Mercurial > hg > js-dsp-test

comparison fft/fftw/fftw-3.3.4/rdft/scalar/r2cf/r2cfII_9.c @ 19:26056e866c29

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW to comparison table
author Chris Cannam
date Tue, 06 Oct 2015 13:08:39 +0100
parents
children
comparison
equal deleted inserted replaced
18:8db794ca3e0b 19:26056e866c29
1 /*
2 * Copyright (c) 2003, 2007-14 Matteo Frigo
3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18 *
19 */
20
21 /* This file was automatically generated --- DO NOT EDIT */
22 /* Generated on Tue Mar 4 13:49:18 EST 2014 */
23
24 #include "codelet-rdft.h"
25
26 #ifdef HAVE_FMA
27
28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 9 -name r2cfII_9 -dft-II -include r2cfII.h */
29
30 /*
31 * This function contains 42 FP additions, 34 FP multiplications,
32 * (or, 12 additions, 4 multiplications, 30 fused multiply/add),
33 * 46 stack variables, 17 constants, and 18 memory accesses
34 */
35 #include "r2cfII.h"
36
37 static void r2cfII_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP939692620, +0.939692620785908384054109277324731469936208134);
40 DK(KP879385241, +0.879385241571816768108218554649462939872416269);
41 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
42 DK(KP852868531, +0.852868531952443209628250963940074071936020296);
43 DK(KP666666666, +0.666666666666666666666666666666666666666666667);
44 DK(KP673648177, +0.673648177666930348851716626769314796000375677);
45 DK(KP898197570, +0.898197570222573798468955502359086394667167570);
46 DK(KP826351822, +0.826351822333069651148283373230685203999624323);
47 DK(KP907603734, +0.907603734547952313649323976213898122064543220);
48 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
49 DK(KP420276625, +0.420276625461206169731530603237061658838781920);
50 DK(KP315207469, +0.315207469095904627298647952427796244129086440);
51 DK(KP203604859, +0.203604859554852403062088995281827210665664861);
52 DK(KP152703644, +0.152703644666139302296566746461370407999248646);
53 DK(KP726681596, +0.726681596905677465811651808188092531873167623);
54 DK(KP968908795, +0.968908795874236621082202410917456709164223497);
55 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
56 {
57 INT i;
58 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(36, rs), MAKE_VOLATILE_STRIDE(36, csr), MAKE_VOLATILE_STRIDE(36, csi)) {
59 E To, T5, Tp, Ta, Ti, Tm, TB, Tq, Tt, Tf, Th;
60 {
61 E T1, T6, T4, Tb, Tk, T9, Tc, Td, Tl, Te;
62 {
63 E T2, T3, T7, T8;
64 T1 = R0[0];
65 T2 = R0[WS(rs, 3)];
66 T3 = R1[WS(rs, 1)];
67 T6 = R0[WS(rs, 1)];
68 T7 = R0[WS(rs, 4)];
69 T8 = R1[WS(rs, 2)];
70 T4 = T2 - T3;
71 To = T2 + T3;
72 Tb = R0[WS(rs, 2)];
73 Tk = T7 + T8;
74 T9 = T7 - T8;
75 Tc = R1[0];
76 Td = R1[WS(rs, 3)];
77 }
78 T5 = T1 + T4;
79 Tp = FNMS(KP500000000, T4, T1);
80 Ta = T6 + T9;
81 Tl = FNMS(KP500000000, T9, T6);
82 Te = Tc + Td;
83 Ti = Tc - Td;
84 Tm = FMA(KP968908795, Tl, Tk);
85 TB = FNMS(KP726681596, Tk, Tl);
86 Tq = FNMS(KP152703644, Tk, Tl);
87 Tt = FMA(KP203604859, Tl, Tk);
88 Tf = Tb - Te;
89 Th = FMA(KP500000000, Te, Tb);
90 }
91 {
92 E Ts, Tr, TA, Tj, Tg;
93 Ts = FMA(KP315207469, Ti, Th);
94 Tr = FNMS(KP420276625, Th, Ti);
95 TA = FMA(KP203604859, Th, Ti);
96 Tj = FNMS(KP152703644, Ti, Th);
97 Tg = Ta + Tf;
98 Ci[WS(csi, 1)] = KP866025403 * (Tf - Ta);
99 {
100 E Tu, Tx, TF, TC;
101 Tu = FNMS(KP907603734, Tt, Ts);
102 Tx = FNMS(KP826351822, Tr, Tq);
103 TF = FMA(KP898197570, TB, TA);
104 TC = FNMS(KP898197570, TB, TA);
105 {
106 E TE, Tn, Tv, Ty;
107 TE = FNMS(KP673648177, Tm, Tj);
108 Tn = FMA(KP673648177, Tm, Tj);
109 Cr[WS(csr, 4)] = T5 + Tg;
110 Cr[WS(csr, 1)] = FNMS(KP500000000, Tg, T5);
111 Tv = FNMS(KP666666666, Tu, Tr);
112 Ty = FNMS(KP666666666, Tx, Tt);
113 Cr[0] = FMA(KP852868531, TF, Tp);
114 {
115 E TG, TD, Tw, Tz;
116 TG = FMA(KP500000000, TF, TE);
117 Ci[0] = -(KP984807753 * (FMA(KP879385241, To, Tn)));
118 TD = FNMS(KP666666666, Tn, TC);
119 Tw = FMA(KP826351822, Tv, Tq);
120 Tz = FMA(KP907603734, Ty, Ts);
121 Cr[WS(csr, 3)] = FNMS(KP852868531, TG, Tp);
122 Ci[WS(csi, 3)] = -(KP866025403 * (FMA(KP852868531, TD, To)));
123 Cr[WS(csr, 2)] = FNMS(KP852868531, Tw, Tp);
124 Ci[WS(csi, 2)] = KP866025403 * (FNMS(KP939692620, Tz, To));
125 }
126 }
127 }
128 }
129 }
130 }
131 }
132
133 static const kr2c_desc desc = { 9, "r2cfII_9", {12, 4, 30, 0}, &GENUS };
134
135 void X(codelet_r2cfII_9) (planner *p) {
136 X(kr2c_register) (p, r2cfII_9, &desc);
137 }
138
139 #else /* HAVE_FMA */
140
141 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 9 -name r2cfII_9 -dft-II -include r2cfII.h */
142
143 /*
144 * This function contains 42 FP additions, 30 FP multiplications,
145 * (or, 25 additions, 13 multiplications, 17 fused multiply/add),
146 * 39 stack variables, 14 constants, and 18 memory accesses
147 */
148 #include "r2cfII.h"
149
150 static void r2cfII_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
151 {
152 DK(KP663413948, +0.663413948168938396205421319635891297216863310);
153 DK(KP642787609, +0.642787609686539326322643409907263432907559884);
154 DK(KP556670399, +0.556670399226419366452912952047023132968291906);
155 DK(KP766044443, +0.766044443118978035202392650555416673935832457);
156 DK(KP852868531, +0.852868531952443209628250963940074071936020296);
157 DK(KP173648177, +0.173648177666930348851716626769314796000375677);
158 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
159 DK(KP150383733, +0.150383733180435296639271897612501926072238258);
160 DK(KP813797681, +0.813797681349373692844693217248393223289101568);
161 DK(KP342020143, +0.342020143325668733044099614682259580763083368);
162 DK(KP939692620, +0.939692620785908384054109277324731469936208134);
163 DK(KP296198132, +0.296198132726023843175338011893050938967728390);
164 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
165 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
166 {
167 INT i;
168 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(36, rs), MAKE_VOLATILE_STRIDE(36, csr), MAKE_VOLATILE_STRIDE(36, csi)) {
169 E T1, T4, To, Ta, Tl, Tk, Tf, Ti, Th, T2, T3, T5, Tg;
170 T1 = R0[0];
171 T2 = R1[WS(rs, 1)];
172 T3 = R0[WS(rs, 3)];
173 T4 = T2 - T3;
174 To = T2 + T3;
175 {
176 E T6, T7, T8, T9;
177 T6 = R0[WS(rs, 1)];
178 T7 = R1[WS(rs, 2)];
179 T8 = R0[WS(rs, 4)];
180 T9 = T7 - T8;
181 Ta = T6 - T9;
182 Tl = T7 + T8;
183 Tk = FMA(KP500000000, T9, T6);
184 }
185 {
186 E Tb, Tc, Td, Te;
187 Tb = R0[WS(rs, 2)];
188 Tc = R1[0];
189 Td = R1[WS(rs, 3)];
190 Te = Tc + Td;
191 Tf = Tb - Te;
192 Ti = FMA(KP500000000, Te, Tb);
193 Th = Tc - Td;
194 }
195 Ci[WS(csi, 1)] = KP866025403 * (Tf - Ta);
196 T5 = T1 - T4;
197 Tg = Ta + Tf;
198 Cr[WS(csr, 1)] = FNMS(KP500000000, Tg, T5);
199 Cr[WS(csr, 4)] = T5 + Tg;
200 {
201 E Tr, Tt, Tw, Tv, Tu, Tp, Tq, Ts, Tj, Tm, Tn;
202 Tr = FMA(KP500000000, T4, T1);
203 Tt = FMA(KP296198132, Th, KP939692620 * Ti);
204 Tw = FNMS(KP813797681, Th, KP342020143 * Ti);
205 Tv = FNMS(KP984807753, Tk, KP150383733 * Tl);
206 Tu = FMA(KP173648177, Tk, KP852868531 * Tl);
207 Tp = FNMS(KP556670399, Tl, KP766044443 * Tk);
208 Tq = FMA(KP852868531, Th, KP173648177 * Ti);
209 Ts = Tp + Tq;
210 Tj = FNMS(KP984807753, Ti, KP150383733 * Th);
211 Tm = FMA(KP642787609, Tk, KP663413948 * Tl);
212 Tn = Tj - Tm;
213 Ci[0] = FNMS(KP866025403, To, Tn);
214 Cr[0] = Tr + Ts;
215 Ci[WS(csi, 3)] = FNMS(KP500000000, Tn, KP866025403 * ((Tp - Tq) - To));
216 Cr[WS(csr, 3)] = FMA(KP866025403, Tm + Tj, Tr) - (KP500000000 * Ts);
217 Ci[WS(csi, 2)] = FMA(KP866025403, To - (Tu + Tt), KP500000000 * (Tw - Tv));
218 Cr[WS(csr, 2)] = FMA(KP500000000, Tt - Tu, Tr) + (KP866025403 * (Tv + Tw));
219 }
220 }
221 }
222 }
223
224 static const kr2c_desc desc = { 9, "r2cfII_9", {25, 13, 17, 0}, &GENUS };
225
226 void X(codelet_r2cfII_9) (planner *p) {
227 X(kr2c_register) (p, r2cfII_9, &desc);
228 }
229
230 #endif /* HAVE_FMA */