Mercurial > hg > sv-dependency-builds
diff src/fftw-3.3.3/rdft/simd/common/hc2cbdftv_8.c @ 10:37bf6b4a2645
Add FFTW3
| author | Chris Cannam |
|---|---|
| date | Wed, 20 Mar 2013 15:35:50 +0000 |
| parents | |
| children |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/src/fftw-3.3.3/rdft/simd/common/hc2cbdftv_8.c Wed Mar 20 15:35:50 2013 +0000 @@ -0,0 +1,228 @@ +/* + * Copyright (c) 2003, 2007-11 Matteo Frigo + * Copyright (c) 2003, 2007-11 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + * + */ + +/* This file was automatically generated --- DO NOT EDIT */ +/* Generated on Sun Nov 25 07:42:29 EST 2012 */ + +#include "codelet-rdft.h" + +#ifdef HAVE_FMA + +/* Generated by: ../../../genfft/gen_hc2cdft_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 8 -dif -sign 1 -name hc2cbdftv_8 -include hc2cbv.h */ + +/* + * This function contains 41 FP additions, 32 FP multiplications, + * (or, 23 additions, 14 multiplications, 18 fused multiply/add), + * 51 stack variables, 1 constants, and 16 memory accesses + */ +#include "hc2cbv.h" + +static void hc2cbdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) +{ + 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(32, rs)) { + V TJ, T4, Tf, TB, TD, TE, Tm, T1, Tj, TF, Tp, Tb, Tg, Tt, Tx; + V T2, T3, Td, Te, T5, T6, T8, T9, Tn, T7, To, Ta, Tk, Tl, TG; + V TL, Tq, Tc, Tu, Th, Tv, Ty, Tw, TC, Ti, TK, TA, Tz, TI, TH; + V Ts, Tr, TN, TM; + T2 = LD(&(Rp[0]), ms, &(Rp[0])); + T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)])); + Td = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0])); + Te = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)])); + T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)])); + T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0])); + T8 = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)])); + T9 = LD(&(Rm[0]), -ms, &(Rm[0])); + TJ = LDW(&(W[0])); + Tk = VFMACONJ(T3, T2); + T4 = VFNMSCONJ(T3, T2); + Tl = VFMACONJ(Te, Td); + Tf = VFNMSCONJ(Te, Td); + Tn = VFMACONJ(T6, T5); + T7 = VFNMSCONJ(T6, T5); + To = VFMACONJ(T9, T8); + Ta = VFMSCONJ(T9, T8); + TB = LDW(&(W[TWVL * 8])); + TD = LDW(&(W[TWVL * 6])); + TE = VADD(Tk, Tl); + Tm = VSUB(Tk, Tl); + T1 = LDW(&(W[TWVL * 12])); + Tj = LDW(&(W[TWVL * 10])); + TF = VADD(Tn, To); + Tp = VSUB(Tn, To); + Tb = VADD(T7, Ta); + Tg = VSUB(T7, Ta); + Tt = LDW(&(W[TWVL * 4])); + Tx = LDW(&(W[TWVL * 2])); + TG = VZMUL(TD, VSUB(TE, TF)); + TL = VADD(TE, TF); + Tq = VZMUL(Tj, VFNMSI(Tp, Tm)); + Tc = VFMA(LDK(KP707106781), Tb, T4); + Tu = VFNMS(LDK(KP707106781), Tb, T4); + Th = VFMA(LDK(KP707106781), Tg, Tf); + Tv = VFNMS(LDK(KP707106781), Tg, Tf); + Ty = VZMUL(Tx, VFMAI(Tp, Tm)); + Tw = VZMULI(Tt, VFNMSI(Tv, Tu)); + TC = VZMULI(TB, VFMAI(Tv, Tu)); + Ti = VZMULI(T1, VFNMSI(Th, Tc)); + TK = VZMULI(TJ, VFMAI(Th, Tc)); + TA = VCONJ(VSUB(Ty, Tw)); + Tz = VADD(Tw, Ty); + TI = VCONJ(VSUB(TG, TC)); + TH = VADD(TC, TG); + Ts = VCONJ(VSUB(Tq, Ti)); + Tr = VADD(Ti, Tq); + TN = VCONJ(VSUB(TL, TK)); + TM = VADD(TK, TL); + ST(&(Rm[WS(rs, 1)]), TA, -ms, &(Rm[WS(rs, 1)])); + ST(&(Rp[WS(rs, 1)]), Tz, ms, &(Rp[WS(rs, 1)])); + ST(&(Rm[WS(rs, 2)]), TI, -ms, &(Rm[0])); + ST(&(Rp[WS(rs, 2)]), TH, ms, &(Rp[0])); + ST(&(Rm[WS(rs, 3)]), Ts, -ms, &(Rm[WS(rs, 1)])); + ST(&(Rp[WS(rs, 3)]), Tr, ms, &(Rp[WS(rs, 1)])); + ST(&(Rm[0]), TN, -ms, &(Rm[0])); + ST(&(Rp[0]), TM, ms, &(Rp[0])); + } + } + VLEAVE(); +} + +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, XSIMD_STRING("hc2cbdftv_8"), twinstr, &GENUS, {23, 14, 18, 0} }; + +void XSIMD(codelet_hc2cbdftv_8) (planner *p) { + X(khc2c_register) (p, hc2cbdftv_8, &desc, HC2C_VIA_DFT); +} +#else /* HAVE_FMA */ + +/* Generated by: ../../../genfft/gen_hc2cdft_c.native -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 8 -dif -sign 1 -name hc2cbdftv_8 -include hc2cbv.h */ + +/* + * This function contains 41 FP additions, 16 FP multiplications, + * (or, 41 additions, 16 multiplications, 0 fused multiply/add), + * 55 stack variables, 1 constants, and 16 memory accesses + */ +#include "hc2cbv.h" + +static void hc2cbdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) +{ + 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(32, rs)) { + V T5, Tj, Tq, TI, Te, Tk, Tt, TJ, T2, Tg, T4, Ti, T3, Th, To; + V Tp, T6, Tc, T8, Tb, T7, Ta, T9, Td, Tr, Ts, TP, Tu, Tm, TO; + V Tn, Tf, Tl, T1, TN, Tv, TR, Tw, TQ, TC, TK, TA, TG, TB, TH; + V Ty, Tz, Tx, TF, TD, TM, TE, TL; + T2 = LD(&(Rp[0]), ms, &(Rp[0])); + Tg = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0])); + T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)])); + T4 = VCONJ(T3); + Th = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)])); + Ti = VCONJ(Th); + T5 = VSUB(T2, T4); + Tj = VSUB(Tg, Ti); + To = VADD(T2, T4); + Tp = VADD(Tg, Ti); + Tq = VSUB(To, Tp); + TI = VADD(To, Tp); + T6 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)])); + Tc = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)])); + T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0])); + T8 = VCONJ(T7); + Ta = LD(&(Rm[0]), -ms, &(Rm[0])); + Tb = VCONJ(Ta); + T9 = VSUB(T6, T8); + Td = VSUB(Tb, Tc); + Te = VMUL(LDK(KP707106781), VADD(T9, Td)); + Tk = VMUL(LDK(KP707106781), VSUB(T9, Td)); + Tr = VADD(T6, T8); + Ts = VADD(Tb, Tc); + Tt = VBYI(VSUB(Tr, Ts)); + TJ = VADD(Tr, Ts); + TP = VADD(TI, TJ); + Tn = LDW(&(W[TWVL * 10])); + Tu = VZMUL(Tn, VSUB(Tq, Tt)); + Tf = VADD(T5, Te); + Tl = VBYI(VADD(Tj, Tk)); + T1 = LDW(&(W[TWVL * 12])); + Tm = VZMULI(T1, VSUB(Tf, Tl)); + TN = LDW(&(W[0])); + TO = VZMULI(TN, VADD(Tl, Tf)); + Tv = VADD(Tm, Tu); + ST(&(Rp[WS(rs, 3)]), Tv, ms, &(Rp[WS(rs, 1)])); + TR = VCONJ(VSUB(TP, TO)); + ST(&(Rm[0]), TR, -ms, &(Rm[0])); + Tw = VCONJ(VSUB(Tu, Tm)); + ST(&(Rm[WS(rs, 3)]), Tw, -ms, &(Rm[WS(rs, 1)])); + TQ = VADD(TO, TP); + ST(&(Rp[0]), TQ, ms, &(Rp[0])); + TB = LDW(&(W[TWVL * 2])); + TC = VZMUL(TB, VADD(Tq, Tt)); + TH = LDW(&(W[TWVL * 6])); + TK = VZMUL(TH, VSUB(TI, TJ)); + Ty = VBYI(VSUB(Tk, Tj)); + Tz = VSUB(T5, Te); + Tx = LDW(&(W[TWVL * 4])); + TA = VZMULI(Tx, VADD(Ty, Tz)); + TF = LDW(&(W[TWVL * 8])); + TG = VZMULI(TF, VSUB(Tz, Ty)); + TD = VADD(TA, TC); + ST(&(Rp[WS(rs, 1)]), TD, ms, &(Rp[WS(rs, 1)])); + TM = VCONJ(VSUB(TK, TG)); + ST(&(Rm[WS(rs, 2)]), TM, -ms, &(Rm[0])); + TE = VCONJ(VSUB(TC, TA)); + ST(&(Rm[WS(rs, 1)]), TE, -ms, &(Rm[WS(rs, 1)])); + TL = VADD(TG, TK); + ST(&(Rp[WS(rs, 2)]), TL, ms, &(Rp[0])); + } + } + VLEAVE(); +} + +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, XSIMD_STRING("hc2cbdftv_8"), twinstr, &GENUS, {41, 16, 0, 0} }; + +void XSIMD(codelet_hc2cbdftv_8) (planner *p) { + X(khc2c_register) (p, hc2cbdftv_8, &desc, HC2C_VIA_DFT); +} +#endif /* HAVE_FMA */