Chris@19: /* Chris@19: * Copyright (c) 2003, 2007-14 Matteo Frigo Chris@19: * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology Chris@19: * Chris@19: * This program is free software; you can redistribute it and/or modify Chris@19: * it under the terms of the GNU General Public License as published by Chris@19: * the Free Software Foundation; either version 2 of the License, or Chris@19: * (at your option) any later version. Chris@19: * Chris@19: * This program is distributed in the hope that it will be useful, Chris@19: * but WITHOUT ANY WARRANTY; without even the implied warranty of Chris@19: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the Chris@19: * GNU General Public License for more details. Chris@19: * Chris@19: * You should have received a copy of the GNU General Public License Chris@19: * along with this program; if not, write to the Free Software Chris@19: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Chris@19: * Chris@19: */ Chris@19: Chris@19: /* This file was automatically generated --- DO NOT EDIT */ Chris@19: /* Generated on Tue Mar 4 13:50:24 EST 2014 */ Chris@19: Chris@19: #include "codelet-rdft.h" Chris@19: Chris@19: #ifdef HAVE_FMA Chris@19: Chris@19: /* Generated by: ../../../genfft/gen_r2cb.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -sign 1 -n 9 -name r2cb_9 -include r2cb.h */ Chris@19: Chris@19: /* Chris@19: * This function contains 32 FP additions, 24 FP multiplications, Chris@19: * (or, 8 additions, 0 multiplications, 24 fused multiply/add), Chris@19: * 40 stack variables, 12 constants, and 18 memory accesses Chris@19: */ Chris@19: #include "r2cb.h" Chris@19: Chris@19: static void r2cb_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) Chris@19: { Chris@19: DK(KP1_326827896, +1.326827896337876792410842639271782594433726619); Chris@19: DK(KP1_705737063, +1.705737063904886419256501927880148143872040591); Chris@19: DK(KP766044443, +0.766044443118978035202392650555416673935832457); Chris@19: DK(KP1_532088886, +1.532088886237956070404785301110833347871664914); Chris@19: DK(KP984807753, +0.984807753012208059366743024589523013670643252); Chris@19: DK(KP1_969615506, +1.969615506024416118733486049179046027341286503); Chris@19: DK(KP839099631, +0.839099631177280011763127298123181364687434283); Chris@19: DK(KP176326980, +0.176326980708464973471090386868618986121633062); Chris@19: DK(KP866025403, +0.866025403784438646763723170752936183471402627); Chris@19: DK(KP500000000, +0.500000000000000000000000000000000000000000000); Chris@19: DK(KP1_732050807, +1.732050807568877293527446341505872366942805254); Chris@19: DK(KP2_000000000, +2.000000000000000000000000000000000000000000000); Chris@19: { Chris@19: INT i; Chris@19: for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(36, rs), MAKE_VOLATILE_STRIDE(36, csr), MAKE_VOLATILE_STRIDE(36, csi)) { Chris@19: E T4, Th, T3, Tb, Tp, Tk, T7, Tf, Ti, Ta, T1, T2; Chris@19: Ta = Ci[WS(csi, 3)]; Chris@19: T1 = Cr[0]; Chris@19: T2 = Cr[WS(csr, 3)]; Chris@19: T4 = Cr[WS(csr, 1)]; Chris@19: Th = Ci[WS(csi, 1)]; Chris@19: { Chris@19: E T5, T9, T6, Td, Te; Chris@19: T5 = Cr[WS(csr, 4)]; Chris@19: T9 = T1 - T2; Chris@19: T3 = FMA(KP2_000000000, T2, T1); Chris@19: T6 = Cr[WS(csr, 2)]; Chris@19: Td = Ci[WS(csi, 4)]; Chris@19: Te = Ci[WS(csi, 2)]; Chris@19: Tb = FNMS(KP1_732050807, Ta, T9); Chris@19: Tp = FMA(KP1_732050807, Ta, T9); Chris@19: Tk = T6 - T5; Chris@19: T7 = T5 + T6; Chris@19: Tf = Td + Te; Chris@19: Ti = Td - Te; Chris@19: } Chris@19: { Chris@19: E Tu, To, Tt, Tn, Tc, T8; Chris@19: Tc = FNMS(KP500000000, T7, T4); Chris@19: T8 = T4 + T7; Chris@19: { Chris@19: E Tw, Tj, Tr, Tg, Tv; Chris@19: Tw = Ti + Th; Chris@19: Tj = FNMS(KP500000000, Ti, Th); Chris@19: Tr = FMA(KP866025403, Tf, Tc); Chris@19: Tg = FNMS(KP866025403, Tf, Tc); Chris@19: Tv = T3 - T8; Chris@19: R0[0] = FMA(KP2_000000000, T8, T3); Chris@19: { Chris@19: E Tq, Tl, Ts, Tm; Chris@19: Tq = FMA(KP866025403, Tk, Tj); Chris@19: Tl = FNMS(KP866025403, Tk, Tj); Chris@19: R0[WS(rs, 3)] = FMA(KP1_732050807, Tw, Tv); Chris@19: R1[WS(rs, 1)] = FNMS(KP1_732050807, Tw, Tv); Chris@19: Ts = FNMS(KP176326980, Tr, Tq); Chris@19: Tu = FMA(KP176326980, Tq, Tr); Chris@19: Tm = FNMS(KP839099631, Tl, Tg); Chris@19: To = FMA(KP839099631, Tg, Tl); Chris@19: R0[WS(rs, 1)] = FNMS(KP1_969615506, Ts, Tp); Chris@19: Tt = FMA(KP984807753, Ts, Tp); Chris@19: R1[0] = FMA(KP1_532088886, Tm, Tb); Chris@19: Tn = FNMS(KP766044443, Tm, Tb); Chris@19: } Chris@19: } Chris@19: R1[WS(rs, 2)] = FNMS(KP1_705737063, Tu, Tt); Chris@19: R0[WS(rs, 4)] = FMA(KP1_705737063, Tu, Tt); Chris@19: R0[WS(rs, 2)] = FNMS(KP1_326827896, To, Tn); Chris@19: R1[WS(rs, 3)] = FMA(KP1_326827896, To, Tn); Chris@19: } Chris@19: } Chris@19: } Chris@19: } Chris@19: Chris@19: static const kr2c_desc desc = { 9, "r2cb_9", {8, 0, 24, 0}, &GENUS }; Chris@19: Chris@19: void X(codelet_r2cb_9) (planner *p) { Chris@19: X(kr2c_register) (p, r2cb_9, &desc); Chris@19: } Chris@19: Chris@19: #else /* HAVE_FMA */ Chris@19: Chris@19: /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 9 -name r2cb_9 -include r2cb.h */ Chris@19: Chris@19: /* Chris@19: * This function contains 32 FP additions, 18 FP multiplications, Chris@19: * (or, 22 additions, 8 multiplications, 10 fused multiply/add), Chris@19: * 35 stack variables, 12 constants, and 18 memory accesses Chris@19: */ Chris@19: #include "r2cb.h" Chris@19: Chris@19: static void r2cb_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) Chris@19: { Chris@19: DK(KP984807753, +0.984807753012208059366743024589523013670643252); Chris@19: DK(KP173648177, +0.173648177666930348851716626769314796000375677); Chris@19: DK(KP300767466, +0.300767466360870593278543795225003852144476517); Chris@19: DK(KP1_705737063, +1.705737063904886419256501927880148143872040591); Chris@19: DK(KP642787609, +0.642787609686539326322643409907263432907559884); Chris@19: DK(KP766044443, +0.766044443118978035202392650555416673935832457); Chris@19: DK(KP1_326827896, +1.326827896337876792410842639271782594433726619); Chris@19: DK(KP1_113340798, +1.113340798452838732905825904094046265936583811); Chris@19: DK(KP500000000, +0.500000000000000000000000000000000000000000000); Chris@19: DK(KP866025403, +0.866025403784438646763723170752936183471402627); Chris@19: DK(KP2_000000000, +2.000000000000000000000000000000000000000000000); Chris@19: DK(KP1_732050807, +1.732050807568877293527446341505872366942805254); Chris@19: { Chris@19: INT i; Chris@19: for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(36, rs), MAKE_VOLATILE_STRIDE(36, csr), MAKE_VOLATILE_STRIDE(36, csi)) { Chris@19: E T3, Tq, Tc, Tk, Tj, T8, Tm, Ts, Th, Tr, Tw, Tx; Chris@19: { Chris@19: E Tb, T1, T2, T9, Ta; Chris@19: Ta = Ci[WS(csi, 3)]; Chris@19: Tb = KP1_732050807 * Ta; Chris@19: T1 = Cr[0]; Chris@19: T2 = Cr[WS(csr, 3)]; Chris@19: T9 = T1 - T2; Chris@19: T3 = FMA(KP2_000000000, T2, T1); Chris@19: Tq = T9 + Tb; Chris@19: Tc = T9 - Tb; Chris@19: } Chris@19: { Chris@19: E T4, T7, Ti, Tg, Tl, Td; Chris@19: T4 = Cr[WS(csr, 1)]; Chris@19: Tk = Ci[WS(csi, 1)]; Chris@19: { Chris@19: E T5, T6, Te, Tf; Chris@19: T5 = Cr[WS(csr, 4)]; Chris@19: T6 = Cr[WS(csr, 2)]; Chris@19: T7 = T5 + T6; Chris@19: Ti = KP866025403 * (T5 - T6); Chris@19: Te = Ci[WS(csi, 4)]; Chris@19: Tf = Ci[WS(csi, 2)]; Chris@19: Tg = KP866025403 * (Te + Tf); Chris@19: Tj = Tf - Te; Chris@19: } Chris@19: T8 = T4 + T7; Chris@19: Tl = FMA(KP500000000, Tj, Tk); Chris@19: Tm = Ti + Tl; Chris@19: Ts = Tl - Ti; Chris@19: Td = FNMS(KP500000000, T7, T4); Chris@19: Th = Td - Tg; Chris@19: Tr = Td + Tg; Chris@19: } Chris@19: R0[0] = FMA(KP2_000000000, T8, T3); Chris@19: Tw = T3 - T8; Chris@19: Tx = KP1_732050807 * (Tk - Tj); Chris@19: R1[WS(rs, 1)] = Tw - Tx; Chris@19: R0[WS(rs, 3)] = Tw + Tx; Chris@19: { Chris@19: E Tp, Tn, To, Tv, Tt, Tu; Chris@19: Tp = FMA(KP1_113340798, Th, KP1_326827896 * Tm); Chris@19: Tn = FNMS(KP642787609, Tm, KP766044443 * Th); Chris@19: To = Tc - Tn; Chris@19: R1[0] = FMA(KP2_000000000, Tn, Tc); Chris@19: R1[WS(rs, 3)] = To + Tp; Chris@19: R0[WS(rs, 2)] = To - Tp; Chris@19: Tv = FMA(KP1_705737063, Tr, KP300767466 * Ts); Chris@19: Tt = FNMS(KP984807753, Ts, KP173648177 * Tr); Chris@19: Tu = Tq - Tt; Chris@19: R0[WS(rs, 1)] = FMA(KP2_000000000, Tt, Tq); Chris@19: R0[WS(rs, 4)] = Tu + Tv; Chris@19: R1[WS(rs, 2)] = Tu - Tv; Chris@19: } Chris@19: } Chris@19: } Chris@19: } Chris@19: Chris@19: static const kr2c_desc desc = { 9, "r2cb_9", {22, 8, 10, 0}, &GENUS }; Chris@19: Chris@19: void X(codelet_r2cb_9) (planner *p) { Chris@19: X(kr2c_register) (p, r2cb_9, &desc); Chris@19: } Chris@19: Chris@19: #endif /* HAVE_FMA */