|
Chris@82
|
1 /*
|
|
Chris@82
|
2 * Copyright (c) 2003, 2007-14 Matteo Frigo
|
|
Chris@82
|
3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
|
|
Chris@82
|
4 *
|
|
Chris@82
|
5 * This program is free software; you can redistribute it and/or modify
|
|
Chris@82
|
6 * it under the terms of the GNU General Public License as published by
|
|
Chris@82
|
7 * the Free Software Foundation; either version 2 of the License, or
|
|
Chris@82
|
8 * (at your option) any later version.
|
|
Chris@82
|
9 *
|
|
Chris@82
|
10 * This program is distributed in the hope that it will be useful,
|
|
Chris@82
|
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
Chris@82
|
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
Chris@82
|
13 * GNU General Public License for more details.
|
|
Chris@82
|
14 *
|
|
Chris@82
|
15 * You should have received a copy of the GNU General Public License
|
|
Chris@82
|
16 * along with this program; if not, write to the Free Software
|
|
Chris@82
|
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
Chris@82
|
18 *
|
|
Chris@82
|
19 */
|
|
Chris@82
|
20
|
|
Chris@82
|
21 /* This file was automatically generated --- DO NOT EDIT */
|
|
Chris@82
|
22 /* Generated on Thu May 24 08:06:43 EDT 2018 */
|
|
Chris@82
|
23
|
|
Chris@82
|
24 #include "rdft/codelet-rdft.h"
|
|
Chris@82
|
25
|
|
Chris@82
|
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
|
|
Chris@82
|
27
|
|
Chris@82
|
28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 9 -name r2cfII_9 -dft-II -include rdft/scalar/r2cfII.h */
|
|
Chris@82
|
29
|
|
Chris@82
|
30 /*
|
|
Chris@82
|
31 * This function contains 42 FP additions, 34 FP multiplications,
|
|
Chris@82
|
32 * (or, 12 additions, 4 multiplications, 30 fused multiply/add),
|
|
Chris@82
|
33 * 48 stack variables, 17 constants, and 18 memory accesses
|
|
Chris@82
|
34 */
|
|
Chris@82
|
35 #include "rdft/scalar/r2cfII.h"
|
|
Chris@82
|
36
|
|
Chris@82
|
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)
|
|
Chris@82
|
38 {
|
|
Chris@82
|
39 DK(KP852868531, +0.852868531952443209628250963940074071936020296);
|
|
Chris@82
|
40 DK(KP879385241, +0.879385241571816768108218554649462939872416269);
|
|
Chris@82
|
41 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
|
|
Chris@82
|
42 DK(KP898197570, +0.898197570222573798468955502359086394667167570);
|
|
Chris@82
|
43 DK(KP673648177, +0.673648177666930348851716626769314796000375677);
|
|
Chris@82
|
44 DK(KP939692620, +0.939692620785908384054109277324731469936208134);
|
|
Chris@82
|
45 DK(KP907603734, +0.907603734547952313649323976213898122064543220);
|
|
Chris@82
|
46 DK(KP666666666, +0.666666666666666666666666666666666666666666667);
|
|
Chris@82
|
47 DK(KP826351822, +0.826351822333069651148283373230685203999624323);
|
|
Chris@82
|
48 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
|
|
Chris@82
|
49 DK(KP315207469, +0.315207469095904627298647952427796244129086440);
|
|
Chris@82
|
50 DK(KP420276625, +0.420276625461206169731530603237061658838781920);
|
|
Chris@82
|
51 DK(KP203604859, +0.203604859554852403062088995281827210665664861);
|
|
Chris@82
|
52 DK(KP152703644, +0.152703644666139302296566746461370407999248646);
|
|
Chris@82
|
53 DK(KP726681596, +0.726681596905677465811651808188092531873167623);
|
|
Chris@82
|
54 DK(KP968908795, +0.968908795874236621082202410917456709164223497);
|
|
Chris@82
|
55 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
|
|
Chris@82
|
56 {
|
|
Chris@82
|
57 INT i;
|
|
Chris@82
|
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)) {
|
|
Chris@82
|
59 E T1, T4, To, Ta, Tm, TB, Tq, Tt, Tf, Tj, TA, Tr, Ts, T2, T3;
|
|
Chris@82
|
60 E T5, Tg;
|
|
Chris@82
|
61 T1 = R0[0];
|
|
Chris@82
|
62 T2 = R0[WS(rs, 3)];
|
|
Chris@82
|
63 T3 = R1[WS(rs, 1)];
|
|
Chris@82
|
64 T4 = T2 - T3;
|
|
Chris@82
|
65 To = T2 + T3;
|
|
Chris@82
|
66 {
|
|
Chris@82
|
67 E T6, T9, Tk, T7, T8, Tl;
|
|
Chris@82
|
68 T6 = R0[WS(rs, 1)];
|
|
Chris@82
|
69 T7 = R0[WS(rs, 4)];
|
|
Chris@82
|
70 T8 = R1[WS(rs, 2)];
|
|
Chris@82
|
71 T9 = T7 - T8;
|
|
Chris@82
|
72 Tk = T7 + T8;
|
|
Chris@82
|
73 Ta = T6 + T9;
|
|
Chris@82
|
74 Tl = FNMS(KP500000000, T9, T6);
|
|
Chris@82
|
75 Tm = FMA(KP968908795, Tl, Tk);
|
|
Chris@82
|
76 TB = FNMS(KP726681596, Tk, Tl);
|
|
Chris@82
|
77 Tq = FNMS(KP152703644, Tk, Tl);
|
|
Chris@82
|
78 Tt = FMA(KP203604859, Tl, Tk);
|
|
Chris@82
|
79 }
|
|
Chris@82
|
80 {
|
|
Chris@82
|
81 E Tb, Te, Ti, Tc, Td, Th;
|
|
Chris@82
|
82 Tb = R0[WS(rs, 2)];
|
|
Chris@82
|
83 Tc = R1[0];
|
|
Chris@82
|
84 Td = R1[WS(rs, 3)];
|
|
Chris@82
|
85 Te = Tc + Td;
|
|
Chris@82
|
86 Ti = Tc - Td;
|
|
Chris@82
|
87 Tf = Tb - Te;
|
|
Chris@82
|
88 Th = FMA(KP500000000, Te, Tb);
|
|
Chris@82
|
89 Tj = FNMS(KP152703644, Ti, Th);
|
|
Chris@82
|
90 TA = FMA(KP203604859, Th, Ti);
|
|
Chris@82
|
91 Tr = FNMS(KP420276625, Th, Ti);
|
|
Chris@82
|
92 Ts = FMA(KP315207469, Ti, Th);
|
|
Chris@82
|
93 }
|
|
Chris@82
|
94 Ci[WS(csi, 1)] = KP866025403 * (Tf - Ta);
|
|
Chris@82
|
95 T5 = T1 + T4;
|
|
Chris@82
|
96 Tg = Ta + Tf;
|
|
Chris@82
|
97 Cr[WS(csr, 1)] = FNMS(KP500000000, Tg, T5);
|
|
Chris@82
|
98 Cr[WS(csr, 4)] = T5 + Tg;
|
|
Chris@82
|
99 {
|
|
Chris@82
|
100 E Ty, Tx, Tz, Tn, TD, TC;
|
|
Chris@82
|
101 Tx = FNMS(KP826351822, Tr, Tq);
|
|
Chris@82
|
102 Ty = FNMS(KP666666666, Tx, Tt);
|
|
Chris@82
|
103 Tz = FMA(KP907603734, Ty, Ts);
|
|
Chris@82
|
104 Ci[WS(csi, 2)] = KP866025403 * (FNMS(KP939692620, Tz, To));
|
|
Chris@82
|
105 Tn = FMA(KP673648177, Tm, Tj);
|
|
Chris@82
|
106 TC = FNMS(KP898197570, TB, TA);
|
|
Chris@82
|
107 TD = FNMS(KP666666666, Tn, TC);
|
|
Chris@82
|
108 Ci[0] = -(KP984807753 * (FMA(KP879385241, To, Tn)));
|
|
Chris@82
|
109 Ci[WS(csi, 3)] = -(KP866025403 * (FMA(KP852868531, TD, To)));
|
|
Chris@82
|
110 {
|
|
Chris@82
|
111 E Tp, Tv, TF, TG, Tu, TE, Tw;
|
|
Chris@82
|
112 Tp = FNMS(KP500000000, T4, T1);
|
|
Chris@82
|
113 Tu = FNMS(KP907603734, Tt, Ts);
|
|
Chris@82
|
114 Tv = FNMS(KP666666666, Tu, Tr);
|
|
Chris@82
|
115 TE = FNMS(KP673648177, Tm, Tj);
|
|
Chris@82
|
116 TF = FMA(KP898197570, TB, TA);
|
|
Chris@82
|
117 TG = FMA(KP500000000, TF, TE);
|
|
Chris@82
|
118 Cr[WS(csr, 3)] = FNMS(KP852868531, TG, Tp);
|
|
Chris@82
|
119 Cr[0] = FMA(KP852868531, TF, Tp);
|
|
Chris@82
|
120 Tw = FMA(KP826351822, Tv, Tq);
|
|
Chris@82
|
121 Cr[WS(csr, 2)] = FNMS(KP852868531, Tw, Tp);
|
|
Chris@82
|
122 }
|
|
Chris@82
|
123 }
|
|
Chris@82
|
124 }
|
|
Chris@82
|
125 }
|
|
Chris@82
|
126 }
|
|
Chris@82
|
127
|
|
Chris@82
|
128 static const kr2c_desc desc = { 9, "r2cfII_9", {12, 4, 30, 0}, &GENUS };
|
|
Chris@82
|
129
|
|
Chris@82
|
130 void X(codelet_r2cfII_9) (planner *p) {
|
|
Chris@82
|
131 X(kr2c_register) (p, r2cfII_9, &desc);
|
|
Chris@82
|
132 }
|
|
Chris@82
|
133
|
|
Chris@82
|
134 #else
|
|
Chris@82
|
135
|
|
Chris@82
|
136 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 9 -name r2cfII_9 -dft-II -include rdft/scalar/r2cfII.h */
|
|
Chris@82
|
137
|
|
Chris@82
|
138 /*
|
|
Chris@82
|
139 * This function contains 42 FP additions, 30 FP multiplications,
|
|
Chris@82
|
140 * (or, 25 additions, 13 multiplications, 17 fused multiply/add),
|
|
Chris@82
|
141 * 39 stack variables, 14 constants, and 18 memory accesses
|
|
Chris@82
|
142 */
|
|
Chris@82
|
143 #include "rdft/scalar/r2cfII.h"
|
|
Chris@82
|
144
|
|
Chris@82
|
145 static void r2cfII_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
|
|
Chris@82
|
146 {
|
|
Chris@82
|
147 DK(KP663413948, +0.663413948168938396205421319635891297216863310);
|
|
Chris@82
|
148 DK(KP642787609, +0.642787609686539326322643409907263432907559884);
|
|
Chris@82
|
149 DK(KP556670399, +0.556670399226419366452912952047023132968291906);
|
|
Chris@82
|
150 DK(KP766044443, +0.766044443118978035202392650555416673935832457);
|
|
Chris@82
|
151 DK(KP852868531, +0.852868531952443209628250963940074071936020296);
|
|
Chris@82
|
152 DK(KP173648177, +0.173648177666930348851716626769314796000375677);
|
|
Chris@82
|
153 DK(KP984807753, +0.984807753012208059366743024589523013670643252);
|
|
Chris@82
|
154 DK(KP150383733, +0.150383733180435296639271897612501926072238258);
|
|
Chris@82
|
155 DK(KP813797681, +0.813797681349373692844693217248393223289101568);
|
|
Chris@82
|
156 DK(KP342020143, +0.342020143325668733044099614682259580763083368);
|
|
Chris@82
|
157 DK(KP939692620, +0.939692620785908384054109277324731469936208134);
|
|
Chris@82
|
158 DK(KP296198132, +0.296198132726023843175338011893050938967728390);
|
|
Chris@82
|
159 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
|
|
Chris@82
|
160 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
|
|
Chris@82
|
161 {
|
|
Chris@82
|
162 INT i;
|
|
Chris@82
|
163 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)) {
|
|
Chris@82
|
164 E T1, T4, To, Ta, Tl, Tk, Tf, Ti, Th, T2, T3, T5, Tg;
|
|
Chris@82
|
165 T1 = R0[0];
|
|
Chris@82
|
166 T2 = R1[WS(rs, 1)];
|
|
Chris@82
|
167 T3 = R0[WS(rs, 3)];
|
|
Chris@82
|
168 T4 = T2 - T3;
|
|
Chris@82
|
169 To = T2 + T3;
|
|
Chris@82
|
170 {
|
|
Chris@82
|
171 E T6, T7, T8, T9;
|
|
Chris@82
|
172 T6 = R0[WS(rs, 1)];
|
|
Chris@82
|
173 T7 = R1[WS(rs, 2)];
|
|
Chris@82
|
174 T8 = R0[WS(rs, 4)];
|
|
Chris@82
|
175 T9 = T7 - T8;
|
|
Chris@82
|
176 Ta = T6 - T9;
|
|
Chris@82
|
177 Tl = T7 + T8;
|
|
Chris@82
|
178 Tk = FMA(KP500000000, T9, T6);
|
|
Chris@82
|
179 }
|
|
Chris@82
|
180 {
|
|
Chris@82
|
181 E Tb, Tc, Td, Te;
|
|
Chris@82
|
182 Tb = R0[WS(rs, 2)];
|
|
Chris@82
|
183 Tc = R1[0];
|
|
Chris@82
|
184 Td = R1[WS(rs, 3)];
|
|
Chris@82
|
185 Te = Tc + Td;
|
|
Chris@82
|
186 Tf = Tb - Te;
|
|
Chris@82
|
187 Ti = FMA(KP500000000, Te, Tb);
|
|
Chris@82
|
188 Th = Tc - Td;
|
|
Chris@82
|
189 }
|
|
Chris@82
|
190 Ci[WS(csi, 1)] = KP866025403 * (Tf - Ta);
|
|
Chris@82
|
191 T5 = T1 - T4;
|
|
Chris@82
|
192 Tg = Ta + Tf;
|
|
Chris@82
|
193 Cr[WS(csr, 1)] = FNMS(KP500000000, Tg, T5);
|
|
Chris@82
|
194 Cr[WS(csr, 4)] = T5 + Tg;
|
|
Chris@82
|
195 {
|
|
Chris@82
|
196 E Tr, Tt, Tw, Tv, Tu, Tp, Tq, Ts, Tj, Tm, Tn;
|
|
Chris@82
|
197 Tr = FMA(KP500000000, T4, T1);
|
|
Chris@82
|
198 Tt = FMA(KP296198132, Th, KP939692620 * Ti);
|
|
Chris@82
|
199 Tw = FNMS(KP813797681, Th, KP342020143 * Ti);
|
|
Chris@82
|
200 Tv = FNMS(KP984807753, Tk, KP150383733 * Tl);
|
|
Chris@82
|
201 Tu = FMA(KP173648177, Tk, KP852868531 * Tl);
|
|
Chris@82
|
202 Tp = FNMS(KP556670399, Tl, KP766044443 * Tk);
|
|
Chris@82
|
203 Tq = FMA(KP852868531, Th, KP173648177 * Ti);
|
|
Chris@82
|
204 Ts = Tp + Tq;
|
|
Chris@82
|
205 Tj = FNMS(KP984807753, Ti, KP150383733 * Th);
|
|
Chris@82
|
206 Tm = FMA(KP642787609, Tk, KP663413948 * Tl);
|
|
Chris@82
|
207 Tn = Tj - Tm;
|
|
Chris@82
|
208 Ci[0] = FNMS(KP866025403, To, Tn);
|
|
Chris@82
|
209 Cr[0] = Tr + Ts;
|
|
Chris@82
|
210 Ci[WS(csi, 3)] = FNMS(KP500000000, Tn, KP866025403 * ((Tp - Tq) - To));
|
|
Chris@82
|
211 Cr[WS(csr, 3)] = FMA(KP866025403, Tm + Tj, Tr) - (KP500000000 * Ts);
|
|
Chris@82
|
212 Ci[WS(csi, 2)] = FMA(KP866025403, To - (Tu + Tt), KP500000000 * (Tw - Tv));
|
|
Chris@82
|
213 Cr[WS(csr, 2)] = FMA(KP500000000, Tt - Tu, Tr) + (KP866025403 * (Tv + Tw));
|
|
Chris@82
|
214 }
|
|
Chris@82
|
215 }
|
|
Chris@82
|
216 }
|
|
Chris@82
|
217 }
|
|
Chris@82
|
218
|
|
Chris@82
|
219 static const kr2c_desc desc = { 9, "r2cfII_9", {25, 13, 17, 0}, &GENUS };
|
|
Chris@82
|
220
|
|
Chris@82
|
221 void X(codelet_r2cfII_9) (planner *p) {
|
|
Chris@82
|
222 X(kr2c_register) (p, r2cfII_9, &desc);
|
|
Chris@82
|
223 }
|
|
Chris@82
|
224
|
|
Chris@82
|
225 #endif
|
