovs, &(xo[2]));
STN2(&(xo[12]), T1P, T1Y, ovs);
}
{
V TT, TW, T1Z, T20;
TT = VSUB(TR, TS);
TW = VBYI(VADD(TU, TV));
T1Z = VSUB(TT, TW);
STM2(&(xo[38]), T1Z, ovs, &(xo[2]));
STN2(&(xo[36]), T1S, T1Z, ovs);
T20 = VADD(TT, TW);
STM2(&(xo[2]), T20, ovs, &(xo[2]));
STN2(&(xo[0]), T1J, T20, ovs);
}
}
}
}
}
}
VLEAVE();
}
static void t1fv_20(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
{
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, T1y, Tk, Tf, T16, T10, TT, T1O, T1w, T1L, T1p, T1M;
V T1s, TZ, TI, T1x, Tp;
{
V T1, Tv, T2, Tt;
T1 = LD(&(x[0]), ms, &(x[0]));
Tv = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
T2 = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
Tt = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V T9, T1n, TN, T1v, TS, Te, T1q, T1u, TE, TG, Tm, T1o, TC, Tn, T1r;
V TH, To;
{
V TP, TR, Ta, Tc;
{
V T5, T7, TJ, TL, T1k, T1l;
T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
TJ = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
TL = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
{
V Tw, T3, Tu, T6, T8, TK, TM, TO, TQ;
TO = LD(&(x[WS(rs, 17)]), ms, &(x[WS(rs, 1)]));
Tw = BYTWJ(&(W[TWVL * 28]), Tv);
T3 = BYTWJ(&(W[TWVL * 18]), T2);
Tu = BYTWJ(&(W[TWVL * 8]), Tt);
T6 = BYTWJ(&(W[TWVL * 6]), T5);
T8 = BYTWJ(&(W[TWVL * 26]), T7);
TK = BYTWJ(&(W[TWVL * 24]), TJ);
TM = BYTWJ(&(W[TWVL * 4]), TL);
TP = BYTWJ(&(W[TWVL * 32]), TO);
TQ = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T4 = VSUB(T1, T3);
T1k = VADD(T1, T3);
Tx = VSUB(Tu, Tw);
T1l = VADD(Tu, Tw);
T9 = VSUB(T6, T8);
T1n = VADD(T6, T8);
TN = VSUB(TK, TM);
T1v = VADD(TK, TM);
TR = BYTWJ(&(W[TWVL * 12]), TQ);
}
Ta = LD(&(x[WS(rs, 16)]), ms, &(x[0]));
T1m = VSUB(T1k, T1l);
T1K = VADD(T1k, T1l);
Tc = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
}
{
V Tb, TA, Td, Th, Tj, Tz, Tg, Ti, Ty;
Tg = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Ti = LD(&(x[WS(rs, 18)]), ms, &(x[0]));
Ty = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
TS = VSUB(TP, TR);
T1y = VADD(TP, TR);
Tb = BYTWJ(&(W[TWVL * 30]), Ta);
TA = LD(&(x[WS(rs, 19)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[TWVL * 10]), Tc);
Th = BYTWJ(&(W[TWVL * 14]), Tg);
Tj = BYTWJ(&(W[TWVL * 34]), Ti);
Tz = BYTWJ(&(W[TWVL * 16]), Ty);
{
V TD, TF, TB, Tl;
TD = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
TF = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tl = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TB = BYTWJ(&(W[TWVL * 36]), TA);
Te = VSUB(Tb, Td);
T1q = VADD(Tb, Td);
Tk = VSUB(Th, Tj);
T1u = VADD(Th, Tj);
TE = BYTWJ(&(W[0]), TD);
TG = BYTWJ(&(W[TWVL * 20]), TF);
Tm = BYTWJ(&(W[TWVL * 22]), Tl);
T1o = VADD(Tz, TB);
TC = VSUB(Tz, TB);
Tn = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
}
}
}
Tf = VADD(T9, Te);
T16 = VSUB(T9, Te);
T10 = VSUB(TS, TN);
TT = VADD(TN, TS);
T1r = VADD(TE, TG);
TH = VSUB(TE, TG);
T1O = VADD(T1u, T1v);
T1w = VSUB(T1u, T1v);
To = BYTWJ(&(W[TWVL * 2]), Tn);
T1L = VADD(T1n, T1o);
T1p = VSUB(T1n, T1o);
T1M = VADD(T1q, T1r);
T1s = VSUB(T1q, T1r);
TZ = VSUB(TH, TC);
TI = VADD(TC, TH);
T1x = VADD(Tm, To);
Tp = VSUB(Tm, To);
}
}
{
V T1V, T1N, T14, T1d, T11, T1G, T1t, T1z, T1P, Tq, T17, T13, TV, TU;
T1V = VSUB(T1L, T1M);
T1N = VADD(T1L, T1M);
T14 = VSUB(TT, TI);
TU = VADD(TI, TT);
T1d = VFNMS(LDK(KP618033988), TZ, T10);
T11 = VFMA(LDK(KP618033988), T10, TZ);
T1G = VSUB(T1p, T1s);
T1t = VADD(T1p, T1s);
T1z = VSUB(T1x, T1y);
T1P = VADD(T1x, T1y);
Tq = VADD(Tk, Tp);
T17 = VSUB(Tk, Tp);
T13 = VFNMS(LDK(KP250000000), TU, Tx);
TV = VADD(Tx, TU);
{
V T1J, T1H, T1D, T1Z, T1X, T1T, T1h, T1j, T1b, T19, T1C, T1S, T1c, TY, T1F;
V T1A;
T1F = VSUB(T1w, T1z);
T1A = VADD(T1w, T1z);
{
V T1W, T1Q, TX, Tr;
T1W = VSUB(T1O, T1P);
T1Q = VADD(T1O, T1P);
TX = VSUB(Tf, Tq);
Tr = VADD(Tf, Tq);
{
V T1g, T18, T1f, T15;
T1g = VFNMS(LDK(KP618033988), T16, T17);
T18 = VFMA(LDK(KP618033988), T17, T16);
T1f = VFMA(LDK(KP559016994), T14, T13);
T15 = VFNMS(LDK(KP559016994), T14, T13);
T1J = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1F, T1G));
T1H = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1G, T1F));
{
V T1B, T1R, TW, Ts;
T1B = VADD(T1t, T1A);
T1D = VSUB(T1t, T1A);
T1Z = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1V, T1W));
T1X = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1W, T1V));
T1R = VADD(T1N, T1Q);
T1T = VSUB(T1N, T1Q);
TW = VFNMS(LDK(KP250000000), Tr, T4);
Ts = VADD(T4, Tr);
T1h = VFNMS(LDK(KP951056516), T1g, T1f);
T1j = VFMA(LDK(KP951056516), T1g, T1f);
T1b = VFNMS(LDK(KP951056516), T18, T15);
T19 = VFMA(LDK(KP951056516), T18, T15);
ST(&(x[WS(rs, 10)]), VADD(T1m, T1B), ms, &(x[0]));
T1C = VFNMS(LDK(KP250000000), T1B, T1m);
ST(&(x[0]), VADD(T1K, T1R), ms, &(x[0]));
T1S = VFNMS(LDK(KP250000000), T1R, T1K);
T1c = VFNMS(LDK(KP559016994), TX, TW);
TY = VFMA(LDK(KP559016994), TX, TW);
ST(&(x[WS(rs, 15)]), VFMAI(TV, Ts), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFNMSI(TV, Ts), ms, &(x[WS(rs, 1)]));
}
}
}
{
V T1E, T1I, T1U, T1Y;
T1E = VFNMS(LDK(KP559016994), T1D, T1C);
T1I = VFMA(LDK(KP559016994), T1D, T1C);
T1U = VFMA(LDK(KP559016994), T1T, T1S);
T1Y = VFNMS(LDK(KP559016994), T1T, T1S);
{
V T1e, T1i, T1a, T12;
T1e = VFNMS(LDK(KP951056516), T1d, T1c);
T1i = VFMA(LDK(KP951056516), T1d, T1c);
T1a = VFNMS(LDK(KP951056516), T11, TY);
T12 = VFMA(LDK(KP951056516), T11, TY);
ST(&(x[WS(rs, 18)]), VFNMSI(T1H, T1E), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(T1H, T1E), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VFMAI(T1J, T1I), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(T1J, T1I), ms, &(x[0]));
ST(&(x[WS(rs, 16)]), VFNMSI(T1X, T1U), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(T1X, T1U), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VFMAI(T1Z, T1Y), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VFNMSI(T1Z, T1Y), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VFMAI(T1h, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 17)]), VFNMSI(T1h, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(T1j, T1i), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VFNMSI(T1j, T1i), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFMAI(T1b, T1a), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFNMSI(T1b, T1a), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 19)]), VFMAI(T19, T12), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(T19, T12), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
VLEAVE();
}
static void n1bv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT i;
const R *xi;
R *xo;
xi = 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 Tl, Ty, T7, Te, Tw, Tt, Tz, TA, TB, Tg, Th, Tm, Tj, Tk;
Tj = LD(&(xi[0]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tl = VSUB(Tj, Tk);
Ty = VADD(Tj, Tk);
{
V T3, Tr, Td, Tv, T6, Ts, Ta, Tu;
{
V T1, T2, Tb, Tc;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tr = VADD(T1, T2);
Tb = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
Tv = VADD(Tb, Tc);
}
{
V T4, T5, T8, T9;
T4 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
Ts = VADD(T4, T5);
T8 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
Tu = VADD(T8, T9);
}
T7 = VSUB(T3, T6);
Te = VSUB(Ta, Td);
Tw = VSUB(Tu, Tv);
Tt = VSUB(Tr, Ts);
Tz = VADD(Tr, Ts);
TA = VADD(Tu, Tv);
TB = VADD(Tz, TA);
Tg = VADD(T3, T6);
Th = VADD(Ta, Td);
Tm = VADD(Tg, Th);
}
ST(&(xo[WS(os, 5)]), VADD(Tl, Tm), ovs, &(xo[WS(os, 1)]));
ST(&(xo[0]), VADD(Ty, TB), ovs, &(xo[0]));
{
V Tf, Tq, To, Tp, Ti, Tn;
Tf = VBYI(VFMA(LDK(KP951056516), T7, VMUL(LDK(KP587785252), Te)));
Tq = VBYI(VFNMS(LDK(KP951056516), Te, VMUL(LDK(KP587785252), T7)));
Ti = VMUL(LDK(KP559016994), VSUB(Tg, Th));
Tn = VFNMS(LDK(KP250000000), Tm, Tl);
To = VADD(Ti, Tn);
Tp = VSUB(Tn, Ti);
ST(&(xo[WS(os, 1)]), VADD(Tf, To), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VADD(Tq, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VSUB(To, Tf), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VSUB(Tp, Tq), ovs, &(xo[WS(os, 1)]));
}
{
V Tx, TG, TE, TF, TC, TD;
Tx = VBYI(VFNMS(LDK(KP951056516), Tw, VMUL(LDK(KP587785252), Tt)));
TG = VBYI(VFMA(LDK(KP951056516), Tt, VMUL(LDK(KP587785252), Tw)));
TC = VFNMS(LDK(KP250000000), TB, Ty);
TD = VMUL(LDK(KP559016994), VSUB(Tz, TA));
TE = VSUB(TC, TD);
TF = VADD(TD, TC);
ST(&(xo[WS(os, 2)]), VADD(Tx, TE), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VADD(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VSUB(TE, Tx), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VSUB(TF, TG), ovs, &(xo[0]));
}
}
}
static void t1fv_20(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 38)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 38), MAKE_VOLATILE_STRIDE(20, rs)) {
V T4, Tx, T1B, T1U, TZ, T16, T17, T10, Tf, Tq, Tr, T1N, T1O, T1S, T1t;
V T1w, T1C, TI, TT, TU, T1K, T1L, T1R, T1m, T1p, T1D, Ts, TV;
{
V T1, Tw, T3, Tu, Tv, T2, Tt, T1z, T1A;
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);
T1z = VADD(T1, T3);
T1A = VADD(Tu, Tw);
T1B = VSUB(T1z, T1A);
T1U = VADD(T1z, T1A);
}
{
V T9, T1r, TN, T1l, TS, T1o, Te, T1u, Tk, T1k, TC, T1s, TH, T1v, Tp;
V T1n;
{
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);
T1r = 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);
T1l = 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);
T1o = 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);
T1u = 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);
T1k = 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);
T1s = 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);
T1v = 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);
T1n = 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);
T1N = VADD(T1k, T1l);
T1O = VADD(T1n, T1o);
T1S = VADD(T1N, T1O);
T1t = VSUB(T1r, T1s);
T1w = VSUB(T1u, T1v);
T1C = VADD(T1t, T1w);
TI = VADD(TC, TH);
TT = VADD(TN, TS);
TU = VADD(TI, TT);
T1K = VADD(T1r, T1s);
T1L = VADD(T1u, T1v);
T1R = VADD(T1K, T1L);
T1m = VSUB(T1k, T1l);
T1p = VSUB(T1n, T1o);
T1D = VADD(T1m, T1p);
}
Ts = VADD(T4, Tr);
TV = VBYI(VADD(Tx, TU));
ST(&(x[WS(rs, 5)]), VSUB(Ts, TV), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 15)]), VADD(Ts, TV), ms, &(x[WS(rs, 1)]));
{
V T1T, T1V, T1W, T1Q, T1Z, T1M, T1P, T1Y, T1X;
T1T = VMUL(LDK(KP559016994), VSUB(T1R, T1S));
T1V = VADD(T1R, T1S);
T1W = VFNMS(LDK(KP250000000), T1V, T1U);
T1M = VSUB(T1K, T1L);
T1P = VSUB(T1N, T1O);
T1Q = VBYI(VFMA(LDK(KP951056516), T1M, VMUL(LDK(KP587785252), T1P)));
T1Z = VBYI(VFNMS(LDK(KP587785252), T1M, VMUL(LDK(KP951056516), T1P)));
ST(&(x[0]), VADD(T1U, T1V), ms, &(x[0]));
T1Y = VSUB(T1W, T1T);
ST(&(x[WS(rs, 8)]), VSUB(T1Y, T1Z), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VADD(T1Z, T1Y), ms, &(x[0]));
T1X = VADD(T1T, T1W);
ST(&(x[WS(rs, 4)]), VADD(T1Q, T1X), ms, &(x[0]));
ST(&(x[WS(rs, 16)]), VSUB(T1X, T1Q), ms, &(x[0]));
}
{
V T1G, T1E, T1F, T1y, T1J, T1q, T1x, T1I, T1H;
T1G = VMUL(LDK(KP559016994), VSUB(T1C, T1D));
T1E = VADD(T1C, T1D);
T1F = VFNMS(LDK(KP250000000), T1E, T1B);
T1q = VSUB(T1m, T1p);
T1x = VSUB(T1t, T1w);
T1y = VBYI(VFNMS(LDK(KP587785252), T1x, VMUL(LDK(KP951056516), T1q)));
T1J = VBYI(VFMA(LDK(KP951056516), T1x, VMUL(LDK(KP587785252), T1q)));
ST(&(x[WS(rs, 10)]), VADD(T1B, T1E), ms, &(x[0]));
T1I = VADD(T1G, T1F);
ST(&(x[WS(rs, 6)]), VSUB(T1I, T1J), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VADD(T1J, T1I), ms, &(x[0]));
T1H = VSUB(T1F, T1G);
ST(&(x[WS(rs, 2)]), VADD(T1y, T1H), ms, &(x[0]));
ST(&(x[WS(rs, 18)]), VSUB(T1H, T1y), ms, &(x[0]));
}
{
V T11, T18, T1g, T1d, T15, T1f, TY, T1c;
T11 = VFMA(LDK(KP951056516), TZ, VMUL(LDK(KP587785252), T10));
T18 = VFMA(LDK(KP951056516), T16, VMUL(LDK(KP587785252), T17));
T1g = VFNMS(LDK(KP587785252), T16, VMUL(LDK(KP951056516), T17));
T1d = VFNMS(LDK(KP587785252), TZ, VMUL(LDK(KP951056516), T10));
{
V T13, T14, TW, TX;
T13 = VFMS(LDK(KP250000000), TU, Tx);
T14 = VMUL(LDK(KP559016994), VSUB(TT, TI));
T15 = VADD(T13, T14);
T1f = VSUB(T14, T13);
TW = VMUL(LDK(KP559016994), VSUB(Tf, Tq));
TX = VFNMS(LDK(KP250000000), Tr, T4);
TY = VADD(TW, TX);
T1c = VSUB(TX, TW);
}
{
V T12, T19, T1i, T1j;
T12 = VADD(TY, T11);
T19 = VBYI(VSUB(T15, T18));
ST(&(x[WS(rs, 19)]), VSUB(T12, T19), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(T12, T19), ms, &(x[WS(rs, 1)]));
T1i = VADD(T1c, T1d);
T1j = VBYI(VADD(T1g, T1f));
ST(&(x[WS(rs, 13)]), VSUB(T1i, T1j), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(T1i, T1j), ms, &(x[WS(rs, 1)]));
}
{
V T1a, T1b, T1e, T1h;
T1a = VSUB(TY, T11);
T1b = VBYI(VADD(T18, T15));
ST(&(x[WS(rs, 11)]), VSUB(T1a, T1b), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(T1a, T1b), ms, &(x[WS(rs, 1)]));
T1e = VSUB(T1c, T1d);
T1h = VBYI(VSUB(T1f, T1g));
ST(&(x[WS(rs, 17)]), VSUB(T1e, T1h), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(T1e, T1h), ms, &(x[WS(rs, 1)]));
}
}
}
}
VLEAVE();
}
static void t3fv_10(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) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(10, rs)) {
V T2, T3, T4, Ta, T5, T6, Tt, Td, Th;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
T4 = VZMUL(T2, T3);
Ta = VZMULJ(T2, T3);
T5 = LDW(&(W[TWVL * 4]));
T6 = VZMULJ(T4, T5);
Tt = VZMULJ(T3, T5);
Td = VZMULJ(Ta, T5);
Th = VZMULJ(T2, T5);
{
V T9, TJ, Ts, Ty, Tz, TN, TO, TP, Tg, Tm, Tn, TK, TL, TM, T1;
V T8, T7;
T1 = LD(&(x[0]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T8 = VZMULJ(T6, T7);
T9 = VSUB(T1, T8);
TJ = VADD(T1, T8);
{
V Tp, Tx, Tr, Tv;
{
V To, Tw, Tq, Tu;
To = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tp = VZMULJ(T4, To);
Tw = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tx = VZMULJ(T2, Tw);
Tq = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tr = VZMULJ(T5, Tq);
Tu = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tv = VZMULJ(Tt, Tu);
}
Ts = VSUB(Tp, Tr);
Ty = VSUB(Tv, Tx);
Tz = VADD(Ts, Ty);
TN = VADD(Tp, Tr);
TO = VADD(Tv, Tx);
TP = VADD(TN, TO);
}
{
V Tc, Tl, Tf, Tj;
{
V Tb, Tk, Te, Ti;
Tb = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tc = VZMULJ(Ta, Tb);
Tk = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tl = VZMULJ(T3, Tk);
Te = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Tf = VZMULJ(Td, Te);
Ti = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tj = VZMULJ(Th, Ti);
}
Tg = VSUB(Tc, Tf);
Tm = VSUB(Tj, Tl);
Tn = VADD(Tg, Tm);
TK = VADD(Tc, Tf);
TL = VADD(Tj, Tl);
TM = VADD(TK, TL);
}
{
V TC, TA, TB, TG, TI, TE, TF, TH, TD;
TC = VSUB(Tn, Tz);
TA = VADD(Tn, Tz);
TB = VFNMS(LDK(KP250000000), TA, T9);
TE = VSUB(Tg, Tm);
TF = VSUB(Ts, Ty);
TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TF, TE));
TI = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TE, TF));
ST(&(x[WS(rs, 5)]), VADD(T9, TA), ms, &(x[WS(rs, 1)]));
TH = VFNMS(LDK(KP559016994), TC, TB);
ST(&(x[WS(rs, 3)]), VFNMSI(TI, TH), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(TI, TH), ms, &(x[WS(rs, 1)]));
TD = VFMA(LDK(KP559016994), TC, TB);
ST(&(x[WS(rs, 1)]), VFNMSI(TG, TD), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFMAI(TG, TD), ms, &(x[WS(rs, 1)]));
}
{
V TS, TQ, TR, TW, TY, TU, TV, TX, TT;
TS = VSUB(TM, TP);
TQ = VADD(TM, TP);
TR = VFNMS(LDK(KP250000000), TQ, TJ);
TU = VSUB(TN, TO);
TV = VSUB(TK, TL);
TW = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TV, TU));
TY = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TU, TV));
ST(&(x[0]), VADD(TJ, TQ), ms, &(x[0]));
TX = VFMA(LDK(KP559016994), TS, TR);
ST(&(x[WS(rs, 4)]), VFMAI(TY, TX), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TY, TX), ms, &(x[0]));
TT = VFNMS(LDK(KP559016994), TS, TR);
ST(&(x[WS(rs, 2)]), VFMAI(TW, TT), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VFNMSI(TW, TT), ms, &(x[0]));
}
}
}
}
VLEAVE();
}
static void t3fv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(10, rs)) {
V T1, T2, T3, Ti, T6, T7, Tx, Tb, To;
T1 = LDW(&(W[0]));
T2 = LDW(&(W[TWVL * 2]));
T3 = VZMULJ(T1, T2);
Ti = VZMUL(T1, T2);
T6 = LDW(&(W[TWVL * 4]));
T7 = VZMULJ(T3, T6);
Tx = VZMULJ(Ti, T6);
Tb = VZMULJ(T1, T6);
To = VZMULJ(T2, T6);
{
V TA, TQ, Tn, Tt, Tu, TJ, TK, TS, Ta, Tg, Th, TM, TN, TR, Tw;
V Tz, Ty;
Tw = LD(&(x[0]), ms, &(x[0]));
Ty = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tz = VZMULJ(Tx, Ty);
TA = VSUB(Tw, Tz);
TQ = VADD(Tw, Tz);
{
V Tk, Ts, Tm, Tq;
{
V Tj, Tr, Tl, Tp;
Tj = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tk = VZMULJ(Ti, Tj);
Tr = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Ts = VZMULJ(T1, Tr);
Tl = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tm = VZMULJ(T6, Tl);
Tp = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tq = VZMULJ(To, Tp);
}
Tn = VSUB(Tk, Tm);
Tt = VSUB(Tq, Ts);
Tu = VADD(Tn, Tt);
TJ = VADD(Tk, Tm);
TK = VADD(Tq, Ts);
TS = VADD(TJ, TK);
}
{
V T5, Tf, T9, Td;
{
V T4, Te, T8, Tc;
T4 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T5 = VZMULJ(T3, T4);
Te = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tf = VZMULJ(T2, Te);
T8 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T9 = VZMULJ(T7, T8);
Tc = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Td = VZMULJ(Tb, Tc);
}
Ta = VSUB(T5, T9);
Tg = VSUB(Td, Tf);
Th = VADD(Ta, Tg);
TM = VADD(T5, T9);
TN = VADD(Td, Tf);
TR = VADD(TM, TN);
}
{
V Tv, TB, TC, TG, TI, TE, TF, TH, TD;
Tv = VMUL(LDK(KP559016994), VSUB(Th, Tu));
TB = VADD(Th, Tu);
TC = VFNMS(LDK(KP250000000), TB, TA);
TE = VSUB(Ta, Tg);
TF = VSUB(Tn, Tt);
TG = VBYI(VFMA(LDK(KP951056516), TE, VMUL(LDK(KP587785252), TF)));
TI = VBYI(VFNMS(LDK(KP587785252), TE, VMUL(LDK(KP951056516), TF)));
ST(&(x[WS(rs, 5)]), VADD(TA, TB), ms, &(x[WS(rs, 1)]));
TH = VSUB(TC, Tv);
ST(&(x[WS(rs, 3)]), VSUB(TH, TI), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(TI, TH), ms, &(x[WS(rs, 1)]));
TD = VADD(Tv, TC);
ST(&(x[WS(rs, 1)]), VSUB(TD, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(TG, TD), ms, &(x[WS(rs, 1)]));
}
{
V TV, TT, TU, TP, TX, TL, TO, TY, TW;
TV = VMUL(LDK(KP559016994), VSUB(TR, TS));
TT = VADD(TR, TS);
TU = VFNMS(LDK(KP250000000), TT, TQ);
TL = VSUB(TJ, TK);
TO = VSUB(TM, TN);
TP = VBYI(VFNMS(LDK(KP587785252), TO, VMUL(LDK(KP951056516), TL)));
TX = VBYI(VFMA(LDK(KP951056516), TO, VMUL(LDK(KP587785252), TL)));
ST(&(x[0]), VADD(TQ, TT), ms, &(x[0]));
TY = VADD(TV, TU);
ST(&(x[WS(rs, 4)]), VADD(TX, TY), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VSUB(TY, TX), ms, &(x[0]));
TW = VSUB(TU, TV);
ST(&(x[WS(rs, 2)]), VADD(TP, TW), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TW, TP), ms, &(x[0]));
}
}
}
}
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);
}
}
}
}
}
}
VLEAVE();
}
static void n2bv_20(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
{
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 T1H, T1I, TS, TA, TN, TV, T1M, T1N, T1O, T1P, T1R, T1S, TK, TU, TR;
V Tl;
{
V T3, TE, T1r, T13, Ta, TL, Tz, TG, Ts, TF, Th, TM, T1u, T1C, T1n;
V T1a, T1m, T1h, T1x, T1D, Tk, Ti;
{
V T1, T2, TC, TD;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
TC = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
TD = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
{
V T14, T6, T1c, Tv, Tm, T1f, Ty, T17, T9, Tn, Tp, T1b, Td, Tq, Te;
V Tf, T15, To;
{
V Tw, Tx, T7, T8, Tb, Tc;
{
V T4, T5, Tt, Tu, T11, T12;
T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Tt = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Tu = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tw = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
T11 = VADD(T1, T2);
TE = VSUB(TC, TD);
T12 = VADD(TC, TD);
T14 = VADD(T4, T5);
T6 = VSUB(T4, T5);
T1c = VADD(Tt, Tu);
Tv = VSUB(Tt, Tu);
Tx = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T1r = VADD(T11, T12);
T13 = VSUB(T11, T12);
}
Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T1f = VADD(Tw, Tx);
Ty = VSUB(Tw, Tx);
T17 = VADD(T7, T8);
T9 = VSUB(T7, T8);
Tn = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
Tp = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T1b = VADD(Tb, Tc);
Td = VSUB(Tb, Tc);
Tq = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
}
Ta = VADD(T6, T9);
TL = VSUB(T6, T9);
T15 = VADD(Tm, Tn);
To = VSUB(Tm, Tn);
Tz = VSUB(Tv, Ty);
TG = VADD(Tv, Ty);
{
V T1d, T1v, T18, Tr, T1e, Tg, T16, T1s;
T1d = VSUB(T1b, T1c);
T1v = VADD(T1b, T1c);
T18 = VADD(Tp, Tq);
Tr = VSUB(Tp, Tq);
T1e = VADD(Te, Tf);
Tg = VSUB(Te, Tf);
T16 = VSUB(T14, T15);
T1s = VADD(T14, T15);
{
V T1t, T19, T1w, T1g;
T1t = VADD(T17, T18);
T19 = VSUB(T17, T18);
Ts = VSUB(To, Tr);
TF = VADD(To, Tr);
T1w = VADD(T1e, T1f);
T1g = VSUB(T1e, T1f);
Th = VADD(Td, Tg);
TM = VSUB(Td, Tg);
T1u = VADD(T1s, T1t);
T1C = VSUB(T1s, T1t);
T1n = VSUB(T16, T19);
T1a = VADD(T16, T19);
T1m = VSUB(T1d, T1g);
T1h = VADD(T1d, T1g);
T1x = VADD(T1v, T1w);
T1D = VSUB(T1v, T1w);
}
}
}
}
Tk = VSUB(Ta, Th);
Ti = VADD(Ta, Th);
{
V TJ, T1k, T1A, TZ, Tj, T1E, T1G, TI, T10, T1j, T1z, T1i, T1y, TH;
TJ = VSUB(TF, TG);
TH = VADD(TF, TG);
T1i = VADD(T1a, T1h);
T1k = VSUB(T1a, T1h);
T1y = VADD(T1u, T1x);
T1A = VSUB(T1u, T1x);
TZ = VADD(T3, Ti);
Tj = VFNMS(LDK(KP250000000), Ti, T3);
T1E = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1D, T1C));
T1G = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1C, T1D));
TI = VFNMS(LDK(KP250000000), TH, TE);
T10 = VADD(TE, TH);
T1j = VFNMS(LDK(KP250000000), T1i, T13);
T1H = VADD(T1r, T1y);
STM2(&(xo[0]), T1H, ovs, &(xo[0]));
T1z = VFNMS(LDK(KP250000000), T1y, T1r);
T1I = VADD(T13, T1i);
STM2(&(xo[20]), T1I, ovs, &(xo[0]));
{
V T1J, T1K, T1p, T1l, T1o, T1q, T1F, T1B, T1L, T1Q;
TS = VFNMS(LDK(KP618033988), Ts, Tz);
TA = VFMA(LDK(KP618033988), Tz, Ts);
TN = VFMA(LDK(KP618033988), TM, TL);
TV = VFNMS(LDK(KP618033988), TL, TM);
T1J = VFMAI(T10, TZ);
STM2(&(xo[10]), T1J, ovs, &(xo[2]));
T1K = VFNMSI(T10, TZ);
STM2(&(xo[30]), T1K, ovs, &(xo[2]));
T1p = VFMA(LDK(KP559016994), T1k, T1j);
T1l = VFNMS(LDK(KP559016994), T1k, T1j);
T1o = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1n, T1m));
T1q = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1m, T1n));
T1F = VFNMS(LDK(KP559016994), T1A, T1z);
T1B = VFMA(LDK(KP559016994), T1A, T1z);
T1L = VFNMSI(T1q, T1p);
STM2(&(xo[28]), T1L, ovs, &(xo[0]));
STN2(&(xo[28]), T1L, T1K, ovs);
T1M = VFMAI(T1q, T1p);
STM2(&(xo[12]), T1M, ovs, &(xo[0]));
T1N = VFMAI(T1o, T1l);
STM2(&(xo[36]), T1N, ovs, &(xo[0]));
T1O = VFNMSI(T1o, T1l);
STM2(&(xo[4]), T1O, ovs, &(xo[0]));
T1P = VFMAI(T1E, T1B);
STM2(&(xo[32]), T1P, ovs, &(xo[0]));
T1Q = VFNMSI(T1E, T1B);
STM2(&(xo[8]), T1Q, ovs, &(xo[0]));
STN2(&(xo[8]), T1Q, T1J, ovs);
T1R = VFNMSI(T1G, T1F);
STM2(&(xo[24]), T1R, ovs, &(xo[0]));
T1S = VFMAI(T1G, T1F);
STM2(&(xo[16]), T1S, ovs, &(xo[0]));
TK = VFMA(LDK(KP559016994), TJ, TI);
TU = VFNMS(LDK(KP559016994), TJ, TI);
TR = VFNMS(LDK(KP559016994), Tk, Tj);
Tl = VFMA(LDK(KP559016994), Tk, Tj);
}
}
}
{
V TY, TW, TO, TQ, TB, TP, TX, TT;
TY = VFMA(LDK(KP951056516), TV, TU);
TW = VFNMS(LDK(KP951056516), TV, TU);
TO = VFMA(LDK(KP951056516), TN, TK);
TQ = VFNMS(LDK(KP951056516), TN, TK);
TB = VFNMS(LDK(KP951056516), TA, Tl);
TP = VFMA(LDK(KP951056516), TA, Tl);
TX = VFNMS(LDK(KP951056516), TS, TR);
TT = VFMA(LDK(KP951056516), TS, TR);
{
V T1T, T1U, T1V, T1W;
T1T = VFMAI(TQ, TP);
STM2(&(xo[18]), T1T, ovs, &(xo[2]));
STN2(&(xo[16]), T1S, T1T, ovs);
T1U = VFNMSI(TQ, TP);
STM2(&(xo[22]), T1U, ovs, &(xo[2]));
STN2(&(xo[20]), T1I, T1U, ovs);
T1V = VFMAI(TO, TB);
STM2(&(xo[2]), T1V, ovs, &(xo[2]));
STN2(&(xo[0]), T1H, T1V, ovs);
T1W = VFNMSI(TO, TB);
STM2(&(xo[38]), T1W, ovs, &(xo[2]));
STN2(&(xo[36]), T1N, T1W, ovs);
{
V T1X, T1Y, T1Z, T20;
T1X = VFMAI(TW, TT);
STM2(&(xo[34]), T1X, ovs, &(xo[2]));
STN2(&(xo[32]), T1P, T1X, ovs);
T1Y = VFNMSI(TW, TT);
STM2(&(xo[6]), T1Y, ovs, &(xo[2]));
STN2(&(xo[4]), T1O, T1Y, ovs);
T1Z = VFMAI(TY, TX);
STM2(&(xo[26]), T1Z, ovs, &(xo[2]));
STN2(&(xo[24]), T1R, T1Z, ovs);
T20 = VFNMSI(TY, TX);
STM2(&(xo[14]), T20, ovs, &(xo[2]));
STN2(&(xo[12]), T1M, T20, ovs);
}
}
}
}
}
VLEAVE();
}
static void n1bv_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
const R *xi;
R *xo;
xi = ii;
xo = io;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
V Ta, Tk, Te, Tj, T7, Tn, Tf, Tm;
{
V T8, T9, Tc, Td;
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Ta = VSUB(T8, T9);
Tk = VADD(T8, T9);
Tc = LD(&(xi[0]), ivs, &(xi[0]));
Td = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Te = VSUB(Tc, Td);
Tj = VADD(Tc, Td);
{
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
T4 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
T7 = VMUL(LDK(KP707106781), VSUB(T3, T6));
Tn = VADD(T4, T5);
Tf = VMUL(LDK(KP707106781), VADD(T3, T6));
Tm = VADD(T1, T2);
}
}
{
V Tb, Tg, Tp, Tq;
Tb = VBYI(VSUB(T7, Ta));
Tg = VSUB(Te, Tf);
ST(&(xo[WS(os, 3)]), VADD(Tb, Tg), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VSUB(Tg, Tb), ovs, &(xo[WS(os, 1)]));
Tp = VADD(Tj, Tk);
Tq = VADD(Tm, Tn);
ST(&(xo[WS(os, 4)]), VSUB(Tp, Tq), ovs, &(xo[0]));
ST(&(xo[0]), VADD(Tp, Tq), ovs, &(xo[0]));
}
{
V Th, Ti, Tl, To;
Th = VBYI(VADD(Ta, T7));
Ti = VADD(Te, Tf);
ST(&(xo[WS(os, 1)]), VADD(Th, Ti), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VSUB(Ti, Th), ovs, &(xo[WS(os, 1)]));
Tl = VSUB(Tj, Tk);
To = VBYI(VSUB(Tm, Tn));
ST(&(xo[WS(os, 6)]), VSUB(Tl, To), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VADD(Tl, To), ovs, &(xo[0]));
}
}
}
VLEAVE();
}
static void hc2cbdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V T5, Tj, Tq, TI, Te, Tk, Tt, TJ, T2, Tg, T4, Ti, T3, Th, To;
V Tp, T6, Tc, T8, Tb, T7, Ta, T9, Td, Tr, Ts, TP, Tu, Tm, TO;
V Tn, Tf, Tl, T1, TN, Tv, TR, Tw, TQ, TC, TK, TA, TG, TB, TH;
V Ty, Tz, Tx, TF, TD, TM, TE, TL;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
Tg = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VCONJ(T3);
Th = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Ti = VCONJ(Th);
T5 = VSUB(T2, T4);
Tj = VSUB(Tg, Ti);
To = VADD(T2, T4);
Tp = VADD(Tg, Ti);
Tq = VSUB(To, Tp);
TI = VADD(To, Tp);
T6 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tc = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
Ta = LD(&(Rm[0]), -ms, &(Rm[0]));
Tb = VCONJ(Ta);
T9 = VSUB(T6, T8);
Td = VSUB(Tb, Tc);
Te = VMUL(LDK(KP707106781), VADD(T9, Td));
Tk = VMUL(LDK(KP707106781), VSUB(T9, Td));
Tr = VADD(T6, T8);
Ts = VADD(Tb, Tc);
Tt = VBYI(VSUB(Tr, Ts));
TJ = VADD(Tr, Ts);
TP = VADD(TI, TJ);
Tn = LDW(&(W[TWVL * 10]));
Tu = VZMUL(Tn, VSUB(Tq, Tt));
Tf = VADD(T5, Te);
Tl = VBYI(VADD(Tj, Tk));
T1 = LDW(&(W[TWVL * 12]));
Tm = VZMULI(T1, VSUB(Tf, Tl));
TN = LDW(&(W[0]));
TO = VZMULI(TN, VADD(Tl, Tf));
Tv = VADD(Tm, Tu);
ST(&(Rp[WS(rs, 3)]), Tv, ms, &(Rp[WS(rs, 1)]));
TR = VCONJ(VSUB(TP, TO));
ST(&(Rm[0]), TR, -ms, &(Rm[0]));
Tw = VCONJ(VSUB(Tu, Tm));
ST(&(Rm[WS(rs, 3)]), Tw, -ms, &(Rm[WS(rs, 1)]));
TQ = VADD(TO, TP);
ST(&(Rp[0]), TQ, ms, &(Rp[0]));
TB = LDW(&(W[TWVL * 2]));
TC = VZMUL(TB, VADD(Tq, Tt));
TH = LDW(&(W[TWVL * 6]));
TK = VZMUL(TH, VSUB(TI, TJ));
Ty = VBYI(VSUB(Tk, Tj));
Tz = VSUB(T5, Te);
Tx = LDW(&(W[TWVL * 4]));
TA = VZMULI(Tx, VADD(Ty, Tz));
TF = LDW(&(W[TWVL * 8]));
TG = VZMULI(TF, VSUB(Tz, Ty));
TD = VADD(TA, TC);
ST(&(Rp[WS(rs, 1)]), TD, ms, &(Rp[WS(rs, 1)]));
TM = VCONJ(VSUB(TK, TG));
ST(&(Rm[WS(rs, 2)]), TM, -ms, &(Rm[0]));
TE = VCONJ(VSUB(TC, TA));
ST(&(Rm[WS(rs, 1)]), TE, -ms, &(Rm[WS(rs, 1)]));
TL = VADD(TG, TK);
ST(&(Rp[WS(rs, 2)]), TL, ms, &(Rp[0]));
}
}
static void n1bv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
int i;
const R *xi;
R *xo;
xi = ii;
xo = io;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T5, Ta, TG, TF, Ty, Tm, Ti, Tp, TJ, TI, Tx, Ts;
{
V T1, T6, T4, Tk, T9, Tl;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T2, T3, T7, T8;
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Tk = VSUB(T2, T3);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Tl = VSUB(T7, T8);
}
T5 = VFNMS(LDK(KP500000000), T4, T1);
Ta = VFNMS(LDK(KP500000000), T9, T6);
TG = VADD(T6, T9);
TF = VADD(T1, T4);
Ty = VADD(Tk, Tl);
Tm = VMUL(LDK(KP866025403), VSUB(Tk, Tl));
}
{
V Tn, Tq, Te, To, Th, Tr;
Tn = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tq = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Tc, Td, Tf, Tg;
Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Te = VSUB(Tc, Td);
To = VADD(Tc, Td);
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tg = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Th = VSUB(Tf, Tg);
Tr = VADD(Tf, Tg);
}
Ti = VMUL(LDK(KP866025403), VSUB(Te, Th));
Tp = VFNMS(LDK(KP500000000), To, Tn);
TJ = VADD(Tq, Tr);
TI = VADD(Tn, To);
Tx = VADD(Te, Th);
Ts = VFNMS(LDK(KP500000000), Tr, Tq);
}
{
V TH, TK, TL, TM;
TH = VSUB(TF, TG);
TK = VBYI(VSUB(TI, TJ));
ST(&(xo[WS(os, 3)]), VSUB(TH, TK), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VADD(TH, TK), ovs, &(xo[WS(os, 1)]));
TL = VADD(TF, TG);
TM = VADD(TI, TJ);
ST(&(xo[WS(os, 6)]), VSUB(TL, TM), ovs, &(xo[0]));
ST(&(xo[0]), VADD(TL, TM), ovs, &(xo[0]));
}
{
V Tj, Tv, Tu, Tw, Tb, Tt;
Tb = VSUB(T5, Ta);
Tj = VSUB(Tb, Ti);
Tv = VADD(Tb, Ti);
Tt = VSUB(Tp, Ts);
Tu = VBYI(VADD(Tm, Tt));
Tw = VBYI(VSUB(Tt, Tm));
ST(&(xo[WS(os, 11)]), VSUB(Tj, Tu), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VADD(Tv, Tw), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VADD(Tj, Tu), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VSUB(Tv, Tw), ovs, &(xo[WS(os, 1)]));
}
{
V Tz, TD, TC, TE, TA, TB;
Tz = VBYI(VMUL(LDK(KP866025403), VSUB(Tx, Ty)));
TD = VBYI(VMUL(LDK(KP866025403), VADD(Ty, Tx)));
TA = VADD(T5, Ta);
TB = VADD(Tp, Ts);
TC = VSUB(TA, TB);
TE = VADD(TA, TB);
ST(&(xo[WS(os, 2)]), VADD(Tz, TC), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VSUB(TE, TD), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VSUB(TC, Tz), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(TD, TE), ovs, &(xo[0]));
}
}
END_SIMD();
}
static void hc2cbdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V TJ, T4, Tf, TB, TD, TE, Tm, T1, Tj, TF, Tp, Tb, Tg, Tt, Tx;
V T2, T3, Td, Te, T5, T6, T8, T9, Tn, T7, To, Ta, Tk, Tl, TG;
V TL, Tq, Tc, Tu, Th, Tv, Ty, Tw, TC, Ti, TK, TA, Tz, TI, TH;
V Ts, Tr, TN, TM;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Td = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Te = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
T9 = LD(&(Rm[0]), -ms, &(Rm[0]));
TJ = LDW(&(W[0]));
Tk = VFMACONJ(T3, T2);
T4 = VFNMSCONJ(T3, T2);
Tl = VFMACONJ(Te, Td);
Tf = VFNMSCONJ(Te, Td);
Tn = VFMACONJ(T6, T5);
T7 = VFNMSCONJ(T6, T5);
To = VFMACONJ(T9, T8);
Ta = VFMSCONJ(T9, T8);
TB = LDW(&(W[TWVL * 8]));
TD = LDW(&(W[TWVL * 6]));
TE = VADD(Tk, Tl);
Tm = VSUB(Tk, Tl);
T1 = LDW(&(W[TWVL * 12]));
Tj = LDW(&(W[TWVL * 10]));
TF = VADD(Tn, To);
Tp = VSUB(Tn, To);
Tb = VADD(T7, Ta);
Tg = VSUB(T7, Ta);
Tt = LDW(&(W[TWVL * 4]));
Tx = LDW(&(W[TWVL * 2]));
TG = VZMUL(TD, VSUB(TE, TF));
TL = VADD(TE, TF);
Tq = VZMUL(Tj, VFNMSI(Tp, Tm));
Tc = VFMA(LDK(KP707106781), Tb, T4);
Tu = VFNMS(LDK(KP707106781), Tb, T4);
Th = VFMA(LDK(KP707106781), Tg, Tf);
Tv = VFNMS(LDK(KP707106781), Tg, Tf);
Ty = VZMUL(Tx, VFMAI(Tp, Tm));
Tw = VZMULI(Tt, VFNMSI(Tv, Tu));
TC = VZMULI(TB, VFMAI(Tv, Tu));
Ti = VZMULI(T1, VFNMSI(Th, Tc));
TK = VZMULI(TJ, VFMAI(Th, Tc));
TA = VCONJ(VSUB(Ty, Tw));
Tz = VADD(Tw, Ty);
TI = VCONJ(VSUB(TG, TC));
TH = VADD(TC, TG);
Ts = VCONJ(VSUB(Tq, Ti));
Tr = VADD(Ti, Tq);
TN = VCONJ(VSUB(TL, TK));
TM = VADD(TK, TL);
ST(&(Rm[WS(rs, 1)]), TA, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 1)]), Tz, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 2)]), TI, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 2)]), TH, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 3)]), Ts, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 3)]), Tr, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[0]), TN, -ms, &(Rm[0]));
ST(&(Rp[0]), TM, ms, &(Rp[0]));
}
}
static void t1fuv_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 12)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 12), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, T4, Te, Tc, T9, T7;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Te = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T7 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
{
V T3, T5, Tf, Td, Ta, T8;
T3 = BYTWJ(&(W[0]), T2);
T5 = BYTWJ(&(W[TWVL * 10]), T4);
Tf = BYTWJ(&(W[TWVL * 6]), Te);
Td = BYTWJ(&(W[TWVL * 4]), Tc);
Ta = BYTWJ(&(W[TWVL * 8]), T9);
T8 = BYTWJ(&(W[TWVL * 2]), T7);
{
V T6, Tk, Tg, Tl, Tb, Tm;
T6 = VADD(T3, T5);
Tk = VSUB(T5, T3);
Tg = VADD(Td, Tf);
Tl = VSUB(Tf, Td);
Tb = VADD(T8, Ta);
Tm = VSUB(Ta, T8);
{
V Th, Ts, Tp, Tu, Tn, Tx, Ti, Tt;
Th = VFNMS(LDK(KP356895867), T6, Tg);
Ts = VFMA(LDK(KP554958132), Tl, Tk);
ST(&(x[0]), VADD(T1, VADD(T6, VADD(Tb, Tg))), ms, &(x[0]));
Tp = VFNMS(LDK(KP356895867), Tb, T6);
Tu = VFNMS(LDK(KP356895867), Tg, Tb);
Tn = VFMA(LDK(KP554958132), Tm, Tl);
Tx = VFNMS(LDK(KP554958132), Tk, Tm);
Ti = VFNMS(LDK(KP692021471), Th, Tb);
Tt = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), Ts, Tm));
{
V Tq, Tv, To, Ty, Tj, Tr, Tw;
Tq = VFNMS(LDK(KP692021471), Tp, Tg);
Tv = VFNMS(LDK(KP692021471), Tu, T6);
To = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tn, Tk));
Ty = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tx, Tl));
Tj = VFNMS(LDK(KP900968867), Ti, T1);
Tr = VFNMS(LDK(KP900968867), Tq, T1);
Tw = VFNMS(LDK(KP900968867), Tv, T1);
ST(&(x[WS(rs, 2)]), VFMAI(To, Tj), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VFNMSI(To, Tj), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(Tt, Tr), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 6)]), VFNMSI(Tt, Tr), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VFMAI(Ty, Tw), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 4)]), VFNMSI(Ty, Tw), ms, &(x[0]));
}
}
}
}
}
}
VLEAVE();
}
static void t2fv_16(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 30)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(16, rs)) {
V TJ, T10, TD, T11, T1b, T1c, Ty, TK, T16, T17, T18, Tb, TN, T13, T14;
V T15, Tm, TM, TG, TI, TH;
TG = LD(&(x[0]), ms, &(x[0]));
TH = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
TI = BYTWJ(&(W[TWVL * 14]), TH);
TJ = VSUB(TG, TI);
T10 = VADD(TG, TI);
{
V TA, TC, Tz, TB;
Tz = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
TA = BYTWJ(&(W[TWVL * 6]), Tz);
TB = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TC = BYTWJ(&(W[TWVL * 22]), TB);
TD = VSUB(TA, TC);
T11 = VADD(TA, TC);
}
{
V Tp, Tw, Tr, Tu, Ts, Tx;
{
V To, Tv, Tq, Tt;
To = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
Tp = BYTWJ(&(W[TWVL * 26]), To);
Tv = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
Tw = BYTWJ(&(W[TWVL * 18]), Tv);
Tq = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tr = BYTWJ(&(W[TWVL * 10]), Tq);
Tt = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tu = BYTWJ(&(W[TWVL * 2]), Tt);
}
T1b = VADD(Tp, Tr);
T1c = VADD(Tu, Tw);
Ts = VSUB(Tp, Tr);
Tx = VSUB(Tu, Tw);
Ty = VMUL(LDK(KP707106781), VSUB(Ts, Tx));
TK = VMUL(LDK(KP707106781), VADD(Tx, Ts));
}
{
V T2, T9, T4, T7, T5, Ta;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
T2 = BYTWJ(&(W[TWVL * 28]), T1);
T8 = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
T9 = BYTWJ(&(W[TWVL * 20]), T8);
T3 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T4 = BYTWJ(&(W[TWVL * 12]), T3);
T6 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T7 = BYTWJ(&(W[TWVL * 4]), T6);
}
T16 = VADD(T2, T4);
T17 = VADD(T7, T9);
T18 = VSUB(T16, T17);
T5 = VSUB(T2, T4);
Ta = VSUB(T7, T9);
Tb = VFNMS(LDK(KP923879532), Ta, VMUL(LDK(KP382683432), T5));
TN = VFMA(LDK(KP923879532), T5, VMUL(LDK(KP382683432), Ta));
}
{
V Td, Tk, Tf, Ti, Tg, Tl;
{
V Tc, Tj, Te, Th;
Tc = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Td = BYTWJ(&(W[0]), Tc);
Tj = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
Tk = BYTWJ(&(W[TWVL * 24]), Tj);
Te = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tf = BYTWJ(&(W[TWVL * 16]), Te);
Th = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Ti = BYTWJ(&(W[TWVL * 8]), Th);
}
T13 = VADD(Td, Tf);
T14 = VADD(Ti, Tk);
T15 = VSUB(T13, T14);
Tg = VSUB(Td, Tf);
Tl = VSUB(Ti, Tk);
Tm = VFMA(LDK(KP382683432), Tg, VMUL(LDK(KP923879532), Tl));
TM = VFNMS(LDK(KP382683432), Tl, VMUL(LDK(KP923879532), Tg));
}
{
V T1a, T1g, T1f, T1h;
{
V T12, T19, T1d, T1e;
T12 = VSUB(T10, T11);
T19 = VMUL(LDK(KP707106781), VADD(T15, T18));
T1a = VADD(T12, T19);
T1g = VSUB(T12, T19);
T1d = VSUB(T1b, T1c);
T1e = VMUL(LDK(KP707106781), VSUB(T18, T15));
T1f = VBYI(VADD(T1d, T1e));
T1h = VBYI(VSUB(T1e, T1d));
}
ST(&(x[WS(rs, 14)]), VSUB(T1a, T1f), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VADD(T1g, T1h), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(T1a, T1f), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VSUB(T1g, T1h), ms, &(x[0]));
}
{
V T1k, T1o, T1n, T1p;
{
V T1i, T1j, T1l, T1m;
T1i = VADD(T10, T11);
T1j = VADD(T1c, T1b);
T1k = VADD(T1i, T1j);
T1o = VSUB(T1i, T1j);
T1l = VADD(T13, T14);
T1m = VADD(T16, T17);
T1n = VADD(T1l, T1m);
T1p = VBYI(VSUB(T1m, T1l));
}
ST(&(x[WS(rs, 8)]), VSUB(T1k, T1n), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VADD(T1o, T1p), ms, &(x[0]));
ST(&(x[0]), VADD(T1k, T1n), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VSUB(T1o, T1p), ms, &(x[0]));
}
{
V TF, TQ, TP, TR;
{
V Tn, TE, TL, TO;
Tn = VSUB(Tb, Tm);
TE = VSUB(Ty, TD);
TF = VBYI(VSUB(Tn, TE));
TQ = VBYI(VADD(TE, Tn));
TL = VADD(TJ, TK);
TO = VADD(TM, TN);
TP = VSUB(TL, TO);
TR = VADD(TL, TO);
}
ST(&(x[WS(rs, 7)]), VADD(TF, TP), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 15)]), VSUB(TR, TQ), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VSUB(TP, TF), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(TQ, TR), ms, &(x[WS(rs, 1)]));
}
{
V TU, TY, TX, TZ;
{
V TS, TT, TV, TW;
TS = VSUB(TJ, TK);
TT = VADD(Tm, Tb);
TU = VADD(TS, TT);
TY = VSUB(TS, TT);
TV = VADD(TD, Ty);
TW = VSUB(TN, TM);
TX = VBYI(VADD(TV, TW));
TZ = VBYI(VSUB(TW, TV));
}
ST(&(x[WS(rs, 13)]), VSUB(TU, TX), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VADD(TY, TZ), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(TU, TX), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VSUB(TY, TZ), ms, &(x[WS(rs, 1)]));
}
}
}
VLEAVE();
}
static void t1fuv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 14)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V 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 void t2fv_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 = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 30)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(16, rs)) {
V TO, Ta, TJ, TP, T14, Tq, T1i, T10, T1b, T1l, T13, T1c, TR, Tl, T15;
V Tv;
{
V Tc, TW, T4, T19, T9, TD, TI, Tj, TZ, T1a, Te, Th, Tn, Tr, Tu;
V Tp;
{
V T1, T2, T5, T7;
T1 = LD(&(x[0]), ms, &(x[0]));
T2 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
{
V Tz, TG, TB, TE;
Tz = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
TG = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
TB = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
TE = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
{
V Ti, TY, TX, Td, Tg, Tm, Tt, To;
{
V T3, T6, T8, TA, TH, TC, TF, Tb;
Tb = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = BYTWJ(&(W[TWVL * 14]), T2);
T6 = BYTWJ(&(W[TWVL * 6]), T5);
T8 = BYTWJ(&(W[TWVL * 22]), T7);
TA = BYTWJ(&(W[TWVL * 26]), Tz);
TH = BYTWJ(&(W[TWVL * 18]), TG);
TC = BYTWJ(&(W[TWVL * 10]), TB);
TF = BYTWJ(&(W[TWVL * 2]), TE);
Tc = BYTWJ(&(W[0]), Tb);
TW = VSUB(T1, T3);
T4 = VADD(T1, T3);
T19 = VSUB(T6, T8);
T9 = VADD(T6, T8);
Ti = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
TD = VADD(TA, TC);
TY = VSUB(TA, TC);
TI = VADD(TF, TH);
TX = VSUB(TF, TH);
}
Td = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tg = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tm = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
Tj = BYTWJ(&(W[TWVL * 24]), Ti);
Tt = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
To = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
TZ = VADD(TX, TY);
T1a = VSUB(TY, TX);
Te = BYTWJ(&(W[TWVL * 16]), Td);
Th = BYTWJ(&(W[TWVL * 8]), Tg);
Tn = BYTWJ(&(W[TWVL * 28]), Tm);
Tr = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tu = BYTWJ(&(W[TWVL * 20]), Tt);
Tp = BYTWJ(&(W[TWVL * 12]), To);
}
}
}
{
V Tf, T11, Tk, T12, Ts;
TO = VADD(T4, T9);
Ta = VSUB(T4, T9);
TJ = VSUB(TD, TI);
TP = VADD(TI, TD);
Tf = VADD(Tc, Te);
T11 = VSUB(Tc, Te);
Tk = VADD(Th, Tj);
T12 = VSUB(Th, Tj);
Ts = BYTWJ(&(W[TWVL * 4]), Tr);
T14 = VSUB(Tn, Tp);
Tq = VADD(Tn, Tp);
T1i = VFNMS(LDK(KP707106781), TZ, TW);
T10 = VFMA(LDK(KP707106781), TZ, TW);
T1b = VFNMS(LDK(KP707106781), T1a, T19);
T1l = VFMA(LDK(KP707106781), T1a, T19);
T13 = VFNMS(LDK(KP414213562), T12, T11);
T1c = VFMA(LDK(KP414213562), T11, T12);
TR = VADD(Tf, Tk);
Tl = VSUB(Tf, Tk);
T15 = VSUB(Tu, Ts);
Tv = VADD(Ts, Tu);
}
}
{
V T1d, T16, TS, Tw, TU, TQ;
T1d = VFMA(LDK(KP414213562), T14, T15);
T16 = VFNMS(LDK(KP414213562), T15, T14);
TS = VADD(Tq, Tv);
Tw = VSUB(Tq, Tv);
TU = VSUB(TO, TP);
TQ = VADD(TO, TP);
{
V T1e, T1j, T17, T1m;
T1e = VSUB(T1c, T1d);
T1j = VADD(T1c, T1d);
T17 = VADD(T13, T16);
T1m = VSUB(T16, T13);
{
V TV, TT, TK, Tx;
TV = VSUB(TS, TR);
TT = VADD(TR, TS);
TK = VSUB(Tw, Tl);
Tx = VADD(Tl, Tw);
{
V T1h, T1f, T1o, T1k;
T1h = VFMA(LDK(KP923879532), T1e, T1b);
T1f = VFNMS(LDK(KP923879532), T1e, T1b);
T1o = VFMA(LDK(KP923879532), T1j, T1i);
T1k = VFNMS(LDK(KP923879532), T1j, T1i);
{
V T1g, T18, T1p, T1n;
T1g = VFMA(LDK(KP923879532), T17, T10);
T18 = VFNMS(LDK(KP923879532), T17, T10);
T1p = VFMA(LDK(KP923879532), T1m, T1l);
T1n = VFNMS(LDK(KP923879532), T1m, T1l);
ST(&(x[WS(rs, 12)]), VFNMSI(TV, TU), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(TV, TU), ms, &(x[0]));
ST(&(x[0]), VADD(TQ, TT), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TQ, TT), ms, &(x[0]));
{
V TN, TL, TM, Ty;
TN = VFMA(LDK(KP707106781), TK, TJ);
TL = VFNMS(LDK(KP707106781), TK, TJ);
TM = VFMA(LDK(KP707106781), Tx, Ta);
Ty = VFNMS(LDK(KP707106781), Tx, Ta);
ST(&(x[WS(rs, 1)]), VFNMSI(T1h, T1g), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 15)]), VFMAI(T1h, T1g), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(T1f, T18), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFNMSI(T1f, T18), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(T1p, T1o), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VFNMSI(T1p, T1o), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 11)]), VFMAI(T1n, T1k), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFNMSI(T1n, T1k), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 14)]), VFNMSI(TN, TM), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TN, TM), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VFMAI(TL, Ty), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TL, Ty), ms, &(x[0]));
}
}
}
}
}
}
}
}
VLEAVE();
}
static void n2fv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
V Ti, Ty, Tm, Tn, Tw, Tt, Tz, TA, TB, T7, Te, Tj, Tg, Th;
Tg = LD(&(xi[0]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Ti = VSUB(Tg, Th);
Ty = VADD(Tg, Th);
{
V T3, Tu, Td, Ts, T6, Tv, Ta, Tr;
{
V T1, T2, Tb, Tc;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tu = VADD(T1, T2);
Tb = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
Ts = VADD(Tb, Tc);
}
{
V T4, T5, T8, T9;
T4 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
Tv = VADD(T4, T5);
T8 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
Tr = VADD(T8, T9);
}
Tm = VSUB(T3, T6);
Tn = VSUB(Ta, Td);
Tw = VSUB(Tu, Tv);
Tt = VSUB(Tr, Ts);
Tz = VADD(Tu, Tv);
TA = VADD(Tr, Ts);
TB = VADD(Tz, TA);
T7 = VADD(T3, T6);
Te = VADD(Ta, Td);
Tj = VADD(T7, Te);
}
{
V TH, TI, TK, TL, TM;
TH = VADD(Ti, Tj);
STM2(&(xo[10]), TH, ovs, &(xo[2]));
TI = VADD(Ty, TB);
STM2(&(xo[0]), TI, ovs, &(xo[0]));
{
V To, Tq, Tl, Tp, Tf, Tk, TJ;
To = VBYI(VFMA(LDK(KP951056516), Tm, VMUL(LDK(KP587785252), Tn)));
Tq = VBYI(VFNMS(LDK(KP587785252), Tm, VMUL(LDK(KP951056516), Tn)));
Tf = VMUL(LDK(KP559016994), VSUB(T7, Te));
Tk = VFNMS(LDK(KP250000000), Tj, Ti);
Tl = VADD(Tf, Tk);
Tp = VSUB(Tk, Tf);
TJ = VSUB(Tl, To);
STM2(&(xo[2]), TJ, ovs, &(xo[2]));
STN2(&(xo[0]), TI, TJ, ovs);
TK = VADD(Tq, Tp);
STM2(&(xo[14]), TK, ovs, &(xo[2]));
TL = VADD(To, Tl);
STM2(&(xo[18]), TL, ovs, &(xo[2]));
TM = VSUB(Tp, Tq);
STM2(&(xo[6]), TM, ovs, &(xo[2]));
}
{
V Tx, TF, TE, TG, TC, TD;
Tx = VBYI(VFNMS(LDK(KP587785252), Tw, VMUL(LDK(KP951056516), Tt)));
TF = VBYI(VFMA(LDK(KP951056516), Tw, VMUL(LDK(KP587785252), Tt)));
TC = VFNMS(LDK(KP250000000), TB, Ty);
TD = VMUL(LDK(KP559016994), VSUB(Tz, TA));
TE = VSUB(TC, TD);
TG = VADD(TD, TC);
{
V TN, TO, TP, TQ;
TN = VADD(Tx, TE);
STM2(&(xo[4]), TN, ovs, &(xo[0]));
STN2(&(xo[4]), TN, TM, ovs);
TO = VSUB(TG, TF);
STM2(&(xo[12]), TO, ovs, &(xo[0]));
STN2(&(xo[12]), TO, TK, ovs);
TP = VSUB(TE, Tx);
STM2(&(xo[16]), TP, ovs, &(xo[0]));
STN2(&(xo[16]), TP, TL, ovs);
TQ = VADD(TF, TG);
STM2(&(xo[8]), TQ, ovs, &(xo[0]));
STN2(&(xo[8]), TQ, TH, ovs);
}
}
}
}
}
static void n1fv_13(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
DVK(KP083333333, +0.083333333333333333333333333333333333333333333);
DVK(KP075902986, +0.075902986037193865983102897245103540356428373);
DVK(KP251768516, +0.251768516431883313623436926934233488546674281);
DVK(KP132983124, +0.132983124607418643793760531921092974399165133);
DVK(KP258260390, +0.258260390311744861420450644284508567852516811);
DVK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
DVK(KP300238635, +0.300238635966332641462884626667381504676006424);
DVK(KP011599105, +0.011599105605768290721655456654083252189827041);
DVK(KP156891391, +0.156891391051584611046832726756003269660212636);
DVK(KP256247671, +0.256247671582936600958684654061725059144125175);
DVK(KP174138601, +0.174138601152135905005660794929264742616964676);
DVK(KP575140729, +0.575140729474003121368385547455453388461001608);
DVK(KP503537032, +0.503537032863766627246873853868466977093348562);
DVK(KP113854479, +0.113854479055790798974654345867655310534642560);
DVK(KP265966249, +0.265966249214837287587521063842185948798330267);
DVK(KP387390585, +0.387390585467617292130675966426762851778775217);
DVK(KP300462606, +0.300462606288665774426601772289207995520941381);
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(26, is), MAKE_VOLATILE_STRIDE(26, os)) {
V TW, Tb, Tm, Tu, TC, TR, TX, TK, TU, Tz, TB, TN, TT;
TW = LD(&(xi[0]), ivs, &(xi[0]));
{
V T3, TH, Tl, Tw, Tp, Tg, Tv, To, T6, Tr, T9, Ts, Ta, TI, T1;
V T2, Tq, Tt;
T1 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
TH = VADD(T1, T2);
{
V Th, Ti, Tj, Tk;
Th = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Ti = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tj = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tk = VADD(Ti, Tj);
Tl = VADD(Th, Tk);
Tw = VSUB(Ti, Tj);
Tp = VFNMS(LDK(KP500000000), Tk, Th);
}
{
V Tc, Td, Te, Tf;
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tf = VADD(Td, Te);
Tg = VADD(Tc, Tf);
Tv = VSUB(Td, Te);
To = VFNMS(LDK(KP500000000), Tf, Tc);
}
{
V T4, T5, T7, T8;
T4 = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
Tr = VADD(T4, T5);
T7 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
Ts = VADD(T7, T8);
}
Ta = VADD(T6, T9);
TI = VADD(Tr, Ts);
Tb = VADD(T3, Ta);
Tm = VSUB(Tg, Tl);
Tq = VSUB(To, Tp);
Tt = VMUL(LDK(KP866025403), VSUB(Tr, Ts));
Tu = VADD(Tq, Tt);
TC = VSUB(Tq, Tt);
{
V TP, TQ, TG, TJ;
TP = VADD(Tg, Tl);
TQ = VADD(TH, TI);
TR = VMUL(LDK(KP300462606), VSUB(TP, TQ));
TX = VADD(TP, TQ);
TG = VADD(To, Tp);
TJ = VFNMS(LDK(KP500000000), TI, TH);
TK = VSUB(TG, TJ);
TU = VADD(TG, TJ);
}
{
V Tx, Ty, TL, TM;
Tx = VMUL(LDK(KP866025403), VSUB(Tv, Tw));
Ty = VFNMS(LDK(KP500000000), Ta, T3);
Tz = VSUB(Tx, Ty);
TB = VADD(Tx, Ty);
TL = VADD(Tv, Tw);
TM = VSUB(T6, T9);
TN = VSUB(TL, TM);
TT = VADD(TL, TM);
}
}
ST(&(xo[0]), VADD(TW, TX), ovs, &(xo[0]));
{
V T19, T1n, T14, T13, T1f, T1k, Tn, TE, T1e, T1j, TS, T1m, TZ, T1c, TA;
V TD;
{
V T17, T18, T11, T12;
T17 = VFMA(LDK(KP387390585), TN, VMUL(LDK(KP265966249), TK));
T18 = VFNMS(LDK(KP503537032), TU, VMUL(LDK(KP113854479), TT));
T19 = VSUB(T17, T18);
T1n = VADD(T17, T18);
T14 = VFMA(LDK(KP575140729), Tm, VMUL(LDK(KP174138601), Tb));
T11 = VFNMS(LDK(KP156891391), TB, VMUL(LDK(KP256247671), TC));
T12 = VFMA(LDK(KP011599105), Tz, VMUL(LDK(KP300238635), Tu));
T13 = VSUB(T11, T12);
T1f = VADD(T14, T13);
T1k = VMUL(LDK(KP1_732050807), VADD(T11, T12));
}
Tn = VFNMS(LDK(KP174138601), Tm, VMUL(LDK(KP575140729), Tb));
TA = VFNMS(LDK(KP300238635), Tz, VMUL(LDK(KP011599105), Tu));
TD = VFMA(LDK(KP256247671), TB, VMUL(LDK(KP156891391), TC));
TE = VSUB(TA, TD);
T1e = VMUL(LDK(KP1_732050807), VADD(TD, TA));
T1j = VSUB(Tn, TE);
{
V TO, T1b, TV, TY, T1a;
TO = VFNMS(LDK(KP132983124), TN, VMUL(LDK(KP258260390), TK));
T1b = VSUB(TR, TO);
TV = VFMA(LDK(KP251768516), TT, VMUL(LDK(KP075902986), TU));
TY = VFNMS(LDK(KP083333333), TX, TW);
T1a = VSUB(TY, TV);
TS = VFMA(LDK(KP2_000000000), TO, TR);
T1m = VADD(T1b, T1a);
TZ = VFMA(LDK(KP2_000000000), TV, TY);
T1c = VSUB(T1a, T1b);
}
{
V TF, T10, T1l, T1o;
TF = VBYI(VFMA(LDK(KP2_000000000), TE, Tn));
T10 = VADD(TS, TZ);
ST(&(xo[WS(os, 1)]), VADD(TF, T10), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 12)]), VSUB(T10, TF), ovs, &(xo[0]));
{
V T15, T16, T1p, T1q;
T15 = VBYI(VFMS(LDK(KP2_000000000), T13, T14));
T16 = VSUB(TZ, TS);
ST(&(xo[WS(os, 5)]), VADD(T15, T16), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VSUB(T16, T15), ovs, &(xo[0]));
T1p = VADD(T1n, T1m);
T1q = VBYI(VADD(T1j, T1k));
ST(&(xo[WS(os, 4)]), VSUB(T1p, T1q), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VADD(T1q, T1p), ovs, &(xo[WS(os, 1)]));
}
T1l = VBYI(VSUB(T1j, T1k));
T1o = VSUB(T1m, T1n);
ST(&(xo[WS(os, 3)]), VADD(T1l, T1o), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 10)]), VSUB(T1o, T1l), ovs, &(xo[0]));
{
V T1h, T1i, T1d, T1g;
T1h = VBYI(VSUB(T1e, T1f));
T1i = VSUB(T1c, T19);
ST(&(xo[WS(os, 6)]), VADD(T1h, T1i), ovs, &(xo[0]));
ST(&(xo[WS(os, 7)]), VSUB(T1i, T1h), ovs, &(xo[WS(os, 1)]));
T1d = VADD(T19, T1c);
T1g = VBYI(VADD(T1e, T1f));
ST(&(xo[WS(os, 2)]), VSUB(T1d, T1g), ovs, &(xo[0]));
ST(&(xo[WS(os, 11)]), VADD(T1g, T1d), ovs, &(xo[WS(os, 1)]));
}
}
}
}
}
VLEAVE();
}
static void t1bv_12(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 22)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(12, rs)) {
V T1, Tt, T6, T7, TB, Tq, TC, TD, T9, Tu, Te, Tf, Tx, Tl, Ty;
V Tz;
{
V T5, T3, T4, T2;
T1 = LD(&(x[0]), ms, &(x[0]));
T4 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T5 = BYTW(&(W[TWVL * 14]), T4);
T2 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T3 = BYTW(&(W[TWVL * 6]), T2);
Tt = VSUB(T3, T5);
T6 = VADD(T3, T5);
T7 = VFNMS(LDK(KP500000000), T6, T1);
}
{
V Tn, Tp, Tm, TA, To;
Tm = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tn = BYTW(&(W[0]), Tm);
TA = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
TB = BYTW(&(W[TWVL * 16]), TA);
To = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tp = BYTW(&(W[TWVL * 8]), To);
Tq = VSUB(Tn, Tp);
TC = VADD(Tn, Tp);
TD = VFNMS(LDK(KP500000000), TC, TB);
}
{
V Td, Tb, T8, Tc, Ta;
T8 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T9 = BYTW(&(W[TWVL * 10]), T8);
Tc = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Td = BYTW(&(W[TWVL * 2]), Tc);
Ta = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
Tb = BYTW(&(W[TWVL * 18]), Ta);
Tu = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Tf = VFNMS(LDK(KP500000000), Te, T9);
}
{
V Ti, Tk, Th, Tw, Tj;
Th = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Ti = BYTW(&(W[TWVL * 12]), Th);
Tw = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tx = BYTW(&(W[TWVL * 4]), Tw);
Tj = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tk = BYTW(&(W[TWVL * 20]), Tj);
Tl = VSUB(Ti, Tk);
Ty = VADD(Ti, Tk);
Tz = VFNMS(LDK(KP500000000), Ty, Tx);
}
{
V Ts, TG, TF, TH;
{
V Tg, Tr, Tv, TE;
Tg = VSUB(T7, Tf);
Tr = VMUL(LDK(KP866025403), VSUB(Tl, Tq));
Ts = VSUB(Tg, Tr);
TG = VADD(Tg, Tr);
Tv = VMUL(LDK(KP866025403), VSUB(Tt, Tu));
TE = VSUB(Tz, TD);
TF = VBYI(VADD(Tv, TE));
TH = VBYI(VSUB(TE, Tv));
}
ST(&(x[WS(rs, 11)]), VSUB(Ts, TF), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VADD(TG, TH), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(Ts, TF), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VSUB(TG, TH), ms, &(x[WS(rs, 1)]));
}
{
V TS, TW, TV, TX;
{
V TQ, TR, TT, TU;
TQ = VADD(T1, T6);
TR = VADD(T9, Te);
TS = VSUB(TQ, TR);
TW = VADD(TQ, TR);
TT = VADD(Tx, Ty);
TU = VADD(TB, TC);
TV = VBYI(VSUB(TT, TU));
TX = VADD(TT, TU);
}
ST(&(x[WS(rs, 3)]), VSUB(TS, TV), ms, &(x[WS(rs, 1)]));
ST(&(x[0]), VADD(TW, TX), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VADD(TS, TV), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 6)]), VSUB(TW, TX), ms, &(x[0]));
}
{
V TK, TO, TN, TP;
{
V TI, TJ, TL, TM;
TI = VADD(Tl, Tq);
TJ = VADD(Tt, Tu);
TK = VBYI(VMUL(LDK(KP866025403), VSUB(TI, TJ)));
TO = VBYI(VMUL(LDK(KP866025403), VADD(TJ, TI)));
TL = VADD(T7, Tf);
TM = VADD(Tz, TD);
TN = VSUB(TL, TM);
TP = VADD(TL, TM);
}
ST(&(x[WS(rs, 2)]), VADD(TK, TN), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TP, TO), ms, &(x[0]));
ST(&(x[WS(rs, 10)]), VSUB(TN, TK), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VADD(TO, TP), ms, &(x[0]));
}
}
}
VLEAVE();
}
static void n1fv_13(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP904176221, +0.904176221990848204433795481776887926501523162);
DVK(KP575140729, +0.575140729474003121368385547455453388461001608);
DVK(KP300462606, +0.300462606288665774426601772289207995520941381);
DVK(KP516520780, +0.516520780623489722840901288569017135705033622);
DVK(KP522026385, +0.522026385161275033714027226654165028300441940);
DVK(KP957805992, +0.957805992594665126462521754605754580515587217);
DVK(KP600477271, +0.600477271932665282925769253334763009352012849);
DVK(KP251768516, +0.251768516431883313623436926934233488546674281);
DVK(KP503537032, +0.503537032863766627246873853868466977093348562);
DVK(KP769338817, +0.769338817572980603471413688209101117038278899);
DVK(KP859542535, +0.859542535098774820163672132761689612766401925);
DVK(KP581704778, +0.581704778510515730456870384989698884939833902);
DVK(KP853480001, +0.853480001859823990758994934970528322872359049);
DVK(KP083333333, +0.083333333333333333333333333333333333333333333);
DVK(KP226109445, +0.226109445035782405468510155372505010481906348);
DVK(KP301479260, +0.301479260047709873958013540496673347309208464);
DVK(KP686558370, +0.686558370781754340655719594850823015421401653);
DVK(KP514918778, +0.514918778086315755491789696138117261566051239);
DVK(KP038632954, +0.038632954644348171955506895830342264440241080);
DVK(KP612264650, +0.612264650376756543746494474777125408779395514);
DVK(KP302775637, +0.302775637731994646559610633735247973125648287);
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(26, is), MAKE_VOLATILE_STRIDE(26, os)) {
V T1, T7, T2, Tg, Tf, TN, Th, Tq, Ta, Tj, T5, Tr, Tk;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
{
V Td, Te, T8, T9, T3, T4;
Td = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T4 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tg = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tf = VADD(Td, Te);
TN = VSUB(Td, Te);
Th = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tq = VSUB(T8, T9);
Ta = VADD(T8, T9);
Tj = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T5 = VADD(T3, T4);
Tr = VSUB(T4, T3);
Tk = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
}
{
V Tt, Ti, Ty, Tb, Ts, TQ, Tx, T6, Tu, Tl;
Tt = VSUB(Tg, Th);
Ti = VADD(Tg, Th);
Ty = VFMS(LDK(KP500000000), Ta, T7);
Tb = VADD(T7, Ta);
Ts = VSUB(Tq, Tr);
TQ = VADD(Tr, Tq);
Tx = VFNMS(LDK(KP500000000), T5, T2);
T6 = VADD(T2, T5);
Tu = VSUB(Tj, Tk);
Tl = VADD(Tj, Tk);
{
V TK, Tz, Tc, TX, Tv, TO, TL, Tm;
TK = VADD(Tx, Ty);
Tz = VSUB(Tx, Ty);
Tc = VADD(T6, Tb);
TX = VSUB(T6, Tb);
Tv = VSUB(Tt, Tu);
TO = VADD(Tt, Tu);
TL = VSUB(Ti, Tl);
Tm = VADD(Ti, Tl);
{
V TF, Tw, TP, TY, TT, TM, TA, Tn;
TF = VSUB(Ts, Tv);
Tw = VADD(Ts, Tv);
TP = VFNMS(LDK(KP500000000), TO, TN);
TY = VADD(TN, TO);
TT = VFNMS(LDK(KP866025403), TL, TK);
TM = VFMA(LDK(KP866025403), TL, TK);
TA = VFNMS(LDK(KP500000000), Tm, Tf);
Tn = VADD(Tf, Tm);
{
V T1f, T1n, TI, T18, T1k, T1c, TD, T17, T10, T1m, T16, T1e, TU, TR;
TU = VFNMS(LDK(KP866025403), TQ, TP);
TR = VFMA(LDK(KP866025403), TQ, TP);
{
V TZ, T15, TE, TB;
TZ = VFMA(LDK(KP302775637), TY, TX);
T15 = VFNMS(LDK(KP302775637), TX, TY);
TE = VSUB(Tz, TA);
TB = VADD(Tz, TA);
{
V TH, To, TV, T13;
TH = VSUB(Tc, Tn);
To = VADD(Tc, Tn);
TV = VFNMS(LDK(KP612264650), TU, TT);
T13 = VFMA(LDK(KP612264650), TT, TU);
{
V TS, T12, TG, T1b;
TS = VFNMS(LDK(KP038632954), TR, TM);
T12 = VFMA(LDK(KP038632954), TM, TR);
TG = VFNMS(LDK(KP514918778), TF, TE);
T1b = VFMA(LDK(KP686558370), TE, TF);
{
V TC, T1a, Tp, TW, T14;
TC = VFMA(LDK(KP301479260), TB, Tw);
T1a = VFNMS(LDK(KP226109445), Tw, TB);
Tp = VFNMS(LDK(KP083333333), To, T1);
ST(&(xo[0]), VADD(T1, To), ovs, &(xo[0]));
T1f = VFMA(LDK(KP853480001), TV, TS);
TW = VFNMS(LDK(KP853480001), TV, TS);
T1n = VFMA(LDK(KP853480001), T13, T12);
T14 = VFNMS(LDK(KP853480001), T13, T12);
TI = VFMA(LDK(KP581704778), TH, TG);
T18 = VFNMS(LDK(KP859542535), TG, TH);
T1k = VFMA(LDK(KP769338817), T1b, T1a);
T1c = VFNMS(LDK(KP769338817), T1b, T1a);
TD = VFMA(LDK(KP503537032), TC, Tp);
T17 = VFNMS(LDK(KP251768516), TC, Tp);
T10 = VMUL(LDK(KP600477271), VFMA(LDK(KP957805992), TZ, TW));
T1m = VFNMS(LDK(KP522026385), TW, TZ);
T16 = VMUL(LDK(KP600477271), VFMA(LDK(KP957805992), T15, T14));
T1e = VFNMS(LDK(KP522026385), T14, T15);
}
}
}
}
{
V T1o, T1q, T1g, T1i, T1d, T1h, T1l, T1p;
{
V T11, TJ, T19, T1j;
T11 = VFMA(LDK(KP516520780), TI, TD);
TJ = VFNMS(LDK(KP516520780), TI, TD);
T19 = VFMA(LDK(KP300462606), T18, T17);
T1j = VFNMS(LDK(KP300462606), T18, T17);
T1o = VMUL(LDK(KP575140729), VFNMS(LDK(KP904176221), T1n, T1m));
T1q = VMUL(LDK(KP575140729), VFMA(LDK(KP904176221), T1n, T1m));
T1g = VMUL(LDK(KP575140729), VFMA(LDK(KP904176221), T1f, T1e));
T1i = VMUL(LDK(KP575140729), VFNMS(LDK(KP904176221), T1f, T1e));
ST(&(xo[WS(os, 12)]), VFNMSI(T16, T11), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VFMAI(T16, T11), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VFMAI(T10, TJ), ovs, &(xo[0]));
ST(&(xo[WS(os, 5)]), VFNMSI(T10, TJ), ovs, &(xo[WS(os, 1)]));
T1d = VFNMS(LDK(KP503537032), T1c, T19);
T1h = VFMA(LDK(KP503537032), T1c, T19);
T1l = VFNMS(LDK(KP503537032), T1k, T1j);
T1p = VFMA(LDK(KP503537032), T1k, T1j);
}
ST(&(xo[WS(os, 9)]), VFMAI(T1g, T1d), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VFNMSI(T1g, T1d), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VFNMSI(T1i, T1h), ovs, &(xo[0]));
ST(&(xo[WS(os, 3)]), VFMAI(T1i, T1h), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VFMAI(T1o, T1l), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VFNMSI(T1o, T1l), ovs, &(xo[0]));
ST(&(xo[WS(os, 11)]), VFMAI(T1q, T1p), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VFNMSI(T1q, T1p), ovs, &(xo[0]));
}
}
}
}
}
}
}
VLEAVE();
}
static void t1buv_9(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP939692620, +0.939692620785908384054109277324731469936208134);
DVK(KP296198132, +0.296198132726023843175338011893050938967728390);
DVK(KP852868531, +0.852868531952443209628250963940074071936020296);
DVK(KP173648177, +0.173648177666930348851716626769314796000375677);
DVK(KP556670399, +0.556670399226419366452912952047023132968291906);
DVK(KP766044443, +0.766044443118978035202392650555416673935832457);
DVK(KP642787609, +0.642787609686539326322643409907263432907559884);
DVK(KP663413948, +0.663413948168938396205421319635891297216863310);
DVK(KP150383733, +0.150383733180435296639271897612501926072238258);
DVK(KP342020143, +0.342020143325668733044099614682259580763083368);
DVK(KP813797681, +0.813797681349373692844693217248393223289101568);
DVK(KP984807753, +0.984807753012208059366743024589523013670643252);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 16)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 16), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T6, Tu, Tg, Tf, TD, Tq, Tp, TE;
T1 = LD(&(x[0]), ms, &(x[0]));
{
V T3, T5, T2, T4;
T2 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T3 = BYTW(&(W[TWVL * 4]), T2);
T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T5 = BYTW(&(W[TWVL * 10]), T4);
T6 = VADD(T3, T5);
Tu = VMUL(LDK(KP866025403), VSUB(T3, T5));
}
{
V T9, Td, Tb, T8, Tc, Ta, Te;
T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T9 = BYTW(&(W[0]), T8);
Tc = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Td = BYTW(&(W[TWVL * 12]), Tc);
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tb = BYTW(&(W[TWVL * 6]), Ta);
Tg = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Tf = VFNMS(LDK(KP500000000), Te, T9);
TD = VADD(T9, Te);
}
{
V Tj, Tn, Tl, Ti, Tm, Tk, To;
Ti = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tj = BYTW(&(W[TWVL * 2]), Ti);
Tm = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tn = BYTW(&(W[TWVL * 14]), Tm);
Tk = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tl = BYTW(&(W[TWVL * 8]), Tk);
Tq = VSUB(Tl, Tn);
To = VADD(Tl, Tn);
Tp = VFNMS(LDK(KP500000000), To, Tj);
TE = VADD(Tj, To);
}
{
V TF, TG, TH, TI;
TF = VBYI(VMUL(LDK(KP866025403), VSUB(TD, TE)));
TG = VADD(T1, T6);
TH = VADD(TD, TE);
TI = VFNMS(LDK(KP500000000), TH, TG);
ST(&(x[WS(rs, 3)]), VADD(TF, TI), ms, &(x[WS(rs, 1)]));
ST(&(x[0]), VADD(TG, TH), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VSUB(TI, TF), ms, &(x[0]));
}
{
V TC, Tv, Tw, Tx, Th, Tr, Ts, T7, TB;
TC = VBYI(VSUB(VFMA(LDK(KP984807753), Tf, VFMA(LDK(KP813797681), Tq, VFNMS(LDK(KP150383733), Tg, VMUL(LDK(KP342020143), Tp)))), Tu));
Tv = VFMA(LDK(KP663413948), Tg, VMUL(LDK(KP642787609), Tf));
Tw = VFMA(LDK(KP150383733), Tq, VMUL(LDK(KP984807753), Tp));
Tx = VADD(Tv, Tw);
Th = VFNMS(LDK(KP556670399), Tg, VMUL(LDK(KP766044443), Tf));
Tr = VFNMS(LDK(KP852868531), Tq, VMUL(LDK(KP173648177), Tp));
Ts = VADD(Th, Tr);
T7 = VFNMS(LDK(KP500000000), T6, T1);
TB = VFMA(LDK(KP852868531), Tg, VFMA(LDK(KP173648177), Tf, VFMA(LDK(KP296198132), Tq, VFNMS(LDK(KP939692620), Tp, T7))));
ST(&(x[WS(rs, 7)]), VSUB(TB, TC), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 2)]), VADD(TB, TC), ms, &(x[0]));
{
V Tt, Ty, Tz, TA;
Tt = VADD(T7, Ts);
Ty = VBYI(VADD(Tu, Tx));
ST(&(x[WS(rs, 8)]), VSUB(Tt, Ty), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VADD(Tt, Ty), ms, &(x[WS(rs, 1)]));
Tz = VBYI(VADD(Tu, VFNMS(LDK(KP500000000), Tx, VMUL(LDK(KP866025403), VSUB(Th, Tr)))));
TA = VFMA(LDK(KP866025403), VSUB(Tw, Tv), VFNMS(LDK(KP500000000), Ts, T7));
ST(&(x[WS(rs, 4)]), VADD(Tz, TA), ms, &(x[0]));
ST(&(x[WS(rs, 5)]), VSUB(TA, Tz), ms, &(x[WS(rs, 1)]));
}
}
}
}
static void t1fv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
V Tr, TH, Tg, Tl, Tm, TA, TB, TJ, T5, Ta, Tb, TD, TE, TI, To;
V Tq, Tp;
To = LD(&(x[0]), ms, &(x[0]));
Tp = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
Tq = BYTWJ(&(W[TWVL * 8]), Tp);
Tr = VSUB(To, Tq);
TH = VADD(To, Tq);
{
V Td, Tk, Tf, Ti;
{
V Tc, Tj, Te, Th;
Tc = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Td = BYTWJ(&(W[TWVL * 6]), Tc);
Tj = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tk = BYTWJ(&(W[0]), Tj);
Te = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tf = BYTWJ(&(W[TWVL * 16]), Te);
Th = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Ti = BYTWJ(&(W[TWVL * 10]), Th);
}
Tg = VSUB(Td, Tf);
Tl = VSUB(Ti, Tk);
Tm = VADD(Tg, Tl);
TA = VADD(Td, Tf);
TB = VADD(Ti, Tk);
TJ = VADD(TA, TB);
}
{
V T2, T9, T4, T7;
{
V T1, T8, T3, T6;
T1 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T2 = BYTWJ(&(W[TWVL * 2]), T1);
T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = BYTWJ(&(W[TWVL * 4]), T8);
T3 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T4 = BYTWJ(&(W[TWVL * 12]), T3);
T6 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T7 = BYTWJ(&(W[TWVL * 14]), T6);
}
T5 = VSUB(T2, T4);
Ta = VSUB(T7, T9);
Tb = VADD(T5, Ta);
TD = VADD(T2, T4);
TE = VADD(T7, T9);
TI = VADD(TD, TE);
}
{
V Tn, Ts, Tt, Tx, Tz, Tv, Tw, Ty, Tu;
Tn = VMUL(LDK(KP559016994), VSUB(Tb, Tm));
Ts = VADD(Tb, Tm);
Tt = VFNMS(LDK(KP250000000), Ts, Tr);
Tv = VSUB(T5, Ta);
Tw = VSUB(Tg, Tl);
Tx = VBYI(VFMA(LDK(KP951056516), Tv, VMUL(LDK(KP587785252), Tw)));
Tz = VBYI(VFNMS(LDK(KP587785252), Tv, VMUL(LDK(KP951056516), Tw)));
ST(&(x[WS(rs, 5)]), VADD(Tr, Ts), ms, &(x[WS(rs, 1)]));
Ty = VSUB(Tt, Tn);
ST(&(x[WS(rs, 3)]), VSUB(Ty, Tz), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(Tz, Ty), ms, &(x[WS(rs, 1)]));
Tu = VADD(Tn, Tt);
ST(&(x[WS(rs, 1)]), VSUB(Tu, Tx), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(Tx, Tu), ms, &(x[WS(rs, 1)]));
}
{
V TM, TK, TL, TG, TO, TC, TF, TP, TN;
TM = VMUL(LDK(KP559016994), VSUB(TI, TJ));
TK = VADD(TI, TJ);
TL = VFNMS(LDK(KP250000000), TK, TH);
TC = VSUB(TA, TB);
TF = VSUB(TD, TE);
TG = VBYI(VFNMS(LDK(KP587785252), TF, VMUL(LDK(KP951056516), TC)));
TO = VBYI(VFMA(LDK(KP951056516), TF, VMUL(LDK(KP587785252), TC)));
ST(&(x[0]), VADD(TH, TK), ms, &(x[0]));
TP = VADD(TM, TL);
ST(&(x[WS(rs, 4)]), VADD(TO, TP), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VSUB(TP, TO), ms, &(x[0]));
TN = VSUB(TL, TM);
ST(&(x[WS(rs, 2)]), VADD(TG, TN), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TN, TG), ms, &(x[0]));
}
}
}
VLEAVE();
}
