Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cb/r2cb_9.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 08:07:28 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_r2cb.native -fma -compact -variables 4 -pipeline-latency 4 -sign 1 -n 9 -name r2cb_9 -include rdft/scalar/r2cb.h */
29
30 /*
31 * This function contains 32 FP additions, 24 FP multiplications,
32 * (or, 8 additions, 0 multiplications, 24 fused multiply/add),
33 * 35 stack variables, 12 constants, and 18 memory accesses
34 */
35 #include "rdft/scalar/r2cb.h"
36
37 static void r2cb_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP1_705737063, +1.705737063904886419256501927880148143872040591);
40 DK(KP1_969615506, +1.969615506024416118733486049179046027341286503);
41 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
42 DK(KP176326980, +0.176326980708464973471090386868618986121633062);
43 DK(KP1_326827896, +1.326827896337876792410842639271782594433726619);
44 DK(KP1_532088886, +1.532088886237956070404785301110833347871664914);
45 DK(KP766044443, +0.766044443118978035202392650555416673935832457);
46 DK(KP839099631, +0.839099631177280011763127298123181364687434283);
47 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
48 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
49 DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
50 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
51 {
52 INT i;
53 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)) {
54 E T3, Tp, Tb, Th, Ti, T8, Tl, Tq, Tg, Tr, Tv, Tw;
55 {
56 E Ta, T1, T2, T9;
57 Ta = Ci[WS(csi, 3)];
58 T1 = Cr[0];
59 T2 = Cr[WS(csr, 3)];
60 T9 = T1 - T2;
61 T3 = FMA(KP2_000000000, T2, T1);
62 Tp = FMA(KP1_732050807, Ta, T9);
63 Tb = FNMS(KP1_732050807, Ta, T9);
64 }
65 {
66 E T4, T7, Tk, Tf, Tj, Tc;
67 T4 = Cr[WS(csr, 1)];
68 Th = Ci[WS(csi, 1)];
69 {
70 E T5, T6, Td, Te;
71 T5 = Cr[WS(csr, 4)];
72 T6 = Cr[WS(csr, 2)];
73 T7 = T5 + T6;
74 Tk = T6 - T5;
75 Td = Ci[WS(csi, 4)];
76 Te = Ci[WS(csi, 2)];
77 Tf = Td + Te;
78 Ti = Td - Te;
79 }
80 T8 = T4 + T7;
81 Tj = FNMS(KP500000000, Ti, Th);
82 Tl = FNMS(KP866025403, Tk, Tj);
83 Tq = FMA(KP866025403, Tk, Tj);
84 Tc = FNMS(KP500000000, T7, T4);
85 Tg = FNMS(KP866025403, Tf, Tc);
86 Tr = FMA(KP866025403, Tf, Tc);
87 }
88 R0[0] = FMA(KP2_000000000, T8, T3);
89 Tv = T3 - T8;
90 Tw = Ti + Th;
91 R1[WS(rs, 1)] = FNMS(KP1_732050807, Tw, Tv);
92 R0[WS(rs, 3)] = FMA(KP1_732050807, Tw, Tv);
93 {
94 E To, Tm, Tn, Tu, Ts, Tt;
95 To = FMA(KP839099631, Tg, Tl);
96 Tm = FNMS(KP839099631, Tl, Tg);
97 Tn = FNMS(KP766044443, Tm, Tb);
98 R1[0] = FMA(KP1_532088886, Tm, Tb);
99 R1[WS(rs, 3)] = FMA(KP1_326827896, To, Tn);
100 R0[WS(rs, 2)] = FNMS(KP1_326827896, To, Tn);
101 Tu = FMA(KP176326980, Tq, Tr);
102 Ts = FNMS(KP176326980, Tr, Tq);
103 Tt = FMA(KP984807753, Ts, Tp);
104 R0[WS(rs, 1)] = FNMS(KP1_969615506, Ts, Tp);
105 R0[WS(rs, 4)] = FMA(KP1_705737063, Tu, Tt);
106 R1[WS(rs, 2)] = FNMS(KP1_705737063, Tu, Tt);
107 }
108 }
109 }
110 }
111
112 static const kr2c_desc desc = { 9, "r2cb_9", {8, 0, 24, 0}, &GENUS };
113
114 void X(codelet_r2cb_9) (planner *p) {
115 X(kr2c_register) (p, r2cb_9, &desc);
116 }
117
118 #else
119
120 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 9 -name r2cb_9 -include rdft/scalar/r2cb.h */
121
122 /*
123 * This function contains 32 FP additions, 18 FP multiplications,
124 * (or, 22 additions, 8 multiplications, 10 fused multiply/add),
125 * 35 stack variables, 12 constants, and 18 memory accesses
126 */
127 #include "rdft/scalar/r2cb.h"
128
129 static void r2cb_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
130 {
131 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
132 DK(KP173648177, +0.173648177666930348851716626769314796000375677);
133 DK(KP300767466, +0.300767466360870593278543795225003852144476517);
134 DK(KP1_705737063, +1.705737063904886419256501927880148143872040591);
135 DK(KP642787609, +0.642787609686539326322643409907263432907559884);
136 DK(KP766044443, +0.766044443118978035202392650555416673935832457);
137 DK(KP1_326827896, +1.326827896337876792410842639271782594433726619);
138 DK(KP1_113340798, +1.113340798452838732905825904094046265936583811);
139 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
140 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
141 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
142 DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
143 {
144 INT i;
145 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)) {
146 E T3, Tq, Tc, Tk, Tj, T8, Tm, Ts, Th, Tr, Tw, Tx;
147 {
148 E Tb, T1, T2, T9, Ta;
149 Ta = Ci[WS(csi, 3)];
150 Tb = KP1_732050807 * Ta;
151 T1 = Cr[0];
152 T2 = Cr[WS(csr, 3)];
153 T9 = T1 - T2;
154 T3 = FMA(KP2_000000000, T2, T1);
155 Tq = T9 + Tb;
156 Tc = T9 - Tb;
157 }
158 {
159 E T4, T7, Ti, Tg, Tl, Td;
160 T4 = Cr[WS(csr, 1)];
161 Tk = Ci[WS(csi, 1)];
162 {
163 E T5, T6, Te, Tf;
164 T5 = Cr[WS(csr, 4)];
165 T6 = Cr[WS(csr, 2)];
166 T7 = T5 + T6;
167 Ti = KP866025403 * (T5 - T6);
168 Te = Ci[WS(csi, 4)];
169 Tf = Ci[WS(csi, 2)];
170 Tg = KP866025403 * (Te + Tf);
171 Tj = Tf - Te;
172 }
173 T8 = T4 + T7;
174 Tl = FMA(KP500000000, Tj, Tk);
175 Tm = Ti + Tl;
176 Ts = Tl - Ti;
177 Td = FNMS(KP500000000, T7, T4);
178 Th = Td - Tg;
179 Tr = Td + Tg;
180 }
181 R0[0] = FMA(KP2_000000000, T8, T3);
182 Tw = T3 - T8;
183 Tx = KP1_732050807 * (Tk - Tj);
184 R1[WS(rs, 1)] = Tw - Tx;
185 R0[WS(rs, 3)] = Tw + Tx;
186 {
187 E Tp, Tn, To, Tv, Tt, Tu;
188 Tp = FMA(KP1_113340798, Th, KP1_326827896 * Tm);
189 Tn = FNMS(KP642787609, Tm, KP766044443 * Th);
190 To = Tc - Tn;
191 R1[0] = FMA(KP2_000000000, Tn, Tc);
192 R1[WS(rs, 3)] = To + Tp;
193 R0[WS(rs, 2)] = To - Tp;
194 Tv = FMA(KP1_705737063, Tr, KP300767466 * Ts);
195 Tt = FNMS(KP984807753, Ts, KP173648177 * Tr);
196 Tu = Tq - Tt;
197 R0[WS(rs, 1)] = FMA(KP2_000000000, Tt, Tq);
198 R0[WS(rs, 4)] = Tu + Tv;
199 R1[WS(rs, 2)] = Tu - Tv;
200 }
201 }
202 }
203 }
204
205 static const kr2c_desc desc = { 9, "r2cb_9", {22, 8, 10, 0}, &GENUS };
206
207 void X(codelet_r2cb_9) (planner *p) {
208 X(kr2c_register) (p, r2cb_9, &desc);
209 }
210
211 #endif