cannam@167: /*
cannam@167:  * Copyright (c) 2003, 2007-14 Matteo Frigo
cannam@167:  * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
cannam@167:  *
cannam@167:  * This program is free software; you can redistribute it and/or modify
cannam@167:  * it under the terms of the GNU General Public License as published by
cannam@167:  * the Free Software Foundation; either version 2 of the License, or
cannam@167:  * (at your option) any later version.
cannam@167:  *
cannam@167:  * This program is distributed in the hope that it will be useful,
cannam@167:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
cannam@167:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
cannam@167:  * GNU General Public License for more details.
cannam@167:  *
cannam@167:  * You should have received a copy of the GNU General Public License
cannam@167:  * along with this program; if not, write to the Free Software
cannam@167:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
cannam@167:  *
cannam@167:  */
cannam@167: 
cannam@167: /* This file was automatically generated --- DO NOT EDIT */
cannam@167: /* Generated on Thu May 24 08:04:51 EDT 2018 */
cannam@167: 
cannam@167: #include "dft/codelet-dft.h"
cannam@167: 
cannam@167: #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
cannam@167: 
cannam@167: /* Generated by: ../../../genfft/gen_notw_c.native -fma -simd -compact -variables 4 -pipeline-latency 8 -n 11 -name n1fv_11 -include dft/simd/n1f.h */
cannam@167: 
cannam@167: /*
cannam@167:  * This function contains 70 FP additions, 60 FP multiplications,
cannam@167:  * (or, 15 additions, 5 multiplications, 55 fused multiply/add),
cannam@167:  * 42 stack variables, 11 constants, and 22 memory accesses
cannam@167:  */
cannam@167: #include "dft/simd/n1f.h"
cannam@167: 
cannam@167: static void n1fv_11(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
cannam@167: {
cannam@167:      DVK(KP959492973, +0.959492973614497389890368057066327699062454848);
cannam@167:      DVK(KP918985947, +0.918985947228994779780736114132655398124909697);
cannam@167:      DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
cannam@167:      DVK(KP830830026, +0.830830026003772851058548298459246407048009821);
cannam@167:      DVK(KP876768831, +0.876768831002589333891339807079336796764054852);
cannam@167:      DVK(KP778434453, +0.778434453334651800608337670740821884709317477);
cannam@167:      DVK(KP372785597, +0.372785597771792209609773152906148328659002598);
cannam@167:      DVK(KP715370323, +0.715370323453429719112414662767260662417897278);
cannam@167:      DVK(KP521108558, +0.521108558113202722944698153526659300680427422);
cannam@167:      DVK(KP634356270, +0.634356270682424498893150776899916060542806975);
cannam@167:      DVK(KP342584725, +0.342584725681637509502641509861112333758894680);
cannam@167:      {
cannam@167: 	  INT i;
cannam@167: 	  const R *xi;
cannam@167: 	  R *xo;
cannam@167: 	  xi = ri;
cannam@167: 	  xo = ro;
cannam@167: 	  for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(22, is), MAKE_VOLATILE_STRIDE(22, os)) {
cannam@167: 	       V T1, T4, Tp, Tg, Tq, T7, Tn, Ta, Tm, Td, To, Ti, Tw, T12, Ts;
cannam@167: 	       V TX, TT, TK, TB, TO, TF, T5, T6;
cannam@167: 	       T1 = LD(&(xi[0]), ivs, &(xi[0]));
cannam@167: 	       {
cannam@167: 		    V T2, T3, Te, Tf;
cannam@167: 		    T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
cannam@167: 		    T4 = VADD(T2, T3);
cannam@167: 		    Tp = VSUB(T3, T2);
cannam@167: 		    Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    Tf = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
cannam@167: 		    Tg = VADD(Te, Tf);
cannam@167: 		    Tq = VSUB(Tf, Te);
cannam@167: 	       }
cannam@167: 	       T5 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
cannam@167: 	       T6 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 	       T7 = VADD(T5, T6);
cannam@167: 	       Tn = VSUB(T6, T5);
cannam@167: 	       {
cannam@167: 		    V T8, T9, Tb, Tc;
cannam@167: 		    T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    T9 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
cannam@167: 		    Ta = VADD(T8, T9);
cannam@167: 		    Tm = VSUB(T9, T8);
cannam@167: 		    Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
cannam@167: 		    Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    Td = VADD(Tb, Tc);
cannam@167: 		    To = VSUB(Tc, Tb);
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V Th, Tv, T11, Tr, TW;
cannam@167: 		    Th = VFNMS(LDK(KP342584725), Ta, T7);
cannam@167: 		    Ti = VFNMS(LDK(KP634356270), Th, Td);
cannam@167: 		    Tv = VFNMS(LDK(KP342584725), Td, T4);
cannam@167: 		    Tw = VFNMS(LDK(KP634356270), Tv, T7);
cannam@167: 		    T11 = VFNMS(LDK(KP521108558), Tp, Tn);
cannam@167: 		    T12 = VFNMS(LDK(KP715370323), T11, Tm);
cannam@167: 		    Tr = VFMA(LDK(KP521108558), Tq, Tp);
cannam@167: 		    Ts = VFMA(LDK(KP715370323), Tr, To);
cannam@167: 		    TW = VFNMS(LDK(KP342584725), Tg, Td);
cannam@167: 		    TX = VFNMS(LDK(KP634356270), TW, Ta);
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V TS, TJ, TA, TN, TE;
cannam@167: 		    TS = VFMA(LDK(KP715370323), Tm, Tp);
cannam@167: 		    TT = VFMA(LDK(KP372785597), To, TS);
cannam@167: 		    TJ = VFNMS(LDK(KP521108558), Tn, To);
cannam@167: 		    TK = VFMA(LDK(KP715370323), TJ, Tq);
cannam@167: 		    TA = VFMA(LDK(KP521108558), Tm, Tq);
cannam@167: 		    TB = VFNMS(LDK(KP715370323), TA, Tn);
cannam@167: 		    TN = VFNMS(LDK(KP342584725), T7, Tg);
cannam@167: 		    TO = VFNMS(LDK(KP634356270), TN, T4);
cannam@167: 		    TE = VFNMS(LDK(KP342584725), T4, Ta);
cannam@167: 		    TF = VFNMS(LDK(KP634356270), TE, Tg);
cannam@167: 	       }
cannam@167: 	       ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, VADD(Ta, VADD(Td, Tg))))), ovs, &(xo[0]));
cannam@167: 	       {
cannam@167: 		    V Tk, Tu, Tj, Tt, Tl;
cannam@167: 		    Tj = VFNMS(LDK(KP778434453), Ti, T4);
cannam@167: 		    Tk = VFNMS(LDK(KP876768831), Tj, Tg);
cannam@167: 		    Tt = VFMA(LDK(KP830830026), Ts, Tn);
cannam@167: 		    Tu = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), Tt, Tm));
cannam@167: 		    Tl = VFNMS(LDK(KP959492973), Tk, T1);
cannam@167: 		    ST(&(xo[WS(os, 10)]), VFNMSI(Tu, Tl), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 1)]), VFMAI(Tu, Tl), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V TZ, T14, TY, T13, T10;
cannam@167: 		    TY = VFNMS(LDK(KP778434453), TX, T7);
cannam@167: 		    TZ = VFNMS(LDK(KP876768831), TY, T4);
cannam@167: 		    T13 = VFNMS(LDK(KP830830026), T12, To);
cannam@167: 		    T14 = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), T13, Tq));
cannam@167: 		    T10 = VFNMS(LDK(KP959492973), TZ, T1);
cannam@167: 		    ST(&(xo[WS(os, 6)]), VFNMSI(T14, T10), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 5)]), VFMAI(T14, T10), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V TQ, TV, TP, TU, TR;
cannam@167: 		    TP = VFNMS(LDK(KP778434453), TO, Ta);
cannam@167: 		    TQ = VFNMS(LDK(KP876768831), TP, Td);
cannam@167: 		    TU = VFNMS(LDK(KP830830026), TT, Tq);
cannam@167: 		    TV = VMUL(LDK(KP989821441), VFMA(LDK(KP918985947), TU, Tn));
cannam@167: 		    TR = VFNMS(LDK(KP959492973), TQ, T1);
cannam@167: 		    ST(&(xo[WS(os, 4)]), VFNMSI(TV, TR), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 7)]), VFMAI(TV, TR), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V TH, TM, TG, TL, TI;
cannam@167: 		    TG = VFNMS(LDK(KP778434453), TF, Td);
cannam@167: 		    TH = VFNMS(LDK(KP876768831), TG, T7);
cannam@167: 		    TL = VFNMS(LDK(KP830830026), TK, Tm);
cannam@167: 		    TM = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TL, Tp));
cannam@167: 		    TI = VFNMS(LDK(KP959492973), TH, T1);
cannam@167: 		    ST(&(xo[WS(os, 8)]), VFNMSI(TM, TI), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 3)]), VFMAI(TM, TI), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	       {
cannam@167: 		    V Ty, TD, Tx, TC, Tz;
cannam@167: 		    Tx = VFNMS(LDK(KP778434453), Tw, Tg);
cannam@167: 		    Ty = VFNMS(LDK(KP876768831), Tx, Ta);
cannam@167: 		    TC = VFMA(LDK(KP830830026), TB, Tp);
cannam@167: 		    TD = VMUL(LDK(KP989821441), VFNMS(LDK(KP918985947), TC, To));
cannam@167: 		    Tz = VFNMS(LDK(KP959492973), Ty, T1);
cannam@167: 		    ST(&(xo[WS(os, 2)]), VFNMSI(TD, Tz), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 9)]), VFMAI(TD, Tz), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	  }
cannam@167:      }
cannam@167:      VLEAVE();
cannam@167: }
cannam@167: 
cannam@167: static const kdft_desc desc = { 11, XSIMD_STRING("n1fv_11"), {15, 5, 55, 0}, &GENUS, 0, 0, 0, 0 };
cannam@167: 
cannam@167: void XSIMD(codelet_n1fv_11) (planner *p) {
cannam@167:      X(kdft_register) (p, n1fv_11, &desc);
cannam@167: }
cannam@167: 
cannam@167: #else
cannam@167: 
cannam@167: /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 11 -name n1fv_11 -include dft/simd/n1f.h */
cannam@167: 
cannam@167: /*
cannam@167:  * This function contains 70 FP additions, 50 FP multiplications,
cannam@167:  * (or, 30 additions, 10 multiplications, 40 fused multiply/add),
cannam@167:  * 32 stack variables, 10 constants, and 22 memory accesses
cannam@167:  */
cannam@167: #include "dft/simd/n1f.h"
cannam@167: 
cannam@167: static void n1fv_11(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
cannam@167: {
cannam@167:      DVK(KP654860733, +0.654860733945285064056925072466293553183791199);
cannam@167:      DVK(KP142314838, +0.142314838273285140443792668616369668791051361);
cannam@167:      DVK(KP959492973, +0.959492973614497389890368057066327699062454848);
cannam@167:      DVK(KP415415013, +0.415415013001886425529274149229623203524004910);
cannam@167:      DVK(KP841253532, +0.841253532831181168861811648919367717513292498);
cannam@167:      DVK(KP989821441, +0.989821441880932732376092037776718787376519372);
cannam@167:      DVK(KP909631995, +0.909631995354518371411715383079028460060241051);
cannam@167:      DVK(KP281732556, +0.281732556841429697711417915346616899035777899);
cannam@167:      DVK(KP540640817, +0.540640817455597582107635954318691695431770608);
cannam@167:      DVK(KP755749574, +0.755749574354258283774035843972344420179717445);
cannam@167:      {
cannam@167: 	  INT i;
cannam@167: 	  const R *xi;
cannam@167: 	  R *xo;
cannam@167: 	  xi = ri;
cannam@167: 	  xo = ro;
cannam@167: 	  for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(22, is), MAKE_VOLATILE_STRIDE(22, os)) {
cannam@167: 	       V T1, T4, Ti, Tg, Tl, Td, Tk, Ta, Tj, T7, Tm, Tb, Tc, Tt, Ts;
cannam@167: 	       T1 = LD(&(xi[0]), ivs, &(xi[0]));
cannam@167: 	       {
cannam@167: 		    V T2, T3, Te, Tf;
cannam@167: 		    T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
cannam@167: 		    T4 = VADD(T2, T3);
cannam@167: 		    Ti = VSUB(T3, T2);
cannam@167: 		    Te = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    Tf = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
cannam@167: 		    Tg = VADD(Te, Tf);
cannam@167: 		    Tl = VSUB(Tf, Te);
cannam@167: 	       }
cannam@167: 	       Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
cannam@167: 	       Tc = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 	       Td = VADD(Tb, Tc);
cannam@167: 	       Tk = VSUB(Tc, Tb);
cannam@167: 	       {
cannam@167: 		    V T8, T9, T5, T6;
cannam@167: 		    T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    T9 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
cannam@167: 		    Ta = VADD(T8, T9);
cannam@167: 		    Tj = VSUB(T9, T8);
cannam@167: 		    T5 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
cannam@167: 		    T6 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
cannam@167: 		    T7 = VADD(T5, T6);
cannam@167: 		    Tm = VSUB(T6, T5);
cannam@167: 	       }
cannam@167: 	       ST(&(xo[0]), VADD(T1, VADD(T4, VADD(T7, VADD(Ta, VADD(Td, Tg))))), ovs, &(xo[0]));
cannam@167: 	       {
cannam@167: 		    V Tn, Th, Tv, Tu;
cannam@167: 		    Tn = VBYI(VFMA(LDK(KP755749574), Ti, VFMA(LDK(KP540640817), Tj, VFNMS(LDK(KP909631995), Tl, VFNMS(LDK(KP989821441), Tm, VMUL(LDK(KP281732556), Tk))))));
cannam@167: 		    Th = VFMA(LDK(KP841253532), Ta, VFMA(LDK(KP415415013), Tg, VFNMS(LDK(KP959492973), Td, VFNMS(LDK(KP142314838), T7, VFNMS(LDK(KP654860733), T4, T1)))));
cannam@167: 		    ST(&(xo[WS(os, 7)]), VSUB(Th, Tn), ovs, &(xo[WS(os, 1)]));
cannam@167: 		    ST(&(xo[WS(os, 4)]), VADD(Th, Tn), ovs, &(xo[0]));
cannam@167: 		    Tv = VBYI(VFMA(LDK(KP281732556), Ti, VFMA(LDK(KP755749574), Tj, VFNMS(LDK(KP909631995), Tk, VFNMS(LDK(KP540640817), Tm, VMUL(LDK(KP989821441), Tl))))));
cannam@167: 		    Tu = VFMA(LDK(KP841253532), T7, VFMA(LDK(KP415415013), Td, VFNMS(LDK(KP142314838), Tg, VFNMS(LDK(KP654860733), Ta, VFNMS(LDK(KP959492973), T4, T1)))));
cannam@167: 		    ST(&(xo[WS(os, 6)]), VSUB(Tu, Tv), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 5)]), VADD(Tu, Tv), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       }
cannam@167: 	       Tt = VBYI(VFMA(LDK(KP989821441), Ti, VFMA(LDK(KP540640817), Tk, VFNMS(LDK(KP909631995), Tj, VFNMS(LDK(KP281732556), Tm, VMUL(LDK(KP755749574), Tl))))));
cannam@167: 	       Ts = VFMA(LDK(KP415415013), Ta, VFMA(LDK(KP841253532), Td, VFNMS(LDK(KP654860733), Tg, VFNMS(LDK(KP959492973), T7, VFNMS(LDK(KP142314838), T4, T1)))));
cannam@167: 	       ST(&(xo[WS(os, 8)]), VSUB(Ts, Tt), ovs, &(xo[0]));
cannam@167: 	       ST(&(xo[WS(os, 3)]), VADD(Ts, Tt), ovs, &(xo[WS(os, 1)]));
cannam@167: 	       {
cannam@167: 		    V Tr, Tq, Tp, To;
cannam@167: 		    Tr = VBYI(VFMA(LDK(KP540640817), Ti, VFMA(LDK(KP909631995), Tm, VFMA(LDK(KP989821441), Tj, VFMA(LDK(KP755749574), Tk, VMUL(LDK(KP281732556), Tl))))));
cannam@167: 		    Tq = VFMA(LDK(KP841253532), T4, VFMA(LDK(KP415415013), T7, VFNMS(LDK(KP959492973), Tg, VFNMS(LDK(KP654860733), Td, VFNMS(LDK(KP142314838), Ta, T1)))));
cannam@167: 		    ST(&(xo[WS(os, 10)]), VSUB(Tq, Tr), ovs, &(xo[0]));
cannam@167: 		    ST(&(xo[WS(os, 1)]), VADD(Tq, Tr), ovs, &(xo[WS(os, 1)]));
cannam@167: 		    Tp = VBYI(VFMA(LDK(KP909631995), Ti, VFNMS(LDK(KP540640817), Tl, VFNMS(LDK(KP989821441), Tk, VFNMS(LDK(KP281732556), Tj, VMUL(LDK(KP755749574), Tm))))));
cannam@167: 		    To = VFMA(LDK(KP415415013), T4, VFMA(LDK(KP841253532), Tg, VFNMS(LDK(KP142314838), Td, VFNMS(LDK(KP959492973), Ta, VFNMS(LDK(KP654860733), T7, T1)))));
cannam@167: 		    ST(&(xo[WS(os, 9)]), VSUB(To, Tp), ovs, &(xo[WS(os, 1)]));
cannam@167: 		    ST(&(xo[WS(os, 2)]), VADD(To, Tp), ovs, &(xo[0]));
cannam@167: 	       }
cannam@167: 	  }
cannam@167:      }
cannam@167:      VLEAVE();
cannam@167: }
cannam@167: 
cannam@167: static const kdft_desc desc = { 11, XSIMD_STRING("n1fv_11"), {30, 10, 40, 0}, &GENUS, 0, 0, 0, 0 };
cannam@167: 
cannam@167: void XSIMD(codelet_n1fv_11) (planner *p) {
cannam@167:      X(kdft_register) (p, n1fv_11, &desc);
cannam@167: }
cannam@167: 
cannam@167: #endif