static const R *t1fv_3(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 4)) {
V T1, T3, T5, T6, T2, T4, T7, T8;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[0]), T2);
T4 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T5 = BYTWJ(&(W[TWVL * 2]), T4);
T6 = VADD(T3, T5);
ST(&(x[0]), VADD(T1, T6), dist, &(x[0]));
T7 = VFNMS(LDK(KP500000000), T6, T1);
T8 = VBYI(VMUL(LDK(KP866025403), VSUB(T5, T3)));
ST(&(x[WS(ios, 2)]), VSUB(T7, T8), dist, &(x[0]));
ST(&(x[WS(ios, 1)]), VADD(T7, T8), dist, &(x[WS(ios, 1)]));
}
END_SIMD();
return W;
}
static void n1bv_4(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
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 T3, T7, T6, T8;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
T7 = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VBYI(VSUB(T4, T5));
T8 = VADD(T4, T5);
}
ST(&(xo[WS(os, 3)]), VSUB(T3, T6), ovs, &(xo[WS(os, 1)]));
ST(&(xo[0]), VADD(T7, T8), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VADD(T3, T6), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VSUB(T7, T8), ovs, &(xo[0]));
}
END_SIMD();
}
static void n2bv_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
int i;
const R *xi;
R *xo;
xi = ii;
xo = io;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
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[6]), VADD(Tb, Tg), ovs, &(xo[2]));
ST(&(xo[10]), VSUB(Tg, Tb), ovs, &(xo[2]));
Tp = VADD(Tj, Tk);
Tq = VADD(Tm, Tn);
ST(&(xo[8]), 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[2]), VADD(Th, Ti), ovs, &(xo[2]));
ST(&(xo[14]), VSUB(Ti, Th), ovs, &(xo[2]));
Tl = VSUB(Tj, Tk);
To = VBYI(VSUB(Tm, Tn));
ST(&(xo[12]), VSUB(Tl, To), ovs, &(xo[0]));
ST(&(xo[4]), VADD(Tl, To), ovs, &(xo[0]));
}
}
END_SIMD();
}
static void m2bv_64(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
int i;
BEGIN_SIMD();
for (i = 0; i < v; i += VL) {
m2bv_64_0(ii, io, is, ivs, ovs);
ii += VL * ivs;
io += VL * ovs;
}
END_SIMD();
}
static void m1fv_32(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
int i;
BEGIN_SIMD();
for (i = 0; i < v; i += VL) {
m1fv_32_0(ri, ro, is, os, ivs, ovs);
ri += VL * ivs;
ro += VL * ovs;
}
END_SIMD();
}
static const R *t1bv_6(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
int i;
R *x;
x = ii;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 10)) {
V Tf, Ti, Ta, Tk, T5, Tj, Tc, Te, Td;
Tc = LD(&(x[0]), dist, &(x[0]));
Td = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Te = BYTW(&(W[TWVL * 4]), Td);
Tf = VSUB(Tc, Te);
Ti = VADD(Tc, Te);
{
V T7, T9, T6, T8;
T6 = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
T7 = BYTW(&(W[TWVL * 6]), T6);
T8 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T9 = BYTW(&(W[0]), T8);
Ta = VSUB(T7, T9);
Tk = VADD(T7, T9);
}
{
V T2, T4, T1, T3;
T1 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T2 = BYTW(&(W[TWVL * 2]), T1);
T3 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
T4 = BYTW(&(W[TWVL * 8]), T3);
T5 = VSUB(T2, T4);
Tj = VADD(T2, T4);
}
{
V Tb, Tg, Th, Tn, Tl, Tm;
Tb = VBYI(VMUL(LDK(KP866025403), VSUB(T5, Ta)));
Tg = VADD(T5, Ta);
Th = VFNMS(LDK(KP500000000), Tg, Tf);
ST(&(x[WS(ios, 1)]), VADD(Tb, Th), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 3)]), VADD(Tf, Tg), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VSUB(Th, Tb), dist, &(x[WS(ios, 1)]));
Tn = VBYI(VMUL(LDK(KP866025403), VSUB(Tj, Tk)));
Tl = VADD(Tj, Tk);
Tm = VFNMS(LDK(KP500000000), Tl, Ti);
ST(&(x[WS(ios, 2)]), VSUB(Tm, Tn), dist, &(x[0]));
ST(&(x[0]), VADD(Ti, Tl), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(Tn, Tm), dist, &(x[0]));
}
}
END_SIMD();
return W;
}
static const R *t1fv_6(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 10)) {
V T4, Ti, Te, Tk, T9, Tj, T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[TWVL * 4]), T2);
T4 = VSUB(T1, T3);
Ti = VADD(T1, T3);
{
V Tb, Td, Ta, Tc;
Ta = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
Tb = BYTWJ(&(W[TWVL * 6]), Ta);
Tc = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Td = BYTWJ(&(W[0]), Tc);
Te = VSUB(Tb, Td);
Tk = VADD(Tb, Td);
}
{
V T6, T8, T5, T7;
T5 = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
T6 = BYTWJ(&(W[TWVL * 2]), T5);
T7 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
T8 = BYTWJ(&(W[TWVL * 8]), T7);
T9 = VSUB(T6, T8);
Tj = VADD(T6, T8);
}
{
V Th, Tf, Tg, Tn, Tl, Tm;
Th = VBYI(VMUL(LDK(KP866025403), VSUB(Te, T9)));
Tf = VADD(T9, Te);
Tg = VFNMS(LDK(KP500000000), Tf, T4);
ST(&(x[WS(ios, 3)]), VADD(T4, Tf), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(Tg, Th), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VSUB(Tg, Th), dist, &(x[WS(ios, 1)]));
Tn = VBYI(VMUL(LDK(KP866025403), VSUB(Tk, Tj)));
Tl = VADD(Tj, Tk);
Tm = VFNMS(LDK(KP500000000), Tl, Ti);
ST(&(x[0]), VADD(Ti, Tl), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(Tm, Tn), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VSUB(Tm, Tn), dist, &(x[0]));
}
}
END_SIMD();
return W;
}
static void n2bv_7(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
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 Tb, T9, Tc, T3, Te, T6, Td, T7, T8, Ti, Tj;
Tb = LD(&(xi[0]), ivs, &(xi[0]));
T7 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T9 = VSUB(T7, T8);
Tc = VADD(T7, T8);
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
Te = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
Td = VADD(T4, T5);
}
ST(&(xo[0]), VADD(Tb, VADD(Te, VADD(Tc, Td))), ovs, &(xo[0]));
Ti = VBYI(VFNMS(LDK(KP781831482), T6, VFNMS(LDK(KP433883739), T9, VMUL(LDK(KP974927912), T3))));
Tj = VFMA(LDK(KP623489801), Td, VFNMS(LDK(KP900968867), Tc, VFNMS(LDK(KP222520933), Te, Tb)));
ST(&(xo[4]), VADD(Ti, Tj), ovs, &(xo[0]));
ST(&(xo[10]), VSUB(Tj, Ti), ovs, &(xo[2]));
{
V Ta, Tf, Tg, Th;
Ta = VBYI(VFMA(LDK(KP433883739), T3, VFNMS(LDK(KP781831482), T9, VMUL(LDK(KP974927912), T6))));
Tf = VFMA(LDK(KP623489801), Tc, VFNMS(LDK(KP222520933), Td, VFNMS(LDK(KP900968867), Te, Tb)));
ST(&(xo[6]), VADD(Ta, Tf), ovs, &(xo[2]));
ST(&(xo[8]), VSUB(Tf, Ta), ovs, &(xo[0]));
Tg = VBYI(VFMA(LDK(KP781831482), T3, VFMA(LDK(KP974927912), T9, VMUL(LDK(KP433883739), T6))));
Th = VFMA(LDK(KP623489801), Te, VFNMS(LDK(KP900968867), Td, VFNMS(LDK(KP222520933), Tc, Tb)));
ST(&(xo[2]), VADD(Tg, Th), ovs, &(xo[2]));
ST(&(xo[12]), VSUB(Th, Tg), ovs, &(xo[0]));
}
}
END_SIMD();
}
static const R *t1fv_2(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 2)) {
V T1, T3, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
T3 = BYTWJ(&(W[0]), T2);
ST(&(x[WS(ios, 1)]), VSUB(T1, T3), dist, &(x[WS(ios, 1)]));
ST(&(x[0]), VADD(T1, T3), dist, &(x[0]));
}
END_SIMD();
return W;
}
static void n2fv_2(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T1, T2;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
ST(&(xo[2]), VSUB(T1, T2), ovs, &(xo[2]));
ST(&(xo[0]), VADD(T1, T2), ovs, &(xo[0]));
}
END_SIMD();
}
static void n2fv_5(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(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T8, T7, Td, T9, Tc;
T8 = LD(&(xi[0]), ivs, &(xi[0]));
{
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VADD(T4, T5);
T7 = VMUL(LDK(KP559016994), VSUB(T3, T6));
Td = VSUB(T4, T5);
T9 = VADD(T3, T6);
Tc = VSUB(T1, T2);
}
ST(&(xo[0]), VADD(T8, T9), ovs, &(xo[0]));
{
V Te, Tf, Tb, Tg, Ta;
Te = VBYI(VFMA(LDK(KP951056516), Tc, VMUL(LDK(KP587785252), Td)));
Tf = VBYI(VFNMS(LDK(KP587785252), Tc, VMUL(LDK(KP951056516), Td)));
Ta = VFNMS(LDK(KP250000000), T9, T8);
Tb = VADD(T7, Ta);
Tg = VSUB(Ta, T7);
ST(&(xo[2]), VSUB(Tb, Te), ovs, &(xo[2]));
ST(&(xo[6]), VSUB(Tg, Tf), ovs, &(xo[2]));
ST(&(xo[8]), VADD(Te, Tb), ovs, &(xo[0]));
ST(&(xo[4]), VADD(Tf, Tg), ovs, &(xo[0]));
}
}
END_SIMD();
}
static const R *q1fv_2(R *ri, R *ii, const R *W, stride is, stride vs, int m, int dist)
{
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = 0; i < m; i = i + VL, x = x + (VL * dist), W = W + (TWVL * 2)) {
V T1, T2, T3, T4, T5, T6;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(is, 1)]), dist, &(x[WS(is, 1)]));
T3 = BYTWJ(&(W[0]), VSUB(T1, T2));
T4 = LD(&(x[WS(vs, 1)]), dist, &(x[WS(vs, 1)]));
T5 = LD(&(x[WS(vs, 1) + WS(is, 1)]), dist, &(x[WS(vs, 1) + WS(is, 1)]));
T6 = BYTWJ(&(W[0]), VSUB(T4, T5));
ST(&(x[WS(vs, 1)]), T3, dist, &(x[WS(vs, 1)]));
ST(&(x[WS(vs, 1) + WS(is, 1)]), T6, dist, &(x[WS(vs, 1) + WS(is, 1)]));
ST(&(x[0]), VADD(T1, T2), dist, &(x[0]));
ST(&(x[WS(is, 1)]), VADD(T4, T5), dist, &(x[WS(is, 1)]));
}
END_SIMD();
return W;
}
static void n2fv_3(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T1, T4, T6, T2, T3, T5;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
T6 = VBYI(VMUL(LDK(KP866025403), VSUB(T3, T2)));
ST(&(xo[0]), VADD(T1, T4), ovs, &(xo[0]));
T5 = VFNMS(LDK(KP500000000), T4, T1);
ST(&(xo[4]), VSUB(T5, T6), ovs, &(xo[0]));
ST(&(xo[2]), VADD(T5, T6), ovs, &(xo[2]));
}
END_SIMD();
}
static void n2bv_9(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP342020143, +0.342020143325668733044099614682259580763083368);
DVK(KP813797681, +0.813797681349373692844693217248393223289101568);
DVK(KP939692620, +0.939692620785908384054109277324731469936208134);
DVK(KP296198132, +0.296198132726023843175338011893050938967728390);
DVK(KP642787609, +0.642787609686539326322643409907263432907559884);
DVK(KP663413948, +0.663413948168938396205421319635891297216863310);
DVK(KP556670399, +0.556670399226419366452912952047023132968291906);
DVK(KP766044443, +0.766044443118978035202392650555416673935832457);
DVK(KP984807753, +0.984807753012208059366743024589523013670643252);
DVK(KP150383733, +0.150383733180435296639271897612501926072238258);
DVK(KP852868531, +0.852868531952443209628250963940074071936020296);
DVK(KP173648177, +0.173648177666930348851716626769314796000375677);
DVK(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, Ty, Tm, Ti, Tw, Th, Tj, To, Tb, Tv, Ta, Tc, Tn;
{
V T1, T2, T3, T4;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
T5 = VFNMS(LDK(KP500000000), T4, T1);
Ty = VADD(T1, T4);
Tm = VMUL(LDK(KP866025403), VSUB(T2, T3));
}
{
V Td, Tg, Te, Tf;
Td = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tf = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tg = VADD(Te, Tf);
Ti = VSUB(Te, Tf);
Tw = VADD(Td, Tg);
Th = VFNMS(LDK(KP500000000), Tg, Td);
Tj = VFNMS(LDK(KP852868531), Ti, VMUL(LDK(KP173648177), Th));
To = VFMA(LDK(KP150383733), Ti, VMUL(LDK(KP984807753), Th));
}
{
V T6, T9, T7, T8;
T6 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T9 = VADD(T7, T8);
Tb = VSUB(T7, T8);
Tv = VADD(T6, T9);
Ta = VFNMS(LDK(KP500000000), T9, T6);
Tc = VFNMS(LDK(KP556670399), Tb, VMUL(LDK(KP766044443), Ta));
Tn = VFMA(LDK(KP663413948), Tb, VMUL(LDK(KP642787609), Ta));
}
{
V Tx, Tz, TA, Tt, Tu;
Tx = VBYI(VMUL(LDK(KP866025403), VSUB(Tv, Tw)));
Tz = VADD(Tv, Tw);
TA = VFNMS(LDK(KP500000000), Tz, Ty);
ST(&(xo[6]), VADD(Tx, TA), ovs, &(xo[2]));
ST(&(xo[0]), VADD(Ty, Tz), ovs, &(xo[0]));
ST(&(xo[12]), VSUB(TA, Tx), ovs, &(xo[0]));
Tt = VFMA(LDK(KP852868531), Tb, VFMA(LDK(KP173648177), Ta, VFMA(LDK(KP296198132), Ti, VFNMS(LDK(KP939692620), Th, T5))));
Tu = VBYI(VSUB(VFMA(LDK(KP984807753), Ta, VFMA(LDK(KP813797681), Ti, VFNMS(LDK(KP150383733), Tb, VMUL(LDK(KP342020143), Th)))), Tm));
ST(&(xo[14]), VSUB(Tt, Tu), ovs, &(xo[2]));
ST(&(xo[4]), VADD(Tt, Tu), ovs, &(xo[0]));
{
V Tl, Ts, Tq, Tr, Tk, Tp;
Tk = VADD(Tc, Tj);
Tl = VADD(T5, Tk);
Ts = VFMA(LDK(KP866025403), VSUB(To, Tn), VFNMS(LDK(KP500000000), Tk, T5));
Tp = VADD(Tn, To);
Tq = VBYI(VADD(Tm, Tp));
Tr = VBYI(VADD(Tm, VFNMS(LDK(KP500000000), Tp, VMUL(LDK(KP866025403), VSUB(Tc, Tj)))));
ST(&(xo[16]), VSUB(Tl, Tq), ovs, &(xo[0]));
ST(&(xo[10]), VSUB(Ts, Tr), ovs, &(xo[2]));
ST(&(xo[2]), VADD(Tl, Tq), ovs, &(xo[2]));
ST(&(xo[8]), VADD(Tr, Ts), ovs, &(xo[0]));
}
}
}
END_SIMD();
}
static void n1fv_15(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP216506350, +0.216506350946109661690930792688234045867850657);
DVK(KP509036960, +0.509036960455127183450980863393907648510733164);
DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP484122918, +0.484122918275927110647408174972799951354115213);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T5, T10, TB, TO, TU, TV, TR, Ta, Tf, Tg, Tl, Tq, Tr, TE, TH;
V TI, TZ, T11, T1f, T1g;
{
V T1, T2, T3, T4;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
T5 = VADD(T1, T4);
T10 = VSUB(T3, T2);
TB = VFNMS(LDK(KP500000000), T4, T1);
}
{
V T6, T9, TC, TP, Tm, Tp, TG, TN, Tb, Te, TD, TQ, Th, Tk, TF;
V TM, TX, TY;
{
V T7, T8, Tn, To;
T6 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T7 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
T9 = VADD(T7, T8);
TC = VFNMS(LDK(KP500000000), T9, T6);
TP = VSUB(T8, T7);
Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tn = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tp = VADD(Tn, To);
TG = VFNMS(LDK(KP500000000), Tp, Tm);
TN = VSUB(To, Tn);
}
{
V Tc, Td, Ti, Tj;
Tb = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = VADD(Tc, Td);
TD = VFNMS(LDK(KP500000000), Te, Tb);
TQ = VSUB(Td, Tc);
Th = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Ti = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tj = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tk = VADD(Ti, Tj);
TF = VFNMS(LDK(KP500000000), Tk, Th);
TM = VSUB(Tj, Ti);
}
TO = VSUB(TM, TN);
TU = VSUB(TF, TG);
TV = VSUB(TC, TD);
TR = VSUB(TP, TQ);
Ta = VADD(T6, T9);
Tf = VADD(Tb, Te);
Tg = VADD(Ta, Tf);
Tl = VADD(Th, Tk);
Tq = VADD(Tm, Tp);
Tr = VADD(Tl, Tq);
TE = VADD(TC, TD);
TH = VADD(TF, TG);
TI = VADD(TE, TH);
TX = VADD(TP, TQ);
TY = VADD(TM, TN);
TZ = VMUL(LDK(KP484122918), VSUB(TX, TY));
T11 = VADD(TX, TY);
}
T1f = VADD(TB, TI);
T1g = VBYI(VMUL(LDK(KP866025403), VADD(T10, T11)));
ST(&(xo[WS(os, 5)]), VSUB(T1f, T1g), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 10)]), VADD(T1f, T1g), ovs, &(xo[0]));
{
V Tu, Ts, Tt, Ty, TA, Tw, Tx, Tz, Tv;
Tu = VMUL(LDK(KP559016994), VSUB(Tg, Tr));
Ts = VADD(Tg, Tr);
Tt = VFNMS(LDK(KP250000000), Ts, T5);
Tw = VSUB(Tl, Tq);
Tx = VSUB(Ta, Tf);
Ty = VBYI(VFNMS(LDK(KP587785252), Tx, VMUL(LDK(KP951056516), Tw)));
TA = VBYI(VFMA(LDK(KP951056516), Tx, VMUL(LDK(KP587785252), Tw)));
ST(&(xo[0]), VADD(T5, Ts), ovs, &(xo[0]));
Tz = VADD(Tu, Tt);
ST(&(xo[WS(os, 6)]), VSUB(Tz, TA), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VADD(TA, Tz), ovs, &(xo[WS(os, 1)]));
Tv = VSUB(Tt, Tu);
ST(&(xo[WS(os, 3)]), VSUB(Tv, Ty), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 12)]), VADD(Ty, Tv), ovs, &(xo[0]));
}
{
V TS, TW, T1b, T18, T13, T1a, TL, T17, T12, TJ, TK;
TS = VFNMS(LDK(KP509036960), TR, VMUL(LDK(KP823639103), TO));
TW = VFNMS(LDK(KP587785252), TV, VMUL(LDK(KP951056516), TU));
T1b = VFMA(LDK(KP951056516), TV, VMUL(LDK(KP587785252), TU));
T18 = VFMA(LDK(KP823639103), TR, VMUL(LDK(KP509036960), TO));
T12 = VFNMS(LDK(KP216506350), T11, VMUL(LDK(KP866025403), T10));
T13 = VSUB(TZ, T12);
T1a = VADD(TZ, T12);
TJ = VFNMS(LDK(KP250000000), TI, TB);
TK = VMUL(LDK(KP559016994), VSUB(TE, TH));
TL = VSUB(TJ, TK);
T17 = VADD(TK, TJ);
{
V TT, T14, T1d, T1e;
TT = VSUB(TL, TS);
T14 = VBYI(VSUB(TW, T13));
ST(&(xo[WS(os, 8)]), VSUB(TT, T14), ovs, &(xo[0]));
ST(&(xo[WS(os, 7)]), VADD(TT, T14), ovs, &(xo[WS(os, 1)]));
T1d = VSUB(T17, T18);
T1e = VBYI(VADD(T1b, T1a));
ST(&(xo[WS(os, 11)]), VSUB(T1d, T1e), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 4)]), VADD(T1d, T1e), ovs, &(xo[0]));
}
{
V T15, T16, T19, T1c;
T15 = VADD(TL, TS);
T16 = VBYI(VADD(TW, T13));
ST(&(xo[WS(os, 13)]), VSUB(T15, T16), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 2)]), VADD(T15, T16), ovs, &(xo[0]));
T19 = VADD(T17, T18);
T1c = VBYI(VSUB(T1a, T1b));
ST(&(xo[WS(os, 14)]), VSUB(T19, T1c), ovs, &(xo[0]));
ST(&(xo[WS(os, 1)]), VADD(T19, T1c), ovs, &(xo[WS(os, 1)]));
}
}
}
END_SIMD();
}
static void n1fv_10(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V Ti, Ty, Tm, Tn, Tw, Tt, Tz, TA, TB, T7, Te, Tj, Tg, Th;
Tg = LD(&(xi[0]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Ti = VSUB(Tg, Th);
Ty = VADD(Tg, Th);
{
V T3, Tu, Td, Ts, T6, Tv, Ta, Tr;
{
V T1, T2, Tb, Tc;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tu = VADD(T1, T2);
Tb = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
Ts = VADD(Tb, Tc);
}
{
V T4, T5, T8, T9;
T4 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
Tv = VADD(T4, T5);
T8 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
Tr = VADD(T8, T9);
}
Tm = VSUB(T3, T6);
Tn = VSUB(Ta, Td);
Tw = VSUB(Tu, Tv);
Tt = VSUB(Tr, Ts);
Tz = VADD(Tu, Tv);
TA = VADD(Tr, Ts);
TB = VADD(Tz, TA);
T7 = VADD(T3, T6);
Te = VADD(Ta, Td);
Tj = VADD(T7, Te);
}
ST(&(xo[WS(os, 5)]), VADD(Ti, Tj), ovs, &(xo[WS(os, 1)]));
ST(&(xo[0]), VADD(Ty, TB), ovs, &(xo[0]));
{
V To, Tq, Tl, Tp, Tf, Tk;
To = VBYI(VFMA(LDK(KP951056516), Tm, VMUL(LDK(KP587785252), Tn)));
Tq = VBYI(VFNMS(LDK(KP587785252), Tm, VMUL(LDK(KP951056516), Tn)));
Tf = VMUL(LDK(KP559016994), VSUB(T7, Te));
Tk = VFNMS(LDK(KP250000000), Tj, Ti);
Tl = VADD(Tf, Tk);
Tp = VSUB(Tk, Tf);
ST(&(xo[WS(os, 1)]), VSUB(Tl, To), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 7)]), VADD(Tq, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VADD(To, Tl), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VSUB(Tp, Tq), ovs, &(xo[WS(os, 1)]));
}
{
V Tx, TF, TE, TG, TC, TD;
Tx = VBYI(VFNMS(LDK(KP587785252), Tw, VMUL(LDK(KP951056516), Tt)));
TF = VBYI(VFMA(LDK(KP951056516), Tw, VMUL(LDK(KP587785252), Tt)));
TC = VFNMS(LDK(KP250000000), TB, Ty);
TD = VMUL(LDK(KP559016994), VSUB(Tz, TA));
TE = VSUB(TC, TD);
TG = VADD(TD, TC);
ST(&(xo[WS(os, 2)]), VADD(Tx, TE), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VSUB(TG, TF), ovs, &(xo[0]));
ST(&(xo[WS(os, 8)]), VSUB(TE, Tx), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(TF, TG), ovs, &(xo[0]));
}
}
END_SIMD();
}
static void n2fv_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;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
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[2]), VADD(TF, T10), ovs, &(xo[2]));
ST(&(xo[24]), 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[10]), VADD(T15, T16), ovs, &(xo[2]));
ST(&(xo[16]), VSUB(T16, T15), ovs, &(xo[0]));
T1p = VADD(T1n, T1m);
T1q = VBYI(VADD(T1j, T1k));
ST(&(xo[8]), VSUB(T1p, T1q), ovs, &(xo[0]));
ST(&(xo[18]), VADD(T1q, T1p), ovs, &(xo[2]));
}
T1l = VBYI(VSUB(T1j, T1k));
T1o = VSUB(T1m, T1n);
ST(&(xo[6]), VADD(T1l, T1o), ovs, &(xo[2]));
ST(&(xo[20]), VSUB(T1o, T1l), ovs, &(xo[0]));
{
V T1h, T1i, T1d, T1g;
T1h = VBYI(VSUB(T1e, T1f));
T1i = VSUB(T1c, T19);
ST(&(xo[12]), VADD(T1h, T1i), ovs, &(xo[0]));
ST(&(xo[14]), VSUB(T1i, T1h), ovs, &(xo[2]));
T1d = VADD(T19, T1c);
T1g = VBYI(VADD(T1e, T1f));
ST(&(xo[4]), VSUB(T1d, T1g), ovs, &(xo[0]));
ST(&(xo[22]), VADD(T1g, T1d), ovs, &(xo[2]));
}
}
}
}
END_SIMD();
}
static void n1bv_11(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP959492973, +0.959492973614497389890368057066327699062454848);
DVK(KP654860733, +0.654860733945285064056925072466293553183791199);
DVK(KP142314838, +0.142314838273285140443792668616369668791051361);
DVK(KP415415013, +0.415415013001886425529274149229623203524004910);
DVK(KP841253532, +0.841253532831181168861811648919367717513292498);
DVK(KP540640817, +0.540640817455597582107635954318691695431770608);
DVK(KP909631995, +0.909631995354518371411715383079028460060241051);
DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
DVK(KP755749574, +0.755749574354258283774035843972344420179717445);
DVK(KP281732556, +0.281732556841429697711417915346616899035777899);
int i;
const R *xi;
R *xo;
xi = ii;
xo = io;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V Th, T3, Tm, Tf, Ti, Tc, Tj, T9, Tk, T6, Tl, Ta, Tb, Ts, Tt;
Th = LD(&(xi[0]), ivs, &(xi[0]));
{
V T1, T2, Td, Te;
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T3 = VSUB(T1, T2);
Tm = VADD(T1, T2);
Td = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Te = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Tf = VSUB(Td, Te);
Ti = VADD(Td, Te);
}
Ta = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tb = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tc = VSUB(Ta, Tb);
Tj = VADD(Ta, Tb);
{
V T7, T8, T4, T5;
T7 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T9 = VSUB(T7, T8);
Tk = VADD(T7, T8);
T4 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T6 = VSUB(T4, T5);
Tl = VADD(T4, T5);
}
ST(&(xo[0]), VADD(Th, VADD(Tm, VADD(Ti, VADD(Tl, VADD(Tj, Tk))))), ovs, &(xo[0]));
{
V Tg, Tn, Tu, Tv;
Tg = VBYI(VFMA(LDK(KP281732556), T3, VFMA(LDK(KP755749574), T6, VFNMS(LDK(KP909631995), Tc, VFNMS(LDK(KP540640817), Tf, VMUL(LDK(KP989821441), T9))))));
Tn = VFMA(LDK(KP841253532), Ti, VFMA(LDK(KP415415013), Tj, VFNMS(LDK(KP142314838), Tk, VFNMS(LDK(KP654860733), Tl, VFNMS(LDK(KP959492973), Tm, Th)))));
ST(&(xo[WS(os, 5)]), VADD(Tg, Tn), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 6)]), VSUB(Tn, Tg), ovs, &(xo[0]));
Tu = VBYI(VFMA(LDK(KP755749574), T3, VFMA(LDK(KP540640817), T6, VFNMS(LDK(KP909631995), T9, VFNMS(LDK(KP989821441), Tf, VMUL(LDK(KP281732556), Tc))))));
Tv = VFMA(LDK(KP841253532), Tl, VFMA(LDK(KP415415013), Tk, VFNMS(LDK(KP959492973), Tj, VFNMS(LDK(KP142314838), Ti, VFNMS(LDK(KP654860733), Tm, Th)))));
ST(&(xo[WS(os, 4)]), VADD(Tu, Tv), ovs, &(xo[0]));
ST(&(xo[WS(os, 7)]), VSUB(Tv, Tu), ovs, &(xo[WS(os, 1)]));
}
Ts = VBYI(VFMA(LDK(KP909631995), T3, VFNMS(LDK(KP540640817), T9, VFNMS(LDK(KP989821441), Tc, VFNMS(LDK(KP281732556), T6, VMUL(LDK(KP755749574), Tf))))));
Tt = VFMA(LDK(KP415415013), Tm, VFMA(LDK(KP841253532), Tk, VFNMS(LDK(KP142314838), Tj, VFNMS(LDK(KP959492973), Tl, VFNMS(LDK(KP654860733), Ti, Th)))));
ST(&(xo[WS(os, 2)]), VADD(Ts, Tt), ovs, &(xo[0]));
ST(&(xo[WS(os, 9)]), VSUB(Tt, Ts), ovs, &(xo[WS(os, 1)]));
{
V Tq, Tr, To, Tp;
Tq = VBYI(VFMA(LDK(KP540640817), T3, VFMA(LDK(KP909631995), Tf, VFMA(LDK(KP989821441), T6, VFMA(LDK(KP755749574), Tc, VMUL(LDK(KP281732556), T9))))));
Tr = VFMA(LDK(KP841253532), Tm, VFMA(LDK(KP415415013), Ti, VFNMS(LDK(KP959492973), Tk, VFNMS(LDK(KP654860733), Tj, VFNMS(LDK(KP142314838), Tl, Th)))));
ST(&(xo[WS(os, 1)]), VADD(Tq, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 10)]), VSUB(Tr, Tq), ovs, &(xo[0]));
To = VBYI(VFMA(LDK(KP989821441), T3, VFMA(LDK(KP540640817), Tc, VFNMS(LDK(KP909631995), T6, VFNMS(LDK(KP281732556), Tf, VMUL(LDK(KP755749574), T9))))));
Tp = VFMA(LDK(KP415415013), Tl, VFMA(LDK(KP841253532), Tj, VFNMS(LDK(KP654860733), Tk, VFNMS(LDK(KP959492973), Ti, VFNMS(LDK(KP142314838), Tm, Th)))));
ST(&(xo[WS(os, 3)]), VADD(To, Tp), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 8)]), VSUB(Tp, To), ovs, &(xo[0]));
}
}
END_SIMD();
}
static void n2fv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V T5, Ta, TJ, Ty, Tq, Tp, Tg, Tl, TI, TA, Tz, Tu;
{
V T1, T6, T4, Tw, T9, Tx;
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);
Tw = VSUB(T3, T2);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Tx = VSUB(T8, T7);
}
T5 = VADD(T1, T4);
Ta = VADD(T6, T9);
TJ = VADD(Tw, Tx);
Ty = VMUL(LDK(KP866025403), VSUB(Tw, Tx));
Tq = VFNMS(LDK(KP500000000), T9, T6);
Tp = VFNMS(LDK(KP500000000), T4, T1);
}
{
V Tc, Th, Tf, Ts, Tk, Tt;
Tc = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Th = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Td, Te, Ti, Tj;
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tf = VADD(Td, Te);
Ts = VSUB(Te, Td);
Ti = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tj = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tk = VADD(Ti, Tj);
Tt = VSUB(Tj, Ti);
}
Tg = VADD(Tc, Tf);
Tl = VADD(Th, Tk);
TI = VADD(Ts, Tt);
TA = VFNMS(LDK(KP500000000), Tk, Th);
Tz = VFNMS(LDK(KP500000000), Tf, Tc);
Tu = VMUL(LDK(KP866025403), VSUB(Ts, Tt));
}
{
V Tb, Tm, Tn, To;
Tb = VSUB(T5, Ta);
Tm = VBYI(VSUB(Tg, Tl));
ST(&(xo[18]), VSUB(Tb, Tm), ovs, &(xo[2]));
ST(&(xo[6]), VADD(Tb, Tm), ovs, &(xo[2]));
Tn = VADD(T5, Ta);
To = VADD(Tg, Tl);
ST(&(xo[12]), VSUB(Tn, To), ovs, &(xo[0]));
ST(&(xo[0]), VADD(Tn, To), ovs, &(xo[0]));
}
{
V Tv, TE, TC, TD, Tr, TB;
Tr = VSUB(Tp, Tq);
Tv = VSUB(Tr, Tu);
TE = VADD(Tr, Tu);
TB = VSUB(Tz, TA);
TC = VBYI(VADD(Ty, TB));
TD = VBYI(VSUB(Ty, TB));
ST(&(xo[10]), VSUB(Tv, TC), ovs, &(xo[2]));
ST(&(xo[22]), VSUB(TE, TD), ovs, &(xo[2]));
ST(&(xo[14]), VADD(TC, Tv), ovs, &(xo[2]));
ST(&(xo[2]), VADD(TD, TE), ovs, &(xo[2]));
}
{
V TK, TM, TH, TL, TF, TG;
TK = VBYI(VMUL(LDK(KP866025403), VSUB(TI, TJ)));
TM = VBYI(VMUL(LDK(KP866025403), VADD(TJ, TI)));
TF = VADD(Tp, Tq);
TG = VADD(Tz, TA);
TH = VSUB(TF, TG);
TL = VADD(TF, TG);
ST(&(xo[20]), VSUB(TH, TK), ovs, &(xo[0]));
ST(&(xo[8]), VADD(TL, TM), ovs, &(xo[0]));
ST(&(xo[4]), VADD(TH, TK), ovs, &(xo[0]));
ST(&(xo[16]), VSUB(TL, TM), ovs, &(xo[0]));
}
}
END_SIMD();
}
static const R *t1bv_32(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP195090322, +0.195090322016128267848284868477022240927691618);
DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
DVK(KP555570233, +0.555570233019602224742830813948532874374937191);
DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
int i;
R *x;
x = ii;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 62)) {
V T4, T1D, T2P, T3h, Tf, T1y, T2K, T3i, TC, T1w, T2G, T3e, Tr, T1v, T2D;
V T3d, T1k, T20, T2y, T3a, T1r, T21, T2v, T39, TV, T1X, T2r, T37, T12, T1Y;
V T2o, T36;
{
V T1, T1C, T3, T1A, T1B, T2, T1z, T2N, T2O;
T1 = LD(&(x[0]), dist, &(x[0]));
T1B = LD(&(x[WS(ios, 24)]), dist, &(x[0]));
T1C = BYTW(&(W[TWVL * 46]), T1B);
T2 = LD(&(x[WS(ios, 16)]), dist, &(x[0]));
T3 = BYTW(&(W[TWVL * 30]), T2);
T1z = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
T1A = BYTW(&(W[TWVL * 14]), T1z);
T4 = VSUB(T1, T3);
T1D = VSUB(T1A, T1C);
T2N = VADD(T1, T3);
T2O = VADD(T1A, T1C);
T2P = VSUB(T2N, T2O);
T3h = VADD(T2N, T2O);
}
{
V T6, Td, T8, Tb;
{
V T5, Tc, T7, Ta;
T5 = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
T6 = BYTW(&(W[TWVL * 6]), T5);
Tc = LD(&(x[WS(ios, 12)]), dist, &(x[0]));
Td = BYTW(&(W[TWVL * 22]), Tc);
T7 = LD(&(x[WS(ios, 20)]), dist, &(x[0]));
T8 = BYTW(&(W[TWVL * 38]), T7);
Ta = LD(&(x[WS(ios, 28)]), dist, &(x[0]));
Tb = BYTW(&(W[TWVL * 54]), Ta);
}
{
V T9, Te, T2I, T2J;
T9 = VSUB(T6, T8);
Te = VSUB(Tb, Td);
Tf = VMUL(LDK(KP707106781), VADD(T9, Te));
T1y = VMUL(LDK(KP707106781), VSUB(T9, Te));
T2I = VADD(T6, T8);
T2J = VADD(Tb, Td);
T2K = VSUB(T2I, T2J);
T3i = VADD(T2I, T2J);
}
}
{
V Tt, TA, Tv, Ty;
{
V Ts, Tz, Tu, Tx;
Ts = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
Tt = BYTW(&(W[TWVL * 10]), Ts);
Tz = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
TA = BYTW(&(W[TWVL * 26]), Tz);
Tu = LD(&(x[WS(ios, 22)]), dist, &(x[0]));
Tv = BYTW(&(W[TWVL * 42]), Tu);
Tx = LD(&(x[WS(ios, 30)]), dist, &(x[0]));
Ty = BYTW(&(W[TWVL * 58]), Tx);
}
{
V Tw, TB, T2E, T2F;
Tw = VSUB(Tt, Tv);
TB = VSUB(Ty, TA);
TC = VFNMS(LDK(KP382683432), TB, VMUL(LDK(KP923879532), Tw));
T1w = VFMA(LDK(KP923879532), TB, VMUL(LDK(KP382683432), Tw));
T2E = VADD(Ty, TA);
T2F = VADD(Tt, Tv);
T2G = VSUB(T2E, T2F);
T3e = VADD(T2E, T2F);
}
}
{
V Ti, Tp, Tk, Tn;
{
V Th, To, Tj, Tm;
Th = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Ti = BYTW(&(W[TWVL * 2]), Th);
To = LD(&(x[WS(ios, 26)]), dist, &(x[0]));
Tp = BYTW(&(W[TWVL * 50]), To);
Tj = LD(&(x[WS(ios, 18)]), dist, &(x[0]));
Tk = BYTW(&(W[TWVL * 34]), Tj);
Tm = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
Tn = BYTW(&(W[TWVL * 18]), Tm);
}
{
V Tl, Tq, T2B, T2C;
Tl = VSUB(Ti, Tk);
Tq = VSUB(Tn, Tp);
Tr = VFMA(LDK(KP382683432), Tl, VMUL(LDK(KP923879532), Tq));
T1v = VFNMS(LDK(KP382683432), Tq, VMUL(LDK(KP923879532), Tl));
T2B = VADD(Ti, Tk);
T2C = VADD(Tn, Tp);
T2D = VSUB(T2B, T2C);
T3d = VADD(T2B, T2C);
}
}
{
V T1g, T1i, T1o, T1m, T1a, T1c, T1d, T15, T17, T18;
{
V T1f, T1h, T1n, T1l;
T1f = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
T1g = BYTW(&(W[TWVL * 12]), T1f);
T1h = LD(&(x[WS(ios, 23)]), dist, &(x[WS(ios, 1)]));
T1i = BYTW(&(W[TWVL * 44]), T1h);
T1n = LD(&(x[WS(ios, 15)]), dist, &(x[WS(ios, 1)]));
T1o = BYTW(&(W[TWVL * 28]), T1n);
T1l = LD(&(x[WS(ios, 31)]), dist, &(x[WS(ios, 1)]));
T1m = BYTW(&(W[TWVL * 60]), T1l);
{
V T19, T1b, T14, T16;
T19 = LD(&(x[WS(ios, 27)]), dist, &(x[WS(ios, 1)]));
T1a = BYTW(&(W[TWVL * 52]), T19);
T1b = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
T1c = BYTW(&(W[TWVL * 20]), T1b);
T1d = VSUB(T1a, T1c);
T14 = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
T15 = BYTW(&(W[TWVL * 4]), T14);
T16 = LD(&(x[WS(ios, 19)]), dist, &(x[WS(ios, 1)]));
T17 = BYTW(&(W[TWVL * 36]), T16);
T18 = VSUB(T15, T17);
}
}
{
V T1e, T1j, T2w, T2x;
T1e = VMUL(LDK(KP707106781), VSUB(T18, T1d));
T1j = VSUB(T1g, T1i);
T1k = VSUB(T1e, T1j);
T20 = VADD(T1j, T1e);
T2w = VADD(T15, T17);
T2x = VADD(T1a, T1c);
T2y = VSUB(T2w, T2x);
T3a = VADD(T2w, T2x);
}
{
V T1p, T1q, T2t, T2u;
T1p = VSUB(T1m, T1o);
T1q = VMUL(LDK(KP707106781), VADD(T18, T1d));
T1r = VSUB(T1p, T1q);
T21 = VADD(T1p, T1q);
T2t = VADD(T1m, T1o);
T2u = VADD(T1g, T1i);
T2v = VSUB(T2t, T2u);
T39 = VADD(T2t, T2u);
}
}
{
V TR, TT, TZ, TX, TL, TN, TO, TG, TI, TJ;
{
V TQ, TS, TY, TW;
TQ = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
TR = BYTW(&(W[TWVL * 16]), TQ);
TS = LD(&(x[WS(ios, 25)]), dist, &(x[WS(ios, 1)]));
TT = BYTW(&(W[TWVL * 48]), TS);
TY = LD(&(x[WS(ios, 17)]), dist, &(x[WS(ios, 1)]));
TZ = BYTW(&(W[TWVL * 32]), TY);
TW = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
TX = BYTW(&(W[0]), TW);
{
V TK, TM, TF, TH;
TK = LD(&(x[WS(ios, 29)]), dist, &(x[WS(ios, 1)]));
TL = BYTW(&(W[TWVL * 56]), TK);
TM = LD(&(x[WS(ios, 13)]), dist, &(x[WS(ios, 1)]));
TN = BYTW(&(W[TWVL * 24]), TM);
TO = VSUB(TL, TN);
TF = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
TG = BYTW(&(W[TWVL * 8]), TF);
TH = LD(&(x[WS(ios, 21)]), dist, &(x[WS(ios, 1)]));
TI = BYTW(&(W[TWVL * 40]), TH);
TJ = VSUB(TG, TI);
}
}
{
V TP, TU, T2p, T2q;
TP = VMUL(LDK(KP707106781), VSUB(TJ, TO));
TU = VSUB(TR, TT);
TV = VSUB(TP, TU);
T1X = VADD(TU, TP);
T2p = VADD(TG, TI);
T2q = VADD(TL, TN);
T2r = VSUB(T2p, T2q);
T37 = VADD(T2p, T2q);
}
{
V T10, T11, T2m, T2n;
T10 = VSUB(TX, TZ);
T11 = VMUL(LDK(KP707106781), VADD(TJ, TO));
T12 = VSUB(T10, T11);
T1Y = VADD(T10, T11);
T2m = VADD(TX, TZ);
T2n = VADD(TR, TT);
T2o = VSUB(T2m, T2n);
T36 = VADD(T2m, T2n);
}
}
{
V T3q, T3u, T3t, T3v;
{
V T3o, T3p, T3r, T3s;
T3o = VADD(T3h, T3i);
T3p = VADD(T3d, T3e);
T3q = VSUB(T3o, T3p);
T3u = VADD(T3o, T3p);
T3r = VADD(T36, T37);
T3s = VADD(T39, T3a);
T3t = VBYI(VSUB(T3r, T3s));
T3v = VADD(T3r, T3s);
}
ST(&(x[WS(ios, 24)]), VSUB(T3q, T3t), dist, &(x[0]));
ST(&(x[0]), VADD(T3u, T3v), dist, &(x[0]));
ST(&(x[WS(ios, 8)]), VADD(T3q, T3t), dist, &(x[0]));
ST(&(x[WS(ios, 16)]), VSUB(T3u, T3v), dist, &(x[0]));
}
{
V T3f, T3j, T3c, T3k, T38, T3b;
T3f = VSUB(T3d, T3e);
T3j = VSUB(T3h, T3i);
T38 = VSUB(T36, T37);
T3b = VSUB(T39, T3a);
T3c = VMUL(LDK(KP707106781), VSUB(T38, T3b));
T3k = VMUL(LDK(KP707106781), VADD(T38, T3b));
{
V T3g, T3l, T3m, T3n;
T3g = VBYI(VSUB(T3c, T3f));
T3l = VSUB(T3j, T3k);
ST(&(x[WS(ios, 12)]), VADD(T3g, T3l), dist, &(x[0]));
ST(&(x[WS(ios, 20)]), VSUB(T3l, T3g), dist, &(x[0]));
T3m = VBYI(VADD(T3f, T3c));
T3n = VADD(T3j, T3k);
ST(&(x[WS(ios, 4)]), VADD(T3m, T3n), dist, &(x[0]));
ST(&(x[WS(ios, 28)]), VSUB(T3n, T3m), dist, &(x[0]));
}
}
{
V T2L, T31, T2R, T2Y, T2A, T2Z, T2U, T32, T2H, T2Q;
T2H = VMUL(LDK(KP707106781), VSUB(T2D, T2G));
T2L = VSUB(T2H, T2K);
T31 = VADD(T2K, T2H);
T2Q = VMUL(LDK(KP707106781), VADD(T2D, T2G));
T2R = VSUB(T2P, T2Q);
T2Y = VADD(T2P, T2Q);
{
V T2s, T2z, T2S, T2T;
T2s = VFNMS(LDK(KP382683432), T2r, VMUL(LDK(KP923879532), T2o));
T2z = VFMA(LDK(KP923879532), T2v, VMUL(LDK(KP382683432), T2y));
T2A = VSUB(T2s, T2z);
T2Z = VADD(T2s, T2z);
T2S = VFMA(LDK(KP382683432), T2o, VMUL(LDK(KP923879532), T2r));
T2T = VFNMS(LDK(KP382683432), T2v, VMUL(LDK(KP923879532), T2y));
T2U = VSUB(T2S, T2T);
T32 = VADD(T2S, T2T);
}
{
V T2M, T2V, T34, T35;
T2M = VBYI(VSUB(T2A, T2L));
T2V = VSUB(T2R, T2U);
ST(&(x[WS(ios, 10)]), VADD(T2M, T2V), dist, &(x[0]));
ST(&(x[WS(ios, 22)]), VSUB(T2V, T2M), dist, &(x[0]));
T34 = VSUB(T2Y, T2Z);
T35 = VBYI(VSUB(T32, T31));
ST(&(x[WS(ios, 18)]), VSUB(T34, T35), dist, &(x[0]));
ST(&(x[WS(ios, 14)]), VADD(T34, T35), dist, &(x[0]));
}
{
V T2W, T2X, T30, T33;
T2W = VBYI(VADD(T2L, T2A));
T2X = VADD(T2R, T2U);
ST(&(x[WS(ios, 6)]), VADD(T2W, T2X), dist, &(x[0]));
ST(&(x[WS(ios, 26)]), VSUB(T2X, T2W), dist, &(x[0]));
T30 = VADD(T2Y, T2Z);
T33 = VBYI(VADD(T31, T32));
ST(&(x[WS(ios, 30)]), VSUB(T30, T33), dist, &(x[0]));
ST(&(x[WS(ios, 2)]), VADD(T30, T33), dist, &(x[0]));
}
}
{
V TE, T1P, T1I, T1Q, T1t, T1M, T1F, T1N;
{
V Tg, TD, T1G, T1H;
Tg = VSUB(T4, Tf);
TD = VSUB(Tr, TC);
TE = VSUB(Tg, TD);
T1P = VADD(Tg, TD);
T1G = VFNMS(LDK(KP555570233), TV, VMUL(LDK(KP831469612), T12));
T1H = VFMA(LDK(KP555570233), T1k, VMUL(LDK(KP831469612), T1r));
T1I = VSUB(T1G, T1H);
T1Q = VADD(T1G, T1H);
}
{
V T13, T1s, T1x, T1E;
T13 = VFMA(LDK(KP831469612), TV, VMUL(LDK(KP555570233), T12));
T1s = VFNMS(LDK(KP555570233), T1r, VMUL(LDK(KP831469612), T1k));
T1t = VSUB(T13, T1s);
T1M = VADD(T13, T1s);
T1x = VSUB(T1v, T1w);
T1E = VSUB(T1y, T1D);
T1F = VSUB(T1x, T1E);
T1N = VADD(T1E, T1x);
}
{
V T1u, T1J, T1S, T1T;
T1u = VADD(TE, T1t);
T1J = VBYI(VADD(T1F, T1I));
ST(&(x[WS(ios, 27)]), VSUB(T1u, T1J), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VADD(T1u, T1J), dist, &(x[WS(ios, 1)]));
T1S = VBYI(VADD(T1N, T1M));
T1T = VADD(T1P, T1Q);
ST(&(x[WS(ios, 3)]), VADD(T1S, T1T), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 29)]), VSUB(T1T, T1S), dist, &(x[WS(ios, 1)]));
}
{
V T1K, T1L, T1O, T1R;
T1K = VSUB(TE, T1t);
T1L = VBYI(VSUB(T1I, T1F));
ST(&(x[WS(ios, 21)]), VSUB(T1K, T1L), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 11)]), VADD(T1K, T1L), dist, &(x[WS(ios, 1)]));
T1O = VBYI(VSUB(T1M, T1N));
T1R = VSUB(T1P, T1Q);
ST(&(x[WS(ios, 13)]), VADD(T1O, T1R), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 19)]), VSUB(T1R, T1O), dist, &(x[WS(ios, 1)]));
}
}
{
V T1W, T2h, T2a, T2i, T23, T2e, T27, T2f;
{
V T1U, T1V, T28, T29;
T1U = VADD(T4, Tf);
T1V = VADD(T1v, T1w);
T1W = VSUB(T1U, T1V);
T2h = VADD(T1U, T1V);
T28 = VFNMS(LDK(KP195090322), T1X, VMUL(LDK(KP980785280), T1Y));
T29 = VFMA(LDK(KP195090322), T20, VMUL(LDK(KP980785280), T21));
T2a = VSUB(T28, T29);
T2i = VADD(T28, T29);
}
{
V T1Z, T22, T25, T26;
T1Z = VFMA(LDK(KP980785280), T1X, VMUL(LDK(KP195090322), T1Y));
T22 = VFNMS(LDK(KP195090322), T21, VMUL(LDK(KP980785280), T20));
T23 = VSUB(T1Z, T22);
T2e = VADD(T1Z, T22);
T25 = VADD(Tr, TC);
T26 = VADD(T1D, T1y);
T27 = VSUB(T25, T26);
T2f = VADD(T26, T25);
}
{
V T24, T2b, T2k, T2l;
T24 = VADD(T1W, T23);
T2b = VBYI(VADD(T27, T2a));
ST(&(x[WS(ios, 25)]), VSUB(T24, T2b), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VADD(T24, T2b), dist, &(x[WS(ios, 1)]));
T2k = VBYI(VADD(T2f, T2e));
T2l = VADD(T2h, T2i);
ST(&(x[WS(ios, 1)]), VADD(T2k, T2l), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 31)]), VSUB(T2l, T2k), dist, &(x[WS(ios, 1)]));
}
{
V T2c, T2d, T2g, T2j;
T2c = VSUB(T1W, T23);
T2d = VBYI(VSUB(T2a, T27));
ST(&(x[WS(ios, 23)]), VSUB(T2c, T2d), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 9)]), VADD(T2c, T2d), dist, &(x[WS(ios, 1)]));
T2g = VBYI(VSUB(T2e, T2f));
T2j = VSUB(T2h, T2i);
ST(&(x[WS(ios, 15)]), VADD(T2g, T2j), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 17)]), VSUB(T2j, T2g), dist, &(x[WS(ios, 1)]));
}
}
}
END_SIMD();
return W;
}
static void n1fv_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
int i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V Tp, T13, Tu, TN, Tm, T14, Tv, TY, T7, T17, Ty, TT, Te, T16, Tx;
V TQ;
{
V Tn, To, TM, Ts, Tt, TL;
Tn = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
TM = VADD(Tn, To);
Ts = LD(&(xi[0]), ivs, &(xi[0]));
Tt = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
TL = VADD(Ts, Tt);
Tp = VSUB(Tn, To);
T13 = VADD(TL, TM);
Tu = VSUB(Ts, Tt);
TN = VSUB(TL, TM);
}
{
V Ti, TW, Tl, TX;
{
V Tg, Th, Tj, Tk;
Tg = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Th = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Ti = VSUB(Tg, Th);
TW = VADD(Tg, Th);
Tj = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Tk = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tl = VSUB(Tj, Tk);
TX = VADD(Tj, Tk);
}
Tm = VMUL(LDK(KP707106781), VSUB(Ti, Tl));
T14 = VADD(TX, TW);
Tv = VMUL(LDK(KP707106781), VADD(Tl, Ti));
TY = VSUB(TW, TX);
}
{
V T3, TR, T6, TS;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
TR = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
T5 = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TS = VADD(T4, T5);
}
T7 = VFNMS(LDK(KP923879532), T6, VMUL(LDK(KP382683432), T3));
T17 = VADD(TR, TS);
Ty = VFMA(LDK(KP923879532), T3, VMUL(LDK(KP382683432), T6));
TT = VSUB(TR, TS);
}
{
V Ta, TO, Td, TP;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
T9 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
TO = VADD(T8, T9);
Tb = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tc = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TP = VADD(Tb, Tc);
}
Te = VFMA(LDK(KP382683432), Ta, VMUL(LDK(KP923879532), Td));
T16 = VADD(TO, TP);
Tx = VFNMS(LDK(KP382683432), Td, VMUL(LDK(KP923879532), Ta));
TQ = VSUB(TO, TP);
}
{
V T15, T18, T19, T1a;
T15 = VADD(T13, T14);
T18 = VADD(T16, T17);
ST(&(xo[WS(os, 8)]), VSUB(T15, T18), ovs, &(xo[0]));
ST(&(xo[0]), VADD(T15, T18), ovs, &(xo[0]));
T19 = VSUB(T13, T14);
T1a = VBYI(VSUB(T17, T16));
ST(&(xo[WS(os, 12)]), VSUB(T19, T1a), ovs, &(xo[0]));
ST(&(xo[WS(os, 4)]), VADD(T19, T1a), ovs, &(xo[0]));
}
{
V TV, T11, T10, T12, TU, TZ;
TU = VMUL(LDK(KP707106781), VADD(TQ, TT));
TV = VADD(TN, TU);
T11 = VSUB(TN, TU);
TZ = VMUL(LDK(KP707106781), VSUB(TT, TQ));
T10 = VBYI(VADD(TY, TZ));
T12 = VBYI(VSUB(TZ, TY));
ST(&(xo[WS(os, 14)]), VSUB(TV, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 6)]), VADD(T11, T12), ovs, &(xo[0]));
ST(&(xo[WS(os, 2)]), VADD(TV, T10), ovs, &(xo[0]));
ST(&(xo[WS(os, 10)]), VSUB(T11, T12), ovs, &(xo[0]));
}
{
V Tr, TB, TA, TC;
{
V Tf, Tq, Tw, Tz;
Tf = VSUB(T7, Te);
Tq = VSUB(Tm, Tp);
Tr = VBYI(VSUB(Tf, Tq));
TB = VBYI(VADD(Tq, Tf));
Tw = VADD(Tu, Tv);
Tz = VADD(Tx, Ty);
TA = VSUB(Tw, Tz);
TC = VADD(Tw, Tz);
}
ST(&(xo[WS(os, 7)]), VADD(Tr, TA), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 15)]), VSUB(TC, TB), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 9)]), VSUB(TA, Tr), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 1)]), VADD(TB, TC), ovs, &(xo[WS(os, 1)]));
}
{
V TF, TJ, TI, TK;
{
V TD, TE, TG, TH;
TD = VSUB(Tu, Tv);
TE = VADD(Te, T7);
TF = VADD(TD, TE);
TJ = VSUB(TD, TE);
TG = VADD(Tp, Tm);
TH = VSUB(Ty, Tx);
TI = VBYI(VADD(TG, TH));
TK = VBYI(VSUB(TH, TG));
}
ST(&(xo[WS(os, 13)]), VSUB(TF, TI), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 5)]), VADD(TJ, TK), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 3)]), VADD(TF, TI), ovs, &(xo[WS(os, 1)]));
ST(&(xo[WS(os, 11)]), VSUB(TJ, TK), ovs, &(xo[WS(os, 1)]));
}
}
END_SIMD();
}
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 const R *t1bv_12(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
int i;
R *x;
x = ii;
BEGIN_SIMD();
for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 22)) {
V T1, Tt, T6, T7, TB, Tq, TC, TD, T9, Tu, Te, Tf, Tx, Tl, Ty;
V Tz;
{
V T5, T3, T4, T2;
T1 = LD(&(x[0]), dist, &(x[0]));
T4 = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
T5 = BYTW(&(W[TWVL * 14]), T4);
T2 = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
T3 = BYTW(&(W[TWVL * 6]), T2);
Tt = VSUB(T3, T5);
T6 = VADD(T3, T5);
T7 = VFNMS(LDK(KP500000000), T6, T1);
}
{
V Tn, Tp, Tm, TA, To;
Tm = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
Tn = BYTW(&(W[0]), Tm);
TA = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
TB = BYTW(&(W[TWVL * 16]), TA);
To = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)]));
Tp = BYTW(&(W[TWVL * 8]), To);
Tq = VSUB(Tn, Tp);
TC = VADD(Tn, Tp);
TD = VFNMS(LDK(KP500000000), TC, TB);
}
{
V Td, Tb, T8, Tc, Ta;
T8 = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
T9 = BYTW(&(W[TWVL * 10]), T8);
Tc = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
Td = BYTW(&(W[TWVL * 2]), Tc);
Ta = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
Tb = BYTW(&(W[TWVL * 18]), Ta);
Tu = VSUB(Tb, Td);
Te = VADD(Tb, Td);
Tf = VFNMS(LDK(KP500000000), Te, T9);
}
{
V Ti, Tk, Th, Tw, Tj;
Th = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
Ti = BYTW(&(W[TWVL * 12]), Th);
Tw = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)]));
Tx = BYTW(&(W[TWVL * 4]), Tw);
Tj = LD(&(x[WS(ios, 11)]), dist, &(x[WS(ios, 1)]));
Tk = BYTW(&(W[TWVL * 20]), Tj);
Tl = VSUB(Ti, Tk);
Ty = VADD(Ti, Tk);
Tz = VFNMS(LDK(KP500000000), Ty, Tx);
}
{
V Ts, TG, TF, TH;
{
V Tg, Tr, Tv, TE;
Tg = VSUB(T7, Tf);
Tr = VMUL(LDK(KP866025403), VSUB(Tl, Tq));
Ts = VSUB(Tg, Tr);
TG = VADD(Tg, Tr);
Tv = VMUL(LDK(KP866025403), VSUB(Tt, Tu));
TE = VSUB(Tz, TD);
TF = VBYI(VADD(Tv, TE));
TH = VBYI(VSUB(TE, Tv));
}
ST(&(x[WS(ios, 11)]), VSUB(Ts, TF), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 5)]), VADD(TG, TH), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 1)]), VADD(Ts, TF), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 7)]), VSUB(TG, TH), dist, &(x[WS(ios, 1)]));
}
{
V TS, TW, TV, TX;
{
V TQ, TR, TT, TU;
TQ = VADD(T1, T6);
TR = VADD(T9, Te);
TS = VSUB(TQ, TR);
TW = VADD(TQ, TR);
TT = VADD(Tx, Ty);
TU = VADD(TB, TC);
TV = VBYI(VSUB(TT, TU));
TX = VADD(TT, TU);
}
ST(&(x[WS(ios, 3)]), VSUB(TS, TV), dist, &(x[WS(ios, 1)]));
ST(&(x[0]), VADD(TW, TX), dist, &(x[0]));
ST(&(x[WS(ios, 9)]), VADD(TS, TV), dist, &(x[WS(ios, 1)]));
ST(&(x[WS(ios, 6)]), VSUB(TW, TX), dist, &(x[0]));
}
{
V TK, TO, TN, TP;
{
V TI, TJ, TL, TM;
TI = VADD(Tl, Tq);
TJ = VADD(Tt, Tu);
TK = VBYI(VMUL(LDK(KP866025403), VSUB(TI, TJ)));
TO = VBYI(VMUL(LDK(KP866025403), VADD(TJ, TI)));
TL = VADD(T7, Tf);
TM = VADD(Tz, TD);
TN = VSUB(TL, TM);
TP = VADD(TL, TM);
}
ST(&(x[WS(ios, 2)]), VADD(TK, TN), dist, &(x[0]));
ST(&(x[WS(ios, 8)]), VSUB(TP, TO), dist, &(x[0]));
ST(&(x[WS(ios, 10)]), VSUB(TN, TK), dist, &(x[0]));
ST(&(x[WS(ios, 4)]), VADD(TO, TP), dist, &(x[0]));
}
}
END_SIMD();
return W;
}
static const R *q1fv_4(R *ri, R *ii, const R *W, stride is, stride vs, int m, int dist)
{
int i;
R *x;
x = ri;
BEGIN_SIMD();
for (i = 0; i < m; i = i + VL, x = x + (VL * dist), W = W + (TWVL * 6)) {
V T3, T9, TA, TG, TD, TH, T6, Ta, Te, Tk, Tp, Tv, Ts, Tw, Th;
V Tl;
{
V T1, T2, Ty, Tz;
T1 = LD(&(x[0]), dist, &(x[0]));
T2 = LD(&(x[WS(is, 2)]), dist, &(x[0]));
T3 = VSUB(T1, T2);
T9 = VADD(T1, T2);
Ty = LD(&(x[WS(vs, 3)]), dist, &(x[WS(vs, 3)]));
Tz = LD(&(x[WS(vs, 3) + WS(is, 2)]), dist, &(x[WS(vs, 3)]));
TA = VSUB(Ty, Tz);
TG = VADD(Ty, Tz);
}
{
V TB, TC, T4, T5;
TB = LD(&(x[WS(vs, 3) + WS(is, 1)]), dist, &(x[WS(vs, 3) + WS(is, 1)]));
TC = LD(&(x[WS(vs, 3) + WS(is, 3)]), dist, &(x[WS(vs, 3) + WS(is, 1)]));
TD = VBYI(VSUB(TB, TC));
TH = VADD(TB, TC);
T4 = LD(&(x[WS(is, 1)]), dist, &(x[WS(is, 1)]));
T5 = LD(&(x[WS(is, 3)]), dist, &(x[WS(is, 1)]));
T6 = VBYI(VSUB(T4, T5));
Ta = VADD(T4, T5);
}
{
V Tc, Td, Tn, To;
Tc = LD(&(x[WS(vs, 1)]), dist, &(x[WS(vs, 1)]));
Td = LD(&(x[WS(vs, 1) + WS(is, 2)]), dist, &(x[WS(vs, 1)]));
Te = VSUB(Tc, Td);
Tk = VADD(Tc, Td);
Tn = LD(&(x[WS(vs, 2)]), dist, &(x[WS(vs, 2)]));
To = LD(&(x[WS(vs, 2) + WS(is, 2)]), dist, &(x[WS(vs, 2)]));
Tp = VSUB(Tn, To);
Tv = VADD(Tn, To);
}
{
V Tq, Tr, Tf, Tg;
Tq = LD(&(x[WS(vs, 2) + WS(is, 1)]), dist, &(x[WS(vs, 2) + WS(is, 1)]));
Tr = LD(&(x[WS(vs, 2) + WS(is, 3)]), dist, &(x[WS(vs, 2) + WS(is, 1)]));
Ts = VBYI(VSUB(Tq, Tr));
Tw = VADD(Tq, Tr);
Tf = LD(&(x[WS(vs, 1) + WS(is, 1)]), dist, &(x[WS(vs, 1) + WS(is, 1)]));
Tg = LD(&(x[WS(vs, 1) + WS(is, 3)]), dist, &(x[WS(vs, 1) + WS(is, 1)]));
Th = VBYI(VSUB(Tf, Tg));
Tl = VADD(Tf, Tg);
}
ST(&(x[0]), VADD(T9, Ta), dist, &(x[0]));
ST(&(x[WS(is, 1)]), VADD(Tk, Tl), dist, &(x[WS(is, 1)]));
ST(&(x[WS(is, 2)]), VADD(Tv, Tw), dist, &(x[0]));
ST(&(x[WS(is, 3)]), VADD(TG, TH), dist, &(x[WS(is, 1)]));
{
V T7, Ti, Tt, TE;
T7 = BYTWJ(&(W[0]), VSUB(T3, T6));
ST(&(x[WS(vs, 1)]), T7, dist, &(x[WS(vs, 1)]));
Ti = BYTWJ(&(W[0]), VSUB(Te, Th));
ST(&(x[WS(vs, 1) + WS(is, 1)]), Ti, dist, &(x[WS(vs, 1) + WS(is, 1)]));
Tt = BYTWJ(&(W[0]), VSUB(Tp, Ts));
ST(&(x[WS(vs, 1) + WS(is, 2)]), Tt, dist, &(x[WS(vs, 1)]));
TE = BYTWJ(&(W[0]), VSUB(TA, TD));
ST(&(x[WS(vs, 1) + WS(is, 3)]), TE, dist, &(x[WS(vs, 1) + WS(is, 1)]));
}
{
V T8, Tj, Tu, TF;
T8 = BYTWJ(&(W[TWVL * 4]), VADD(T3, T6));
ST(&(x[WS(vs, 3)]), T8, dist, &(x[WS(vs, 3)]));
Tj = BYTWJ(&(W[TWVL * 4]), VADD(Te, Th));
ST(&(x[WS(vs, 3) + WS(is, 1)]), Tj, dist, &(x[WS(vs, 3) + WS(is, 1)]));
Tu = BYTWJ(&(W[TWVL * 4]), VADD(Tp, Ts));
ST(&(x[WS(vs, 3) + WS(is, 2)]), Tu, dist, &(x[WS(vs, 3)]));
TF = BYTWJ(&(W[TWVL * 4]), VADD(TA, TD));
ST(&(x[WS(vs, 3) + WS(is, 3)]), TF, dist, &(x[WS(vs, 3) + WS(is, 1)]));
}
{
V Tb, Tm, Tx, TI;
Tb = BYTWJ(&(W[TWVL * 2]), VSUB(T9, Ta));
ST(&(x[WS(vs, 2)]), Tb, dist, &(x[WS(vs, 2)]));
Tm = BYTWJ(&(W[TWVL * 2]), VSUB(Tk, Tl));
ST(&(x[WS(vs, 2) + WS(is, 1)]), Tm, dist, &(x[WS(vs, 2) + WS(is, 1)]));
Tx = BYTWJ(&(W[TWVL * 2]), VSUB(Tv, Tw));
ST(&(x[WS(vs, 2) + WS(is, 2)]), Tx, dist, &(x[WS(vs, 2)]));
TI = BYTWJ(&(W[TWVL * 2]), VSUB(TG, TH));
ST(&(x[WS(vs, 2) + WS(is, 3)]), TI, dist, &(x[WS(vs, 2) + WS(is, 1)]));
}
}
END_SIMD();
return W;
}
static void n2bv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, int v, int ivs, int ovs)
{
DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
int i;
const R *xi;
R *xo;
xi = ii;
xo = io;
BEGIN_SIMD();
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs)) {
V Tp, Ty, Tl, TL, Tq, TE, T7, TJ, Ts, TB, Te, TK, Tr, TH, Tn;
V To;
Tn = LD(&(xi[0]), ivs, &(xi[0]));
To = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Tp = VSUB(Tn, To);
Ty = VADD(Tn, To);
{
V Th, TC, Tk, TD;
{
V Tf, Tg, Ti, Tj;
Tf = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Tg = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Th = VSUB(Tf, Tg);
TC = VADD(Tf, Tg);
Ti = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Tj = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tk = VSUB(Ti, Tj);
TD = VADD(Ti, Tj);
}
Tl = VSUB(Th, Tk);
TL = VSUB(TD, TC);
Tq = VADD(Th, Tk);
TE = VADD(TC, TD);
}
{
V T3, Tz, T6, TA;
{
V T1, T2, T4, T5;
T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T2 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
T3 = VSUB(T1, T2);
Tz = VADD(T1, T2);
T4 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
T6 = VSUB(T4, T5);
TA = VADD(T4, T5);
}
T7 = VSUB(T3, T6);
TJ = VSUB(Tz, TA);
Ts = VADD(T3, T6);
TB = VADD(Tz, TA);
}
{
V Ta, TF, Td, TG;
{
V T8, T9, Tb, Tc;
T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
T9 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Ta = VSUB(T8, T9);
TF = VADD(T8, T9);
Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Tc = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Td = VSUB(Tb, Tc);
TG = VADD(Tb, Tc);
}
Te = VSUB(Ta, Td);
TK = VSUB(TG, TF);
Tr = VADD(Ta, Td);
TH = VADD(TF, TG);
}
ST(&(xo[14]), VADD(Tp, VADD(Ts, VADD(Tq, Tr))), ovs, &(xo[2]));
ST(&(xo[0]), VADD(Ty, VADD(TB, VADD(TE, TH))), ovs, &(xo[0]));
{
V Tm, Tt, TQ, TP;
Tm = VBYI(VFMA(LDK(KP433883739), T7, VFNMS(LDK(KP781831482), Tl, VMUL(LDK(KP974927912), Te))));
Tt = VFMA(LDK(KP623489801), Tq, VFNMS(LDK(KP222520933), Tr, VFNMS(LDK(KP900968867), Ts, Tp)));
ST(&(xo[6]), VADD(Tm, Tt), ovs, &(xo[2]));
ST(&(xo[22]), VSUB(Tt, Tm), ovs, &(xo[2]));
TQ = VBYI(VFMA(LDK(KP974927912), TJ, VFMA(LDK(KP433883739), TL, VMUL(LDK(KP781831482), TK))));
TP = VFMA(LDK(KP623489801), TH, VFNMS(LDK(KP900968867), TE, VFNMS(LDK(KP222520933), TB, Ty)));
ST(&(xo[24]), VSUB(TP, TQ), ovs, &(xo[0]));
ST(&(xo[4]), VADD(TP, TQ), ovs, &(xo[0]));
}
{
V Tu, Tv, TM, TI;
Tu = VBYI(VFMA(LDK(KP781831482), T7, VFMA(LDK(KP974927912), Tl, VMUL(LDK(KP433883739), Te))));
Tv = VFMA(LDK(KP623489801), Ts, VFNMS(LDK(KP900968867), Tr, VFNMS(LDK(KP222520933), Tq, Tp)));
ST(&(xo[2]), VADD(Tu, Tv), ovs, &(xo[2]));
ST(&(xo[26]), VSUB(Tv, Tu), ovs, &(xo[2]));
TM = VBYI(VFNMS(LDK(KP433883739), TK, VFNMS(LDK(KP974927912), TL, VMUL(LDK(KP781831482), TJ))));
TI = VFMA(LDK(KP623489801), TB, VFNMS(LDK(KP900968867), TH, VFNMS(LDK(KP222520933), TE, Ty)));
ST(&(xo[12]), VSUB(TI, TM), ovs, &(xo[0]));
ST(&(xo[16]), VADD(TI, TM), ovs, &(xo[0]));
}
{
V TO, TN, Tx, Tw;
TO = VBYI(VFMA(LDK(KP433883739), TJ, VFNMS(LDK(KP974927912), TK, VMUL(LDK(KP781831482), TL))));
TN = VFMA(LDK(KP623489801), TE, VFNMS(LDK(KP222520933), TH, VFNMS(LDK(KP900968867), TB, Ty)));
ST(&(xo[8]), VSUB(TN, TO), ovs, &(xo[0]));
ST(&(xo[20]), VADD(TN, TO), ovs, &(xo[0]));
Tx = VBYI(VFNMS(LDK(KP781831482), Te, VFNMS(LDK(KP433883739), Tl, VMUL(LDK(KP974927912), T7))));
Tw = VFMA(LDK(KP623489801), Tr, VFNMS(LDK(KP900968867), Tq, VFNMS(LDK(KP222520933), Ts, Tp)));
ST(&(xo[10]), VSUB(Tw, Tx), ovs, &(xo[2]));
ST(&(xo[18]), VADD(Tx, Tw), ovs, &(xo[2]));
}
}
END_SIMD();
}
