cannam@167: /* cannam@167: * Copyright (c) 2003, 2007-14 Matteo Frigo cannam@167: * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology cannam@167: * cannam@167: * This program is free software; you can redistribute it and/or modify cannam@167: * it under the terms of the GNU General Public License as published by cannam@167: * the Free Software Foundation; either version 2 of the License, or cannam@167: * (at your option) any later version. cannam@167: * cannam@167: * This program is distributed in the hope that it will be useful, cannam@167: * but WITHOUT ANY WARRANTY; without even the implied warranty of cannam@167: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the cannam@167: * GNU General Public License for more details. cannam@167: * cannam@167: * You should have received a copy of the GNU General Public License cannam@167: * along with this program; if not, write to the Free Software cannam@167: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA cannam@167: * cannam@167: */ cannam@167: cannam@167: /* This file was automatically generated --- DO NOT EDIT */ cannam@167: /* Generated on Thu May 24 08:06:26 EDT 2018 */ cannam@167: cannam@167: #include "rdft/codelet-rdft.h" cannam@167: cannam@167: #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA) cannam@167: cannam@167: /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include rdft/scalar/r2cf.h */ cannam@167: cannam@167: /* cannam@167: * This function contains 34 FP additions, 14 FP multiplications, cannam@167: * (or, 24 additions, 4 multiplications, 10 fused multiply/add), cannam@167: * 26 stack variables, 4 constants, and 20 memory accesses cannam@167: */ cannam@167: #include "rdft/scalar/r2cf.h" cannam@167: cannam@167: static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@167: { cannam@167: DK(KP559016994, +0.559016994374947424102293417182819058860154590); cannam@167: DK(KP250000000, +0.250000000000000000000000000000000000000000000); cannam@167: DK(KP618033988, +0.618033988749894848204586834365638117720309180); cannam@167: DK(KP951056516, +0.951056516295153572116439333379382143405698634); cannam@167: { cannam@167: INT i; cannam@167: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) { cannam@167: E T3, Tt, Td, Tn, Tg, To, Th, Tv, T6, Tq, T9, Tr, Ta, Tu, T1; cannam@167: E T2; cannam@167: T1 = R0[0]; cannam@167: T2 = R1[WS(rs, 2)]; cannam@167: T3 = T1 - T2; cannam@167: Tt = T1 + T2; cannam@167: { cannam@167: E Tb, Tc, Te, Tf; cannam@167: Tb = R0[WS(rs, 2)]; cannam@167: Tc = R1[WS(rs, 4)]; cannam@167: Td = Tb - Tc; cannam@167: Tn = Tb + Tc; cannam@167: Te = R0[WS(rs, 3)]; cannam@167: Tf = R1[0]; cannam@167: Tg = Te - Tf; cannam@167: To = Te + Tf; cannam@167: } cannam@167: Th = Td + Tg; cannam@167: Tv = Tn + To; cannam@167: { cannam@167: E T4, T5, T7, T8; cannam@167: T4 = R0[WS(rs, 1)]; cannam@167: T5 = R1[WS(rs, 3)]; cannam@167: T6 = T4 - T5; cannam@167: Tq = T4 + T5; cannam@167: T7 = R0[WS(rs, 4)]; cannam@167: T8 = R1[WS(rs, 1)]; cannam@167: T9 = T7 - T8; cannam@167: Tr = T7 + T8; cannam@167: } cannam@167: Ta = T6 + T9; cannam@167: Tu = Tq + Tr; cannam@167: { cannam@167: E Tl, Tm, Tk, Ti, Tj; cannam@167: Tl = T6 - T9; cannam@167: Tm = Tg - Td; cannam@167: Ci[WS(csi, 1)] = -(KP951056516 * (FNMS(KP618033988, Tm, Tl))); cannam@167: Ci[WS(csi, 3)] = KP951056516 * (FMA(KP618033988, Tl, Tm)); cannam@167: Tk = Ta - Th; cannam@167: Ti = Ta + Th; cannam@167: Tj = FNMS(KP250000000, Ti, T3); cannam@167: Cr[WS(csr, 1)] = FMA(KP559016994, Tk, Tj); cannam@167: Cr[WS(csr, 5)] = T3 + Ti; cannam@167: Cr[WS(csr, 3)] = FNMS(KP559016994, Tk, Tj); cannam@167: } cannam@167: { cannam@167: E Tp, Ts, Ty, Tw, Tx; cannam@167: Tp = Tn - To; cannam@167: Ts = Tq - Tr; cannam@167: Ci[WS(csi, 2)] = KP951056516 * (FNMS(KP618033988, Ts, Tp)); cannam@167: Ci[WS(csi, 4)] = KP951056516 * (FMA(KP618033988, Tp, Ts)); cannam@167: Ty = Tu - Tv; cannam@167: Tw = Tu + Tv; cannam@167: Tx = FNMS(KP250000000, Tw, Tt); cannam@167: Cr[WS(csr, 2)] = FNMS(KP559016994, Ty, Tx); cannam@167: Cr[0] = Tt + Tw; cannam@167: Cr[WS(csr, 4)] = FMA(KP559016994, Ty, Tx); cannam@167: } cannam@167: } cannam@167: } cannam@167: } cannam@167: cannam@167: static const kr2c_desc desc = { 10, "r2cf_10", {24, 4, 10, 0}, &GENUS }; cannam@167: cannam@167: void X(codelet_r2cf_10) (planner *p) { cannam@167: X(kr2c_register) (p, r2cf_10, &desc); cannam@167: } cannam@167: cannam@167: #else cannam@167: cannam@167: /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include rdft/scalar/r2cf.h */ cannam@167: cannam@167: /* cannam@167: * This function contains 34 FP additions, 12 FP multiplications, cannam@167: * (or, 28 additions, 6 multiplications, 6 fused multiply/add), cannam@167: * 26 stack variables, 4 constants, and 20 memory accesses cannam@167: */ cannam@167: #include "rdft/scalar/r2cf.h" cannam@167: cannam@167: static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@167: { cannam@167: DK(KP250000000, +0.250000000000000000000000000000000000000000000); cannam@167: DK(KP559016994, +0.559016994374947424102293417182819058860154590); cannam@167: DK(KP951056516, +0.951056516295153572116439333379382143405698634); cannam@167: DK(KP587785252, +0.587785252292473129168705954639072768597652438); cannam@167: { cannam@167: INT i; cannam@167: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) { cannam@167: E Ti, Tt, Ta, Tn, Td, To, Te, Tv, T3, Tq, T6, Tr, T7, Tu, Tg; cannam@167: E Th; cannam@167: Tg = R0[0]; cannam@167: Th = R1[WS(rs, 2)]; cannam@167: Ti = Tg - Th; cannam@167: Tt = Tg + Th; cannam@167: { cannam@167: E T8, T9, Tb, Tc; cannam@167: T8 = R0[WS(rs, 2)]; cannam@167: T9 = R1[WS(rs, 4)]; cannam@167: Ta = T8 - T9; cannam@167: Tn = T8 + T9; cannam@167: Tb = R0[WS(rs, 3)]; cannam@167: Tc = R1[0]; cannam@167: Td = Tb - Tc; cannam@167: To = Tb + Tc; cannam@167: } cannam@167: Te = Ta + Td; cannam@167: Tv = Tn + To; cannam@167: { cannam@167: E T1, T2, T4, T5; cannam@167: T1 = R0[WS(rs, 1)]; cannam@167: T2 = R1[WS(rs, 3)]; cannam@167: T3 = T1 - T2; cannam@167: Tq = T1 + T2; cannam@167: T4 = R0[WS(rs, 4)]; cannam@167: T5 = R1[WS(rs, 1)]; cannam@167: T6 = T4 - T5; cannam@167: Tr = T4 + T5; cannam@167: } cannam@167: T7 = T3 + T6; cannam@167: Tu = Tq + Tr; cannam@167: { cannam@167: E Tl, Tm, Tf, Tj, Tk; cannam@167: Tl = Td - Ta; cannam@167: Tm = T3 - T6; cannam@167: Ci[WS(csi, 1)] = FNMS(KP951056516, Tm, KP587785252 * Tl); cannam@167: Ci[WS(csi, 3)] = FMA(KP587785252, Tm, KP951056516 * Tl); cannam@167: Tf = KP559016994 * (T7 - Te); cannam@167: Tj = T7 + Te; cannam@167: Tk = FNMS(KP250000000, Tj, Ti); cannam@167: Cr[WS(csr, 1)] = Tf + Tk; cannam@167: Cr[WS(csr, 5)] = Ti + Tj; cannam@167: Cr[WS(csr, 3)] = Tk - Tf; cannam@167: } cannam@167: { cannam@167: E Tp, Ts, Ty, Tw, Tx; cannam@167: Tp = Tn - To; cannam@167: Ts = Tq - Tr; cannam@167: Ci[WS(csi, 2)] = FNMS(KP587785252, Ts, KP951056516 * Tp); cannam@167: Ci[WS(csi, 4)] = FMA(KP951056516, Ts, KP587785252 * Tp); cannam@167: Ty = KP559016994 * (Tu - Tv); cannam@167: Tw = Tu + Tv; cannam@167: Tx = FNMS(KP250000000, Tw, Tt); cannam@167: Cr[WS(csr, 2)] = Tx - Ty; cannam@167: Cr[0] = Tt + Tw; cannam@167: Cr[WS(csr, 4)] = Ty + Tx; cannam@167: } cannam@167: } cannam@167: } cannam@167: } cannam@167: cannam@167: static const kr2c_desc desc = { 10, "r2cf_10", {28, 6, 6, 0}, &GENUS }; cannam@167: cannam@167: void X(codelet_r2cf_10) (planner *p) { cannam@167: X(kr2c_register) (p, r2cf_10, &desc); cannam@167: } cannam@167: cannam@167: #endif