cannam@95: /* cannam@95: * Copyright (c) 2003, 2007-11 Matteo Frigo cannam@95: * Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology cannam@95: * cannam@95: * This program is free software; you can redistribute it and/or modify cannam@95: * it under the terms of the GNU General Public License as published by cannam@95: * the Free Software Foundation; either version 2 of the License, or cannam@95: * (at your option) any later version. cannam@95: * cannam@95: * This program is distributed in the hope that it will be useful, cannam@95: * but WITHOUT ANY WARRANTY; without even the implied warranty of cannam@95: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the cannam@95: * GNU General Public License for more details. cannam@95: * cannam@95: * You should have received a copy of the GNU General Public License cannam@95: * along with this program; if not, write to the Free Software cannam@95: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA cannam@95: * cannam@95: */ cannam@95: cannam@95: /* This file was automatically generated --- DO NOT EDIT */ cannam@95: /* Generated on Sun Nov 25 07:39:47 EST 2012 */ cannam@95: cannam@95: #include "codelet-rdft.h" cannam@95: cannam@95: #ifdef HAVE_FMA cannam@95: cannam@95: /* Generated by: ../../../genfft/gen_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 128 -name r2cf_128 -include r2cf.h */ cannam@95: cannam@95: /* cannam@95: * This function contains 956 FP additions, 516 FP multiplications, cannam@95: * (or, 440 additions, 0 multiplications, 516 fused multiply/add), cannam@95: * 229 stack variables, 31 constants, and 256 memory accesses cannam@95: */ cannam@95: #include "r2cf.h" cannam@95: cannam@95: static void r2cf_128(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@95: { cannam@95: DK(KP989176509, +0.989176509964780973451673738016243063983689533); cannam@95: DK(KP803207531, +0.803207531480644909806676512963141923879569427); cannam@95: DK(KP148335987, +0.148335987538347428753676511486911367000625355); cannam@95: DK(KP741650546, +0.741650546272035369581266691172079863842265220); cannam@95: DK(KP998795456, +0.998795456205172392714771604759100694443203615); cannam@95: DK(KP740951125, +0.740951125354959091175616897495162729728955309); cannam@95: DK(KP049126849, +0.049126849769467254105343321271313617079695752); cannam@95: DK(KP906347169, +0.906347169019147157946142717268914412664134293); cannam@95: DK(KP857728610, +0.857728610000272069902269984284770137042490799); cannam@95: DK(KP970031253, +0.970031253194543992603984207286100251456865962); cannam@95: DK(KP599376933, +0.599376933681923766271389869014404232837890546); cannam@95: DK(KP250486960, +0.250486960191305461595702160124721208578685568); cannam@95: DK(KP941544065, +0.941544065183020778412509402599502357185589796); cannam@95: DK(KP903989293, +0.903989293123443331586200297230537048710132025); cannam@95: DK(KP472964775, +0.472964775891319928124438237972992463904131113); cannam@95: DK(KP357805721, +0.357805721314524104672487743774474392487532769); cannam@95: DK(KP773010453, +0.773010453362736960810906609758469800971041293); cannam@95: DK(KP995184726, +0.995184726672196886244836953109479921575474869); cannam@95: DK(KP098491403, +0.098491403357164253077197521291327432293052451); cannam@95: DK(KP820678790, +0.820678790828660330972281985331011598767386482); cannam@95: DK(KP956940335, +0.956940335732208864935797886980269969482849206); cannam@95: DK(KP881921264, +0.881921264348355029712756863660388349508442621); cannam@95: DK(KP534511135, +0.534511135950791641089685961295362908582039528); cannam@95: DK(KP303346683, +0.303346683607342391675883946941299872384187453); cannam@95: DK(KP831469612, +0.831469612302545237078788377617905756738560812); cannam@95: DK(KP980785280, +0.980785280403230449126182236134239036973933731); cannam@95: DK(KP923879532, +0.923879532511286756128183189396788286822416626); cannam@95: DK(KP198912367, +0.198912367379658006911597622644676228597850501); cannam@95: DK(KP668178637, +0.668178637919298919997757686523080761552472251); cannam@95: DK(KP414213562, +0.414213562373095048801688724209698078569671875); cannam@95: DK(KP707106781, +0.707106781186547524400844362104849039284835938); cannam@95: { cannam@95: INT i; cannam@95: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(512, rs), MAKE_VOLATILE_STRIDE(512, csr), MAKE_VOLATILE_STRIDE(512, csi)) { cannam@95: E T95, T96; cannam@95: { cannam@95: E TcD, TdR, T5P, T8v, T27, T7r, Tf, Ta5, T7s, T5S, T8w, T2e, TdS, TcG, Tbn; cannam@95: E Tu, TcK, TdU, TK, Ta6, T7w, T8y, T2o, T5U, TcN, TdV, TZ, Ta7, T7z, T8z; cannam@95: E T2x, T5V, T1g, Taa, Tab, T1v, Tew, TcX, Tex, TcU, T6A, T2M, T9b, T7E, T9a; cannam@95: E T7H, T6z, T2T, TeO, TdK, TeL, Tdz, T9p, T8d, T6O, T5G, T6L, T4X, Tc3, TaV; cannam@95: E Tc4, Tbi, T9s, T8o, TeH, Tdp, TeE, Tde, T9i, T7U, T6H, T4r, T6E, T3I, TbW; cannam@95: E Tao, TbX, TaL, T9l, T85, T1L, Tad, Tae, T20, Tez, Td6, TeA, Td3, T6x, T37; cannam@95: E T9e, T7L, T9d, T7O, T6w, T3e, TbZ, T3Z, T4s, Tc0, TeF, Tds, T4t, T4g, T87; cannam@95: E T80, TeI, Tdl, T86, T7X, TaM, TaD, Tb2, Tc6, T8e, T8f, T5e, T5H, Tb9, Tc7; cannam@95: E TeM, TdN, T5I, T5v, T8q, T8j, TeP, TdG; cannam@95: { cannam@95: E T7G, T2S, T2P, T7F; cannam@95: { cannam@95: E T28, Ti, Tn, T2c, Ts, T29, Tl, To; cannam@95: { cannam@95: E T4, T23, T3, T25, Td, T5, T8, T9; cannam@95: { cannam@95: E T1, T2, Tb, Tc; cannam@95: T1 = R0[0]; cannam@95: T2 = R0[WS(rs, 32)]; cannam@95: Tb = R0[WS(rs, 56)]; cannam@95: Tc = R0[WS(rs, 24)]; cannam@95: T4 = R0[WS(rs, 16)]; cannam@95: T23 = T1 - T2; cannam@95: T3 = T1 + T2; cannam@95: T25 = Tb - Tc; cannam@95: Td = Tb + Tc; cannam@95: T5 = R0[WS(rs, 48)]; cannam@95: T8 = R0[WS(rs, 8)]; cannam@95: T9 = R0[WS(rs, 40)]; cannam@95: } cannam@95: { cannam@95: E Tq, Tr, Tj, Tk; cannam@95: { cannam@95: E Tg, T5N, T6, T24, Ta, Th; cannam@95: Tg = R0[WS(rs, 4)]; cannam@95: T5N = T4 - T5; cannam@95: T6 = T4 + T5; cannam@95: T24 = T8 - T9; cannam@95: Ta = T8 + T9; cannam@95: Th = R0[WS(rs, 36)]; cannam@95: { cannam@95: E T7, T26, T5O, Te; cannam@95: TcD = T3 - T6; cannam@95: T7 = T3 + T6; cannam@95: T26 = T24 + T25; cannam@95: T5O = T25 - T24; cannam@95: TdR = Td - Ta; cannam@95: Te = Ta + Td; cannam@95: T5P = FNMS(KP707106781, T5O, T5N); cannam@95: T8v = FMA(KP707106781, T5O, T5N); cannam@95: T27 = FMA(KP707106781, T26, T23); cannam@95: T7r = FNMS(KP707106781, T26, T23); cannam@95: Tf = T7 + Te; cannam@95: Ta5 = T7 - Te; cannam@95: T28 = Tg - Th; cannam@95: Ti = Tg + Th; cannam@95: } cannam@95: } cannam@95: Tq = R0[WS(rs, 12)]; cannam@95: Tr = R0[WS(rs, 44)]; cannam@95: Tj = R0[WS(rs, 20)]; cannam@95: Tk = R0[WS(rs, 52)]; cannam@95: Tn = R0[WS(rs, 60)]; cannam@95: T2c = Tq - Tr; cannam@95: Ts = Tq + Tr; cannam@95: T29 = Tj - Tk; cannam@95: Tl = Tj + Tk; cannam@95: To = R0[WS(rs, 28)]; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T2g, T2l, T2h, TF, TcI, TC, T2i, TI; cannam@95: { cannam@95: E Ty, TG, TB, TH; cannam@95: { cannam@95: E Tw, T5Q, T2a, TcE, Tm, T2b, Tp, Tx; cannam@95: Tw = R0[WS(rs, 2)]; cannam@95: T5Q = FMA(KP414213562, T28, T29); cannam@95: T2a = FNMS(KP414213562, T29, T28); cannam@95: TcE = Ti - Tl; cannam@95: Tm = Ti + Tl; cannam@95: T2b = Tn - To; cannam@95: Tp = Tn + To; cannam@95: Tx = R0[WS(rs, 34)]; cannam@95: { cannam@95: E Tz, TA, TD, TE; cannam@95: Tz = R0[WS(rs, 18)]; cannam@95: { cannam@95: E T5R, T2d, TcF, Tt; cannam@95: T5R = FNMS(KP414213562, T2b, T2c); cannam@95: T2d = FMA(KP414213562, T2c, T2b); cannam@95: TcF = Tp - Ts; cannam@95: Tt = Tp + Ts; cannam@95: T2g = Tw - Tx; cannam@95: Ty = Tw + Tx; cannam@95: T7s = T5Q - T5R; cannam@95: T5S = T5Q + T5R; cannam@95: T8w = T2d - T2a; cannam@95: T2e = T2a + T2d; cannam@95: TdS = TcF - TcE; cannam@95: TcG = TcE + TcF; cannam@95: Tbn = Tt - Tm; cannam@95: Tu = Tm + Tt; cannam@95: TA = R0[WS(rs, 50)]; cannam@95: } cannam@95: TD = R0[WS(rs, 10)]; cannam@95: TE = R0[WS(rs, 42)]; cannam@95: TG = R0[WS(rs, 58)]; cannam@95: T2l = Tz - TA; cannam@95: TB = Tz + TA; cannam@95: T2h = TD - TE; cannam@95: TF = TD + TE; cannam@95: TH = R0[WS(rs, 26)]; cannam@95: } cannam@95: } cannam@95: TcI = Ty - TB; cannam@95: TC = Ty + TB; cannam@95: T2i = TG - TH; cannam@95: TI = TG + TH; cannam@95: } cannam@95: { cannam@95: E T2p, T2u, T2q, TU, TcL, TR, T2r, TX; cannam@95: { cannam@95: E TN, TV, TQ, TW; cannam@95: { cannam@95: E T2k, T7u, T2n, T7v, TL, TM; cannam@95: TL = R0[WS(rs, 62)]; cannam@95: TM = R0[WS(rs, 30)]; cannam@95: { cannam@95: E TJ, TcJ, T2m, T2j; cannam@95: TJ = TF + TI; cannam@95: TcJ = TI - TF; cannam@95: T2m = T2h - T2i; cannam@95: T2j = T2h + T2i; cannam@95: TcK = FMA(KP414213562, TcJ, TcI); cannam@95: TdU = FNMS(KP414213562, TcI, TcJ); cannam@95: TK = TC + TJ; cannam@95: Ta6 = TC - TJ; cannam@95: T2k = FMA(KP707106781, T2j, T2g); cannam@95: T7u = FNMS(KP707106781, T2j, T2g); cannam@95: T2n = FMA(KP707106781, T2m, T2l); cannam@95: T7v = FNMS(KP707106781, T2m, T2l); cannam@95: T2p = TL - TM; cannam@95: TN = TL + TM; cannam@95: } cannam@95: T7w = FMA(KP668178637, T7v, T7u); cannam@95: T8y = FNMS(KP668178637, T7u, T7v); cannam@95: T2o = FNMS(KP198912367, T2n, T2k); cannam@95: T5U = FMA(KP198912367, T2k, T2n); cannam@95: { cannam@95: E TO, TP, TS, TT; cannam@95: TO = R0[WS(rs, 14)]; cannam@95: TP = R0[WS(rs, 46)]; cannam@95: TS = R0[WS(rs, 6)]; cannam@95: TT = R0[WS(rs, 38)]; cannam@95: TV = R0[WS(rs, 54)]; cannam@95: T2u = TO - TP; cannam@95: TQ = TO + TP; cannam@95: T2q = TS - TT; cannam@95: TU = TS + TT; cannam@95: TW = R0[WS(rs, 22)]; cannam@95: } cannam@95: } cannam@95: TcL = TN - TQ; cannam@95: TR = TN + TQ; cannam@95: T2r = TV - TW; cannam@95: TX = TV + TW; cannam@95: } cannam@95: { cannam@95: E T2A, T14, T2N, T17, T1b, T1e, T2D, T2O, T1r, T2I, T1q, T2Q, T2H, TcR, T1n; cannam@95: E T1s, T15, T16; cannam@95: { cannam@95: E T2t, T7x, T2w, T7y, T12, T13; cannam@95: T12 = R0[WS(rs, 1)]; cannam@95: T13 = R0[WS(rs, 33)]; cannam@95: { cannam@95: E TY, TcM, T2v, T2s; cannam@95: TY = TU + TX; cannam@95: TcM = TX - TU; cannam@95: T2v = T2q - T2r; cannam@95: T2s = T2q + T2r; cannam@95: TcN = FNMS(KP414213562, TcM, TcL); cannam@95: TdV = FMA(KP414213562, TcL, TcM); cannam@95: TZ = TR + TY; cannam@95: Ta7 = TR - TY; cannam@95: T2t = FMA(KP707106781, T2s, T2p); cannam@95: T7x = FNMS(KP707106781, T2s, T2p); cannam@95: T2w = FMA(KP707106781, T2v, T2u); cannam@95: T7y = FNMS(KP707106781, T2v, T2u); cannam@95: T2A = T12 - T13; cannam@95: T14 = T12 + T13; cannam@95: } cannam@95: T7z = FNMS(KP668178637, T7y, T7x); cannam@95: T8z = FMA(KP668178637, T7x, T7y); cannam@95: T2x = FMA(KP198912367, T2w, T2t); cannam@95: T5V = FNMS(KP198912367, T2t, T2w); cannam@95: T15 = R0[WS(rs, 17)]; cannam@95: T16 = R0[WS(rs, 49)]; cannam@95: } cannam@95: { cannam@95: E T1c, T2B, T1d, T19, T1a; cannam@95: T19 = R0[WS(rs, 9)]; cannam@95: T1a = R0[WS(rs, 41)]; cannam@95: T1c = R0[WS(rs, 57)]; cannam@95: T2N = T15 - T16; cannam@95: T17 = T15 + T16; cannam@95: T2B = T19 - T1a; cannam@95: T1b = T19 + T1a; cannam@95: T1d = R0[WS(rs, 25)]; cannam@95: { cannam@95: E T1k, T2F, T1j, T1l, T1h, T1i, T2C; cannam@95: T1h = R0[WS(rs, 5)]; cannam@95: T1i = R0[WS(rs, 37)]; cannam@95: T2C = T1c - T1d; cannam@95: T1e = T1c + T1d; cannam@95: T1k = R0[WS(rs, 21)]; cannam@95: T2F = T1h - T1i; cannam@95: T1j = T1h + T1i; cannam@95: T2D = T2B + T2C; cannam@95: T2O = T2B - T2C; cannam@95: T1l = R0[WS(rs, 53)]; cannam@95: { cannam@95: E T1o, T1p, T2G, T1m; cannam@95: T1o = R0[WS(rs, 61)]; cannam@95: T1p = R0[WS(rs, 29)]; cannam@95: T1r = R0[WS(rs, 13)]; cannam@95: T2G = T1k - T1l; cannam@95: T1m = T1k + T1l; cannam@95: T2I = T1o - T1p; cannam@95: T1q = T1o + T1p; cannam@95: T2Q = FMA(KP414213562, T2F, T2G); cannam@95: T2H = FNMS(KP414213562, T2G, T2F); cannam@95: TcR = T1j - T1m; cannam@95: T1n = T1j + T1m; cannam@95: T1s = R0[WS(rs, 45)]; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E TcQ, TcV, T2K, T2R, T1u, TcT, TcW, TcS; cannam@95: { cannam@95: E T18, T1f, T1t, T2J; cannam@95: T18 = T14 + T17; cannam@95: TcQ = T14 - T17; cannam@95: TcV = T1e - T1b; cannam@95: T1f = T1b + T1e; cannam@95: T1t = T1r + T1s; cannam@95: T2J = T1r - T1s; cannam@95: T1g = T18 + T1f; cannam@95: Taa = T18 - T1f; cannam@95: T2K = FMA(KP414213562, T2J, T2I); cannam@95: T2R = FNMS(KP414213562, T2I, T2J); cannam@95: T1u = T1q + T1t; cannam@95: TcS = T1q - T1t; cannam@95: } cannam@95: TcT = TcR + TcS; cannam@95: TcW = TcS - TcR; cannam@95: { cannam@95: E T7C, T2E, T2L, T7D; cannam@95: T7C = FNMS(KP707106781, T2D, T2A); cannam@95: T2E = FMA(KP707106781, T2D, T2A); cannam@95: Tab = T1u - T1n; cannam@95: T1v = T1n + T1u; cannam@95: Tew = FNMS(KP707106781, TcW, TcV); cannam@95: TcX = FMA(KP707106781, TcW, TcV); cannam@95: Tex = FNMS(KP707106781, TcT, TcQ); cannam@95: TcU = FMA(KP707106781, TcT, TcQ); cannam@95: T2L = T2H + T2K; cannam@95: T7G = T2K - T2H; cannam@95: T7D = T2Q - T2R; cannam@95: T2S = T2Q + T2R; cannam@95: T2P = FMA(KP707106781, T2O, T2N); cannam@95: T7F = FNMS(KP707106781, T2O, T2N); cannam@95: T6A = FNMS(KP923879532, T2L, T2E); cannam@95: T2M = FMA(KP923879532, T2L, T2E); cannam@95: T9b = FNMS(KP923879532, T7D, T7C); cannam@95: T7E = FMA(KP923879532, T7D, T7C); cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T83, T84, T8m, T8n; cannam@95: { cannam@95: E TaP, T4z, TaQ, T5A, TaS, TaT, T4G, T5B, T4O, T5D, Tbh, Tdw, T4R, Tbc, T4S; cannam@95: E T4T; cannam@95: { cannam@95: E T4x, T4y, T5y, T5z; cannam@95: T4x = R1[WS(rs, 63)]; cannam@95: T9a = FNMS(KP923879532, T7G, T7F); cannam@95: T7H = FMA(KP923879532, T7G, T7F); cannam@95: T6z = FNMS(KP923879532, T2S, T2P); cannam@95: T2T = FMA(KP923879532, T2S, T2P); cannam@95: T4y = R1[WS(rs, 31)]; cannam@95: T5y = R1[WS(rs, 47)]; cannam@95: T5z = R1[WS(rs, 15)]; cannam@95: { cannam@95: E T4A, T4B, T4D, T4E; cannam@95: T4A = R1[WS(rs, 7)]; cannam@95: TaP = T4x + T4y; cannam@95: T4z = T4x - T4y; cannam@95: TaQ = T5z + T5y; cannam@95: T5A = T5y - T5z; cannam@95: T4B = R1[WS(rs, 39)]; cannam@95: T4D = R1[WS(rs, 55)]; cannam@95: T4E = R1[WS(rs, 23)]; cannam@95: { cannam@95: E T4K, Tbf, Tbg, T4N, T4P, T4Q; cannam@95: { cannam@95: E T4I, T4C, T4F, T4J, T4L, T4M; cannam@95: T4I = R1[WS(rs, 3)]; cannam@95: TaS = T4A + T4B; cannam@95: T4C = T4A - T4B; cannam@95: TaT = T4D + T4E; cannam@95: T4F = T4D - T4E; cannam@95: T4J = R1[WS(rs, 35)]; cannam@95: T4L = R1[WS(rs, 51)]; cannam@95: T4M = R1[WS(rs, 19)]; cannam@95: T4G = T4C + T4F; cannam@95: T5B = T4F - T4C; cannam@95: T4K = T4I - T4J; cannam@95: Tbf = T4I + T4J; cannam@95: Tbg = T4M + T4L; cannam@95: T4N = T4L - T4M; cannam@95: } cannam@95: T4P = R1[WS(rs, 59)]; cannam@95: T4Q = R1[WS(rs, 27)]; cannam@95: T4O = FMA(KP414213562, T4N, T4K); cannam@95: T5D = FNMS(KP414213562, T4K, T4N); cannam@95: Tbh = Tbf + Tbg; cannam@95: Tdw = Tbf - Tbg; cannam@95: T4R = T4P - T4Q; cannam@95: Tbc = T4P + T4Q; cannam@95: T4S = R1[WS(rs, 43)]; cannam@95: T4T = R1[WS(rs, 11)]; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T4H, T8b, TaR, Tdv, TdI, TaU, T4U, Tbd, T5C; cannam@95: T4H = FMA(KP707106781, T4G, T4z); cannam@95: T8b = FNMS(KP707106781, T4G, T4z); cannam@95: TaR = TaP + TaQ; cannam@95: Tdv = TaP - TaQ; cannam@95: TdI = TaT - TaS; cannam@95: TaU = TaS + TaT; cannam@95: T4U = T4S - T4T; cannam@95: Tbd = T4T + T4S; cannam@95: T8m = FNMS(KP707106781, T5B, T5A); cannam@95: T5C = FMA(KP707106781, T5B, T5A); cannam@95: { cannam@95: E Tbe, Tdx, T5E, T4V; cannam@95: Tbe = Tbc + Tbd; cannam@95: Tdx = Tbc - Tbd; cannam@95: T5E = FMA(KP414213562, T4R, T4U); cannam@95: T4V = FNMS(KP414213562, T4U, T4R); cannam@95: { cannam@95: E Tdy, TdJ, T5F, T8c, T4W; cannam@95: Tdy = Tdw + Tdx; cannam@95: TdJ = Tdx - Tdw; cannam@95: T5F = T5D + T5E; cannam@95: T8c = T5E - T5D; cannam@95: T8n = T4V - T4O; cannam@95: T4W = T4O + T4V; cannam@95: TeO = FNMS(KP707106781, TdJ, TdI); cannam@95: TdK = FMA(KP707106781, TdJ, TdI); cannam@95: TeL = FNMS(KP707106781, Tdy, Tdv); cannam@95: Tdz = FMA(KP707106781, Tdy, Tdv); cannam@95: T9p = FNMS(KP923879532, T8c, T8b); cannam@95: T8d = FMA(KP923879532, T8c, T8b); cannam@95: T6O = FNMS(KP923879532, T5F, T5C); cannam@95: T5G = FMA(KP923879532, T5F, T5C); cannam@95: T6L = FNMS(KP923879532, T4W, T4H); cannam@95: T4X = FMA(KP923879532, T4W, T4H); cannam@95: } cannam@95: Tc3 = TaR + TaU; cannam@95: TaV = TaR - TaU; cannam@95: Tc4 = Tbh + Tbe; cannam@95: Tbi = Tbe - Tbh; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tai, T3k, Taj, T4l, Tal, Tam, T4m, T3r, T3D, TaF, T3C, Tdb, TaK, T3z, T4o; cannam@95: E T3E; cannam@95: { cannam@95: E T4j, T4k, T3i, T3j; cannam@95: T3i = R1[0]; cannam@95: T3j = R1[WS(rs, 32)]; cannam@95: T4j = R1[WS(rs, 16)]; cannam@95: T9s = FMA(KP923879532, T8n, T8m); cannam@95: T8o = FNMS(KP923879532, T8n, T8m); cannam@95: Tai = T3i + T3j; cannam@95: T3k = T3i - T3j; cannam@95: T4k = R1[WS(rs, 48)]; cannam@95: { cannam@95: E T3o, T3n, T3p, T3l, T3m; cannam@95: T3l = R1[WS(rs, 8)]; cannam@95: T3m = R1[WS(rs, 40)]; cannam@95: T3o = R1[WS(rs, 56)]; cannam@95: Taj = T4j + T4k; cannam@95: T4l = T4j - T4k; cannam@95: T3n = T3l - T3m; cannam@95: Tal = T3l + T3m; cannam@95: T3p = R1[WS(rs, 24)]; cannam@95: { cannam@95: E T3w, TaI, T3v, T3x, T3t, T3u, T3q; cannam@95: T3t = R1[WS(rs, 4)]; cannam@95: T3u = R1[WS(rs, 36)]; cannam@95: T3q = T3o - T3p; cannam@95: Tam = T3o + T3p; cannam@95: T3w = R1[WS(rs, 20)]; cannam@95: TaI = T3t + T3u; cannam@95: T3v = T3t - T3u; cannam@95: T4m = T3n - T3q; cannam@95: T3r = T3n + T3q; cannam@95: T3x = R1[WS(rs, 52)]; cannam@95: { cannam@95: E T3A, T3B, TaJ, T3y; cannam@95: T3A = R1[WS(rs, 60)]; cannam@95: T3B = R1[WS(rs, 28)]; cannam@95: T3D = R1[WS(rs, 12)]; cannam@95: TaJ = T3w + T3x; cannam@95: T3y = T3w - T3x; cannam@95: TaF = T3A + T3B; cannam@95: T3C = T3A - T3B; cannam@95: Tdb = TaI - TaJ; cannam@95: TaK = TaI + TaJ; cannam@95: T3z = FNMS(KP414213562, T3y, T3v); cannam@95: T4o = FMA(KP414213562, T3v, T3y); cannam@95: T3E = R1[WS(rs, 44)]; cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T3s, T7S, Tak, Tda, Tdn, Tan, T3F, TaG, T4n; cannam@95: T3s = FMA(KP707106781, T3r, T3k); cannam@95: T7S = FNMS(KP707106781, T3r, T3k); cannam@95: Tak = Tai + Taj; cannam@95: Tda = Tai - Taj; cannam@95: Tdn = Tam - Tal; cannam@95: Tan = Tal + Tam; cannam@95: T3F = T3D - T3E; cannam@95: TaG = T3D + T3E; cannam@95: T83 = FNMS(KP707106781, T4m, T4l); cannam@95: T4n = FMA(KP707106781, T4m, T4l); cannam@95: { cannam@95: E TaH, Tdc, T4p, T3G; cannam@95: TaH = TaF + TaG; cannam@95: Tdc = TaF - TaG; cannam@95: T4p = FNMS(KP414213562, T3C, T3F); cannam@95: T3G = FMA(KP414213562, T3F, T3C); cannam@95: { cannam@95: E Tdd, Tdo, T4q, T7T, T3H; cannam@95: Tdd = Tdb + Tdc; cannam@95: Tdo = Tdc - Tdb; cannam@95: T4q = T4o + T4p; cannam@95: T7T = T4o - T4p; cannam@95: T84 = T3G - T3z; cannam@95: T3H = T3z + T3G; cannam@95: TeH = FNMS(KP707106781, Tdo, Tdn); cannam@95: Tdp = FMA(KP707106781, Tdo, Tdn); cannam@95: TeE = FNMS(KP707106781, Tdd, Tda); cannam@95: Tde = FMA(KP707106781, Tdd, Tda); cannam@95: T9i = FNMS(KP923879532, T7T, T7S); cannam@95: T7U = FMA(KP923879532, T7T, T7S); cannam@95: T6H = FNMS(KP923879532, T4q, T4n); cannam@95: T4r = FMA(KP923879532, T4q, T4n); cannam@95: T6E = FNMS(KP923879532, T3H, T3s); cannam@95: T3I = FMA(KP923879532, T3H, T3s); cannam@95: } cannam@95: TbW = Tak + Tan; cannam@95: Tao = Tak - Tan; cannam@95: TbX = TaK + TaH; cannam@95: TaL = TaH - TaK; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T7N, T3d, T3a, T7M; cannam@95: { cannam@95: E T2V, T1z, T38, T1C, T1G, T1J, T2Y, T39, T1W, T33, T1V, T3b, T32, Td0, T1S; cannam@95: E T1X; cannam@95: { cannam@95: E T1A, T1B, T1x, T1y; cannam@95: T1x = R0[WS(rs, 63)]; cannam@95: T1y = R0[WS(rs, 31)]; cannam@95: T1A = R0[WS(rs, 15)]; cannam@95: T9l = FNMS(KP923879532, T84, T83); cannam@95: T85 = FMA(KP923879532, T84, T83); cannam@95: T2V = T1x - T1y; cannam@95: T1z = T1x + T1y; cannam@95: T1B = R0[WS(rs, 47)]; cannam@95: { cannam@95: E T1H, T2W, T1I, T1E, T1F; cannam@95: T1E = R0[WS(rs, 7)]; cannam@95: T1F = R0[WS(rs, 39)]; cannam@95: T1H = R0[WS(rs, 55)]; cannam@95: T38 = T1A - T1B; cannam@95: T1C = T1A + T1B; cannam@95: T2W = T1E - T1F; cannam@95: T1G = T1E + T1F; cannam@95: T1I = R0[WS(rs, 23)]; cannam@95: { cannam@95: E T1P, T30, T1O, T1Q, T1M, T1N, T2X; cannam@95: T1M = R0[WS(rs, 3)]; cannam@95: T1N = R0[WS(rs, 35)]; cannam@95: T2X = T1H - T1I; cannam@95: T1J = T1H + T1I; cannam@95: T1P = R0[WS(rs, 19)]; cannam@95: T30 = T1M - T1N; cannam@95: T1O = T1M + T1N; cannam@95: T2Y = T2W + T2X; cannam@95: T39 = T2W - T2X; cannam@95: T1Q = R0[WS(rs, 51)]; cannam@95: { cannam@95: E T1T, T1U, T31, T1R; cannam@95: T1T = R0[WS(rs, 59)]; cannam@95: T1U = R0[WS(rs, 27)]; cannam@95: T1W = R0[WS(rs, 11)]; cannam@95: T31 = T1P - T1Q; cannam@95: T1R = T1P + T1Q; cannam@95: T33 = T1T - T1U; cannam@95: T1V = T1T + T1U; cannam@95: T3b = FMA(KP414213562, T30, T31); cannam@95: T32 = FNMS(KP414213562, T31, T30); cannam@95: Td0 = T1O - T1R; cannam@95: T1S = T1O + T1R; cannam@95: T1X = R0[WS(rs, 43)]; cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E TcZ, Td4, T35, T3c, T1Z, Td2, Td5, Td1; cannam@95: { cannam@95: E T1D, T1K, T1Y, T34; cannam@95: T1D = T1z + T1C; cannam@95: TcZ = T1z - T1C; cannam@95: Td4 = T1J - T1G; cannam@95: T1K = T1G + T1J; cannam@95: T1Y = T1W + T1X; cannam@95: T34 = T1W - T1X; cannam@95: T1L = T1D + T1K; cannam@95: Tad = T1D - T1K; cannam@95: T35 = FMA(KP414213562, T34, T33); cannam@95: T3c = FNMS(KP414213562, T33, T34); cannam@95: T1Z = T1V + T1Y; cannam@95: Td1 = T1V - T1Y; cannam@95: } cannam@95: Td2 = Td0 + Td1; cannam@95: Td5 = Td1 - Td0; cannam@95: { cannam@95: E T7J, T2Z, T36, T7K; cannam@95: T7J = FNMS(KP707106781, T2Y, T2V); cannam@95: T2Z = FMA(KP707106781, T2Y, T2V); cannam@95: Tae = T1Z - T1S; cannam@95: T20 = T1S + T1Z; cannam@95: Tez = FNMS(KP707106781, Td5, Td4); cannam@95: Td6 = FMA(KP707106781, Td5, Td4); cannam@95: TeA = FNMS(KP707106781, Td2, TcZ); cannam@95: Td3 = FMA(KP707106781, Td2, TcZ); cannam@95: T36 = T32 + T35; cannam@95: T7N = T35 - T32; cannam@95: T7K = T3b - T3c; cannam@95: T3d = T3b + T3c; cannam@95: T3a = FMA(KP707106781, T39, T38); cannam@95: T7M = FNMS(KP707106781, T39, T38); cannam@95: T6x = FNMS(KP923879532, T36, T2Z); cannam@95: T37 = FMA(KP923879532, T36, T2Z); cannam@95: T9e = FNMS(KP923879532, T7K, T7J); cannam@95: T7L = FMA(KP923879532, T7K, T7J); cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tav, T7V, T7W, TaC; cannam@95: { cannam@95: E T3L, T3W, Tdf, Tar, T42, T4d, Tay, Tdi, T46, Tau, Tdg, T3X, T3S, Taz, T45; cannam@95: E T47, Taw, Tax; cannam@95: { cannam@95: E T3J, T3K, T3U, T3V; cannam@95: T3J = R1[WS(rs, 2)]; cannam@95: T9d = FNMS(KP923879532, T7N, T7M); cannam@95: T7O = FMA(KP923879532, T7N, T7M); cannam@95: T6w = FNMS(KP923879532, T3d, T3a); cannam@95: T3e = FMA(KP923879532, T3d, T3a); cannam@95: T3K = R1[WS(rs, 34)]; cannam@95: T3U = R1[WS(rs, 18)]; cannam@95: T3V = R1[WS(rs, 50)]; cannam@95: { cannam@95: E T40, Tap, Taq, T41, T4b, T4c; cannam@95: T40 = R1[WS(rs, 62)]; cannam@95: T3L = T3J - T3K; cannam@95: Tap = T3J + T3K; cannam@95: T3W = T3U - T3V; cannam@95: Taq = T3U + T3V; cannam@95: T41 = R1[WS(rs, 30)]; cannam@95: T4b = R1[WS(rs, 14)]; cannam@95: T4c = R1[WS(rs, 46)]; cannam@95: Tdf = Tap - Taq; cannam@95: Tar = Tap + Taq; cannam@95: T42 = T40 - T41; cannam@95: Taw = T40 + T41; cannam@95: Tax = T4b + T4c; cannam@95: T4d = T4b - T4c; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T3M, T3N, T3P, T3Q; cannam@95: T3M = R1[WS(rs, 10)]; cannam@95: Tay = Taw + Tax; cannam@95: Tdi = Taw - Tax; cannam@95: T3N = R1[WS(rs, 42)]; cannam@95: T3P = R1[WS(rs, 58)]; cannam@95: T3Q = R1[WS(rs, 26)]; cannam@95: { cannam@95: E T43, Tas, T3O, Tat, T3R, T44; cannam@95: T43 = R1[WS(rs, 6)]; cannam@95: Tas = T3M + T3N; cannam@95: T3O = T3M - T3N; cannam@95: Tat = T3P + T3Q; cannam@95: T3R = T3P - T3Q; cannam@95: T44 = R1[WS(rs, 38)]; cannam@95: T46 = R1[WS(rs, 54)]; cannam@95: Tau = Tas + Tat; cannam@95: Tdg = Tat - Tas; cannam@95: T3X = T3O - T3R; cannam@95: T3S = T3O + T3R; cannam@95: Taz = T43 + T44; cannam@95: T45 = T43 - T44; cannam@95: T47 = R1[WS(rs, 22)]; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tdq, Tdh, T49, T4e, Tdr, Tdk; cannam@95: Tav = Tar - Tau; cannam@95: TbZ = Tar + Tau; cannam@95: { cannam@95: E T3T, T3Y, TaA, T48, Tdj, TaB; cannam@95: T3T = FMA(KP707106781, T3S, T3L); cannam@95: T7V = FNMS(KP707106781, T3S, T3L); cannam@95: T7W = FNMS(KP707106781, T3X, T3W); cannam@95: T3Y = FMA(KP707106781, T3X, T3W); cannam@95: TaA = T46 + T47; cannam@95: T48 = T46 - T47; cannam@95: Tdq = FNMS(KP414213562, Tdf, Tdg); cannam@95: Tdh = FMA(KP414213562, Tdg, Tdf); cannam@95: T3Z = FNMS(KP198912367, T3Y, T3T); cannam@95: T4s = FMA(KP198912367, T3T, T3Y); cannam@95: Tdj = TaA - Taz; cannam@95: TaB = Taz + TaA; cannam@95: T49 = T45 + T48; cannam@95: T4e = T45 - T48; cannam@95: TaC = Tay - TaB; cannam@95: Tc0 = Tay + TaB; cannam@95: Tdr = FMA(KP414213562, Tdi, Tdj); cannam@95: Tdk = FNMS(KP414213562, Tdj, Tdi); cannam@95: } cannam@95: { cannam@95: E T7Z, T7Y, T4f, T4a; cannam@95: T7Z = FNMS(KP707106781, T4e, T4d); cannam@95: T4f = FMA(KP707106781, T4e, T4d); cannam@95: T4a = FMA(KP707106781, T49, T42); cannam@95: T7Y = FNMS(KP707106781, T49, T42); cannam@95: TeF = Tdr - Tdq; cannam@95: Tds = Tdq + Tdr; cannam@95: T4t = FNMS(KP198912367, T4a, T4f); cannam@95: T4g = FMA(KP198912367, T4f, T4a); cannam@95: T87 = FMA(KP668178637, T7Y, T7Z); cannam@95: T80 = FNMS(KP668178637, T7Z, T7Y); cannam@95: TeI = Tdh - Tdk; cannam@95: Tdl = Tdh + Tdk; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T50, T5b, TdA, TaY, T5h, T5s, Tb5, TdD, T5l, Tb1, TdB, T5c, T57, Tb6, T5k; cannam@95: E T5m, Tb3, Tb4; cannam@95: { cannam@95: E T4Y, T4Z, T59, T5a; cannam@95: T4Y = R1[WS(rs, 1)]; cannam@95: T86 = FNMS(KP668178637, T7V, T7W); cannam@95: T7X = FMA(KP668178637, T7W, T7V); cannam@95: TaM = TaC - Tav; cannam@95: TaD = Tav + TaC; cannam@95: T4Z = R1[WS(rs, 33)]; cannam@95: T59 = R1[WS(rs, 49)]; cannam@95: T5a = R1[WS(rs, 17)]; cannam@95: { cannam@95: E T5f, TaW, TaX, T5g, T5q, T5r; cannam@95: T5f = R1[WS(rs, 61)]; cannam@95: T50 = T4Y - T4Z; cannam@95: TaW = T4Y + T4Z; cannam@95: T5b = T59 - T5a; cannam@95: TaX = T5a + T59; cannam@95: T5g = R1[WS(rs, 29)]; cannam@95: T5q = R1[WS(rs, 45)]; cannam@95: T5r = R1[WS(rs, 13)]; cannam@95: TdA = TaW - TaX; cannam@95: TaY = TaW + TaX; cannam@95: T5h = T5f - T5g; cannam@95: Tb3 = T5f + T5g; cannam@95: Tb4 = T5r + T5q; cannam@95: T5s = T5q - T5r; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T51, T52, T54, T55; cannam@95: T51 = R1[WS(rs, 9)]; cannam@95: Tb5 = Tb3 + Tb4; cannam@95: TdD = Tb3 - Tb4; cannam@95: T52 = R1[WS(rs, 41)]; cannam@95: T54 = R1[WS(rs, 57)]; cannam@95: T55 = R1[WS(rs, 25)]; cannam@95: { cannam@95: E T5i, TaZ, T53, Tb0, T56, T5j; cannam@95: T5i = R1[WS(rs, 5)]; cannam@95: TaZ = T51 + T52; cannam@95: T53 = T51 - T52; cannam@95: Tb0 = T54 + T55; cannam@95: T56 = T54 - T55; cannam@95: T5j = R1[WS(rs, 37)]; cannam@95: T5l = R1[WS(rs, 53)]; cannam@95: Tb1 = TaZ + Tb0; cannam@95: TdB = Tb0 - TaZ; cannam@95: T5c = T56 - T53; cannam@95: T57 = T53 + T56; cannam@95: Tb6 = T5i + T5j; cannam@95: T5k = T5i - T5j; cannam@95: T5m = R1[WS(rs, 21)]; cannam@95: } cannam@95: } cannam@95: { cannam@95: E TdL, TdC, T5o, T5t, TdM, TdF; cannam@95: Tb2 = TaY - Tb1; cannam@95: Tc6 = TaY + Tb1; cannam@95: { cannam@95: E T58, T5d, Tb7, T5n, TdE, Tb8; cannam@95: T58 = FMA(KP707106781, T57, T50); cannam@95: T8e = FNMS(KP707106781, T57, T50); cannam@95: T8f = FNMS(KP707106781, T5c, T5b); cannam@95: T5d = FMA(KP707106781, T5c, T5b); cannam@95: Tb7 = T5l + T5m; cannam@95: T5n = T5l - T5m; cannam@95: TdL = FNMS(KP414213562, TdA, TdB); cannam@95: TdC = FMA(KP414213562, TdB, TdA); cannam@95: T5e = FMA(KP198912367, T5d, T58); cannam@95: T5H = FNMS(KP198912367, T58, T5d); cannam@95: TdE = Tb7 - Tb6; cannam@95: Tb8 = Tb6 + Tb7; cannam@95: T5o = T5k + T5n; cannam@95: T5t = T5n - T5k; cannam@95: Tb9 = Tb5 - Tb8; cannam@95: Tc7 = Tb5 + Tb8; cannam@95: TdM = FMA(KP414213562, TdD, TdE); cannam@95: TdF = FNMS(KP414213562, TdE, TdD); cannam@95: } cannam@95: { cannam@95: E T8i, T8h, T5u, T5p; cannam@95: T8i = FNMS(KP707106781, T5t, T5s); cannam@95: T5u = FMA(KP707106781, T5t, T5s); cannam@95: T5p = FMA(KP707106781, T5o, T5h); cannam@95: T8h = FNMS(KP707106781, T5o, T5h); cannam@95: TeM = TdM - TdL; cannam@95: TdN = TdL + TdM; cannam@95: T5I = FMA(KP198912367, T5p, T5u); cannam@95: T5v = FNMS(KP198912367, T5u, T5p); cannam@95: T8q = FNMS(KP668178637, T8h, T8i); cannam@95: T8j = FMA(KP668178637, T8i, T8h); cannam@95: TeP = TdF - TdC; cannam@95: TdG = TdC + TdF; cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T8p, T8g, TcH, TdW, TdT, TcO, Tfp, Tfk, Tfj, Tfq; cannam@95: { cannam@95: E Tbj, Tba, Tcy, Tco, TcB, Tcl, Tcx, Tcv, Tcz, Tcr; cannam@95: { cannam@95: E Tch, Tct, Tcp, Tcq, Tci, T1w, TbV, T11, Tcf, Tc9, T21, Tcj, Tcm, TbY, Tc1; cannam@95: E Tcn, Tcu, Tck; cannam@95: { cannam@95: E Tv, T10, Tc5, Tc8; cannam@95: Tch = Tf - Tu; cannam@95: Tv = Tf + Tu; cannam@95: T8p = FMA(KP668178637, T8e, T8f); cannam@95: T8g = FNMS(KP668178637, T8f, T8e); cannam@95: Tbj = Tb9 - Tb2; cannam@95: Tba = Tb2 + Tb9; cannam@95: T10 = TK + TZ; cannam@95: Tct = TZ - TK; cannam@95: Tcp = Tc3 - Tc4; cannam@95: Tc5 = Tc3 + Tc4; cannam@95: Tc8 = Tc6 + Tc7; cannam@95: Tcq = Tc7 - Tc6; cannam@95: Tci = T1g - T1v; cannam@95: T1w = T1g + T1v; cannam@95: TbV = Tv - T10; cannam@95: T11 = Tv + T10; cannam@95: Tcf = Tc5 + Tc8; cannam@95: Tc9 = Tc5 - Tc8; cannam@95: T21 = T1L + T20; cannam@95: Tcj = T1L - T20; cannam@95: Tcm = TbW - TbX; cannam@95: TbY = TbW + TbX; cannam@95: Tc1 = TbZ + Tc0; cannam@95: Tcn = Tc0 - TbZ; cannam@95: } cannam@95: { cannam@95: E Tcb, T22, Tce, Tc2; cannam@95: Tcb = T21 - T1w; cannam@95: T22 = T1w + T21; cannam@95: Tce = TbY + Tc1; cannam@95: Tc2 = TbY - Tc1; cannam@95: { cannam@95: E Tcd, Tcg, Tca, Tcc; cannam@95: Tcd = T11 + T22; cannam@95: Cr[WS(csr, 32)] = T11 - T22; cannam@95: Tcg = Tce + Tcf; cannam@95: Ci[WS(csi, 32)] = Tcf - Tce; cannam@95: Tca = Tc2 + Tc9; cannam@95: Tcc = Tc9 - Tc2; cannam@95: Cr[0] = Tcd + Tcg; cannam@95: Cr[WS(csr, 64)] = Tcd - Tcg; cannam@95: Ci[WS(csi, 48)] = FMS(KP707106781, Tcc, Tcb); cannam@95: Ci[WS(csi, 16)] = FMA(KP707106781, Tcc, Tcb); cannam@95: Cr[WS(csr, 16)] = FMA(KP707106781, Tca, TbV); cannam@95: Cr[WS(csr, 48)] = FNMS(KP707106781, Tca, TbV); cannam@95: Tcu = Tcj - Tci; cannam@95: Tck = Tci + Tcj; cannam@95: Tcy = FNMS(KP414213562, Tcm, Tcn); cannam@95: Tco = FMA(KP414213562, Tcn, Tcm); cannam@95: } cannam@95: } cannam@95: TcB = FNMS(KP707106781, Tck, Tch); cannam@95: Tcl = FMA(KP707106781, Tck, Tch); cannam@95: Tcx = FMA(KP707106781, Tcu, Tct); cannam@95: Tcv = FNMS(KP707106781, Tcu, Tct); cannam@95: Tcz = FMA(KP414213562, Tcp, Tcq); cannam@95: Tcr = FNMS(KP414213562, Tcq, Tcp); cannam@95: } cannam@95: { cannam@95: E TbT, TbO, TbN, TbU; cannam@95: { cannam@95: E Ta9, TbB, Tbb, TbL, Tbp, TbM, Tag, Tbk, TbR, TbJ, Tbw, TaO, TbC, Tbs, TbQ; cannam@95: E TbG; cannam@95: { cannam@95: E Tbq, Tbr, TbH, TbI; cannam@95: { cannam@95: E Tbo, Ta8, Tac, Taf; cannam@95: Tbo = Ta7 - Ta6; cannam@95: Ta8 = Ta6 + Ta7; cannam@95: { cannam@95: E TcC, TcA, Tcw, Tcs; cannam@95: TcC = Tcz - Tcy; cannam@95: TcA = Tcy + Tcz; cannam@95: Tcw = Tcr - Tco; cannam@95: Tcs = Tco + Tcr; cannam@95: Cr[WS(csr, 24)] = FMA(KP923879532, TcC, TcB); cannam@95: Cr[WS(csr, 40)] = FNMS(KP923879532, TcC, TcB); cannam@95: Ci[WS(csi, 56)] = FMS(KP923879532, TcA, Tcx); cannam@95: Ci[WS(csi, 8)] = FMA(KP923879532, TcA, Tcx); cannam@95: Ci[WS(csi, 40)] = FMA(KP923879532, Tcw, Tcv); cannam@95: Ci[WS(csi, 24)] = FMS(KP923879532, Tcw, Tcv); cannam@95: Cr[WS(csr, 8)] = FMA(KP923879532, Tcs, Tcl); cannam@95: Cr[WS(csr, 56)] = FNMS(KP923879532, Tcs, Tcl); cannam@95: Ta9 = FMA(KP707106781, Ta8, Ta5); cannam@95: TbB = FNMS(KP707106781, Ta8, Ta5); cannam@95: } cannam@95: Tbq = FNMS(KP414213562, Taa, Tab); cannam@95: Tac = FMA(KP414213562, Tab, Taa); cannam@95: Taf = FNMS(KP414213562, Tae, Tad); cannam@95: Tbr = FMA(KP414213562, Tad, Tae); cannam@95: Tbb = FMA(KP707106781, Tba, TaV); cannam@95: TbH = FNMS(KP707106781, Tba, TaV); cannam@95: TbL = FNMS(KP707106781, Tbo, Tbn); cannam@95: Tbp = FMA(KP707106781, Tbo, Tbn); cannam@95: TbM = Taf - Tac; cannam@95: Tag = Tac + Taf; cannam@95: TbI = FNMS(KP707106781, Tbj, Tbi); cannam@95: Tbk = FMA(KP707106781, Tbj, Tbi); cannam@95: } cannam@95: { cannam@95: E TbE, TbF, TaE, TaN; cannam@95: TbE = FNMS(KP707106781, TaD, Tao); cannam@95: TaE = FMA(KP707106781, TaD, Tao); cannam@95: TaN = FMA(KP707106781, TaM, TaL); cannam@95: TbF = FNMS(KP707106781, TaM, TaL); cannam@95: TbR = FNMS(KP668178637, TbH, TbI); cannam@95: TbJ = FMA(KP668178637, TbI, TbH); cannam@95: Tbw = FNMS(KP198912367, TaE, TaN); cannam@95: TaO = FMA(KP198912367, TaN, TaE); cannam@95: TbC = Tbr - Tbq; cannam@95: Tbs = Tbq + Tbr; cannam@95: TbQ = FMA(KP668178637, TbE, TbF); cannam@95: TbG = FNMS(KP668178637, TbF, TbE); cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tbz, Tah, Tbv, Tbt, Tbx, Tbl; cannam@95: Tbz = FNMS(KP923879532, Tag, Ta9); cannam@95: Tah = FMA(KP923879532, Tag, Ta9); cannam@95: Tbv = FMA(KP923879532, Tbs, Tbp); cannam@95: Tbt = FNMS(KP923879532, Tbs, Tbp); cannam@95: Tbx = FMA(KP198912367, Tbb, Tbk); cannam@95: Tbl = FNMS(KP198912367, Tbk, Tbb); cannam@95: { cannam@95: E TbD, TbK, TbP, TbS; cannam@95: TbT = FNMS(KP923879532, TbC, TbB); cannam@95: TbD = FMA(KP923879532, TbC, TbB); cannam@95: { cannam@95: E TbA, Tby, Tbu, Tbm; cannam@95: TbA = Tbx - Tbw; cannam@95: Tby = Tbw + Tbx; cannam@95: Tbu = Tbl - TaO; cannam@95: Tbm = TaO + Tbl; cannam@95: Cr[WS(csr, 28)] = FMA(KP980785280, TbA, Tbz); cannam@95: Cr[WS(csr, 36)] = FNMS(KP980785280, TbA, Tbz); cannam@95: Ci[WS(csi, 60)] = FMS(KP980785280, Tby, Tbv); cannam@95: Ci[WS(csi, 4)] = FMA(KP980785280, Tby, Tbv); cannam@95: Ci[WS(csi, 36)] = FMA(KP980785280, Tbu, Tbt); cannam@95: Ci[WS(csi, 28)] = FMS(KP980785280, Tbu, Tbt); cannam@95: Cr[WS(csr, 4)] = FMA(KP980785280, Tbm, Tah); cannam@95: Cr[WS(csr, 60)] = FNMS(KP980785280, Tbm, Tah); cannam@95: TbK = TbG + TbJ; cannam@95: TbO = TbJ - TbG; cannam@95: } cannam@95: TbN = FMA(KP923879532, TbM, TbL); cannam@95: TbP = FNMS(KP923879532, TbM, TbL); cannam@95: TbS = TbQ + TbR; cannam@95: TbU = TbQ - TbR; cannam@95: Cr[WS(csr, 12)] = FMA(KP831469612, TbK, TbD); cannam@95: Cr[WS(csr, 52)] = FNMS(KP831469612, TbK, TbD); cannam@95: Ci[WS(csi, 52)] = FNMS(KP831469612, TbS, TbP); cannam@95: Ci[WS(csi, 12)] = -(FMA(KP831469612, TbS, TbP)); cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E TeN, Tf7, Tev, Tfm, Tfc, TeQ, TeX, TeW, Tfn, Tff, Tfi, TeC, Tf2, TeK, Tfh; cannam@95: E TeV, Tf8; cannam@95: { cannam@95: E TeG, TeJ, Tfd, Tfe, Tey, TeB, TeT, TeU; cannam@95: { cannam@95: E Tet, Teu, Tfa, Tfb; cannam@95: TcH = FMA(KP707106781, TcG, TcD); cannam@95: Tet = FNMS(KP707106781, TcG, TcD); cannam@95: Ci[WS(csi, 44)] = FMS(KP831469612, TbO, TbN); cannam@95: Ci[WS(csi, 20)] = FMA(KP831469612, TbO, TbN); cannam@95: Cr[WS(csr, 20)] = FMA(KP831469612, TbU, TbT); cannam@95: Cr[WS(csr, 44)] = FNMS(KP831469612, TbU, TbT); cannam@95: Teu = TdV - TdU; cannam@95: TdW = TdU + TdV; cannam@95: TeG = FNMS(KP923879532, TeF, TeE); cannam@95: Tfa = FMA(KP923879532, TeF, TeE); cannam@95: Tfb = FMA(KP923879532, TeI, TeH); cannam@95: TeJ = FNMS(KP923879532, TeI, TeH); cannam@95: TeN = FNMS(KP923879532, TeM, TeL); cannam@95: Tfd = FMA(KP923879532, TeM, TeL); cannam@95: Tf7 = FMA(KP923879532, Teu, Tet); cannam@95: Tev = FNMS(KP923879532, Teu, Tet); cannam@95: Tfm = FMA(KP303346683, Tfa, Tfb); cannam@95: Tfc = FNMS(KP303346683, Tfb, Tfa); cannam@95: Tfe = FNMS(KP923879532, TeP, TeO); cannam@95: TeQ = FMA(KP923879532, TeP, TeO); cannam@95: TeX = FNMS(KP668178637, Tew, Tex); cannam@95: Tey = FMA(KP668178637, Tex, Tew); cannam@95: TeB = FNMS(KP668178637, TeA, Tez); cannam@95: TeW = FMA(KP668178637, Tez, TeA); cannam@95: } cannam@95: Tfn = FNMS(KP303346683, Tfd, Tfe); cannam@95: Tff = FMA(KP303346683, Tfe, Tfd); cannam@95: Tfi = Tey + TeB; cannam@95: TeC = Tey - TeB; cannam@95: TdT = FMA(KP707106781, TdS, TdR); cannam@95: TeT = FNMS(KP707106781, TdS, TdR); cannam@95: TeU = TcN - TcK; cannam@95: TcO = TcK + TcN; cannam@95: Tf2 = FNMS(KP534511135, TeG, TeJ); cannam@95: TeK = FMA(KP534511135, TeJ, TeG); cannam@95: Tfh = FNMS(KP923879532, TeU, TeT); cannam@95: TeV = FMA(KP923879532, TeU, TeT); cannam@95: } cannam@95: { cannam@95: E Tf5, TeD, TeY, Tf3, TeR; cannam@95: Tf5 = FNMS(KP831469612, TeC, Tev); cannam@95: TeD = FMA(KP831469612, TeC, Tev); cannam@95: Tf8 = TeX + TeW; cannam@95: TeY = TeW - TeX; cannam@95: Tf3 = FMA(KP534511135, TeN, TeQ); cannam@95: TeR = FNMS(KP534511135, TeQ, TeN); cannam@95: { cannam@95: E Tf1, TeZ, Tf6, Tf4, Tf0, TeS; cannam@95: Tf1 = FMA(KP831469612, TeY, TeV); cannam@95: TeZ = FNMS(KP831469612, TeY, TeV); cannam@95: Tf6 = Tf3 - Tf2; cannam@95: Tf4 = Tf2 + Tf3; cannam@95: Tf0 = TeR - TeK; cannam@95: TeS = TeK + TeR; cannam@95: Ci[WS(csi, 54)] = FMS(KP881921264, Tf4, Tf1); cannam@95: Ci[WS(csi, 10)] = FMA(KP881921264, Tf4, Tf1); cannam@95: Ci[WS(csi, 42)] = FMA(KP881921264, Tf0, TeZ); cannam@95: Ci[WS(csi, 22)] = FMS(KP881921264, Tf0, TeZ); cannam@95: Cr[WS(csr, 10)] = FMA(KP881921264, TeS, TeD); cannam@95: Cr[WS(csr, 54)] = FNMS(KP881921264, TeS, TeD); cannam@95: Cr[WS(csr, 42)] = FNMS(KP881921264, Tf6, Tf5); cannam@95: Cr[WS(csr, 22)] = FMA(KP881921264, Tf6, Tf5); cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tf9, Tfg, Tfl, Tfo; cannam@95: Tfp = FNMS(KP831469612, Tf8, Tf7); cannam@95: Tf9 = FMA(KP831469612, Tf8, Tf7); cannam@95: Tfg = Tfc + Tff; cannam@95: Tfk = Tff - Tfc; cannam@95: Tfj = FNMS(KP831469612, Tfi, Tfh); cannam@95: Tfl = FMA(KP831469612, Tfi, Tfh); cannam@95: Tfo = Tfm + Tfn; cannam@95: Tfq = Tfm - Tfn; cannam@95: Cr[WS(csr, 6)] = FMA(KP956940335, Tfg, Tf9); cannam@95: Cr[WS(csr, 58)] = FNMS(KP956940335, Tfg, Tf9); cannam@95: Ci[WS(csi, 58)] = FNMS(KP956940335, Tfo, Tfl); cannam@95: Ci[WS(csi, 6)] = -(FMA(KP956940335, Tfo, Tfl)); cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T2f, T5W, T5T, T2y, T5J, T5w, T4u, T4h, T7p, T7q; cannam@95: { cannam@95: E Ter, Tem, Tel, Tes; cannam@95: { cannam@95: E TdH, Te9, TcP, Teo, Tee, TdO, TdY, TdZ, Tep, Teh, Tek, Td8, Te4, Tdu, Tej; cannam@95: E TdX, Tea; cannam@95: { cannam@95: E Tdm, Tdt, Tef, Teg, TcY, Td7, Tec, Ted; cannam@95: Ci[WS(csi, 38)] = FMS(KP956940335, Tfk, Tfj); cannam@95: Ci[WS(csi, 26)] = FMA(KP956940335, Tfk, Tfj); cannam@95: Cr[WS(csr, 26)] = FMA(KP956940335, Tfq, Tfp); cannam@95: Cr[WS(csr, 38)] = FNMS(KP956940335, Tfq, Tfp); cannam@95: Tdm = FMA(KP923879532, Tdl, Tde); cannam@95: Tec = FNMS(KP923879532, Tdl, Tde); cannam@95: Ted = FNMS(KP923879532, Tds, Tdp); cannam@95: Tdt = FMA(KP923879532, Tds, Tdp); cannam@95: TdH = FMA(KP923879532, TdG, Tdz); cannam@95: Tef = FNMS(KP923879532, TdG, Tdz); cannam@95: Te9 = FNMS(KP923879532, TcO, TcH); cannam@95: TcP = FMA(KP923879532, TcO, TcH); cannam@95: Teo = FMA(KP820678790, Tec, Ted); cannam@95: Tee = FNMS(KP820678790, Ted, Tec); cannam@95: Teg = FNMS(KP923879532, TdN, TdK); cannam@95: TdO = FMA(KP923879532, TdN, TdK); cannam@95: TdY = FNMS(KP198912367, TcU, TcX); cannam@95: TcY = FMA(KP198912367, TcX, TcU); cannam@95: Td7 = FNMS(KP198912367, Td6, Td3); cannam@95: TdZ = FMA(KP198912367, Td3, Td6); cannam@95: Tep = FNMS(KP820678790, Tef, Teg); cannam@95: Teh = FMA(KP820678790, Teg, Tef); cannam@95: Tek = Td7 - TcY; cannam@95: Td8 = TcY + Td7; cannam@95: Te4 = FNMS(KP098491403, Tdm, Tdt); cannam@95: Tdu = FMA(KP098491403, Tdt, Tdm); cannam@95: Tej = FNMS(KP923879532, TdW, TdT); cannam@95: TdX = FMA(KP923879532, TdW, TdT); cannam@95: } cannam@95: { cannam@95: E Te7, Td9, Te0, Te5, TdP; cannam@95: Te7 = FNMS(KP980785280, Td8, TcP); cannam@95: Td9 = FMA(KP980785280, Td8, TcP); cannam@95: Tea = TdZ - TdY; cannam@95: Te0 = TdY + TdZ; cannam@95: Te5 = FMA(KP098491403, TdH, TdO); cannam@95: TdP = FNMS(KP098491403, TdO, TdH); cannam@95: { cannam@95: E Te3, Te1, Te8, Te6, Te2, TdQ; cannam@95: Te3 = FMA(KP980785280, Te0, TdX); cannam@95: Te1 = FNMS(KP980785280, Te0, TdX); cannam@95: Te8 = Te5 - Te4; cannam@95: Te6 = Te4 + Te5; cannam@95: Te2 = TdP - Tdu; cannam@95: TdQ = Tdu + TdP; cannam@95: Ci[WS(csi, 62)] = FMS(KP995184726, Te6, Te3); cannam@95: Ci[WS(csi, 2)] = FMA(KP995184726, Te6, Te3); cannam@95: Ci[WS(csi, 34)] = FMA(KP995184726, Te2, Te1); cannam@95: Ci[WS(csi, 30)] = FMS(KP995184726, Te2, Te1); cannam@95: Cr[WS(csr, 2)] = FMA(KP995184726, TdQ, Td9); cannam@95: Cr[WS(csr, 62)] = FNMS(KP995184726, TdQ, Td9); cannam@95: Cr[WS(csr, 34)] = FNMS(KP995184726, Te8, Te7); cannam@95: Cr[WS(csr, 30)] = FMA(KP995184726, Te8, Te7); cannam@95: } cannam@95: } cannam@95: { cannam@95: E Teb, Tei, Ten, Teq; cannam@95: Ter = FNMS(KP980785280, Tea, Te9); cannam@95: Teb = FMA(KP980785280, Tea, Te9); cannam@95: Tei = Tee + Teh; cannam@95: Tem = Teh - Tee; cannam@95: Tel = FMA(KP980785280, Tek, Tej); cannam@95: Ten = FNMS(KP980785280, Tek, Tej); cannam@95: Teq = Teo + Tep; cannam@95: Tes = Teo - Tep; cannam@95: Cr[WS(csr, 14)] = FMA(KP773010453, Tei, Teb); cannam@95: Cr[WS(csr, 50)] = FNMS(KP773010453, Tei, Teb); cannam@95: Ci[WS(csi, 50)] = FNMS(KP773010453, Teq, Ten); cannam@95: Ci[WS(csi, 14)] = -(FMA(KP773010453, Teq, Ten)); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T77, T6v, T7i, T6C, T78, T6Y, T7h, T6V, T6N, T7d, T6P, T6F, T6I; cannam@95: { cannam@95: E T6W, T6X, T6T, T6U, T6M; cannam@95: { cannam@95: E T6t, T6u, T6y, T6B; cannam@95: T2f = FMA(KP923879532, T2e, T27); cannam@95: T6t = FNMS(KP923879532, T2e, T27); cannam@95: Ci[WS(csi, 46)] = FMS(KP773010453, Tem, Tel); cannam@95: Ci[WS(csi, 18)] = FMA(KP773010453, Tem, Tel); cannam@95: Cr[WS(csr, 18)] = FMA(KP773010453, Tes, Ter); cannam@95: Cr[WS(csr, 46)] = FNMS(KP773010453, Tes, Ter); cannam@95: T6u = T5U - T5V; cannam@95: T5W = T5U + T5V; cannam@95: T6W = FNMS(KP820678790, T6w, T6x); cannam@95: T6y = FMA(KP820678790, T6x, T6w); cannam@95: T6B = FNMS(KP820678790, T6A, T6z); cannam@95: T6X = FMA(KP820678790, T6z, T6A); cannam@95: T77 = FMA(KP980785280, T6u, T6t); cannam@95: T6v = FNMS(KP980785280, T6u, T6t); cannam@95: T7i = T6B + T6y; cannam@95: T6C = T6y - T6B; cannam@95: } cannam@95: T5T = FMA(KP923879532, T5S, T5P); cannam@95: T6T = FNMS(KP923879532, T5S, T5P); cannam@95: T6U = T2x - T2o; cannam@95: T2y = T2o + T2x; cannam@95: T5J = T5H + T5I; cannam@95: T6M = T5I - T5H; cannam@95: T78 = T6X + T6W; cannam@95: T6Y = T6W - T6X; cannam@95: T7h = FMA(KP980785280, T6U, T6T); cannam@95: T6V = FNMS(KP980785280, T6U, T6T); cannam@95: T6N = FNMS(KP980785280, T6M, T6L); cannam@95: T7d = FMA(KP980785280, T6M, T6L); cannam@95: T6P = T5v - T5e; cannam@95: T5w = T5e + T5v; cannam@95: T4u = T4s + T4t; cannam@95: T6F = T4s - T4t; cannam@95: T6I = T4g - T3Z; cannam@95: T4h = T3Z + T4g; cannam@95: } cannam@95: { cannam@95: E T75, T7f, T7n, T7c, T7m, T76; cannam@95: { cannam@95: E T6D, T72, T6R, T73, T6K, T71, T6Z, T7e, T6Q, T74, T70, T6S; cannam@95: T75 = FNMS(KP773010453, T6C, T6v); cannam@95: T6D = FMA(KP773010453, T6C, T6v); cannam@95: T7e = FNMS(KP980785280, T6P, T6O); cannam@95: T6Q = FMA(KP980785280, T6P, T6O); cannam@95: { cannam@95: E T7a, T6G, T7b, T6J; cannam@95: T7a = FMA(KP980785280, T6F, T6E); cannam@95: T6G = FNMS(KP980785280, T6F, T6E); cannam@95: T7b = FMA(KP980785280, T6I, T6H); cannam@95: T6J = FNMS(KP980785280, T6I, T6H); cannam@95: T7f = FMA(KP357805721, T7e, T7d); cannam@95: T7n = FNMS(KP357805721, T7d, T7e); cannam@95: T72 = FMA(KP472964775, T6N, T6Q); cannam@95: T6R = FNMS(KP472964775, T6Q, T6N); cannam@95: T7c = FMA(KP357805721, T7b, T7a); cannam@95: T7m = FNMS(KP357805721, T7a, T7b); cannam@95: T73 = FMA(KP472964775, T6G, T6J); cannam@95: T6K = FNMS(KP472964775, T6J, T6G); cannam@95: } cannam@95: T71 = FNMS(KP773010453, T6Y, T6V); cannam@95: T6Z = FMA(KP773010453, T6Y, T6V); cannam@95: T74 = T72 - T73; cannam@95: T76 = T73 + T72; cannam@95: T70 = T6R - T6K; cannam@95: T6S = T6K + T6R; cannam@95: Ci[WS(csi, 55)] = FMA(KP903989293, T74, T71); cannam@95: Ci[WS(csi, 9)] = FMS(KP903989293, T74, T71); cannam@95: Cr[WS(csr, 9)] = FMA(KP903989293, T6S, T6D); cannam@95: Cr[WS(csr, 55)] = FNMS(KP903989293, T6S, T6D); cannam@95: Ci[WS(csi, 41)] = FMS(KP903989293, T70, T6Z); cannam@95: Ci[WS(csi, 23)] = FMA(KP903989293, T70, T6Z); cannam@95: } cannam@95: { cannam@95: E T7k, T7j, T7l, T7o, T79, T7g; cannam@95: T7p = FNMS(KP773010453, T78, T77); cannam@95: T79 = FMA(KP773010453, T78, T77); cannam@95: T7g = T7c + T7f; cannam@95: T7k = T7f - T7c; cannam@95: T7j = FNMS(KP773010453, T7i, T7h); cannam@95: T7l = FMA(KP773010453, T7i, T7h); cannam@95: Cr[WS(csr, 23)] = FMA(KP903989293, T76, T75); cannam@95: Cr[WS(csr, 41)] = FNMS(KP903989293, T76, T75); cannam@95: Cr[WS(csr, 7)] = FMA(KP941544065, T7g, T79); cannam@95: Cr[WS(csr, 57)] = FNMS(KP941544065, T7g, T79); cannam@95: T7o = T7m - T7n; cannam@95: T7q = T7m + T7n; cannam@95: Ci[WS(csi, 57)] = FMS(KP941544065, T7o, T7l); cannam@95: Ci[WS(csi, 7)] = FMA(KP941544065, T7o, T7l); cannam@95: Ci[WS(csi, 39)] = FMA(KP941544065, T7k, T7j); cannam@95: Ci[WS(csi, 25)] = FMS(KP941544065, T7k, T7j); cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T7t, T8A, T8x, T7A, T8r, T8k, T88, T81, Ta3, Ta4, T6r, T6s; cannam@95: { cannam@95: E T9L, T99, T9W, T9g, T9M, T9C, T9V, T9z, T9k, T9O, T9T, Ta0, T9H, T9v, T9m; cannam@95: { cannam@95: E T9B, T9c, T9f, T9A, T97, T98; cannam@95: T7t = FMA(KP923879532, T7s, T7r); cannam@95: T97 = FNMS(KP923879532, T7s, T7r); cannam@95: T98 = T8z - T8y; cannam@95: T8A = T8y + T8z; cannam@95: T9B = FNMS(KP534511135, T9a, T9b); cannam@95: T9c = FMA(KP534511135, T9b, T9a); cannam@95: Cr[WS(csr, 25)] = FNMS(KP941544065, T7q, T7p); cannam@95: Cr[WS(csr, 39)] = FMA(KP941544065, T7q, T7p); cannam@95: T9L = FMA(KP831469612, T98, T97); cannam@95: T99 = FNMS(KP831469612, T98, T97); cannam@95: T9f = FNMS(KP534511135, T9e, T9d); cannam@95: T9A = FMA(KP534511135, T9d, T9e); cannam@95: { cannam@95: E T9x, T9y, T9q, T9t; cannam@95: T8x = FMA(KP923879532, T8w, T8v); cannam@95: T9x = FNMS(KP923879532, T8w, T8v); cannam@95: T9W = T9c + T9f; cannam@95: T9g = T9c - T9f; cannam@95: T9M = T9B + T9A; cannam@95: T9C = T9A - T9B; cannam@95: T9y = T7z - T7w; cannam@95: T7A = T7w + T7z; cannam@95: T8r = T8p + T8q; cannam@95: T9q = T8p - T8q; cannam@95: T9t = T8j - T8g; cannam@95: T8k = T8g + T8j; cannam@95: { cannam@95: E T9R, T9r, T9S, T9u, T9j; cannam@95: T88 = T86 + T87; cannam@95: T9j = T87 - T86; cannam@95: T9V = FNMS(KP831469612, T9y, T9x); cannam@95: T9z = FMA(KP831469612, T9y, T9x); cannam@95: T9R = FMA(KP831469612, T9q, T9p); cannam@95: T9r = FNMS(KP831469612, T9q, T9p); cannam@95: T9S = FMA(KP831469612, T9t, T9s); cannam@95: T9u = FNMS(KP831469612, T9t, T9s); cannam@95: T9k = FNMS(KP831469612, T9j, T9i); cannam@95: T9O = FMA(KP831469612, T9j, T9i); cannam@95: T9T = FNMS(KP250486960, T9S, T9R); cannam@95: Ta0 = FMA(KP250486960, T9R, T9S); cannam@95: T9H = FNMS(KP599376933, T9r, T9u); cannam@95: T9v = FMA(KP599376933, T9u, T9r); cannam@95: T9m = T7X - T80; cannam@95: T81 = T7X + T80; cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T9J, T9h, T9F, T9D, T9P, T9n; cannam@95: T9J = FNMS(KP881921264, T9g, T99); cannam@95: T9h = FMA(KP881921264, T9g, T99); cannam@95: T9F = FMA(KP881921264, T9C, T9z); cannam@95: T9D = FNMS(KP881921264, T9C, T9z); cannam@95: T9P = FMA(KP831469612, T9m, T9l); cannam@95: T9n = FNMS(KP831469612, T9m, T9l); cannam@95: { cannam@95: E T9Y, T9X, T9Z, Ta2; cannam@95: { cannam@95: E T9N, Ta1, T9G, T9o, T9U, T9Q; cannam@95: Ta3 = FNMS(KP881921264, T9M, T9L); cannam@95: T9N = FMA(KP881921264, T9M, T9L); cannam@95: T9Q = FNMS(KP250486960, T9P, T9O); cannam@95: Ta1 = FMA(KP250486960, T9O, T9P); cannam@95: T9G = FNMS(KP599376933, T9k, T9n); cannam@95: T9o = FMA(KP599376933, T9n, T9k); cannam@95: T9U = T9Q + T9T; cannam@95: T9Y = T9T - T9Q; cannam@95: T9X = FNMS(KP881921264, T9W, T9V); cannam@95: T9Z = FMA(KP881921264, T9W, T9V); cannam@95: { cannam@95: E T9K, T9I, T9E, T9w; cannam@95: T9K = T9G + T9H; cannam@95: T9I = T9G - T9H; cannam@95: T9E = T9v - T9o; cannam@95: T9w = T9o + T9v; cannam@95: Cr[WS(csr, 5)] = FMA(KP970031253, T9U, T9N); cannam@95: Cr[WS(csr, 59)] = FNMS(KP970031253, T9U, T9N); cannam@95: Cr[WS(csr, 21)] = FNMS(KP857728610, T9K, T9J); cannam@95: Cr[WS(csr, 43)] = FMA(KP857728610, T9K, T9J); cannam@95: Ci[WS(csi, 53)] = FMS(KP857728610, T9I, T9F); cannam@95: Ci[WS(csi, 11)] = FMA(KP857728610, T9I, T9F); cannam@95: Ci[WS(csi, 43)] = FMA(KP857728610, T9E, T9D); cannam@95: Ci[WS(csi, 21)] = FMS(KP857728610, T9E, T9D); cannam@95: Cr[WS(csr, 11)] = FMA(KP857728610, T9w, T9h); cannam@95: Cr[WS(csr, 53)] = FNMS(KP857728610, T9w, T9h); cannam@95: Ta2 = Ta0 - Ta1; cannam@95: Ta4 = Ta1 + Ta0; cannam@95: } cannam@95: } cannam@95: Ci[WS(csi, 59)] = FMA(KP970031253, Ta2, T9Z); cannam@95: Ci[WS(csi, 5)] = FMS(KP970031253, Ta2, T9Z); cannam@95: Ci[WS(csi, 37)] = FMS(KP970031253, T9Y, T9X); cannam@95: Ci[WS(csi, 27)] = FMA(KP970031253, T9Y, T9X); cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T69, T2z, T6k, T3g, T6a, T60, T6j, T5X, T4i, T6c, T6h, T6p, T64, T5L; cannam@95: { cannam@95: E T5Y, T2U, T3f, T5Z; cannam@95: T5Y = FMA(KP098491403, T2M, T2T); cannam@95: T2U = FNMS(KP098491403, T2T, T2M); cannam@95: Cr[WS(csr, 27)] = FMA(KP970031253, Ta4, Ta3); cannam@95: Cr[WS(csr, 37)] = FNMS(KP970031253, Ta4, Ta3); cannam@95: T69 = FNMS(KP980785280, T2y, T2f); cannam@95: T2z = FMA(KP980785280, T2y, T2f); cannam@95: T3f = FMA(KP098491403, T3e, T37); cannam@95: T5Z = FNMS(KP098491403, T37, T3e); cannam@95: T6k = T3f - T2U; cannam@95: T3g = T2U + T3f; cannam@95: T6a = T5Y - T5Z; cannam@95: T60 = T5Y + T5Z; cannam@95: { cannam@95: E T6f, T5x, T6g, T5K; cannam@95: T6j = FNMS(KP980785280, T5W, T5T); cannam@95: T5X = FMA(KP980785280, T5W, T5T); cannam@95: T6f = FNMS(KP980785280, T5w, T4X); cannam@95: T5x = FMA(KP980785280, T5w, T4X); cannam@95: T6g = FNMS(KP980785280, T5J, T5G); cannam@95: T5K = FMA(KP980785280, T5J, T5G); cannam@95: T4i = FMA(KP980785280, T4h, T3I); cannam@95: T6c = FNMS(KP980785280, T4h, T3I); cannam@95: T6h = FMA(KP906347169, T6g, T6f); cannam@95: T6p = FNMS(KP906347169, T6f, T6g); cannam@95: T64 = FMA(KP049126849, T5x, T5K); cannam@95: T5L = FNMS(KP049126849, T5K, T5x); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T67, T3h, T63, T61, T6d, T4v; cannam@95: T67 = FNMS(KP995184726, T3g, T2z); cannam@95: T3h = FMA(KP995184726, T3g, T2z); cannam@95: T63 = FMA(KP995184726, T60, T5X); cannam@95: T61 = FNMS(KP995184726, T60, T5X); cannam@95: T6d = FNMS(KP980785280, T4u, T4r); cannam@95: T4v = FMA(KP980785280, T4u, T4r); cannam@95: { cannam@95: E T6m, T6l, T6n, T6q; cannam@95: { cannam@95: E T6b, T6o, T65, T4w, T6i, T6e; cannam@95: T6r = FNMS(KP995184726, T6a, T69); cannam@95: T6b = FMA(KP995184726, T6a, T69); cannam@95: T6e = FMA(KP906347169, T6d, T6c); cannam@95: T6o = FNMS(KP906347169, T6c, T6d); cannam@95: T65 = FMA(KP049126849, T4i, T4v); cannam@95: T4w = FNMS(KP049126849, T4v, T4i); cannam@95: T6i = T6e + T6h; cannam@95: T6m = T6h - T6e; cannam@95: T6l = FNMS(KP995184726, T6k, T6j); cannam@95: T6n = FMA(KP995184726, T6k, T6j); cannam@95: { cannam@95: E T68, T66, T62, T5M; cannam@95: T68 = T65 + T64; cannam@95: T66 = T64 - T65; cannam@95: T62 = T5L - T4w; cannam@95: T5M = T4w + T5L; cannam@95: Cr[WS(csr, 15)] = FMA(KP740951125, T6i, T6b); cannam@95: Cr[WS(csr, 49)] = FNMS(KP740951125, T6i, T6b); cannam@95: Cr[WS(csr, 31)] = FMA(KP998795456, T68, T67); cannam@95: Cr[WS(csr, 33)] = FNMS(KP998795456, T68, T67); cannam@95: Ci[WS(csi, 63)] = FMA(KP998795456, T66, T63); cannam@95: Ci[WS(csi, 1)] = FMS(KP998795456, T66, T63); cannam@95: Ci[WS(csi, 33)] = FMS(KP998795456, T62, T61); cannam@95: Ci[WS(csi, 31)] = FMA(KP998795456, T62, T61); cannam@95: Cr[WS(csr, 1)] = FMA(KP998795456, T5M, T3h); cannam@95: Cr[WS(csr, 63)] = FNMS(KP998795456, T5M, T3h); cannam@95: T6q = T6o - T6p; cannam@95: T6s = T6o + T6p; cannam@95: } cannam@95: } cannam@95: Ci[WS(csi, 49)] = FMS(KP740951125, T6q, T6n); cannam@95: Ci[WS(csi, 15)] = FMA(KP740951125, T6q, T6n); cannam@95: Ci[WS(csi, 47)] = FMA(KP740951125, T6m, T6l); cannam@95: Ci[WS(csi, 17)] = FMS(KP740951125, T6m, T6l); cannam@95: } cannam@95: } cannam@95: } cannam@95: { cannam@95: E T8N, T7B, T8Y, T7Q, T8O, T8E, T8X, T8B, T82, T8Q, T8V, T92, T8J, T8t; cannam@95: { cannam@95: E T8C, T7I, T7P, T8D; cannam@95: T8C = FNMS(KP303346683, T7E, T7H); cannam@95: T7I = FMA(KP303346683, T7H, T7E); cannam@95: Cr[WS(csr, 17)] = FNMS(KP740951125, T6s, T6r); cannam@95: Cr[WS(csr, 47)] = FMA(KP740951125, T6s, T6r); cannam@95: T8N = FNMS(KP831469612, T7A, T7t); cannam@95: T7B = FMA(KP831469612, T7A, T7t); cannam@95: T7P = FNMS(KP303346683, T7O, T7L); cannam@95: T8D = FMA(KP303346683, T7L, T7O); cannam@95: T8Y = T7P - T7I; cannam@95: T7Q = T7I + T7P; cannam@95: T8O = T8D - T8C; cannam@95: T8E = T8C + T8D; cannam@95: { cannam@95: E T8T, T8l, T8U, T8s; cannam@95: T8X = FNMS(KP831469612, T8A, T8x); cannam@95: T8B = FMA(KP831469612, T8A, T8x); cannam@95: T8T = FNMS(KP831469612, T8k, T8d); cannam@95: T8l = FMA(KP831469612, T8k, T8d); cannam@95: T8U = FNMS(KP831469612, T8r, T8o); cannam@95: T8s = FMA(KP831469612, T8r, T8o); cannam@95: T82 = FMA(KP831469612, T81, T7U); cannam@95: T8Q = FNMS(KP831469612, T81, T7U); cannam@95: T8V = FNMS(KP741650546, T8U, T8T); cannam@95: T92 = FMA(KP741650546, T8T, T8U); cannam@95: T8J = FNMS(KP148335987, T8l, T8s); cannam@95: T8t = FMA(KP148335987, T8s, T8l); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T8L, T7R, T8H, T8F, T8R, T89; cannam@95: T8L = FNMS(KP956940335, T7Q, T7B); cannam@95: T7R = FMA(KP956940335, T7Q, T7B); cannam@95: T8H = FMA(KP956940335, T8E, T8B); cannam@95: T8F = FNMS(KP956940335, T8E, T8B); cannam@95: T8R = FNMS(KP831469612, T88, T85); cannam@95: T89 = FMA(KP831469612, T88, T85); cannam@95: { cannam@95: E T90, T8Z, T91, T94; cannam@95: { cannam@95: E T8P, T93, T8I, T8a, T8W, T8S; cannam@95: T95 = FNMS(KP956940335, T8O, T8N); cannam@95: T8P = FMA(KP956940335, T8O, T8N); cannam@95: T8S = FNMS(KP741650546, T8R, T8Q); cannam@95: T93 = FMA(KP741650546, T8Q, T8R); cannam@95: T8I = FNMS(KP148335987, T82, T89); cannam@95: T8a = FMA(KP148335987, T89, T82); cannam@95: T8W = T8S + T8V; cannam@95: T90 = T8V - T8S; cannam@95: T8Z = FMA(KP956940335, T8Y, T8X); cannam@95: T91 = FNMS(KP956940335, T8Y, T8X); cannam@95: { cannam@95: E T8M, T8K, T8G, T8u; cannam@95: T8M = T8I + T8J; cannam@95: T8K = T8I - T8J; cannam@95: T8G = T8t - T8a; cannam@95: T8u = T8a + T8t; cannam@95: Cr[WS(csr, 13)] = FMA(KP803207531, T8W, T8P); cannam@95: Cr[WS(csr, 51)] = FNMS(KP803207531, T8W, T8P); cannam@95: Cr[WS(csr, 29)] = FNMS(KP989176509, T8M, T8L); cannam@95: Cr[WS(csr, 35)] = FMA(KP989176509, T8M, T8L); cannam@95: Ci[WS(csi, 61)] = FMS(KP989176509, T8K, T8H); cannam@95: Ci[WS(csi, 3)] = FMA(KP989176509, T8K, T8H); cannam@95: Ci[WS(csi, 35)] = FMA(KP989176509, T8G, T8F); cannam@95: Ci[WS(csi, 29)] = FMS(KP989176509, T8G, T8F); cannam@95: Cr[WS(csr, 3)] = FMA(KP989176509, T8u, T7R); cannam@95: Cr[WS(csr, 61)] = FNMS(KP989176509, T8u, T7R); cannam@95: T94 = T92 - T93; cannam@95: T96 = T93 + T92; cannam@95: } cannam@95: } cannam@95: Ci[WS(csi, 51)] = FMA(KP803207531, T94, T91); cannam@95: Ci[WS(csi, 13)] = FMS(KP803207531, T94, T91); cannam@95: Ci[WS(csi, 45)] = FMS(KP803207531, T90, T8Z); cannam@95: Ci[WS(csi, 19)] = FMA(KP803207531, T90, T8Z); cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: Cr[WS(csr, 19)] = FMA(KP803207531, T96, T95); cannam@95: Cr[WS(csr, 45)] = FNMS(KP803207531, T96, T95); cannam@95: } cannam@95: } cannam@95: } cannam@95: cannam@95: static const kr2c_desc desc = { 128, "r2cf_128", {440, 0, 516, 0}, &GENUS }; cannam@95: cannam@95: void X(codelet_r2cf_128) (planner *p) { cannam@95: X(kr2c_register) (p, r2cf_128, &desc); cannam@95: } cannam@95: cannam@95: #else /* HAVE_FMA */ cannam@95: cannam@95: /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 128 -name r2cf_128 -include r2cf.h */ cannam@95: cannam@95: /* cannam@95: * This function contains 956 FP additions, 330 FP multiplications, cannam@95: * (or, 812 additions, 186 multiplications, 144 fused multiply/add), cannam@95: * 186 stack variables, 31 constants, and 256 memory accesses cannam@95: */ cannam@95: #include "r2cf.h" cannam@95: cannam@95: static void r2cf_128(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@95: { cannam@95: DK(KP803207531, +0.803207531480644909806676512963141923879569427); cannam@95: DK(KP595699304, +0.595699304492433343467036528829969889511926338); cannam@95: DK(KP146730474, +0.146730474455361751658850129646717819706215317); cannam@95: DK(KP989176509, +0.989176509964780973451673738016243063983689533); cannam@95: DK(KP740951125, +0.740951125354959091175616897495162729728955309); cannam@95: DK(KP671558954, +0.671558954847018400625376850427421803228750632); cannam@95: DK(KP049067674, +0.049067674327418014254954976942682658314745363); cannam@95: DK(KP998795456, +0.998795456205172392714771604759100694443203615); cannam@95: DK(KP242980179, +0.242980179903263889948274162077471118320990783); cannam@95: DK(KP970031253, +0.970031253194543992603984207286100251456865962); cannam@95: DK(KP514102744, +0.514102744193221726593693838968815772608049120); cannam@95: DK(KP857728610, +0.857728610000272069902269984284770137042490799); cannam@95: DK(KP336889853, +0.336889853392220050689253212619147570477766780); cannam@95: DK(KP941544065, +0.941544065183020778412509402599502357185589796); cannam@95: DK(KP427555093, +0.427555093430282094320966856888798534304578629); cannam@95: DK(KP903989293, +0.903989293123443331586200297230537048710132025); cannam@95: DK(KP098017140, +0.098017140329560601994195563888641845861136673); cannam@95: DK(KP995184726, +0.995184726672196886244836953109479921575474869); cannam@95: DK(KP634393284, +0.634393284163645498215171613225493370675687095); cannam@95: DK(KP773010453, +0.773010453362736960810906609758469800971041293); cannam@95: DK(KP881921264, +0.881921264348355029712756863660388349508442621); cannam@95: DK(KP471396736, +0.471396736825997648556387625905254377657460319); cannam@95: DK(KP956940335, +0.956940335732208864935797886980269969482849206); cannam@95: DK(KP290284677, +0.290284677254462367636192375817395274691476278); cannam@95: DK(KP555570233, +0.555570233019602224742830813948532874374937191); cannam@95: DK(KP831469612, +0.831469612302545237078788377617905756738560812); cannam@95: DK(KP195090322, +0.195090322016128267848284868477022240927691618); cannam@95: DK(KP980785280, +0.980785280403230449126182236134239036973933731); cannam@95: DK(KP382683432, +0.382683432365089771728459984030398866761344562); cannam@95: DK(KP923879532, +0.923879532511286756128183189396788286822416626); cannam@95: DK(KP707106781, +0.707106781186547524400844362104849039284835938); cannam@95: { cannam@95: INT i; cannam@95: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(512, rs), MAKE_VOLATILE_STRIDE(512, csr), MAKE_VOLATILE_STRIDE(512, csi)) { cannam@95: E TcD, TdU, T27, T7r, T5S, T8y, Tf, Ta5, Tu, Tbq, TcG, TdV, T2e, T8z, T5V; cannam@95: E T7s, TK, Ta6, TcK, TdX, T2o, T5X, T7w, T8B, TZ, Ta7, TcN, TdY, T2x, T5Y; cannam@95: E T7z, T8C, T1g, Taa, TcU, TeA, TcX, Tez, T1v, Tab, T2M, T6z, T7E, T9e, T7H; cannam@95: E T9d, T2T, T6A, T4X, T6L, Tdz, TeL, TdK, TeP, T5G, T6P, T8d, T9p, TaV, Tc3; cannam@95: E Tbi, Tc4, T8o, T9t, T3I, T6H, Tde, TeH, Tdp, TeF, T4r, T6F, T7U, T9l, Tao; cannam@95: E TbW, TaL, TbX, T85, T9j, T1L, Tad, Td3, Tew, Td6, Tex, T20, Tae, T37, T6x; cannam@95: E T7L, T9a, T7O, T9b, T3e, T6w, TbZ, Tc0, T3Z, T4s, Tds, TeI, T4g, T4t, T80; cannam@95: E T87, Tdl, TeE, T7X, T86, TaD, TaM, Tc6, Tc7, T5e, T5H, TdN, TeM, T5v, T5I; cannam@95: E T8j, T8q, TdG, TeO, T8g, T8p, Tba, Tbj; cannam@95: { cannam@95: E T3, T23, Td, T25, T6, T5R, Ta, T24; cannam@95: { cannam@95: E T1, T2, Tb, Tc; cannam@95: T1 = R0[0]; cannam@95: T2 = R0[WS(rs, 32)]; cannam@95: T3 = T1 + T2; cannam@95: T23 = T1 - T2; cannam@95: Tb = R0[WS(rs, 56)]; cannam@95: Tc = R0[WS(rs, 24)]; cannam@95: Td = Tb + Tc; cannam@95: T25 = Tb - Tc; cannam@95: } cannam@95: { cannam@95: E T4, T5, T8, T9; cannam@95: T4 = R0[WS(rs, 16)]; cannam@95: T5 = R0[WS(rs, 48)]; cannam@95: T6 = T4 + T5; cannam@95: T5R = T4 - T5; cannam@95: T8 = R0[WS(rs, 8)]; cannam@95: T9 = R0[WS(rs, 40)]; cannam@95: Ta = T8 + T9; cannam@95: T24 = T8 - T9; cannam@95: } cannam@95: TcD = T3 - T6; cannam@95: TdU = Td - Ta; cannam@95: { cannam@95: E T26, T5Q, T7, Te; cannam@95: T26 = KP707106781 * (T24 + T25); cannam@95: T27 = T23 + T26; cannam@95: T7r = T23 - T26; cannam@95: T5Q = KP707106781 * (T25 - T24); cannam@95: T5S = T5Q - T5R; cannam@95: T8y = T5R + T5Q; cannam@95: T7 = T3 + T6; cannam@95: Te = Ta + Td; cannam@95: Tf = T7 + Te; cannam@95: Ta5 = T7 - Te; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Ti, T28, Ts, T2c, Tl, T29, Tp, T2b; cannam@95: { cannam@95: E Tg, Th, Tq, Tr; cannam@95: Tg = R0[WS(rs, 4)]; cannam@95: Th = R0[WS(rs, 36)]; cannam@95: Ti = Tg + Th; cannam@95: T28 = Tg - Th; cannam@95: Tq = R0[WS(rs, 12)]; cannam@95: Tr = R0[WS(rs, 44)]; cannam@95: Ts = Tq + Tr; cannam@95: T2c = Tq - Tr; cannam@95: } cannam@95: { cannam@95: E Tj, Tk, Tn, To; cannam@95: Tj = R0[WS(rs, 20)]; cannam@95: Tk = R0[WS(rs, 52)]; cannam@95: Tl = Tj + Tk; cannam@95: T29 = Tj - Tk; cannam@95: Tn = R0[WS(rs, 60)]; cannam@95: To = R0[WS(rs, 28)]; cannam@95: Tp = Tn + To; cannam@95: T2b = Tn - To; cannam@95: } cannam@95: { cannam@95: E Tm, Tt, TcE, TcF; cannam@95: Tm = Ti + Tl; cannam@95: Tt = Tp + Ts; cannam@95: Tu = Tm + Tt; cannam@95: Tbq = Tt - Tm; cannam@95: TcE = Ti - Tl; cannam@95: TcF = Tp - Ts; cannam@95: TcG = KP707106781 * (TcE + TcF); cannam@95: TdV = KP707106781 * (TcF - TcE); cannam@95: } cannam@95: { cannam@95: E T2a, T2d, T5T, T5U; cannam@95: T2a = FNMS(KP382683432, T29, KP923879532 * T28); cannam@95: T2d = FMA(KP923879532, T2b, KP382683432 * T2c); cannam@95: T2e = T2a + T2d; cannam@95: T8z = T2d - T2a; cannam@95: T5T = FNMS(KP923879532, T2c, KP382683432 * T2b); cannam@95: T5U = FMA(KP382683432, T28, KP923879532 * T29); cannam@95: T5V = T5T - T5U; cannam@95: T7s = T5U + T5T; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Ty, T2g, TB, T2m, TF, T2l, TI, T2j; cannam@95: { cannam@95: E Tw, Tx, Tz, TA; cannam@95: Tw = R0[WS(rs, 2)]; cannam@95: Tx = R0[WS(rs, 34)]; cannam@95: Ty = Tw + Tx; cannam@95: T2g = Tw - Tx; cannam@95: Tz = R0[WS(rs, 18)]; cannam@95: TA = R0[WS(rs, 50)]; cannam@95: TB = Tz + TA; cannam@95: T2m = Tz - TA; cannam@95: { cannam@95: E TD, TE, T2h, TG, TH, T2i; cannam@95: TD = R0[WS(rs, 10)]; cannam@95: TE = R0[WS(rs, 42)]; cannam@95: T2h = TD - TE; cannam@95: TG = R0[WS(rs, 58)]; cannam@95: TH = R0[WS(rs, 26)]; cannam@95: T2i = TG - TH; cannam@95: TF = TD + TE; cannam@95: T2l = KP707106781 * (T2i - T2h); cannam@95: TI = TG + TH; cannam@95: T2j = KP707106781 * (T2h + T2i); cannam@95: } cannam@95: } cannam@95: { cannam@95: E TC, TJ, TcI, TcJ; cannam@95: TC = Ty + TB; cannam@95: TJ = TF + TI; cannam@95: TK = TC + TJ; cannam@95: Ta6 = TC - TJ; cannam@95: TcI = Ty - TB; cannam@95: TcJ = TI - TF; cannam@95: TcK = FMA(KP923879532, TcI, KP382683432 * TcJ); cannam@95: TdX = FNMS(KP382683432, TcI, KP923879532 * TcJ); cannam@95: } cannam@95: { cannam@95: E T2k, T2n, T7u, T7v; cannam@95: T2k = T2g + T2j; cannam@95: T2n = T2l - T2m; cannam@95: T2o = FMA(KP980785280, T2k, KP195090322 * T2n); cannam@95: T5X = FNMS(KP195090322, T2k, KP980785280 * T2n); cannam@95: T7u = T2g - T2j; cannam@95: T7v = T2m + T2l; cannam@95: T7w = FMA(KP831469612, T7u, KP555570233 * T7v); cannam@95: T8B = FNMS(KP555570233, T7u, KP831469612 * T7v); cannam@95: } cannam@95: } cannam@95: { cannam@95: E TN, T2p, TQ, T2v, TU, T2u, TX, T2s; cannam@95: { cannam@95: E TL, TM, TO, TP; cannam@95: TL = R0[WS(rs, 62)]; cannam@95: TM = R0[WS(rs, 30)]; cannam@95: TN = TL + TM; cannam@95: T2p = TL - TM; cannam@95: TO = R0[WS(rs, 14)]; cannam@95: TP = R0[WS(rs, 46)]; cannam@95: TQ = TO + TP; cannam@95: T2v = TO - TP; cannam@95: { cannam@95: E TS, TT, T2q, TV, TW, T2r; cannam@95: TS = R0[WS(rs, 6)]; cannam@95: TT = R0[WS(rs, 38)]; cannam@95: T2q = TS - TT; cannam@95: TV = R0[WS(rs, 54)]; cannam@95: TW = R0[WS(rs, 22)]; cannam@95: T2r = TV - TW; cannam@95: TU = TS + TT; cannam@95: T2u = KP707106781 * (T2r - T2q); cannam@95: TX = TV + TW; cannam@95: T2s = KP707106781 * (T2q + T2r); cannam@95: } cannam@95: } cannam@95: { cannam@95: E TR, TY, TcL, TcM; cannam@95: TR = TN + TQ; cannam@95: TY = TU + TX; cannam@95: TZ = TR + TY; cannam@95: Ta7 = TR - TY; cannam@95: TcL = TN - TQ; cannam@95: TcM = TX - TU; cannam@95: TcN = FNMS(KP382683432, TcM, KP923879532 * TcL); cannam@95: TdY = FMA(KP382683432, TcL, KP923879532 * TcM); cannam@95: } cannam@95: { cannam@95: E T2t, T2w, T7x, T7y; cannam@95: T2t = T2p + T2s; cannam@95: T2w = T2u - T2v; cannam@95: T2x = FNMS(KP195090322, T2w, KP980785280 * T2t); cannam@95: T5Y = FMA(KP195090322, T2t, KP980785280 * T2w); cannam@95: T7x = T2p - T2s; cannam@95: T7y = T2v + T2u; cannam@95: T7z = FNMS(KP555570233, T7y, KP831469612 * T7x); cannam@95: T8C = FMA(KP555570233, T7x, KP831469612 * T7y); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T14, T2N, T17, T2D, T1b, T2O, T1e, T2C, T1j, T1m, T2K, TcR, T2Q, T1q, T1t; cannam@95: E T2H, TcS, T2R; cannam@95: { cannam@95: E T12, T13, T15, T16; cannam@95: T12 = R0[WS(rs, 1)]; cannam@95: T13 = R0[WS(rs, 33)]; cannam@95: T14 = T12 + T13; cannam@95: T2N = T12 - T13; cannam@95: T15 = R0[WS(rs, 17)]; cannam@95: T16 = R0[WS(rs, 49)]; cannam@95: T17 = T15 + T16; cannam@95: T2D = T15 - T16; cannam@95: } cannam@95: { cannam@95: E T19, T1a, T2B, T1c, T1d, T2A; cannam@95: T19 = R0[WS(rs, 9)]; cannam@95: T1a = R0[WS(rs, 41)]; cannam@95: T2B = T19 - T1a; cannam@95: T1c = R0[WS(rs, 57)]; cannam@95: T1d = R0[WS(rs, 25)]; cannam@95: T2A = T1c - T1d; cannam@95: T1b = T19 + T1a; cannam@95: T2O = KP707106781 * (T2B + T2A); cannam@95: T1e = T1c + T1d; cannam@95: T2C = KP707106781 * (T2A - T2B); cannam@95: } cannam@95: { cannam@95: E T2I, T2J, T2F, T2G; cannam@95: { cannam@95: E T1h, T1i, T1k, T1l; cannam@95: T1h = R0[WS(rs, 5)]; cannam@95: T1i = R0[WS(rs, 37)]; cannam@95: T1j = T1h + T1i; cannam@95: T2I = T1h - T1i; cannam@95: T1k = R0[WS(rs, 21)]; cannam@95: T1l = R0[WS(rs, 53)]; cannam@95: T1m = T1k + T1l; cannam@95: T2J = T1k - T1l; cannam@95: } cannam@95: T2K = FMA(KP382683432, T2I, KP923879532 * T2J); cannam@95: TcR = T1j - T1m; cannam@95: T2Q = FNMS(KP382683432, T2J, KP923879532 * T2I); cannam@95: { cannam@95: E T1o, T1p, T1r, T1s; cannam@95: T1o = R0[WS(rs, 61)]; cannam@95: T1p = R0[WS(rs, 29)]; cannam@95: T1q = T1o + T1p; cannam@95: T2F = T1o - T1p; cannam@95: T1r = R0[WS(rs, 13)]; cannam@95: T1s = R0[WS(rs, 45)]; cannam@95: T1t = T1r + T1s; cannam@95: T2G = T1r - T1s; cannam@95: } cannam@95: T2H = FNMS(KP923879532, T2G, KP382683432 * T2F); cannam@95: TcS = T1q - T1t; cannam@95: T2R = FMA(KP923879532, T2F, KP382683432 * T2G); cannam@95: } cannam@95: { cannam@95: E T18, T1f, TcQ, TcT; cannam@95: T18 = T14 + T17; cannam@95: T1f = T1b + T1e; cannam@95: T1g = T18 + T1f; cannam@95: Taa = T18 - T1f; cannam@95: TcQ = T14 - T17; cannam@95: TcT = KP707106781 * (TcR + TcS); cannam@95: TcU = TcQ + TcT; cannam@95: TeA = TcQ - TcT; cannam@95: } cannam@95: { cannam@95: E TcV, TcW, T1n, T1u; cannam@95: TcV = T1e - T1b; cannam@95: TcW = KP707106781 * (TcS - TcR); cannam@95: TcX = TcV + TcW; cannam@95: Tez = TcW - TcV; cannam@95: T1n = T1j + T1m; cannam@95: T1u = T1q + T1t; cannam@95: T1v = T1n + T1u; cannam@95: Tab = T1u - T1n; cannam@95: } cannam@95: { cannam@95: E T2E, T2L, T7C, T7D; cannam@95: T2E = T2C - T2D; cannam@95: T2L = T2H - T2K; cannam@95: T2M = T2E + T2L; cannam@95: T6z = T2L - T2E; cannam@95: T7C = T2N - T2O; cannam@95: T7D = T2K + T2H; cannam@95: T7E = T7C + T7D; cannam@95: T9e = T7C - T7D; cannam@95: } cannam@95: { cannam@95: E T7F, T7G, T2P, T2S; cannam@95: T7F = T2D + T2C; cannam@95: T7G = T2R - T2Q; cannam@95: T7H = T7F + T7G; cannam@95: T9d = T7G - T7F; cannam@95: T2P = T2N + T2O; cannam@95: T2S = T2Q + T2R; cannam@95: T2T = T2P + T2S; cannam@95: T6A = T2P - T2S; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T4z, TaP, T5B, TaQ, T4G, TaT, T5y, TaS, Tbf, Tbg, T4O, Tdw, T5E, Tbc, Tbd; cannam@95: E T4V, Tdx, T5D; cannam@95: { cannam@95: E T4x, T4y, T5z, T5A; cannam@95: T4x = R1[WS(rs, 63)]; cannam@95: T4y = R1[WS(rs, 31)]; cannam@95: T4z = T4x - T4y; cannam@95: TaP = T4x + T4y; cannam@95: T5z = R1[WS(rs, 15)]; cannam@95: T5A = R1[WS(rs, 47)]; cannam@95: T5B = T5z - T5A; cannam@95: TaQ = T5z + T5A; cannam@95: } cannam@95: { cannam@95: E T4A, T4B, T4C, T4D, T4E, T4F; cannam@95: T4A = R1[WS(rs, 7)]; cannam@95: T4B = R1[WS(rs, 39)]; cannam@95: T4C = T4A - T4B; cannam@95: T4D = R1[WS(rs, 55)]; cannam@95: T4E = R1[WS(rs, 23)]; cannam@95: T4F = T4D - T4E; cannam@95: T4G = KP707106781 * (T4C + T4F); cannam@95: TaT = T4D + T4E; cannam@95: T5y = KP707106781 * (T4F - T4C); cannam@95: TaS = T4A + T4B; cannam@95: } cannam@95: { cannam@95: E T4K, T4N, T4R, T4U; cannam@95: { cannam@95: E T4I, T4J, T4L, T4M; cannam@95: T4I = R1[WS(rs, 3)]; cannam@95: T4J = R1[WS(rs, 35)]; cannam@95: T4K = T4I - T4J; cannam@95: Tbf = T4I + T4J; cannam@95: T4L = R1[WS(rs, 19)]; cannam@95: T4M = R1[WS(rs, 51)]; cannam@95: T4N = T4L - T4M; cannam@95: Tbg = T4L + T4M; cannam@95: } cannam@95: T4O = FNMS(KP382683432, T4N, KP923879532 * T4K); cannam@95: Tdw = Tbf - Tbg; cannam@95: T5E = FMA(KP382683432, T4K, KP923879532 * T4N); cannam@95: { cannam@95: E T4P, T4Q, T4S, T4T; cannam@95: T4P = R1[WS(rs, 59)]; cannam@95: T4Q = R1[WS(rs, 27)]; cannam@95: T4R = T4P - T4Q; cannam@95: Tbc = T4P + T4Q; cannam@95: T4S = R1[WS(rs, 11)]; cannam@95: T4T = R1[WS(rs, 43)]; cannam@95: T4U = T4S - T4T; cannam@95: Tbd = T4S + T4T; cannam@95: } cannam@95: T4V = FMA(KP923879532, T4R, KP382683432 * T4U); cannam@95: Tdx = Tbc - Tbd; cannam@95: T5D = FNMS(KP923879532, T4U, KP382683432 * T4R); cannam@95: } cannam@95: { cannam@95: E T4H, T4W, Tdv, Tdy; cannam@95: T4H = T4z + T4G; cannam@95: T4W = T4O + T4V; cannam@95: T4X = T4H + T4W; cannam@95: T6L = T4H - T4W; cannam@95: Tdv = TaP - TaQ; cannam@95: Tdy = KP707106781 * (Tdw + Tdx); cannam@95: Tdz = Tdv + Tdy; cannam@95: TeL = Tdv - Tdy; cannam@95: } cannam@95: { cannam@95: E TdI, TdJ, T5C, T5F; cannam@95: TdI = TaT - TaS; cannam@95: TdJ = KP707106781 * (Tdx - Tdw); cannam@95: TdK = TdI + TdJ; cannam@95: TeP = TdJ - TdI; cannam@95: T5C = T5y - T5B; cannam@95: T5F = T5D - T5E; cannam@95: T5G = T5C + T5F; cannam@95: T6P = T5F - T5C; cannam@95: } cannam@95: { cannam@95: E T8b, T8c, TaR, TaU; cannam@95: T8b = T4z - T4G; cannam@95: T8c = T5E + T5D; cannam@95: T8d = T8b + T8c; cannam@95: T9p = T8b - T8c; cannam@95: TaR = TaP + TaQ; cannam@95: TaU = TaS + TaT; cannam@95: TaV = TaR - TaU; cannam@95: Tc3 = TaR + TaU; cannam@95: } cannam@95: { cannam@95: E Tbe, Tbh, T8m, T8n; cannam@95: Tbe = Tbc + Tbd; cannam@95: Tbh = Tbf + Tbg; cannam@95: Tbi = Tbe - Tbh; cannam@95: Tc4 = Tbh + Tbe; cannam@95: T8m = T5B + T5y; cannam@95: T8n = T4V - T4O; cannam@95: T8o = T8m + T8n; cannam@95: T9t = T8n - T8m; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T3k, Tai, T4m, Taj, T3r, Tam, T4j, Tal, TaI, TaJ, T3z, Tdb, T4p, TaF, TaG; cannam@95: E T3G, Tdc, T4o; cannam@95: { cannam@95: E T3i, T3j, T4k, T4l; cannam@95: T3i = R1[0]; cannam@95: T3j = R1[WS(rs, 32)]; cannam@95: T3k = T3i - T3j; cannam@95: Tai = T3i + T3j; cannam@95: T4k = R1[WS(rs, 16)]; cannam@95: T4l = R1[WS(rs, 48)]; cannam@95: T4m = T4k - T4l; cannam@95: Taj = T4k + T4l; cannam@95: } cannam@95: { cannam@95: E T3l, T3m, T3n, T3o, T3p, T3q; cannam@95: T3l = R1[WS(rs, 8)]; cannam@95: T3m = R1[WS(rs, 40)]; cannam@95: T3n = T3l - T3m; cannam@95: T3o = R1[WS(rs, 56)]; cannam@95: T3p = R1[WS(rs, 24)]; cannam@95: T3q = T3o - T3p; cannam@95: T3r = KP707106781 * (T3n + T3q); cannam@95: Tam = T3o + T3p; cannam@95: T4j = KP707106781 * (T3q - T3n); cannam@95: Tal = T3l + T3m; cannam@95: } cannam@95: { cannam@95: E T3v, T3y, T3C, T3F; cannam@95: { cannam@95: E T3t, T3u, T3w, T3x; cannam@95: T3t = R1[WS(rs, 4)]; cannam@95: T3u = R1[WS(rs, 36)]; cannam@95: T3v = T3t - T3u; cannam@95: TaI = T3t + T3u; cannam@95: T3w = R1[WS(rs, 20)]; cannam@95: T3x = R1[WS(rs, 52)]; cannam@95: T3y = T3w - T3x; cannam@95: TaJ = T3w + T3x; cannam@95: } cannam@95: T3z = FNMS(KP382683432, T3y, KP923879532 * T3v); cannam@95: Tdb = TaI - TaJ; cannam@95: T4p = FMA(KP382683432, T3v, KP923879532 * T3y); cannam@95: { cannam@95: E T3A, T3B, T3D, T3E; cannam@95: T3A = R1[WS(rs, 60)]; cannam@95: T3B = R1[WS(rs, 28)]; cannam@95: T3C = T3A - T3B; cannam@95: TaF = T3A + T3B; cannam@95: T3D = R1[WS(rs, 12)]; cannam@95: T3E = R1[WS(rs, 44)]; cannam@95: T3F = T3D - T3E; cannam@95: TaG = T3D + T3E; cannam@95: } cannam@95: T3G = FMA(KP923879532, T3C, KP382683432 * T3F); cannam@95: Tdc = TaF - TaG; cannam@95: T4o = FNMS(KP923879532, T3F, KP382683432 * T3C); cannam@95: } cannam@95: { cannam@95: E T3s, T3H, Tda, Tdd; cannam@95: T3s = T3k + T3r; cannam@95: T3H = T3z + T3G; cannam@95: T3I = T3s + T3H; cannam@95: T6H = T3s - T3H; cannam@95: Tda = Tai - Taj; cannam@95: Tdd = KP707106781 * (Tdb + Tdc); cannam@95: Tde = Tda + Tdd; cannam@95: TeH = Tda - Tdd; cannam@95: } cannam@95: { cannam@95: E Tdn, Tdo, T4n, T4q; cannam@95: Tdn = Tam - Tal; cannam@95: Tdo = KP707106781 * (Tdc - Tdb); cannam@95: Tdp = Tdn + Tdo; cannam@95: TeF = Tdo - Tdn; cannam@95: T4n = T4j - T4m; cannam@95: T4q = T4o - T4p; cannam@95: T4r = T4n + T4q; cannam@95: T6F = T4q - T4n; cannam@95: } cannam@95: { cannam@95: E T7S, T7T, Tak, Tan; cannam@95: T7S = T3k - T3r; cannam@95: T7T = T4p + T4o; cannam@95: T7U = T7S + T7T; cannam@95: T9l = T7S - T7T; cannam@95: Tak = Tai + Taj; cannam@95: Tan = Tal + Tam; cannam@95: Tao = Tak - Tan; cannam@95: TbW = Tak + Tan; cannam@95: } cannam@95: { cannam@95: E TaH, TaK, T83, T84; cannam@95: TaH = TaF + TaG; cannam@95: TaK = TaI + TaJ; cannam@95: TaL = TaH - TaK; cannam@95: TbX = TaK + TaH; cannam@95: T83 = T4m + T4j; cannam@95: T84 = T3G - T3z; cannam@95: T85 = T83 + T84; cannam@95: T9j = T84 - T83; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T1z, T2V, T1C, T39, T1G, T38, T1J, T2Y, T1O, T1R, T32, Td0, T3c, T1V, T1Y; cannam@95: E T35, Td1, T3b; cannam@95: { cannam@95: E T1x, T1y, T1A, T1B; cannam@95: T1x = R0[WS(rs, 63)]; cannam@95: T1y = R0[WS(rs, 31)]; cannam@95: T1z = T1x + T1y; cannam@95: T2V = T1x - T1y; cannam@95: T1A = R0[WS(rs, 15)]; cannam@95: T1B = R0[WS(rs, 47)]; cannam@95: T1C = T1A + T1B; cannam@95: T39 = T1A - T1B; cannam@95: } cannam@95: { cannam@95: E T1E, T1F, T2W, T1H, T1I, T2X; cannam@95: T1E = R0[WS(rs, 7)]; cannam@95: T1F = R0[WS(rs, 39)]; cannam@95: T2W = T1E - T1F; cannam@95: T1H = R0[WS(rs, 55)]; cannam@95: T1I = R0[WS(rs, 23)]; cannam@95: T2X = T1H - T1I; cannam@95: T1G = T1E + T1F; cannam@95: T38 = KP707106781 * (T2X - T2W); cannam@95: T1J = T1H + T1I; cannam@95: T2Y = KP707106781 * (T2W + T2X); cannam@95: } cannam@95: { cannam@95: E T30, T31, T33, T34; cannam@95: { cannam@95: E T1M, T1N, T1P, T1Q; cannam@95: T1M = R0[WS(rs, 3)]; cannam@95: T1N = R0[WS(rs, 35)]; cannam@95: T1O = T1M + T1N; cannam@95: T30 = T1M - T1N; cannam@95: T1P = R0[WS(rs, 19)]; cannam@95: T1Q = R0[WS(rs, 51)]; cannam@95: T1R = T1P + T1Q; cannam@95: T31 = T1P - T1Q; cannam@95: } cannam@95: T32 = FNMS(KP382683432, T31, KP923879532 * T30); cannam@95: Td0 = T1O - T1R; cannam@95: T3c = FMA(KP382683432, T30, KP923879532 * T31); cannam@95: { cannam@95: E T1T, T1U, T1W, T1X; cannam@95: T1T = R0[WS(rs, 59)]; cannam@95: T1U = R0[WS(rs, 27)]; cannam@95: T1V = T1T + T1U; cannam@95: T33 = T1T - T1U; cannam@95: T1W = R0[WS(rs, 11)]; cannam@95: T1X = R0[WS(rs, 43)]; cannam@95: T1Y = T1W + T1X; cannam@95: T34 = T1W - T1X; cannam@95: } cannam@95: T35 = FMA(KP923879532, T33, KP382683432 * T34); cannam@95: Td1 = T1V - T1Y; cannam@95: T3b = FNMS(KP923879532, T34, KP382683432 * T33); cannam@95: } cannam@95: { cannam@95: E T1D, T1K, TcZ, Td2; cannam@95: T1D = T1z + T1C; cannam@95: T1K = T1G + T1J; cannam@95: T1L = T1D + T1K; cannam@95: Tad = T1D - T1K; cannam@95: TcZ = T1z - T1C; cannam@95: Td2 = KP707106781 * (Td0 + Td1); cannam@95: Td3 = TcZ + Td2; cannam@95: Tew = TcZ - Td2; cannam@95: } cannam@95: { cannam@95: E Td4, Td5, T1S, T1Z; cannam@95: Td4 = T1J - T1G; cannam@95: Td5 = KP707106781 * (Td1 - Td0); cannam@95: Td6 = Td4 + Td5; cannam@95: Tex = Td5 - Td4; cannam@95: T1S = T1O + T1R; cannam@95: T1Z = T1V + T1Y; cannam@95: T20 = T1S + T1Z; cannam@95: Tae = T1Z - T1S; cannam@95: } cannam@95: { cannam@95: E T2Z, T36, T7J, T7K; cannam@95: T2Z = T2V + T2Y; cannam@95: T36 = T32 + T35; cannam@95: T37 = T2Z + T36; cannam@95: T6x = T2Z - T36; cannam@95: T7J = T2V - T2Y; cannam@95: T7K = T3c + T3b; cannam@95: T7L = T7J + T7K; cannam@95: T9a = T7J - T7K; cannam@95: } cannam@95: { cannam@95: E T7M, T7N, T3a, T3d; cannam@95: T7M = T39 + T38; cannam@95: T7N = T35 - T32; cannam@95: T7O = T7M + T7N; cannam@95: T9b = T7N - T7M; cannam@95: T3a = T38 - T39; cannam@95: T3d = T3b - T3c; cannam@95: T3e = T3a + T3d; cannam@95: T6w = T3d - T3a; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T3L, Tdf, T3X, Tar, T42, Tdi, T4e, Tay, T3S, Tdg, T3U, Tau, T49, Tdj, T4b; cannam@95: E TaB, Tdh, Tdk; cannam@95: { cannam@95: E T3J, T3K, Tap, T3V, T3W, Taq; cannam@95: T3J = R1[WS(rs, 2)]; cannam@95: T3K = R1[WS(rs, 34)]; cannam@95: Tap = T3J + T3K; cannam@95: T3V = R1[WS(rs, 18)]; cannam@95: T3W = R1[WS(rs, 50)]; cannam@95: Taq = T3V + T3W; cannam@95: T3L = T3J - T3K; cannam@95: Tdf = Tap - Taq; cannam@95: T3X = T3V - T3W; cannam@95: Tar = Tap + Taq; cannam@95: } cannam@95: { cannam@95: E T40, T41, Taw, T4c, T4d, Tax; cannam@95: T40 = R1[WS(rs, 62)]; cannam@95: T41 = R1[WS(rs, 30)]; cannam@95: Taw = T40 + T41; cannam@95: T4c = R1[WS(rs, 14)]; cannam@95: T4d = R1[WS(rs, 46)]; cannam@95: Tax = T4c + T4d; cannam@95: T42 = T40 - T41; cannam@95: Tdi = Taw - Tax; cannam@95: T4e = T4c - T4d; cannam@95: Tay = Taw + Tax; cannam@95: } cannam@95: { cannam@95: E T3O, Tas, T3R, Tat; cannam@95: { cannam@95: E T3M, T3N, T3P, T3Q; cannam@95: T3M = R1[WS(rs, 10)]; cannam@95: T3N = R1[WS(rs, 42)]; cannam@95: T3O = T3M - T3N; cannam@95: Tas = T3M + T3N; cannam@95: T3P = R1[WS(rs, 58)]; cannam@95: T3Q = R1[WS(rs, 26)]; cannam@95: T3R = T3P - T3Q; cannam@95: Tat = T3P + T3Q; cannam@95: } cannam@95: T3S = KP707106781 * (T3O + T3R); cannam@95: Tdg = Tat - Tas; cannam@95: T3U = KP707106781 * (T3R - T3O); cannam@95: Tau = Tas + Tat; cannam@95: } cannam@95: { cannam@95: E T45, Taz, T48, TaA; cannam@95: { cannam@95: E T43, T44, T46, T47; cannam@95: T43 = R1[WS(rs, 6)]; cannam@95: T44 = R1[WS(rs, 38)]; cannam@95: T45 = T43 - T44; cannam@95: Taz = T43 + T44; cannam@95: T46 = R1[WS(rs, 54)]; cannam@95: T47 = R1[WS(rs, 22)]; cannam@95: T48 = T46 - T47; cannam@95: TaA = T46 + T47; cannam@95: } cannam@95: T49 = KP707106781 * (T45 + T48); cannam@95: Tdj = TaA - Taz; cannam@95: T4b = KP707106781 * (T48 - T45); cannam@95: TaB = Taz + TaA; cannam@95: } cannam@95: TbZ = Tar + Tau; cannam@95: Tc0 = Tay + TaB; cannam@95: { cannam@95: E T3T, T3Y, Tdq, Tdr; cannam@95: T3T = T3L + T3S; cannam@95: T3Y = T3U - T3X; cannam@95: T3Z = FMA(KP980785280, T3T, KP195090322 * T3Y); cannam@95: T4s = FNMS(KP195090322, T3T, KP980785280 * T3Y); cannam@95: Tdq = FNMS(KP382683432, Tdf, KP923879532 * Tdg); cannam@95: Tdr = FMA(KP382683432, Tdi, KP923879532 * Tdj); cannam@95: Tds = Tdq + Tdr; cannam@95: TeI = Tdr - Tdq; cannam@95: } cannam@95: { cannam@95: E T4a, T4f, T7Y, T7Z; cannam@95: T4a = T42 + T49; cannam@95: T4f = T4b - T4e; cannam@95: T4g = FNMS(KP195090322, T4f, KP980785280 * T4a); cannam@95: T4t = FMA(KP195090322, T4a, KP980785280 * T4f); cannam@95: T7Y = T42 - T49; cannam@95: T7Z = T4e + T4b; cannam@95: T80 = FNMS(KP555570233, T7Z, KP831469612 * T7Y); cannam@95: T87 = FMA(KP555570233, T7Y, KP831469612 * T7Z); cannam@95: } cannam@95: Tdh = FMA(KP923879532, Tdf, KP382683432 * Tdg); cannam@95: Tdk = FNMS(KP382683432, Tdj, KP923879532 * Tdi); cannam@95: Tdl = Tdh + Tdk; cannam@95: TeE = Tdk - Tdh; cannam@95: { cannam@95: E T7V, T7W, Tav, TaC; cannam@95: T7V = T3L - T3S; cannam@95: T7W = T3X + T3U; cannam@95: T7X = FMA(KP831469612, T7V, KP555570233 * T7W); cannam@95: T86 = FNMS(KP555570233, T7V, KP831469612 * T7W); cannam@95: Tav = Tar - Tau; cannam@95: TaC = Tay - TaB; cannam@95: TaD = KP707106781 * (Tav + TaC); cannam@95: TaM = KP707106781 * (TaC - Tav); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T50, TdA, T5c, TaY, T5h, TdD, T5t, Tb5, T57, TdB, T59, Tb1, T5o, TdE, T5q; cannam@95: E Tb8, TdC, TdF; cannam@95: { cannam@95: E T4Y, T4Z, TaW, T5a, T5b, TaX; cannam@95: T4Y = R1[WS(rs, 1)]; cannam@95: T4Z = R1[WS(rs, 33)]; cannam@95: TaW = T4Y + T4Z; cannam@95: T5a = R1[WS(rs, 17)]; cannam@95: T5b = R1[WS(rs, 49)]; cannam@95: TaX = T5a + T5b; cannam@95: T50 = T4Y - T4Z; cannam@95: TdA = TaW - TaX; cannam@95: T5c = T5a - T5b; cannam@95: TaY = TaW + TaX; cannam@95: } cannam@95: { cannam@95: E T5f, T5g, Tb3, T5r, T5s, Tb4; cannam@95: T5f = R1[WS(rs, 61)]; cannam@95: T5g = R1[WS(rs, 29)]; cannam@95: Tb3 = T5f + T5g; cannam@95: T5r = R1[WS(rs, 13)]; cannam@95: T5s = R1[WS(rs, 45)]; cannam@95: Tb4 = T5r + T5s; cannam@95: T5h = T5f - T5g; cannam@95: TdD = Tb3 - Tb4; cannam@95: T5t = T5r - T5s; cannam@95: Tb5 = Tb3 + Tb4; cannam@95: } cannam@95: { cannam@95: E T53, TaZ, T56, Tb0; cannam@95: { cannam@95: E T51, T52, T54, T55; cannam@95: T51 = R1[WS(rs, 9)]; cannam@95: T52 = R1[WS(rs, 41)]; cannam@95: T53 = T51 - T52; cannam@95: TaZ = T51 + T52; cannam@95: T54 = R1[WS(rs, 57)]; cannam@95: T55 = R1[WS(rs, 25)]; cannam@95: T56 = T54 - T55; cannam@95: Tb0 = T54 + T55; cannam@95: } cannam@95: T57 = KP707106781 * (T53 + T56); cannam@95: TdB = Tb0 - TaZ; cannam@95: T59 = KP707106781 * (T56 - T53); cannam@95: Tb1 = TaZ + Tb0; cannam@95: } cannam@95: { cannam@95: E T5k, Tb6, T5n, Tb7; cannam@95: { cannam@95: E T5i, T5j, T5l, T5m; cannam@95: T5i = R1[WS(rs, 5)]; cannam@95: T5j = R1[WS(rs, 37)]; cannam@95: T5k = T5i - T5j; cannam@95: Tb6 = T5i + T5j; cannam@95: T5l = R1[WS(rs, 53)]; cannam@95: T5m = R1[WS(rs, 21)]; cannam@95: T5n = T5l - T5m; cannam@95: Tb7 = T5l + T5m; cannam@95: } cannam@95: T5o = KP707106781 * (T5k + T5n); cannam@95: TdE = Tb7 - Tb6; cannam@95: T5q = KP707106781 * (T5n - T5k); cannam@95: Tb8 = Tb6 + Tb7; cannam@95: } cannam@95: Tc6 = TaY + Tb1; cannam@95: Tc7 = Tb5 + Tb8; cannam@95: { cannam@95: E T58, T5d, TdL, TdM; cannam@95: T58 = T50 + T57; cannam@95: T5d = T59 - T5c; cannam@95: T5e = FMA(KP980785280, T58, KP195090322 * T5d); cannam@95: T5H = FNMS(KP195090322, T58, KP980785280 * T5d); cannam@95: TdL = FNMS(KP382683432, TdA, KP923879532 * TdB); cannam@95: TdM = FMA(KP382683432, TdD, KP923879532 * TdE); cannam@95: TdN = TdL + TdM; cannam@95: TeM = TdM - TdL; cannam@95: } cannam@95: { cannam@95: E T5p, T5u, T8h, T8i; cannam@95: T5p = T5h + T5o; cannam@95: T5u = T5q - T5t; cannam@95: T5v = FNMS(KP195090322, T5u, KP980785280 * T5p); cannam@95: T5I = FMA(KP195090322, T5p, KP980785280 * T5u); cannam@95: T8h = T5h - T5o; cannam@95: T8i = T5t + T5q; cannam@95: T8j = FNMS(KP555570233, T8i, KP831469612 * T8h); cannam@95: T8q = FMA(KP555570233, T8h, KP831469612 * T8i); cannam@95: } cannam@95: TdC = FMA(KP923879532, TdA, KP382683432 * TdB); cannam@95: TdF = FNMS(KP382683432, TdE, KP923879532 * TdD); cannam@95: TdG = TdC + TdF; cannam@95: TeO = TdF - TdC; cannam@95: { cannam@95: E T8e, T8f, Tb2, Tb9; cannam@95: T8e = T50 - T57; cannam@95: T8f = T5c + T59; cannam@95: T8g = FMA(KP831469612, T8e, KP555570233 * T8f); cannam@95: T8p = FNMS(KP555570233, T8e, KP831469612 * T8f); cannam@95: Tb2 = TaY - Tb1; cannam@95: Tb9 = Tb5 - Tb8; cannam@95: Tba = KP707106781 * (Tb2 + Tb9); cannam@95: Tbj = KP707106781 * (Tb9 - Tb2); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T11, TbV, Tc9, Tcf, T22, Tcb, Tc2, Tce; cannam@95: { cannam@95: E Tv, T10, Tc5, Tc8; cannam@95: Tv = Tf + Tu; cannam@95: T10 = TK + TZ; cannam@95: T11 = Tv + T10; cannam@95: TbV = Tv - T10; cannam@95: Tc5 = Tc3 + Tc4; cannam@95: Tc8 = Tc6 + Tc7; cannam@95: Tc9 = Tc5 - Tc8; cannam@95: Tcf = Tc5 + Tc8; cannam@95: } cannam@95: { cannam@95: E T1w, T21, TbY, Tc1; cannam@95: T1w = T1g + T1v; cannam@95: T21 = T1L + T20; cannam@95: T22 = T1w + T21; cannam@95: Tcb = T21 - T1w; cannam@95: TbY = TbW + TbX; cannam@95: Tc1 = TbZ + Tc0; cannam@95: Tc2 = TbY - Tc1; cannam@95: Tce = TbY + Tc1; cannam@95: } cannam@95: Cr[WS(csr, 32)] = T11 - T22; cannam@95: Ci[WS(csi, 32)] = Tcf - Tce; cannam@95: { cannam@95: E Tca, Tcc, Tcd, Tcg; cannam@95: Tca = KP707106781 * (Tc2 + Tc9); cannam@95: Cr[WS(csr, 48)] = TbV - Tca; cannam@95: Cr[WS(csr, 16)] = TbV + Tca; cannam@95: Tcc = KP707106781 * (Tc9 - Tc2); cannam@95: Ci[WS(csi, 16)] = Tcb + Tcc; cannam@95: Ci[WS(csi, 48)] = Tcc - Tcb; cannam@95: Tcd = T11 + T22; cannam@95: Tcg = Tce + Tcf; cannam@95: Cr[WS(csr, 64)] = Tcd - Tcg; cannam@95: Cr[0] = Tcd + Tcg; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tch, Tcu, Tck, Tct, Tco, Tcy, Tcr, Tcz, Tci, Tcj; cannam@95: Tch = Tf - Tu; cannam@95: Tcu = TZ - TK; cannam@95: Tci = T1g - T1v; cannam@95: Tcj = T1L - T20; cannam@95: Tck = KP707106781 * (Tci + Tcj); cannam@95: Tct = KP707106781 * (Tcj - Tci); cannam@95: { cannam@95: E Tcm, Tcn, Tcp, Tcq; cannam@95: Tcm = TbW - TbX; cannam@95: Tcn = Tc0 - TbZ; cannam@95: Tco = FMA(KP923879532, Tcm, KP382683432 * Tcn); cannam@95: Tcy = FNMS(KP382683432, Tcm, KP923879532 * Tcn); cannam@95: Tcp = Tc3 - Tc4; cannam@95: Tcq = Tc7 - Tc6; cannam@95: Tcr = FNMS(KP382683432, Tcq, KP923879532 * Tcp); cannam@95: Tcz = FMA(KP382683432, Tcp, KP923879532 * Tcq); cannam@95: } cannam@95: { cannam@95: E Tcl, Tcs, Tcx, TcA; cannam@95: Tcl = Tch + Tck; cannam@95: Tcs = Tco + Tcr; cannam@95: Cr[WS(csr, 56)] = Tcl - Tcs; cannam@95: Cr[WS(csr, 8)] = Tcl + Tcs; cannam@95: Tcx = Tcu + Tct; cannam@95: TcA = Tcy + Tcz; cannam@95: Ci[WS(csi, 8)] = Tcx + TcA; cannam@95: Ci[WS(csi, 56)] = TcA - Tcx; cannam@95: } cannam@95: { cannam@95: E Tcv, Tcw, TcB, TcC; cannam@95: Tcv = Tct - Tcu; cannam@95: Tcw = Tcr - Tco; cannam@95: Ci[WS(csi, 24)] = Tcv + Tcw; cannam@95: Ci[WS(csi, 40)] = Tcw - Tcv; cannam@95: TcB = Tch - Tck; cannam@95: TcC = Tcz - Tcy; cannam@95: Cr[WS(csr, 40)] = TcB - TcC; cannam@95: Cr[WS(csr, 24)] = TcB + TcC; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Ta9, TbB, Tbs, TbM, Tag, TbL, TbJ, TbR, TaO, Tbw, Tbp, TbC, TbG, TbQ, Tbl; cannam@95: E Tbx, Ta8, Tbr; cannam@95: Ta8 = KP707106781 * (Ta6 + Ta7); cannam@95: Ta9 = Ta5 + Ta8; cannam@95: TbB = Ta5 - Ta8; cannam@95: Tbr = KP707106781 * (Ta7 - Ta6); cannam@95: Tbs = Tbq + Tbr; cannam@95: TbM = Tbr - Tbq; cannam@95: { cannam@95: E Tac, Taf, TbH, TbI; cannam@95: Tac = FMA(KP923879532, Taa, KP382683432 * Tab); cannam@95: Taf = FNMS(KP382683432, Tae, KP923879532 * Tad); cannam@95: Tag = Tac + Taf; cannam@95: TbL = Taf - Tac; cannam@95: TbH = TaV - Tba; cannam@95: TbI = Tbj - Tbi; cannam@95: TbJ = FNMS(KP555570233, TbI, KP831469612 * TbH); cannam@95: TbR = FMA(KP555570233, TbH, KP831469612 * TbI); cannam@95: } cannam@95: { cannam@95: E TaE, TaN, Tbn, Tbo; cannam@95: TaE = Tao + TaD; cannam@95: TaN = TaL + TaM; cannam@95: TaO = FMA(KP980785280, TaE, KP195090322 * TaN); cannam@95: Tbw = FNMS(KP195090322, TaE, KP980785280 * TaN); cannam@95: Tbn = FNMS(KP382683432, Taa, KP923879532 * Tab); cannam@95: Tbo = FMA(KP382683432, Tad, KP923879532 * Tae); cannam@95: Tbp = Tbn + Tbo; cannam@95: TbC = Tbo - Tbn; cannam@95: } cannam@95: { cannam@95: E TbE, TbF, Tbb, Tbk; cannam@95: TbE = Tao - TaD; cannam@95: TbF = TaM - TaL; cannam@95: TbG = FMA(KP831469612, TbE, KP555570233 * TbF); cannam@95: TbQ = FNMS(KP555570233, TbE, KP831469612 * TbF); cannam@95: Tbb = TaV + Tba; cannam@95: Tbk = Tbi + Tbj; cannam@95: Tbl = FNMS(KP195090322, Tbk, KP980785280 * Tbb); cannam@95: Tbx = FMA(KP195090322, Tbb, KP980785280 * Tbk); cannam@95: } cannam@95: { cannam@95: E Tah, Tbm, Tbv, Tby; cannam@95: Tah = Ta9 + Tag; cannam@95: Tbm = TaO + Tbl; cannam@95: Cr[WS(csr, 60)] = Tah - Tbm; cannam@95: Cr[WS(csr, 4)] = Tah + Tbm; cannam@95: Tbv = Tbs + Tbp; cannam@95: Tby = Tbw + Tbx; cannam@95: Ci[WS(csi, 4)] = Tbv + Tby; cannam@95: Ci[WS(csi, 60)] = Tby - Tbv; cannam@95: } cannam@95: { cannam@95: E Tbt, Tbu, Tbz, TbA; cannam@95: Tbt = Tbp - Tbs; cannam@95: Tbu = Tbl - TaO; cannam@95: Ci[WS(csi, 28)] = Tbt + Tbu; cannam@95: Ci[WS(csi, 36)] = Tbu - Tbt; cannam@95: Tbz = Ta9 - Tag; cannam@95: TbA = Tbx - Tbw; cannam@95: Cr[WS(csr, 36)] = Tbz - TbA; cannam@95: Cr[WS(csr, 28)] = Tbz + TbA; cannam@95: } cannam@95: { cannam@95: E TbD, TbK, TbP, TbS; cannam@95: TbD = TbB + TbC; cannam@95: TbK = TbG + TbJ; cannam@95: Cr[WS(csr, 52)] = TbD - TbK; cannam@95: Cr[WS(csr, 12)] = TbD + TbK; cannam@95: TbP = TbM + TbL; cannam@95: TbS = TbQ + TbR; cannam@95: Ci[WS(csi, 12)] = TbP + TbS; cannam@95: Ci[WS(csi, 52)] = TbS - TbP; cannam@95: } cannam@95: { cannam@95: E TbN, TbO, TbT, TbU; cannam@95: TbN = TbL - TbM; cannam@95: TbO = TbJ - TbG; cannam@95: Ci[WS(csi, 20)] = TbN + TbO; cannam@95: Ci[WS(csi, 44)] = TbO - TbN; cannam@95: TbT = TbB - TbC; cannam@95: TbU = TbR - TbQ; cannam@95: Cr[WS(csr, 44)] = TbT - TbU; cannam@95: Cr[WS(csr, 20)] = TbT + TbU; cannam@95: } cannam@95: } cannam@95: { cannam@95: E Tev, Tf7, Tfc, Tfm, Tff, Tfn, TeC, Tfh, TeK, Tf2, TeV, Tf8, TeY, Tfi, TeR; cannam@95: E Tf3; cannam@95: { cannam@95: E Tet, Teu, Tfa, Tfb; cannam@95: Tet = TcD - TcG; cannam@95: Teu = TdY - TdX; cannam@95: Tev = Tet - Teu; cannam@95: Tf7 = Tet + Teu; cannam@95: Tfa = TeF + TeE; cannam@95: Tfb = TeH + TeI; cannam@95: Tfc = FMA(KP290284677, Tfa, KP956940335 * Tfb); cannam@95: Tfm = FNMS(KP290284677, Tfb, KP956940335 * Tfa); cannam@95: } cannam@95: { cannam@95: E Tfd, Tfe, Tey, TeB; cannam@95: Tfd = TeL + TeM; cannam@95: Tfe = TeP + TeO; cannam@95: Tff = FNMS(KP290284677, Tfe, KP956940335 * Tfd); cannam@95: Tfn = FMA(KP956940335, Tfe, KP290284677 * Tfd); cannam@95: Tey = FMA(KP555570233, Tew, KP831469612 * Tex); cannam@95: TeB = FNMS(KP555570233, TeA, KP831469612 * Tez); cannam@95: TeC = Tey - TeB; cannam@95: Tfh = TeB + Tey; cannam@95: } cannam@95: { cannam@95: E TeG, TeJ, TeT, TeU; cannam@95: TeG = TeE - TeF; cannam@95: TeJ = TeH - TeI; cannam@95: TeK = FMA(KP471396736, TeG, KP881921264 * TeJ); cannam@95: Tf2 = FNMS(KP471396736, TeJ, KP881921264 * TeG); cannam@95: TeT = FNMS(KP555570233, Tex, KP831469612 * Tew); cannam@95: TeU = FMA(KP831469612, TeA, KP555570233 * Tez); cannam@95: TeV = TeT - TeU; cannam@95: Tf8 = TeU + TeT; cannam@95: } cannam@95: { cannam@95: E TeW, TeX, TeN, TeQ; cannam@95: TeW = TcN - TcK; cannam@95: TeX = TdV - TdU; cannam@95: TeY = TeW - TeX; cannam@95: Tfi = TeX + TeW; cannam@95: TeN = TeL - TeM; cannam@95: TeQ = TeO - TeP; cannam@95: TeR = FNMS(KP471396736, TeQ, KP881921264 * TeN); cannam@95: Tf3 = FMA(KP881921264, TeQ, KP471396736 * TeN); cannam@95: } cannam@95: { cannam@95: E TeD, TeS, Tf1, Tf4; cannam@95: TeD = Tev + TeC; cannam@95: TeS = TeK + TeR; cannam@95: Cr[WS(csr, 54)] = TeD - TeS; cannam@95: Cr[WS(csr, 10)] = TeD + TeS; cannam@95: Tf1 = TeY + TeV; cannam@95: Tf4 = Tf2 + Tf3; cannam@95: Ci[WS(csi, 10)] = Tf1 + Tf4; cannam@95: Ci[WS(csi, 54)] = Tf4 - Tf1; cannam@95: } cannam@95: { cannam@95: E TeZ, Tf0, Tf5, Tf6; cannam@95: TeZ = TeV - TeY; cannam@95: Tf0 = TeR - TeK; cannam@95: Ci[WS(csi, 22)] = TeZ + Tf0; cannam@95: Ci[WS(csi, 42)] = Tf0 - TeZ; cannam@95: Tf5 = Tev - TeC; cannam@95: Tf6 = Tf3 - Tf2; cannam@95: Cr[WS(csr, 42)] = Tf5 - Tf6; cannam@95: Cr[WS(csr, 22)] = Tf5 + Tf6; cannam@95: } cannam@95: { cannam@95: E Tf9, Tfg, Tfl, Tfo; cannam@95: Tf9 = Tf7 + Tf8; cannam@95: Tfg = Tfc + Tff; cannam@95: Cr[WS(csr, 58)] = Tf9 - Tfg; cannam@95: Cr[WS(csr, 6)] = Tf9 + Tfg; cannam@95: Tfl = Tfi + Tfh; cannam@95: Tfo = Tfm + Tfn; cannam@95: Ci[WS(csi, 6)] = Tfl + Tfo; cannam@95: Ci[WS(csi, 58)] = Tfo - Tfl; cannam@95: } cannam@95: { cannam@95: E Tfj, Tfk, Tfp, Tfq; cannam@95: Tfj = Tfh - Tfi; cannam@95: Tfk = Tff - Tfc; cannam@95: Ci[WS(csi, 26)] = Tfj + Tfk; cannam@95: Ci[WS(csi, 38)] = Tfk - Tfj; cannam@95: Tfp = Tf7 - Tf8; cannam@95: Tfq = Tfn - Tfm; cannam@95: Cr[WS(csr, 38)] = Tfp - Tfq; cannam@95: Cr[WS(csr, 26)] = Tfp + Tfq; cannam@95: } cannam@95: } cannam@95: { cannam@95: E TcP, Te9, Tee, Teo, Teh, Tep, Td8, Tej, Tdu, Te4, TdT, Tea, Te0, Tek, TdP; cannam@95: E Te5; cannam@95: { cannam@95: E TcH, TcO, Tec, Ted; cannam@95: TcH = TcD + TcG; cannam@95: TcO = TcK + TcN; cannam@95: TcP = TcH + TcO; cannam@95: Te9 = TcH - TcO; cannam@95: Tec = Tde - Tdl; cannam@95: Ted = Tds - Tdp; cannam@95: Tee = FMA(KP773010453, Tec, KP634393284 * Ted); cannam@95: Teo = FNMS(KP634393284, Tec, KP773010453 * Ted); cannam@95: } cannam@95: { cannam@95: E Tef, Teg, TcY, Td7; cannam@95: Tef = Tdz - TdG; cannam@95: Teg = TdN - TdK; cannam@95: Teh = FNMS(KP634393284, Teg, KP773010453 * Tef); cannam@95: Tep = FMA(KP634393284, Tef, KP773010453 * Teg); cannam@95: TcY = FMA(KP980785280, TcU, KP195090322 * TcX); cannam@95: Td7 = FNMS(KP195090322, Td6, KP980785280 * Td3); cannam@95: Td8 = TcY + Td7; cannam@95: Tej = Td7 - TcY; cannam@95: } cannam@95: { cannam@95: E Tdm, Tdt, TdR, TdS; cannam@95: Tdm = Tde + Tdl; cannam@95: Tdt = Tdp + Tds; cannam@95: Tdu = FMA(KP995184726, Tdm, KP098017140 * Tdt); cannam@95: Te4 = FNMS(KP098017140, Tdm, KP995184726 * Tdt); cannam@95: TdR = FNMS(KP195090322, TcU, KP980785280 * TcX); cannam@95: TdS = FMA(KP195090322, Td3, KP980785280 * Td6); cannam@95: TdT = TdR + TdS; cannam@95: Tea = TdS - TdR; cannam@95: } cannam@95: { cannam@95: E TdW, TdZ, TdH, TdO; cannam@95: TdW = TdU + TdV; cannam@95: TdZ = TdX + TdY; cannam@95: Te0 = TdW + TdZ; cannam@95: Tek = TdZ - TdW; cannam@95: TdH = Tdz + TdG; cannam@95: TdO = TdK + TdN; cannam@95: TdP = FNMS(KP098017140, TdO, KP995184726 * TdH); cannam@95: Te5 = FMA(KP098017140, TdH, KP995184726 * TdO); cannam@95: } cannam@95: { cannam@95: E Td9, TdQ, Te3, Te6; cannam@95: Td9 = TcP + Td8; cannam@95: TdQ = Tdu + TdP; cannam@95: Cr[WS(csr, 62)] = Td9 - TdQ; cannam@95: Cr[WS(csr, 2)] = Td9 + TdQ; cannam@95: Te3 = Te0 + TdT; cannam@95: Te6 = Te4 + Te5; cannam@95: Ci[WS(csi, 2)] = Te3 + Te6; cannam@95: Ci[WS(csi, 62)] = Te6 - Te3; cannam@95: } cannam@95: { cannam@95: E Te1, Te2, Te7, Te8; cannam@95: Te1 = TdT - Te0; cannam@95: Te2 = TdP - Tdu; cannam@95: Ci[WS(csi, 30)] = Te1 + Te2; cannam@95: Ci[WS(csi, 34)] = Te2 - Te1; cannam@95: Te7 = TcP - Td8; cannam@95: Te8 = Te5 - Te4; cannam@95: Cr[WS(csr, 34)] = Te7 - Te8; cannam@95: Cr[WS(csr, 30)] = Te7 + Te8; cannam@95: } cannam@95: { cannam@95: E Teb, Tei, Ten, Teq; cannam@95: Teb = Te9 + Tea; cannam@95: Tei = Tee + Teh; cannam@95: Cr[WS(csr, 50)] = Teb - Tei; cannam@95: Cr[WS(csr, 14)] = Teb + Tei; cannam@95: Ten = Tek + Tej; cannam@95: Teq = Teo + Tep; cannam@95: Ci[WS(csi, 14)] = Ten + Teq; cannam@95: Ci[WS(csi, 50)] = Teq - Ten; cannam@95: } cannam@95: { cannam@95: E Tel, Tem, Ter, Tes; cannam@95: Tel = Tej - Tek; cannam@95: Tem = Teh - Tee; cannam@95: Ci[WS(csi, 18)] = Tel + Tem; cannam@95: Ci[WS(csi, 46)] = Tem - Tel; cannam@95: Ter = Te9 - Tea; cannam@95: Tes = Tep - Teo; cannam@95: Cr[WS(csr, 46)] = Ter - Tes; cannam@95: Cr[WS(csr, 18)] = Ter + Tes; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T6v, T77, T6C, T7h, T6Y, T7i, T6V, T78, T6R, T7n, T73, T7f, T6K, T7m, T72; cannam@95: E T7c; cannam@95: { cannam@95: E T6t, T6u, T6T, T6U; cannam@95: T6t = T27 - T2e; cannam@95: T6u = T5Y - T5X; cannam@95: T6v = T6t - T6u; cannam@95: T77 = T6t + T6u; cannam@95: { cannam@95: E T6y, T6B, T6W, T6X; cannam@95: T6y = FMA(KP773010453, T6w, KP634393284 * T6x); cannam@95: T6B = FNMS(KP634393284, T6A, KP773010453 * T6z); cannam@95: T6C = T6y - T6B; cannam@95: T7h = T6B + T6y; cannam@95: T6W = T2x - T2o; cannam@95: T6X = T5V - T5S; cannam@95: T6Y = T6W - T6X; cannam@95: T7i = T6X + T6W; cannam@95: } cannam@95: T6T = FNMS(KP634393284, T6w, KP773010453 * T6x); cannam@95: T6U = FMA(KP634393284, T6z, KP773010453 * T6A); cannam@95: T6V = T6T - T6U; cannam@95: T78 = T6U + T6T; cannam@95: { cannam@95: E T6N, T7d, T6Q, T7e, T6M, T6O; cannam@95: T6M = T5I - T5H; cannam@95: T6N = T6L - T6M; cannam@95: T7d = T6L + T6M; cannam@95: T6O = T5v - T5e; cannam@95: T6Q = T6O - T6P; cannam@95: T7e = T6P + T6O; cannam@95: T6R = FNMS(KP427555093, T6Q, KP903989293 * T6N); cannam@95: T7n = FMA(KP941544065, T7e, KP336889853 * T7d); cannam@95: T73 = FMA(KP903989293, T6Q, KP427555093 * T6N); cannam@95: T7f = FNMS(KP336889853, T7e, KP941544065 * T7d); cannam@95: } cannam@95: { cannam@95: E T6G, T7a, T6J, T7b, T6E, T6I; cannam@95: T6E = T4g - T3Z; cannam@95: T6G = T6E - T6F; cannam@95: T7a = T6F + T6E; cannam@95: T6I = T4t - T4s; cannam@95: T6J = T6H - T6I; cannam@95: T7b = T6H + T6I; cannam@95: T6K = FMA(KP427555093, T6G, KP903989293 * T6J); cannam@95: T7m = FNMS(KP336889853, T7b, KP941544065 * T7a); cannam@95: T72 = FNMS(KP427555093, T6J, KP903989293 * T6G); cannam@95: T7c = FMA(KP336889853, T7a, KP941544065 * T7b); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T6D, T6S, T71, T74; cannam@95: T6D = T6v + T6C; cannam@95: T6S = T6K + T6R; cannam@95: Cr[WS(csr, 55)] = T6D - T6S; cannam@95: Cr[WS(csr, 9)] = T6D + T6S; cannam@95: T71 = T6Y + T6V; cannam@95: T74 = T72 + T73; cannam@95: Ci[WS(csi, 9)] = T71 + T74; cannam@95: Ci[WS(csi, 55)] = T74 - T71; cannam@95: } cannam@95: { cannam@95: E T6Z, T70, T75, T76; cannam@95: T6Z = T6V - T6Y; cannam@95: T70 = T6R - T6K; cannam@95: Ci[WS(csi, 23)] = T6Z + T70; cannam@95: Ci[WS(csi, 41)] = T70 - T6Z; cannam@95: T75 = T6v - T6C; cannam@95: T76 = T73 - T72; cannam@95: Cr[WS(csr, 41)] = T75 - T76; cannam@95: Cr[WS(csr, 23)] = T75 + T76; cannam@95: } cannam@95: { cannam@95: E T79, T7g, T7l, T7o; cannam@95: T79 = T77 + T78; cannam@95: T7g = T7c + T7f; cannam@95: Cr[WS(csr, 57)] = T79 - T7g; cannam@95: Cr[WS(csr, 7)] = T79 + T7g; cannam@95: T7l = T7i + T7h; cannam@95: T7o = T7m + T7n; cannam@95: Ci[WS(csi, 7)] = T7l + T7o; cannam@95: Ci[WS(csi, 57)] = T7o - T7l; cannam@95: } cannam@95: { cannam@95: E T7j, T7k, T7p, T7q; cannam@95: T7j = T7h - T7i; cannam@95: T7k = T7f - T7c; cannam@95: Ci[WS(csi, 25)] = T7j + T7k; cannam@95: Ci[WS(csi, 39)] = T7k - T7j; cannam@95: T7p = T77 - T78; cannam@95: T7q = T7n - T7m; cannam@95: Cr[WS(csr, 39)] = T7p - T7q; cannam@95: Cr[WS(csr, 25)] = T7p + T7q; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T99, T9L, T9g, T9V, T9C, T9W, T9z, T9M, T9v, Ta1, T9H, T9T, T9o, Ta0, T9G; cannam@95: E T9Q; cannam@95: { cannam@95: E T97, T98, T9x, T9y; cannam@95: T97 = T7r - T7s; cannam@95: T98 = T8C - T8B; cannam@95: T99 = T97 - T98; cannam@95: T9L = T97 + T98; cannam@95: { cannam@95: E T9c, T9f, T9A, T9B; cannam@95: T9c = FMA(KP471396736, T9a, KP881921264 * T9b); cannam@95: T9f = FNMS(KP471396736, T9e, KP881921264 * T9d); cannam@95: T9g = T9c - T9f; cannam@95: T9V = T9f + T9c; cannam@95: T9A = T7z - T7w; cannam@95: T9B = T8z - T8y; cannam@95: T9C = T9A - T9B; cannam@95: T9W = T9B + T9A; cannam@95: } cannam@95: T9x = FNMS(KP471396736, T9b, KP881921264 * T9a); cannam@95: T9y = FMA(KP881921264, T9e, KP471396736 * T9d); cannam@95: T9z = T9x - T9y; cannam@95: T9M = T9y + T9x; cannam@95: { cannam@95: E T9r, T9R, T9u, T9S, T9q, T9s; cannam@95: T9q = T8q - T8p; cannam@95: T9r = T9p - T9q; cannam@95: T9R = T9p + T9q; cannam@95: T9s = T8j - T8g; cannam@95: T9u = T9s - T9t; cannam@95: T9S = T9t + T9s; cannam@95: T9v = FNMS(KP514102744, T9u, KP857728610 * T9r); cannam@95: Ta1 = FMA(KP970031253, T9S, KP242980179 * T9R); cannam@95: T9H = FMA(KP857728610, T9u, KP514102744 * T9r); cannam@95: T9T = FNMS(KP242980179, T9S, KP970031253 * T9R); cannam@95: } cannam@95: { cannam@95: E T9k, T9O, T9n, T9P, T9i, T9m; cannam@95: T9i = T80 - T7X; cannam@95: T9k = T9i - T9j; cannam@95: T9O = T9j + T9i; cannam@95: T9m = T87 - T86; cannam@95: T9n = T9l - T9m; cannam@95: T9P = T9l + T9m; cannam@95: T9o = FMA(KP514102744, T9k, KP857728610 * T9n); cannam@95: Ta0 = FNMS(KP242980179, T9P, KP970031253 * T9O); cannam@95: T9G = FNMS(KP514102744, T9n, KP857728610 * T9k); cannam@95: T9Q = FMA(KP242980179, T9O, KP970031253 * T9P); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T9h, T9w, T9F, T9I; cannam@95: T9h = T99 + T9g; cannam@95: T9w = T9o + T9v; cannam@95: Cr[WS(csr, 53)] = T9h - T9w; cannam@95: Cr[WS(csr, 11)] = T9h + T9w; cannam@95: T9F = T9C + T9z; cannam@95: T9I = T9G + T9H; cannam@95: Ci[WS(csi, 11)] = T9F + T9I; cannam@95: Ci[WS(csi, 53)] = T9I - T9F; cannam@95: } cannam@95: { cannam@95: E T9D, T9E, T9J, T9K; cannam@95: T9D = T9z - T9C; cannam@95: T9E = T9v - T9o; cannam@95: Ci[WS(csi, 21)] = T9D + T9E; cannam@95: Ci[WS(csi, 43)] = T9E - T9D; cannam@95: T9J = T99 - T9g; cannam@95: T9K = T9H - T9G; cannam@95: Cr[WS(csr, 43)] = T9J - T9K; cannam@95: Cr[WS(csr, 21)] = T9J + T9K; cannam@95: } cannam@95: { cannam@95: E T9N, T9U, T9Z, Ta2; cannam@95: T9N = T9L + T9M; cannam@95: T9U = T9Q + T9T; cannam@95: Cr[WS(csr, 59)] = T9N - T9U; cannam@95: Cr[WS(csr, 5)] = T9N + T9U; cannam@95: T9Z = T9W + T9V; cannam@95: Ta2 = Ta0 + Ta1; cannam@95: Ci[WS(csi, 5)] = T9Z + Ta2; cannam@95: Ci[WS(csi, 59)] = Ta2 - T9Z; cannam@95: } cannam@95: { cannam@95: E T9X, T9Y, Ta3, Ta4; cannam@95: T9X = T9V - T9W; cannam@95: T9Y = T9T - T9Q; cannam@95: Ci[WS(csi, 27)] = T9X + T9Y; cannam@95: Ci[WS(csi, 37)] = T9Y - T9X; cannam@95: Ta3 = T9L - T9M; cannam@95: Ta4 = Ta1 - Ta0; cannam@95: Cr[WS(csr, 37)] = Ta3 - Ta4; cannam@95: Cr[WS(csr, 27)] = Ta3 + Ta4; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T2z, T69, T3g, T6j, T60, T6k, T5P, T6a, T5L, T6p, T65, T6h, T4w, T6o, T64; cannam@95: E T6e; cannam@95: { cannam@95: E T2f, T2y, T5N, T5O; cannam@95: T2f = T27 + T2e; cannam@95: T2y = T2o + T2x; cannam@95: T2z = T2f + T2y; cannam@95: T69 = T2f - T2y; cannam@95: { cannam@95: E T2U, T3f, T5W, T5Z; cannam@95: T2U = FMA(KP098017140, T2M, KP995184726 * T2T); cannam@95: T3f = FNMS(KP098017140, T3e, KP995184726 * T37); cannam@95: T3g = T2U + T3f; cannam@95: T6j = T3f - T2U; cannam@95: T5W = T5S + T5V; cannam@95: T5Z = T5X + T5Y; cannam@95: T60 = T5W + T5Z; cannam@95: T6k = T5Z - T5W; cannam@95: } cannam@95: T5N = FNMS(KP098017140, T2T, KP995184726 * T2M); cannam@95: T5O = FMA(KP995184726, T3e, KP098017140 * T37); cannam@95: T5P = T5N + T5O; cannam@95: T6a = T5O - T5N; cannam@95: { cannam@95: E T5x, T6f, T5K, T6g, T5w, T5J; cannam@95: T5w = T5e + T5v; cannam@95: T5x = T4X + T5w; cannam@95: T6f = T4X - T5w; cannam@95: T5J = T5H + T5I; cannam@95: T5K = T5G + T5J; cannam@95: T6g = T5J - T5G; cannam@95: T5L = FNMS(KP049067674, T5K, KP998795456 * T5x); cannam@95: T6p = FMA(KP671558954, T6f, KP740951125 * T6g); cannam@95: T65 = FMA(KP049067674, T5x, KP998795456 * T5K); cannam@95: T6h = FNMS(KP671558954, T6g, KP740951125 * T6f); cannam@95: } cannam@95: { cannam@95: E T4i, T6c, T4v, T6d, T4h, T4u; cannam@95: T4h = T3Z + T4g; cannam@95: T4i = T3I + T4h; cannam@95: T6c = T3I - T4h; cannam@95: T4u = T4s + T4t; cannam@95: T4v = T4r + T4u; cannam@95: T6d = T4u - T4r; cannam@95: T4w = FMA(KP998795456, T4i, KP049067674 * T4v); cannam@95: T6o = FNMS(KP671558954, T6c, KP740951125 * T6d); cannam@95: T64 = FNMS(KP049067674, T4i, KP998795456 * T4v); cannam@95: T6e = FMA(KP740951125, T6c, KP671558954 * T6d); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T3h, T5M, T63, T66; cannam@95: T3h = T2z + T3g; cannam@95: T5M = T4w + T5L; cannam@95: Cr[WS(csr, 63)] = T3h - T5M; cannam@95: Cr[WS(csr, 1)] = T3h + T5M; cannam@95: T63 = T60 + T5P; cannam@95: T66 = T64 + T65; cannam@95: Ci[WS(csi, 1)] = T63 + T66; cannam@95: Ci[WS(csi, 63)] = T66 - T63; cannam@95: } cannam@95: { cannam@95: E T61, T62, T67, T68; cannam@95: T61 = T5P - T60; cannam@95: T62 = T5L - T4w; cannam@95: Ci[WS(csi, 31)] = T61 + T62; cannam@95: Ci[WS(csi, 33)] = T62 - T61; cannam@95: T67 = T2z - T3g; cannam@95: T68 = T65 - T64; cannam@95: Cr[WS(csr, 33)] = T67 - T68; cannam@95: Cr[WS(csr, 31)] = T67 + T68; cannam@95: } cannam@95: { cannam@95: E T6b, T6i, T6n, T6q; cannam@95: T6b = T69 + T6a; cannam@95: T6i = T6e + T6h; cannam@95: Cr[WS(csr, 49)] = T6b - T6i; cannam@95: Cr[WS(csr, 15)] = T6b + T6i; cannam@95: T6n = T6k + T6j; cannam@95: T6q = T6o + T6p; cannam@95: Ci[WS(csi, 15)] = T6n + T6q; cannam@95: Ci[WS(csi, 49)] = T6q - T6n; cannam@95: } cannam@95: { cannam@95: E T6l, T6m, T6r, T6s; cannam@95: T6l = T6j - T6k; cannam@95: T6m = T6h - T6e; cannam@95: Ci[WS(csi, 17)] = T6l + T6m; cannam@95: Ci[WS(csi, 47)] = T6m - T6l; cannam@95: T6r = T69 - T6a; cannam@95: T6s = T6p - T6o; cannam@95: Cr[WS(csr, 47)] = T6r - T6s; cannam@95: Cr[WS(csr, 17)] = T6r + T6s; cannam@95: } cannam@95: } cannam@95: { cannam@95: E T7B, T8N, T7Q, T8X, T8E, T8Y, T8x, T8O, T8t, T93, T8J, T8V, T8a, T92, T8I; cannam@95: E T8S; cannam@95: { cannam@95: E T7t, T7A, T8v, T8w; cannam@95: T7t = T7r + T7s; cannam@95: T7A = T7w + T7z; cannam@95: T7B = T7t + T7A; cannam@95: T8N = T7t - T7A; cannam@95: { cannam@95: E T7I, T7P, T8A, T8D; cannam@95: T7I = FMA(KP956940335, T7E, KP290284677 * T7H); cannam@95: T7P = FNMS(KP290284677, T7O, KP956940335 * T7L); cannam@95: T7Q = T7I + T7P; cannam@95: T8X = T7P - T7I; cannam@95: T8A = T8y + T8z; cannam@95: T8D = T8B + T8C; cannam@95: T8E = T8A + T8D; cannam@95: T8Y = T8D - T8A; cannam@95: } cannam@95: T8v = FNMS(KP290284677, T7E, KP956940335 * T7H); cannam@95: T8w = FMA(KP290284677, T7L, KP956940335 * T7O); cannam@95: T8x = T8v + T8w; cannam@95: T8O = T8w - T8v; cannam@95: { cannam@95: E T8l, T8T, T8s, T8U, T8k, T8r; cannam@95: T8k = T8g + T8j; cannam@95: T8l = T8d + T8k; cannam@95: T8T = T8d - T8k; cannam@95: T8r = T8p + T8q; cannam@95: T8s = T8o + T8r; cannam@95: T8U = T8r - T8o; cannam@95: T8t = FNMS(KP146730474, T8s, KP989176509 * T8l); cannam@95: T93 = FMA(KP595699304, T8T, KP803207531 * T8U); cannam@95: T8J = FMA(KP146730474, T8l, KP989176509 * T8s); cannam@95: T8V = FNMS(KP595699304, T8U, KP803207531 * T8T); cannam@95: } cannam@95: { cannam@95: E T82, T8Q, T89, T8R, T81, T88; cannam@95: T81 = T7X + T80; cannam@95: T82 = T7U + T81; cannam@95: T8Q = T7U - T81; cannam@95: T88 = T86 + T87; cannam@95: T89 = T85 + T88; cannam@95: T8R = T88 - T85; cannam@95: T8a = FMA(KP989176509, T82, KP146730474 * T89); cannam@95: T92 = FNMS(KP595699304, T8Q, KP803207531 * T8R); cannam@95: T8I = FNMS(KP146730474, T82, KP989176509 * T89); cannam@95: T8S = FMA(KP803207531, T8Q, KP595699304 * T8R); cannam@95: } cannam@95: } cannam@95: { cannam@95: E T7R, T8u, T8H, T8K; cannam@95: T7R = T7B + T7Q; cannam@95: T8u = T8a + T8t; cannam@95: Cr[WS(csr, 61)] = T7R - T8u; cannam@95: Cr[WS(csr, 3)] = T7R + T8u; cannam@95: T8H = T8E + T8x; cannam@95: T8K = T8I + T8J; cannam@95: Ci[WS(csi, 3)] = T8H + T8K; cannam@95: Ci[WS(csi, 61)] = T8K - T8H; cannam@95: } cannam@95: { cannam@95: E T8F, T8G, T8L, T8M; cannam@95: T8F = T8x - T8E; cannam@95: T8G = T8t - T8a; cannam@95: Ci[WS(csi, 29)] = T8F + T8G; cannam@95: Ci[WS(csi, 35)] = T8G - T8F; cannam@95: T8L = T7B - T7Q; cannam@95: T8M = T8J - T8I; cannam@95: Cr[WS(csr, 35)] = T8L - T8M; cannam@95: Cr[WS(csr, 29)] = T8L + T8M; cannam@95: } cannam@95: { cannam@95: E T8P, T8W, T91, T94; cannam@95: T8P = T8N + T8O; cannam@95: T8W = T8S + T8V; cannam@95: Cr[WS(csr, 51)] = T8P - T8W; cannam@95: Cr[WS(csr, 13)] = T8P + T8W; cannam@95: T91 = T8Y + T8X; cannam@95: T94 = T92 + T93; cannam@95: Ci[WS(csi, 13)] = T91 + T94; cannam@95: Ci[WS(csi, 51)] = T94 - T91; cannam@95: } cannam@95: { cannam@95: E T8Z, T90, T95, T96; cannam@95: T8Z = T8X - T8Y; cannam@95: T90 = T8V - T8S; cannam@95: Ci[WS(csi, 19)] = T8Z + T90; cannam@95: Ci[WS(csi, 45)] = T90 - T8Z; cannam@95: T95 = T8N - T8O; cannam@95: T96 = T93 - T92; cannam@95: Cr[WS(csr, 45)] = T95 - T96; cannam@95: Cr[WS(csr, 19)] = T95 + T96; cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: } cannam@95: cannam@95: static const kr2c_desc desc = { 128, "r2cf_128", {812, 186, 144, 0}, &GENUS }; cannam@95: cannam@95: void X(codelet_r2cf_128) (planner *p) { cannam@95: X(kr2c_register) (p, r2cf_128, &desc); cannam@95: } cannam@95: cannam@95: #endif /* HAVE_FMA */