Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cf_11.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 08:06:26 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 11 -name r2cf_11 -include rdft/scalar/r2cf.h */
29
30 /*
31 * This function contains 60 FP additions, 50 FP multiplications,
32 * (or, 15 additions, 5 multiplications, 45 fused multiply/add),
33 * 42 stack variables, 10 constants, and 22 memory accesses
34 */
35 #include "rdft/scalar/r2cf.h"
36
37 static void r2cf_11(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP918985947, +0.918985947228994779780736114132655398124909697);
40 DK(KP989821441, +0.989821441880932732376092037776718787376519372);
41 DK(KP830830026, +0.830830026003772851058548298459246407048009821);
42 DK(KP715370323, +0.715370323453429719112414662767260662417897278);
43 DK(KP959492973, +0.959492973614497389890368057066327699062454848);
44 DK(KP876768831, +0.876768831002589333891339807079336796764054852);
45 DK(KP778434453, +0.778434453334651800608337670740821884709317477);
46 DK(KP634356270, +0.634356270682424498893150776899916060542806975);
47 DK(KP342584725, +0.342584725681637509502641509861112333758894680);
48 DK(KP521108558, +0.521108558113202722944698153526659300680427422);
49 {
50 INT i;
51 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
52 E T1, T4, TC, Tg, TE, T7, TD, Ta, TF, Td, TB, TG, TM, TS, TJ;
53 E TP, Ty, Tq, Ti, Tu, Tm, T5, T6;
54 T1 = R0[0];
55 {
56 E T2, T3, Te, Tf;
57 T2 = R1[0];
58 T3 = R0[WS(rs, 5)];
59 T4 = T2 + T3;
60 TC = T3 - T2;
61 Te = R1[WS(rs, 2)];
62 Tf = R0[WS(rs, 3)];
63 Tg = Te + Tf;
64 TE = Tf - Te;
65 }
66 T5 = R0[WS(rs, 1)];
67 T6 = R1[WS(rs, 4)];
68 T7 = T5 + T6;
69 TD = T5 - T6;
70 {
71 E T8, T9, Tb, Tc;
72 T8 = R1[WS(rs, 1)];
73 T9 = R0[WS(rs, 4)];
74 Ta = T8 + T9;
75 TF = T9 - T8;
76 Tb = R0[WS(rs, 2)];
77 Tc = R1[WS(rs, 3)];
78 Td = Tb + Tc;
79 TB = Tb - Tc;
80 }
81 TG = FMA(KP521108558, TF, TE);
82 TM = FNMS(KP521108558, TD, TB);
83 TS = FMA(KP521108558, TC, TD);
84 TJ = FMA(KP521108558, TE, TC);
85 TP = FNMS(KP521108558, TB, TF);
86 {
87 E Tx, Tp, Th, Tt, Tl;
88 Tx = FNMS(KP342584725, Ta, T7);
89 Ty = FNMS(KP634356270, Tx, Td);
90 Tp = FNMS(KP342584725, T4, Ta);
91 Tq = FNMS(KP634356270, Tp, Tg);
92 Th = FNMS(KP342584725, Tg, Td);
93 Ti = FNMS(KP634356270, Th, Ta);
94 Tt = FNMS(KP342584725, Td, T4);
95 Tu = FNMS(KP634356270, Tt, T7);
96 Tl = FNMS(KP342584725, T7, Tg);
97 Tm = FNMS(KP634356270, Tl, T4);
98 }
99 {
100 E To, Tn, TI, TH;
101 {
102 E Tk, Tj, TU, TT;
103 Tj = FNMS(KP778434453, Ti, T7);
104 Tk = FNMS(KP876768831, Tj, T4);
105 Cr[WS(csr, 5)] = FNMS(KP959492973, Tk, T1);
106 TT = FMA(KP715370323, TS, TF);
107 TU = FMA(KP830830026, TT, TB);
108 Ci[WS(csi, 5)] = KP989821441 * (FMA(KP918985947, TU, TE));
109 }
110 Tn = FNMS(KP778434453, Tm, Ta);
111 To = FNMS(KP876768831, Tn, Td);
112 Cr[WS(csr, 4)] = FNMS(KP959492973, To, T1);
113 {
114 E TR, TQ, Ts, Tr;
115 TQ = FMA(KP715370323, TP, TC);
116 TR = FNMS(KP830830026, TQ, TE);
117 Ci[WS(csi, 4)] = KP989821441 * (FNMS(KP918985947, TR, TD));
118 Tr = FNMS(KP778434453, Tq, Td);
119 Ts = FNMS(KP876768831, Tr, T7);
120 Cr[WS(csr, 3)] = FNMS(KP959492973, Ts, T1);
121 }
122 {
123 E TO, TN, Tw, Tv;
124 TN = FNMS(KP715370323, TM, TE);
125 TO = FNMS(KP830830026, TN, TF);
126 Ci[WS(csi, 3)] = KP989821441 * (FNMS(KP918985947, TO, TC));
127 Tv = FNMS(KP778434453, Tu, Tg);
128 Tw = FNMS(KP876768831, Tv, Ta);
129 Cr[WS(csr, 2)] = FNMS(KP959492973, Tw, T1);
130 Cr[0] = T1 + T4 + T7 + Ta + Td + Tg;
131 }
132 TH = FMA(KP715370323, TG, TD);
133 TI = FNMS(KP830830026, TH, TC);
134 Ci[WS(csi, 2)] = KP989821441 * (FMA(KP918985947, TI, TB));
135 {
136 E TL, TK, TA, Tz;
137 TK = FNMS(KP715370323, TJ, TB);
138 TL = FMA(KP830830026, TK, TD);
139 Ci[WS(csi, 1)] = KP989821441 * (FNMS(KP918985947, TL, TF));
140 Tz = FNMS(KP778434453, Ty, T4);
141 TA = FNMS(KP876768831, Tz, Tg);
142 Cr[WS(csr, 1)] = FNMS(KP959492973, TA, T1);
143 }
144 }
145 }
146 }
147 }
148
149 static const kr2c_desc desc = { 11, "r2cf_11", {15, 5, 45, 0}, &GENUS };
150
151 void X(codelet_r2cf_11) (planner *p) {
152 X(kr2c_register) (p, r2cf_11, &desc);
153 }
154
155 #else
156
157 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 11 -name r2cf_11 -include rdft/scalar/r2cf.h */
158
159 /*
160 * This function contains 60 FP additions, 50 FP multiplications,
161 * (or, 20 additions, 10 multiplications, 40 fused multiply/add),
162 * 28 stack variables, 10 constants, and 22 memory accesses
163 */
164 #include "rdft/scalar/r2cf.h"
165
166 static void r2cf_11(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
167 {
168 DK(KP654860733, +0.654860733945285064056925072466293553183791199);
169 DK(KP142314838, +0.142314838273285140443792668616369668791051361);
170 DK(KP959492973, +0.959492973614497389890368057066327699062454848);
171 DK(KP415415013, +0.415415013001886425529274149229623203524004910);
172 DK(KP841253532, +0.841253532831181168861811648919367717513292498);
173 DK(KP989821441, +0.989821441880932732376092037776718787376519372);
174 DK(KP909631995, +0.909631995354518371411715383079028460060241051);
175 DK(KP281732556, +0.281732556841429697711417915346616899035777899);
176 DK(KP540640817, +0.540640817455597582107635954318691695431770608);
177 DK(KP755749574, +0.755749574354258283774035843972344420179717445);
178 {
179 INT i;
180 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
181 E T1, T4, Tl, Tg, Th, Td, Ti, Ta, Tk, T7, Tj, Tb, Tc;
182 T1 = R0[0];
183 {
184 E T2, T3, Te, Tf;
185 T2 = R0[WS(rs, 1)];
186 T3 = R1[WS(rs, 4)];
187 T4 = T2 + T3;
188 Tl = T3 - T2;
189 Te = R1[0];
190 Tf = R0[WS(rs, 5)];
191 Tg = Te + Tf;
192 Th = Tf - Te;
193 }
194 Tb = R1[WS(rs, 1)];
195 Tc = R0[WS(rs, 4)];
196 Td = Tb + Tc;
197 Ti = Tc - Tb;
198 {
199 E T8, T9, T5, T6;
200 T8 = R1[WS(rs, 2)];
201 T9 = R0[WS(rs, 3)];
202 Ta = T8 + T9;
203 Tk = T9 - T8;
204 T5 = R0[WS(rs, 2)];
205 T6 = R1[WS(rs, 3)];
206 T7 = T5 + T6;
207 Tj = T6 - T5;
208 }
209 Ci[WS(csi, 4)] = FMA(KP755749574, Th, KP540640817 * Ti) + FNMS(KP909631995, Tk, KP281732556 * Tj) - (KP989821441 * Tl);
210 Cr[WS(csr, 4)] = FMA(KP841253532, Td, T1) + FNMS(KP959492973, T7, KP415415013 * Ta) + FNMA(KP142314838, T4, KP654860733 * Tg);
211 Ci[WS(csi, 2)] = FMA(KP909631995, Th, KP755749574 * Tl) + FNMA(KP540640817, Tk, KP989821441 * Tj) - (KP281732556 * Ti);
212 Ci[WS(csi, 5)] = FMA(KP281732556, Th, KP755749574 * Ti) + FNMS(KP909631995, Tj, KP989821441 * Tk) - (KP540640817 * Tl);
213 Ci[WS(csi, 1)] = FMA(KP540640817, Th, KP909631995 * Tl) + FMA(KP989821441, Ti, KP755749574 * Tj) + (KP281732556 * Tk);
214 Ci[WS(csi, 3)] = FMA(KP989821441, Th, KP540640817 * Tj) + FNMS(KP909631995, Ti, KP755749574 * Tk) - (KP281732556 * Tl);
215 Cr[WS(csr, 3)] = FMA(KP415415013, Td, T1) + FNMS(KP654860733, Ta, KP841253532 * T7) + FNMA(KP959492973, T4, KP142314838 * Tg);
216 Cr[WS(csr, 1)] = FMA(KP841253532, Tg, T1) + FNMS(KP959492973, Ta, KP415415013 * T4) + FNMA(KP654860733, T7, KP142314838 * Td);
217 Cr[0] = T1 + Tg + T4 + Td + T7 + Ta;
218 Cr[WS(csr, 2)] = FMA(KP415415013, Tg, T1) + FNMS(KP142314838, T7, KP841253532 * Ta) + FNMA(KP959492973, Td, KP654860733 * T4);
219 Cr[WS(csr, 5)] = FMA(KP841253532, T4, T1) + FNMS(KP142314838, Ta, KP415415013 * T7) + FNMA(KP654860733, Td, KP959492973 * Tg);
220 }
221 }
222 }
223
224 static const kr2c_desc desc = { 11, "r2cf_11", {20, 10, 40, 0}, &GENUS };
225
226 void X(codelet_r2cf_11) (planner *p) {
227 X(kr2c_register) (p, r2cf_11, &desc);
228 }
229
230 #endif