Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cfII_12.c @ 167:bd3cc4d1df30

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW 3.3.8 source, and a Linux build
author Chris Cannam <cannam@all-day-breakfast.com>
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
comparison
equal deleted inserted replaced
166:cbd6d7e562c7 167:bd3cc4d1df30
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 Thu May 24 08:06:43 EDT 2018 */
23
24 #include "rdft/codelet-rdft.h"
25
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
27
28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 12 -name r2cfII_12 -dft-II -include rdft/scalar/r2cfII.h */
29
30 /*
31 * This function contains 45 FP additions, 24 FP multiplications,
32 * (or, 21 additions, 0 multiplications, 24 fused multiply/add),
33 * 28 stack variables, 3 constants, and 24 memory accesses
34 */
35 #include "rdft/scalar/r2cfII.h"
36
37 static void r2cfII_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
40 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
41 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
42 {
43 INT i;
44 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
45 E Tx, Ty, T8, Tz, Tl, Tm, Tv, T5, TA, Tt, Te, Tf, Tu, T6, T7;
46 E Tw, TF, TG;
47 Tx = R0[WS(rs, 3)];
48 T6 = R0[WS(rs, 5)];
49 T7 = R0[WS(rs, 1)];
50 Ty = T6 + T7;
51 T8 = T6 - T7;
52 Tz = FMA(KP500000000, Ty, Tx);
53 {
54 E Th, Ti, Tj, Tk;
55 Th = R1[WS(rs, 4)];
56 Ti = R1[WS(rs, 2)];
57 Tj = R1[0];
58 Tk = Ti - Tj;
59 Tl = FMA(KP500000000, Tk, Th);
60 Tm = Ti + Tj;
61 Tv = Ti - Tj - Th;
62 }
63 {
64 E T1, T2, T3, T4;
65 T1 = R0[0];
66 T2 = R0[WS(rs, 2)];
67 T3 = R0[WS(rs, 4)];
68 T4 = T2 - T3;
69 T5 = FMA(KP500000000, T4, T1);
70 TA = T3 + T2;
71 Tt = T1 + T3 - T2;
72 }
73 {
74 E Ta, Tb, Tc, Td;
75 Ta = R1[WS(rs, 1)];
76 Tb = R1[WS(rs, 3)];
77 Tc = R1[WS(rs, 5)];
78 Td = Tb - Tc;
79 Te = FMA(KP500000000, Td, Ta);
80 Tf = Tc + Tb;
81 Tu = Ta + Tc - Tb;
82 }
83 Tw = Tu + Tv;
84 Cr[WS(csr, 1)] = FNMS(KP707106781, Tw, Tt);
85 Cr[WS(csr, 4)] = FMA(KP707106781, Tw, Tt);
86 TF = Tx - Ty;
87 TG = Tv - Tu;
88 Ci[WS(csi, 4)] = FMS(KP707106781, TG, TF);
89 Ci[WS(csi, 1)] = FMA(KP707106781, TG, TF);
90 {
91 E T9, TD, To, TE, Tg, Tn;
92 T9 = FNMS(KP866025403, T8, T5);
93 TD = FNMS(KP866025403, TA, Tz);
94 Tg = FNMS(KP866025403, Tf, Te);
95 Tn = FNMS(KP866025403, Tm, Tl);
96 To = Tg - Tn;
97 TE = Tg + Tn;
98 Cr[WS(csr, 5)] = FNMS(KP707106781, To, T9);
99 Ci[WS(csi, 3)] = FMA(KP707106781, TE, TD);
100 Cr[0] = FMA(KP707106781, To, T9);
101 Ci[WS(csi, 2)] = FMS(KP707106781, TE, TD);
102 }
103 {
104 E Tp, TB, Ts, TC, Tq, Tr;
105 Tp = FMA(KP866025403, T8, T5);
106 TB = FMA(KP866025403, TA, Tz);
107 Tq = FMA(KP866025403, Tm, Tl);
108 Tr = FMA(KP866025403, Tf, Te);
109 Ts = Tq - Tr;
110 TC = Tr + Tq;
111 Cr[WS(csr, 3)] = FNMS(KP707106781, Ts, Tp);
112 Ci[WS(csi, 5)] = FNMS(KP707106781, TC, TB);
113 Cr[WS(csr, 2)] = FMA(KP707106781, Ts, Tp);
114 Ci[0] = -(FMA(KP707106781, TC, TB));
115 }
116 }
117 }
118 }
119
120 static const kr2c_desc desc = { 12, "r2cfII_12", {21, 0, 24, 0}, &GENUS };
121
122 void X(codelet_r2cfII_12) (planner *p) {
123 X(kr2c_register) (p, r2cfII_12, &desc);
124 }
125
126 #else
127
128 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 12 -name r2cfII_12 -dft-II -include rdft/scalar/r2cfII.h */
129
130 /*
131 * This function contains 43 FP additions, 12 FP multiplications,
132 * (or, 39 additions, 8 multiplications, 4 fused multiply/add),
133 * 28 stack variables, 5 constants, and 24 memory accesses
134 */
135 #include "rdft/scalar/r2cfII.h"
136
137 static void r2cfII_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
138 {
139 DK(KP353553390, +0.353553390593273762200422181052424519642417969);
140 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
141 DK(KP612372435, +0.612372435695794524549321018676472847991486870);
142 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
143 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
144 {
145 INT i;
146 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
147 E Tx, Tg, T4, Tz, Ty, Tj, TA, T9, Tm, Tl, Te, Tp, To, Tf, TE;
148 E TF;
149 {
150 E T1, T3, T2, Th, Ti;
151 T1 = R0[0];
152 T3 = R0[WS(rs, 2)];
153 T2 = R0[WS(rs, 4)];
154 Tx = KP866025403 * (T2 + T3);
155 Tg = FMA(KP500000000, T3 - T2, T1);
156 T4 = T1 + T2 - T3;
157 Tz = R0[WS(rs, 3)];
158 Th = R0[WS(rs, 5)];
159 Ti = R0[WS(rs, 1)];
160 Ty = Th + Ti;
161 Tj = KP866025403 * (Th - Ti);
162 TA = FMA(KP500000000, Ty, Tz);
163 }
164 {
165 E T5, T6, T7, T8;
166 T5 = R1[WS(rs, 1)];
167 T6 = R1[WS(rs, 5)];
168 T7 = R1[WS(rs, 3)];
169 T8 = T6 - T7;
170 T9 = T5 + T8;
171 Tm = KP612372435 * (T6 + T7);
172 Tl = FNMS(KP353553390, T8, KP707106781 * T5);
173 }
174 {
175 E Td, Ta, Tb, Tc;
176 Td = R1[WS(rs, 4)];
177 Ta = R1[WS(rs, 2)];
178 Tb = R1[0];
179 Tc = Ta - Tb;
180 Te = Tc - Td;
181 Tp = FMA(KP353553390, Tc, KP707106781 * Td);
182 To = KP612372435 * (Ta + Tb);
183 }
184 Tf = KP707106781 * (T9 + Te);
185 Cr[WS(csr, 1)] = T4 - Tf;
186 Cr[WS(csr, 4)] = T4 + Tf;
187 TE = KP707106781 * (Te - T9);
188 TF = Tz - Ty;
189 Ci[WS(csi, 4)] = TE - TF;
190 Ci[WS(csi, 1)] = TE + TF;
191 {
192 E Tk, TB, Tr, Tw, Tn, Tq;
193 Tk = Tg - Tj;
194 TB = Tx - TA;
195 Tn = Tl - Tm;
196 Tq = To - Tp;
197 Tr = Tn + Tq;
198 Tw = Tn - Tq;
199 Cr[WS(csr, 5)] = Tk - Tr;
200 Ci[WS(csi, 2)] = Tw + TB;
201 Cr[0] = Tk + Tr;
202 Ci[WS(csi, 3)] = Tw - TB;
203 }
204 {
205 E Ts, TD, Tv, TC, Tt, Tu;
206 Ts = Tg + Tj;
207 TD = Tx + TA;
208 Tt = To + Tp;
209 Tu = Tm + Tl;
210 Tv = Tt - Tu;
211 TC = Tu + Tt;
212 Cr[WS(csr, 3)] = Ts - Tv;
213 Ci[WS(csi, 5)] = TD - TC;
214 Cr[WS(csr, 2)] = Ts + Tv;
215 Ci[0] = -(TC + TD);
216 }
217 }
218 }
219 }
220
221 static const kr2c_desc desc = { 12, "r2cfII_12", {39, 8, 4, 0}, &GENUS };
222
223 void X(codelet_r2cfII_12) (planner *p) {
224 X(kr2c_register) (p, r2cfII_12, &desc);
225 }
226
227 #endif