Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cb/r2cb_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 08:07:28 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 11 -name r2cb_11 -include rdft/scalar/r2cb.h */
29
30 /*
31 * This function contains 60 FP additions, 56 FP multiplications,
32 * (or, 4 additions, 0 multiplications, 56 fused multiply/add),
33 * 44 stack variables, 11 constants, and 22 memory accesses
34 */
35 #include "rdft/scalar/r2cb.h"
36
37 static void r2cb_11(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP1_979642883, +1.979642883761865464752184075553437574753038744);
40 DK(KP918985947, +0.918985947228994779780736114132655398124909697);
41 DK(KP830830026, +0.830830026003772851058548298459246407048009821);
42 DK(KP1_918985947, +1.918985947228994779780736114132655398124909697);
43 DK(KP876768831, +0.876768831002589333891339807079336796764054852);
44 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
45 DK(KP778434453, +0.778434453334651800608337670740821884709317477);
46 DK(KP634356270, +0.634356270682424498893150776899916060542806975);
47 DK(KP342584725, +0.342584725681637509502641509861112333758894680);
48 DK(KP715370323, +0.715370323453429719112414662767260662417897278);
49 DK(KP521108558, +0.521108558113202722944698153526659300680427422);
50 {
51 INT i;
52 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
53 E T1, Td, Th, Te, Tf, Tg, Tj, TT, Ts, TB, TK, T2, T6, T3, T4;
54 E T5, Ta, To, TP, TG, Tx, T7;
55 T1 = Cr[0];
56 {
57 E Ti, TS, Tr, TA, TJ;
58 Td = Ci[WS(csi, 3)];
59 Th = Ci[WS(csi, 5)];
60 Te = Ci[WS(csi, 2)];
61 Tf = Ci[WS(csi, 4)];
62 Tg = Ci[WS(csi, 1)];
63 Ti = FMA(KP521108558, Th, Tg);
64 TS = FMS(KP521108558, Tg, Te);
65 Tr = FMA(KP521108558, Td, Th);
66 TA = FNMS(KP521108558, Te, Tf);
67 TJ = FMA(KP521108558, Tf, Td);
68 Tj = FMA(KP715370323, Ti, Tf);
69 TT = FMA(KP715370323, TS, Td);
70 Ts = FNMS(KP715370323, Tr, Te);
71 TB = FMA(KP715370323, TA, Th);
72 TK = FMA(KP715370323, TJ, Tg);
73 }
74 {
75 E T8, TN, Tm, Tv, TE;
76 T2 = Cr[WS(csr, 1)];
77 T6 = Cr[WS(csr, 5)];
78 T3 = Cr[WS(csr, 2)];
79 T4 = Cr[WS(csr, 3)];
80 T5 = Cr[WS(csr, 4)];
81 T8 = FNMS(KP342584725, T4, T3);
82 TN = FNMS(KP342584725, T6, T5);
83 Tm = FNMS(KP342584725, T5, T2);
84 Tv = FNMS(KP342584725, T2, T4);
85 TE = FNMS(KP342584725, T3, T6);
86 {
87 E T9, Tn, TO, TF, Tw;
88 T9 = FNMS(KP634356270, T8, T5);
89 Ta = FNMS(KP778434453, T9, T2);
90 Tn = FNMS(KP634356270, Tm, T3);
91 To = FNMS(KP778434453, Tn, T6);
92 TO = FNMS(KP634356270, TN, T4);
93 TP = FNMS(KP778434453, TO, T3);
94 TF = FNMS(KP634356270, TE, T2);
95 TG = FNMS(KP778434453, TF, T4);
96 Tw = FNMS(KP634356270, Tv, T6);
97 Tx = FNMS(KP778434453, Tw, T5);
98 T7 = T2 + T3 + T4 + T5 + T6;
99 }
100 }
101 R0[0] = FMA(KP2_000000000, T7, T1);
102 {
103 E Tc, Tl, Tb, Tk;
104 Tb = FNMS(KP876768831, Ta, T6);
105 Tc = FNMS(KP1_918985947, Tb, T1);
106 Tk = FMA(KP830830026, Tj, Te);
107 Tl = FMA(KP918985947, Tk, Td);
108 R1[0] = FNMS(KP1_979642883, Tl, Tc);
109 R0[WS(rs, 5)] = FMA(KP1_979642883, Tl, Tc);
110 }
111 {
112 E TR, TV, TQ, TU;
113 TQ = FNMS(KP876768831, TP, T2);
114 TR = FNMS(KP1_918985947, TQ, T1);
115 TU = FNMS(KP830830026, TT, Tf);
116 TV = FNMS(KP918985947, TU, Th);
117 R1[WS(rs, 2)] = FNMS(KP1_979642883, TV, TR);
118 R0[WS(rs, 3)] = FMA(KP1_979642883, TV, TR);
119 }
120 {
121 E TI, TM, TH, TL;
122 TH = FNMS(KP876768831, TG, T5);
123 TI = FNMS(KP1_918985947, TH, T1);
124 TL = FNMS(KP830830026, TK, Th);
125 TM = FMA(KP918985947, TL, Te);
126 R1[WS(rs, 3)] = FNMS(KP1_979642883, TM, TI);
127 R0[WS(rs, 2)] = FMA(KP1_979642883, TM, TI);
128 }
129 {
130 E Tz, TD, Ty, TC;
131 Ty = FNMS(KP876768831, Tx, T3);
132 Tz = FNMS(KP1_918985947, Ty, T1);
133 TC = FNMS(KP830830026, TB, Td);
134 TD = FNMS(KP918985947, TC, Tg);
135 R1[WS(rs, 1)] = FNMS(KP1_979642883, TD, Tz);
136 R0[WS(rs, 4)] = FMA(KP1_979642883, TD, Tz);
137 }
138 {
139 E Tq, Tu, Tp, Tt;
140 Tp = FNMS(KP876768831, To, T4);
141 Tq = FNMS(KP1_918985947, Tp, T1);
142 Tt = FMA(KP830830026, Ts, Tg);
143 Tu = FNMS(KP918985947, Tt, Tf);
144 R1[WS(rs, 4)] = FNMS(KP1_979642883, Tu, Tq);
145 R0[WS(rs, 1)] = FMA(KP1_979642883, Tu, Tq);
146 }
147 }
148 }
149 }
150
151 static const kr2c_desc desc = { 11, "r2cb_11", {4, 0, 56, 0}, &GENUS };
152
153 void X(codelet_r2cb_11) (planner *p) {
154 X(kr2c_register) (p, r2cb_11, &desc);
155 }
156
157 #else
158
159 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 11 -name r2cb_11 -include rdft/scalar/r2cb.h */
160
161 /*
162 * This function contains 60 FP additions, 51 FP multiplications,
163 * (or, 19 additions, 10 multiplications, 41 fused multiply/add),
164 * 33 stack variables, 11 constants, and 22 memory accesses
165 */
166 #include "rdft/scalar/r2cb.h"
167
168 static void r2cb_11(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
169 {
170 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
171 DK(KP1_918985947, +1.918985947228994779780736114132655398124909697);
172 DK(KP1_309721467, +1.309721467890570128113850144932587106367582399);
173 DK(KP284629676, +0.284629676546570280887585337232739337582102722);
174 DK(KP830830026, +0.830830026003772851058548298459246407048009821);
175 DK(KP1_682507065, +1.682507065662362337723623297838735435026584997);
176 DK(KP563465113, +0.563465113682859395422835830693233798071555798);
177 DK(KP1_511499148, +1.511499148708516567548071687944688840359434890);
178 DK(KP1_979642883, +1.979642883761865464752184075553437574753038744);
179 DK(KP1_819263990, +1.819263990709036742823430766158056920120482102);
180 DK(KP1_081281634, +1.081281634911195164215271908637383390863541216);
181 {
182 INT i;
183 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
184 E Td, Tl, Tf, Th, Tj, T1, T2, T6, T5, T4, T3, T7, Tk, Te, Tg;
185 E Ti;
186 {
187 E T8, Tc, T9, Ta, Tb;
188 T8 = Ci[WS(csi, 2)];
189 Tc = Ci[WS(csi, 1)];
190 T9 = Ci[WS(csi, 4)];
191 Ta = Ci[WS(csi, 5)];
192 Tb = Ci[WS(csi, 3)];
193 Td = FMA(KP1_081281634, T8, KP1_819263990 * T9) + FNMA(KP1_979642883, Ta, KP1_511499148 * Tb) - (KP563465113 * Tc);
194 Tl = FMA(KP1_979642883, T8, KP1_819263990 * Ta) + FNMA(KP563465113, T9, KP1_081281634 * Tb) - (KP1_511499148 * Tc);
195 Tf = FMA(KP563465113, T8, KP1_819263990 * Tb) + FNMA(KP1_511499148, Ta, KP1_081281634 * T9) - (KP1_979642883 * Tc);
196 Th = FMA(KP1_081281634, Tc, KP1_819263990 * T8) + FMA(KP1_979642883, Tb, KP1_511499148 * T9) + (KP563465113 * Ta);
197 Tj = FMA(KP563465113, Tb, KP1_979642883 * T9) + FNMS(KP1_511499148, T8, KP1_081281634 * Ta) - (KP1_819263990 * Tc);
198 }
199 T1 = Cr[0];
200 T2 = Cr[WS(csr, 1)];
201 T6 = Cr[WS(csr, 5)];
202 T5 = Cr[WS(csr, 4)];
203 T4 = Cr[WS(csr, 3)];
204 T3 = Cr[WS(csr, 2)];
205 T7 = FMA(KP1_682507065, T3, T1) + FNMS(KP284629676, T6, KP830830026 * T5) + FNMA(KP1_309721467, T4, KP1_918985947 * T2);
206 Tk = FMA(KP1_682507065, T4, T1) + FNMS(KP1_918985947, T5, KP830830026 * T6) + FNMA(KP284629676, T3, KP1_309721467 * T2);
207 Te = FMA(KP830830026, T4, T1) + FNMS(KP1_309721467, T6, KP1_682507065 * T5) + FNMA(KP1_918985947, T3, KP284629676 * T2);
208 Tg = FMA(KP1_682507065, T2, T1) + FNMS(KP1_918985947, T6, KP830830026 * T3) + FNMA(KP1_309721467, T5, KP284629676 * T4);
209 Ti = FMA(KP830830026, T2, T1) + FNMS(KP284629676, T5, KP1_682507065 * T6) + FNMA(KP1_918985947, T4, KP1_309721467 * T3);
210 R0[WS(rs, 3)] = T7 - Td;
211 R0[WS(rs, 4)] = Te - Tf;
212 R0[WS(rs, 2)] = Tk + Tl;
213 R1[WS(rs, 2)] = T7 + Td;
214 R1[WS(rs, 3)] = Tk - Tl;
215 R0[WS(rs, 1)] = Ti + Tj;
216 R1[WS(rs, 1)] = Te + Tf;
217 R0[WS(rs, 5)] = Tg + Th;
218 R1[0] = Tg - Th;
219 R1[WS(rs, 4)] = Ti - Tj;
220 R0[0] = FMA(KP2_000000000, T2 + T3 + T4 + T5 + T6, T1);
221 }
222 }
223 }
224
225 static const kr2c_desc desc = { 11, "r2cb_11", {19, 10, 41, 0}, &GENUS };
226
227 void X(codelet_r2cb_11) (planner *p) {
228 X(kr2c_register) (p, r2cb_11, &desc);
229 }
230
231 #endif