/* * Copyright (c) 2003, 2007-8 Matteo Frigo * Copyright (c) 2003, 2007-8 Massachusetts Institute of Technology * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * */ /* This file was automatically generated --- DO NOT EDIT */ /* Generated on Sun Jul 12 06:47:45 EDT 2009 */ #include "codelet-rdft.h" #ifdef HAVE_FMA /* Generated by: ../../../genfft/gen_hc2cdft_c -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 8 -dit -name hc2cfdftv_8 -include hc2cfv.h */ /* * This function contains 41 FP additions, 40 FP multiplications, * (or, 23 additions, 22 multiplications, 18 fused multiply/add), * 52 stack variables, 2 constants, and 16 memory accesses */ #include "hc2cfv.h" static void hc2cfdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) { DVK(KP500000000, +0.500000000000000000000000000000000000000000000); DVK(KP707106781, +0.707106781186547524400844362104849039284835938); INT m; for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) { V T3, Tc, Tl, Ts, Tf, Tg, Te, Tp, T7, Ta, T1, T2, Tb, Tj, Tk; V Ti, Tr, T5, T6, T4, T9, Th, Tq, TC, T8, Td, TF, Tm, TG, TD; V Tt, Tu, Tn, TH, TL, TE, TK, Tz, Tv, Ty, To, TJ, TI, TN, TM; V TB, TA, Tx, Tw; T1 = LD(&(Rp[0]), ms, &(Rp[0])); T2 = LD(&(Rm[0]), -ms, &(Rm[0])); Tb = LDW(&(W[0])); Tj = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)])); Tk = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)])); Ti = LDW(&(W[TWVL * 12])); Tr = LDW(&(W[TWVL * 10])); T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0])); T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0])); T3 = VFMACONJ(T2, T1); Tc = VZMULIJ(Tb, VFNMSCONJ(T2, T1)); T4 = LDW(&(W[TWVL * 6])); T9 = LDW(&(W[TWVL * 8])); Tl = VZMULIJ(Ti, VFNMSCONJ(Tk, Tj)); Ts = VZMULJ(Tr, VFMACONJ(Tk, Tj)); Tf = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)])); Tg = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)])); Te = LDW(&(W[TWVL * 4])); Tp = LDW(&(W[TWVL * 2])); T7 = VZMULJ(T4, VFMACONJ(T6, T5)); Ta = VZMULIJ(T9, VFNMSCONJ(T6, T5)); Th = VZMULIJ(Te, VFNMSCONJ(Tg, Tf)); Tq = VZMULJ(Tp, VFMACONJ(Tg, Tf)); TC = VADD(T3, T7); T8 = VSUB(T3, T7); Td = VSUB(Ta, Tc); TF = VADD(Tc, Ta); Tm = VSUB(Th, Tl); TG = VADD(Th, Tl); TD = VADD(Tq, Ts); Tt = VSUB(Tq, Ts); Tu = VSUB(Tm, Td); Tn = VADD(Td, Tm); TH = VSUB(TF, TG); TL = VADD(TF, TG); TE = VSUB(TC, TD); TK = VADD(TC, TD); Tz = VFMA(LDK(KP707106781), Tu, Tt); Tv = VFNMS(LDK(KP707106781), Tu, Tt); Ty = VFNMS(LDK(KP707106781), Tn, T8); To = VFMA(LDK(KP707106781), Tn, T8); TJ = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TH, TE))); TI = VMUL(LDK(KP500000000), VFMAI(TH, TE)); TN = VCONJ(VMUL(LDK(KP500000000), VADD(TL, TK))); TM = VMUL(LDK(KP500000000), VSUB(TK, TL)); TB = VMUL(LDK(KP500000000), VFMAI(Tz, Ty)); TA = VCONJ(VMUL(LDK(KP500000000), VFNMSI(Tz, Ty))); Tx = VCONJ(VMUL(LDK(KP500000000), VFMAI(Tv, To))); Tw = VMUL(LDK(KP500000000), VFNMSI(Tv, To)); ST(&(Rm[WS(rs, 1)]), TJ, -ms, &(Rm[WS(rs, 1)])); ST(&(Rp[WS(rs, 2)]), TI, ms, &(Rp[0])); ST(&(Rm[WS(rs, 3)]), TN, -ms, &(Rm[WS(rs, 1)])); ST(&(Rp[0]), TM, ms, &(Rp[0])); ST(&(Rp[WS(rs, 3)]), TB, ms, &(Rp[WS(rs, 1)])); ST(&(Rm[WS(rs, 2)]), TA, -ms, &(Rm[0])); ST(&(Rm[0]), Tx, -ms, &(Rm[0])); ST(&(Rp[WS(rs, 1)]), Tw, ms, &(Rp[WS(rs, 1)])); } } static const tw_instr twinstr[] = { VTW(1, 1), VTW(1, 2), VTW(1, 3), VTW(1, 4), VTW(1, 5), VTW(1, 6), VTW(1, 7), {TW_NEXT, VL, 0} }; static const hc2c_desc desc = { 8, "hc2cfdftv_8", twinstr, &GENUS, {23, 22, 18, 0} }; void X(codelet_hc2cfdftv_8) (planner *p) { X(khc2c_register) (p, hc2cfdftv_8, &desc, HC2C_VIA_DFT); } #else /* HAVE_FMA */ /* Generated by: ../../../genfft/gen_hc2cdft_c -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 8 -dit -name hc2cfdftv_8 -include hc2cfv.h */ /* * This function contains 41 FP additions, 23 FP multiplications, * (or, 41 additions, 23 multiplications, 0 fused multiply/add), * 57 stack variables, 3 constants, and 16 memory accesses */ #include "hc2cfv.h" static void hc2cfdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) { DVK(KP707106781, +0.707106781186547524400844362104849039284835938); DVK(KP353553390, +0.353553390593273762200422181052424519642417969); DVK(KP500000000, +0.500000000000000000000000000000000000000000000); INT m; for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) { V Ta, TE, Tr, TF, Tl, TK, Tw, TG, T1, T6, T3, T8, T2, T7, T4; V T9, T5, To, Tq, Tn, Tp, Tc, Th, Te, Tj, Td, Ti, Tf, Tk, Tb; V Tg, Tt, Tv, Ts, Tu, Ty, Tz, Tm, Tx, TC, TD, TA, TB, TI, TO; V TL, TP, TH, TJ, TM, TR, TN, TQ; T1 = LD(&(Rp[0]), ms, &(Rp[0])); T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0])); T2 = LD(&(Rm[0]), -ms, &(Rm[0])); T3 = VCONJ(T2); T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0])); T8 = VCONJ(T7); T4 = VADD(T1, T3); T5 = LDW(&(W[TWVL * 6])); T9 = VZMULJ(T5, VADD(T6, T8)); Ta = VADD(T4, T9); TE = VMUL(LDK(KP500000000), VSUB(T4, T9)); Tn = LDW(&(W[0])); To = VZMULIJ(Tn, VSUB(T3, T1)); Tp = LDW(&(W[TWVL * 8])); Tq = VZMULIJ(Tp, VSUB(T8, T6)); Tr = VADD(To, Tq); TF = VSUB(To, Tq); Tc = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)])); Th = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)])); Td = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)])); Te = VCONJ(Td); Ti = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)])); Tj = VCONJ(Ti); Tb = LDW(&(W[TWVL * 2])); Tf = VZMULJ(Tb, VADD(Tc, Te)); Tg = LDW(&(W[TWVL * 10])); Tk = VZMULJ(Tg, VADD(Th, Tj)); Tl = VADD(Tf, Tk); TK = VSUB(Tf, Tk); Ts = LDW(&(W[TWVL * 4])); Tt = VZMULIJ(Ts, VSUB(Te, Tc)); Tu = LDW(&(W[TWVL * 12])); Tv = VZMULIJ(Tu, VSUB(Tj, Th)); Tw = VADD(Tt, Tv); TG = VSUB(Tv, Tt); Tm = VADD(Ta, Tl); Tx = VADD(Tr, Tw); Ty = VCONJ(VMUL(LDK(KP500000000), VSUB(Tm, Tx))); Tz = VMUL(LDK(KP500000000), VADD(Tm, Tx)); ST(&(Rm[WS(rs, 3)]), Ty, -ms, &(Rm[WS(rs, 1)])); ST(&(Rp[0]), Tz, ms, &(Rp[0])); TA = VSUB(Ta, Tl); TB = VBYI(VSUB(Tw, Tr)); TC = VCONJ(VMUL(LDK(KP500000000), VSUB(TA, TB))); TD = VMUL(LDK(KP500000000), VADD(TA, TB)); ST(&(Rm[WS(rs, 1)]), TC, -ms, &(Rm[WS(rs, 1)])); ST(&(Rp[WS(rs, 2)]), TD, ms, &(Rp[0])); TH = VMUL(LDK(KP353553390), VADD(TF, TG)); TI = VADD(TE, TH); TO = VSUB(TE, TH); TJ = VMUL(LDK(KP707106781), VSUB(TG, TF)); TL = VMUL(LDK(KP500000000), VBYI(VSUB(TJ, TK))); TP = VMUL(LDK(KP500000000), VBYI(VADD(TK, TJ))); TM = VCONJ(VSUB(TI, TL)); ST(&(Rm[0]), TM, -ms, &(Rm[0])); TR = VADD(TO, TP); ST(&(Rp[WS(rs, 3)]), TR, ms, &(Rp[WS(rs, 1)])); TN = VADD(TI, TL); ST(&(Rp[WS(rs, 1)]), TN, ms, &(Rp[WS(rs, 1)])); TQ = VCONJ(VSUB(TO, TP)); ST(&(Rm[WS(rs, 2)]), TQ, -ms, &(Rm[0])); } } static const tw_instr twinstr[] = { VTW(1, 1), VTW(1, 2), VTW(1, 3), VTW(1, 4), VTW(1, 5), VTW(1, 6), VTW(1, 7), {TW_NEXT, VL, 0} }; static const hc2c_desc desc = { 8, "hc2cfdftv_8", twinstr, &GENUS, {41, 23, 0, 0} }; void X(codelet_hc2cfdftv_8) (planner *p) { X(khc2c_register) (p, hc2cfdftv_8, &desc, HC2C_VIA_DFT); } #endif /* HAVE_FMA */