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 Dec 10 07:05:49 EST 2020 */
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 9 -name r2cf_9 -include rdft/scalar/r2cf.h */
29 
30 /*
31  * This function contains 38 FP additions, 30 FP multiplications,
32  * (or, 12 additions, 4 multiplications, 26 fused multiply/add),
33  * 48 stack variables, 18 constants, and 18 memory accesses
34  */
35 #include "rdft/scalar/r2cf.h"
36 
r2cf_9(R * R0,R * R1,R * Cr,R * Ci,stride rs,stride csr,stride csi,INT v,INT ivs,INT ovs)37 static void r2cf_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39      DK(KP907603734, +0.907603734547952313649323976213898122064543220);
40      DK(KP347296355, +0.347296355333860697703433253538629592000751354);
41      DK(KP852868531, +0.852868531952443209628250963940074071936020296);
42      DK(KP666666666, +0.666666666666666666666666666666666666666666667);
43      DK(KP898197570, +0.898197570222573798468955502359086394667167570);
44      DK(KP673648177, +0.673648177666930348851716626769314796000375677);
45      DK(KP879385241, +0.879385241571816768108218554649462939872416269);
46      DK(KP984807753, +0.984807753012208059366743024589523013670643252);
47      DK(KP939692620, +0.939692620785908384054109277324731469936208134);
48      DK(KP394930843, +0.394930843634698457567117349190734585290304520);
49      DK(KP866025403, +0.866025403784438646763723170752936183471402627);
50      DK(KP586256827, +0.586256827714544512072145703099641959914944179);
51      DK(KP726681596, +0.726681596905677465811651808188092531873167623);
52      DK(KP968908795, +0.968908795874236621082202410917456709164223497);
53      DK(KP203604859, +0.203604859554852403062088995281827210665664861);
54      DK(KP152703644, +0.152703644666139302296566746461370407999248646);
55      DK(KP500000000, +0.500000000000000000000000000000000000000000000);
56      DK(KP184792530, +0.184792530904095372701352047572203755870913560);
57      {
58 	  INT i;
59 	  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)) {
60 	       E T1, T4, To, Tk, Ta, Tu, Tf, Th, Tj, Tx, Tl, Tm, Ty, Tq, T2;
61 	       E T3, T5, Tg;
62 	       T1 = R0[0];
63 	       T2 = R1[WS(rs, 1)];
64 	       T3 = R0[WS(rs, 3)];
65 	       T4 = T2 + T3;
66 	       To = T3 - T2;
67 	       {
68 		    E T6, Tb, T9, Te, Ti;
69 		    T6 = R1[0];
70 		    Tb = R0[WS(rs, 1)];
71 		    {
72 			 E T7, T8, Tc, Td;
73 			 T7 = R0[WS(rs, 2)];
74 			 T8 = R1[WS(rs, 3)];
75 			 T9 = T7 + T8;
76 			 Tk = T7 - T8;
77 			 Tc = R1[WS(rs, 2)];
78 			 Td = R0[WS(rs, 4)];
79 			 Te = Tc + Td;
80 			 Ti = Td - Tc;
81 		    }
82 		    Ta = T6 + T9;
83 		    Tu = FMA(KP184792530, Tk, Ti);
84 		    Tf = Tb + Te;
85 		    Th = FNMS(KP500000000, Te, Tb);
86 		    Tj = FNMS(KP152703644, Ti, Th);
87 		    Tx = FMA(KP203604859, Th, Ti);
88 		    Tl = FMS(KP500000000, T9, T6);
89 		    Tm = FNMS(KP968908795, Tl, Tk);
90 		    Ty = FMA(KP726681596, Tk, Tl);
91 		    Tq = FMA(KP586256827, Tl, Ti);
92 	       }
93 	       Ci[WS(csi, 3)] = KP866025403 * (Tf - Ta);
94 	       T5 = T1 + T4;
95 	       Tg = Ta + Tf;
96 	       Cr[WS(csr, 3)] = FNMS(KP500000000, Tg, T5);
97 	       Cr[0] = T5 + Tg;
98 	       {
99 		    E Tv, Tt, Tn, TC, TB;
100 		    Tt = FMA(KP394930843, Th, To);
101 		    Tv = FNMS(KP939692620, Tu, Tt);
102 		    Ci[WS(csi, 2)] = KP984807753 * (FNMS(KP879385241, Tv, Tl));
103 		    Tn = FMA(KP673648177, Tm, Tj);
104 		    TB = FMA(KP898197570, Ty, Tx);
105 		    TC = FMA(KP666666666, Tn, TB);
106 		    Ci[WS(csi, 1)] = -(KP984807753 * (FNMS(KP879385241, To, Tn)));
107 		    Ci[WS(csi, 4)] = KP866025403 * (FMA(KP852868531, TC, To));
108 		    {
109 			 E Tp, Ts, Tz, TA, Tr, Tw;
110 			 Tp = FNMS(KP500000000, T4, T1);
111 			 Tr = FNMS(KP347296355, Tq, Tk);
112 			 Ts = FNMS(KP907603734, Tr, Th);
113 			 Tw = FNMS(KP673648177, Tm, Tj);
114 			 Tz = FNMS(KP898197570, Ty, Tx);
115 			 TA = FNMS(KP500000000, Tz, Tw);
116 			 Cr[WS(csr, 2)] = FNMS(KP939692620, Ts, Tp);
117 			 Cr[WS(csr, 1)] = FMA(KP852868531, Tz, Tp);
118 			 Cr[WS(csr, 4)] = FMA(KP852868531, TA, Tp);
119 		    }
120 	       }
121 	  }
122      }
123 }
124 
125 static const kr2c_desc desc = { 9, "r2cf_9", { 12, 4, 26, 0 }, &GENUS };
126 
X(codelet_r2cf_9)127 void X(codelet_r2cf_9) (planner *p) { X(kr2c_register) (p, r2cf_9, &desc);
128 }
129 
130 #else
131 
132 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 9 -name r2cf_9 -include rdft/scalar/r2cf.h */
133 
134 /*
135  * This function contains 38 FP additions, 26 FP multiplications,
136  * (or, 21 additions, 9 multiplications, 17 fused multiply/add),
137  * 36 stack variables, 14 constants, and 18 memory accesses
138  */
139 #include "rdft/scalar/r2cf.h"
140 
r2cf_9(R * R0,R * R1,R * Cr,R * Ci,stride rs,stride csr,stride csi,INT v,INT ivs,INT ovs)141 static void r2cf_9(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
142 {
143      DK(KP939692620, +0.939692620785908384054109277324731469936208134);
144      DK(KP296198132, +0.296198132726023843175338011893050938967728390);
145      DK(KP342020143, +0.342020143325668733044099614682259580763083368);
146      DK(KP813797681, +0.813797681349373692844693217248393223289101568);
147      DK(KP984807753, +0.984807753012208059366743024589523013670643252);
148      DK(KP150383733, +0.150383733180435296639271897612501926072238258);
149      DK(KP642787609, +0.642787609686539326322643409907263432907559884);
150      DK(KP663413948, +0.663413948168938396205421319635891297216863310);
151      DK(KP852868531, +0.852868531952443209628250963940074071936020296);
152      DK(KP173648177, +0.173648177666930348851716626769314796000375677);
153      DK(KP556670399, +0.556670399226419366452912952047023132968291906);
154      DK(KP766044443, +0.766044443118978035202392650555416673935832457);
155      DK(KP866025403, +0.866025403784438646763723170752936183471402627);
156      DK(KP500000000, +0.500000000000000000000000000000000000000000000);
157      {
158 	  INT i;
159 	  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)) {
160 	       E T1, T4, Tr, Ta, Tl, Ti, Tf, Tk, Tj, T2, T3, T5, Tg;
161 	       T1 = R0[0];
162 	       T2 = R1[WS(rs, 1)];
163 	       T3 = R0[WS(rs, 3)];
164 	       T4 = T2 + T3;
165 	       Tr = T3 - T2;
166 	       {
167 		    E T6, T7, T8, T9;
168 		    T6 = R1[0];
169 		    T7 = R0[WS(rs, 2)];
170 		    T8 = R1[WS(rs, 3)];
171 		    T9 = T7 + T8;
172 		    Ta = T6 + T9;
173 		    Tl = T8 - T7;
174 		    Ti = FNMS(KP500000000, T9, T6);
175 	       }
176 	       {
177 		    E Tb, Tc, Td, Te;
178 		    Tb = R0[WS(rs, 1)];
179 		    Tc = R1[WS(rs, 2)];
180 		    Td = R0[WS(rs, 4)];
181 		    Te = Tc + Td;
182 		    Tf = Tb + Te;
183 		    Tk = FNMS(KP500000000, Te, Tb);
184 		    Tj = Td - Tc;
185 	       }
186 	       Ci[WS(csi, 3)] = KP866025403 * (Tf - Ta);
187 	       T5 = T1 + T4;
188 	       Tg = Ta + Tf;
189 	       Cr[WS(csr, 3)] = FNMS(KP500000000, Tg, T5);
190 	       Cr[0] = T5 + Tg;
191 	       {
192 		    E Tt, Th, Tm, Tn, To, Tp, Tq, Ts;
193 		    Tt = KP866025403 * Tr;
194 		    Th = FNMS(KP500000000, T4, T1);
195 		    Tm = FMA(KP766044443, Ti, KP556670399 * Tl);
196 		    Tn = FMA(KP173648177, Tk, KP852868531 * Tj);
197 		    To = Tm + Tn;
198 		    Tp = FNMS(KP642787609, Ti, KP663413948 * Tl);
199 		    Tq = FNMS(KP984807753, Tk, KP150383733 * Tj);
200 		    Ts = Tp + Tq;
201 		    Cr[WS(csr, 1)] = Th + To;
202 		    Ci[WS(csi, 1)] = Tt + Ts;
203 		    Cr[WS(csr, 4)] = FMA(KP866025403, Tp - Tq, Th) - (KP500000000 * To);
204 		    Ci[WS(csi, 4)] = FNMS(KP500000000, Ts, KP866025403 * (Tr + (Tn - Tm)));
205 		    Ci[WS(csi, 2)] = FNMS(KP342020143, Tk, KP813797681 * Tj) + FNMA(KP150383733, Tl, KP984807753 * Ti) - Tt;
206 		    Cr[WS(csr, 2)] = FMA(KP173648177, Ti, Th) + FNMA(KP296198132, Tj, KP939692620 * Tk) - (KP852868531 * Tl);
207 	       }
208 	  }
209      }
210 }
211 
212 static const kr2c_desc desc = { 9, "r2cf_9", { 21, 9, 17, 0 }, &GENUS };
213 
X(codelet_r2cf_9)214 void X(codelet_r2cf_9) (planner *p) { X(kr2c_register) (p, r2cf_9, &desc);
215 }
216 
217 #endif
218