1 #include "simint/boys/boys.h"
2 #include "simint/ostei/gen/ostei_generated.h"
3 #include "simint/vectorization/vectorization.h"
4 #include <math.h>
5 #include <string.h>
6
7
ostei_i_p_k_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_p_k_p)8 int ostei_i_p_k_p(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__i_p_k_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_p_k_p);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__i_s_k_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__i_s_l_s = work + (SIMINT_NSHELL_SIMD * 1008);
31 double * const INT__k_s_k_s = work + (SIMINT_NSHELL_SIMD * 2268);
32 double * const INT__k_s_l_s = work + (SIMINT_NSHELL_SIMD * 3564);
33 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*5184);
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 16;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 61;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 145;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 275;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 455;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 686;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 966;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 1290;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 1650;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 1713;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1839;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 2049;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 2364;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2805;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 3393;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 4149;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 5094;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 5310;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 5670;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 6210;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 6966;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 7974;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 9270;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 10890;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 11390;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 12140;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 13190;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 14590;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 16390;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 18640;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 19540;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 20800;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 22480;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_l_s = primwork + 24640;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 27340;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 28663;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 30427;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_l_s = primwork + 32695;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_i_s = primwork + 35530;
74 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_k_s = primwork + 37098;
75 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_l_s = primwork + 39114;
76 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_k_s = primwork + 41634;
77 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_l_s = primwork + 42930;
78 double * const hrrwork = (double *)(primwork + 44550);
79 double * const HRR_INT__i_p_k_s = hrrwork + 0;
80 double * const HRR_INT__i_p_l_s = hrrwork + 3024;
81
82
83 // Create constants
84 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
85 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
86 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
87 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
88 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
89 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
90 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
91 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
92 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
93
94
95 ////////////////////////////////////////
96 // Loop over shells and primitives
97 ////////////////////////////////////////
98
99 real_abcd = 0;
100 istart = 0;
101 for(ab = 0; ab < P.nshell12_clip; ++ab)
102 {
103 const int iend = istart + P.nprim12[ab];
104
105 cd = 0;
106 jstart = 0;
107
108 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
109 {
110 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
111 int jend = jstart;
112 for(i = 0; i < nshellbatch; i++)
113 jend += Q.nprim12[cd+i];
114
115 // Clear the beginning of the workspace (where we are accumulating integrals)
116 memset(work, 0, SIMINT_NSHELL_SIMD * 5184 * sizeof(double));
117 abcd = 0;
118
119
120 for(i = istart; i < iend; ++i)
121 {
122 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
123
124 if(check_screen)
125 {
126 // Skip this whole thing if always insignificant
127 if((P.screen[i] * Q.screen_max) < screen_tol)
128 continue;
129 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
130 }
131
132 icd = 0;
133 iprimcd = 0;
134 nprim_icd = Q.nprim12[cd];
135 double * restrict PRIM_PTR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
136 double * restrict PRIM_PTR_INT__i_s_l_s = INT__i_s_l_s + abcd * 1260;
137 double * restrict PRIM_PTR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
138 double * restrict PRIM_PTR_INT__k_s_l_s = INT__k_s_l_s + abcd * 1620;
139
140
141
142 // Load these one per loop over i
143 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
144 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
145 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
146
147 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
148
149 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
150 {
151 // calculate the shell offsets
152 // these are the offset from the shell pointed to by cd
153 // for each element
154 int shelloffsets[SIMINT_SIMD_LEN] = {0};
155 int lastoffset = 0;
156 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
157
158 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
159 {
160 // Handle if the first element of the vector is a new shell
161 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
162 {
163 nprim_icd += Q.nprim12[cd + (++icd)];
164 PRIM_PTR_INT__i_s_k_s += 1008;
165 PRIM_PTR_INT__i_s_l_s += 1260;
166 PRIM_PTR_INT__k_s_k_s += 1296;
167 PRIM_PTR_INT__k_s_l_s += 1620;
168 }
169 iprimcd++;
170 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
171 {
172 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
173 {
174 shelloffsets[n] = shelloffsets[n-1] + 1;
175 lastoffset++;
176 nprim_icd += Q.nprim12[cd + (++icd)];
177 }
178 else
179 shelloffsets[n] = shelloffsets[n-1];
180 iprimcd++;
181 }
182 }
183 else
184 iprimcd += SIMINT_SIMD_LEN;
185
186 // Do we have to compute this vector (or has it been screened out)?
187 // (not_screened != 0 means we have to do this vector)
188 if(check_screen)
189 {
190 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
191 if(vmax < screen_tol)
192 {
193 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
194 PRIM_PTR_INT__i_s_l_s += lastoffset*1260;
195 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
196 PRIM_PTR_INT__k_s_l_s += lastoffset*1620;
197 continue;
198 }
199 }
200
201 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
202 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
203 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
204 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
205
206
207 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
208 SIMINT_DBLTYPE PQ[3];
209 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
210 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
211 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
212 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
213 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
214 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
215
216 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
217 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
218 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
219 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
220 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
221 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
222 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
223
224 // NOTE: Minus sign!
225 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
226 SIMINT_DBLTYPE aop_PQ[3];
227 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
228 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
229 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
230
231 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
232 SIMINT_DBLTYPE aoq_PQ[3];
233 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
234 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
235 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
236 // Put a minus sign here so we don't have to in RR routines
237 a_over_q = SIMINT_NEG(a_over_q);
238
239
240 //////////////////////////////////////////////
241 // Fjt function section
242 // Maximum v value: 15
243 //////////////////////////////////////////////
244 // The parameter to the Fjt function
245 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
246
247
248 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
249
250
251 boys_F_split(PRIM_INT__s_s_s_s, F_x, 15);
252 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
253 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
254 for(n = 0; n <= 15; n++)
255 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
256
257 //////////////////////////////////////////////
258 // Primitive integrals: Vertical recurrance
259 //////////////////////////////////////////////
260
261 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
262 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
263 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
264 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
265 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
266 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
267 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
268 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
269 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
270 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
271 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
272 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
273 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
274 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
275 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
276 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
277 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
278 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
279 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
280 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
281 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
282
283
284
285 // Forming PRIM_INT__s_s_p_s[15 * 3];
286 for(n = 0; n < 15; ++n) // loop over orders of auxiliary function
287 {
288
289 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
290 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]);
291
292 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
293 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 1]);
294
295 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
296 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 2]);
297
298 }
299
300
301
302 // Forming PRIM_INT__s_s_d_s[14 * 6];
303 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
304 {
305
306 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
307 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 0]);
308 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 0]);
309
310 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
311 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 1]);
312
313 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
314 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 2]);
315
316 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
317 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 3]);
318 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 3]);
319
320 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
321 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 4]);
322
323 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
324 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_d_s[n * 6 + 5]);
325 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 5]);
326
327 }
328
329
330
331 // Forming PRIM_INT__s_s_f_s[13 * 10];
332 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
333 {
334
335 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
336 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 0]);
337 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]), PRIM_INT__s_s_f_s[n * 10 + 0]);
338
339 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
340 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 1]);
341
342 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
343 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 2]);
344
345 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
346 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 3]);
347
348 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
349 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__s_s_f_s[n * 10 + 4]);
350
351 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
352 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 5]);
353
354 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
355 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 6]);
356 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_p_s[n * 3 + 1]), PRIM_INT__s_s_f_s[n * 10 + 6]);
357
358 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
359 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 7]);
360
361 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
362 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 8]);
363
364 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
365 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 9]);
366 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_p_s[n * 3 + 2]), PRIM_INT__s_s_f_s[n * 10 + 9]);
367
368 }
369
370
371 VRR_K_s_s_g_s(
372 PRIM_INT__s_s_g_s,
373 PRIM_INT__s_s_f_s,
374 PRIM_INT__s_s_d_s,
375 Q_PA,
376 a_over_q,
377 aoq_PQ,
378 one_over_2q,
379 12);
380
381
382 VRR_K_s_s_h_s(
383 PRIM_INT__s_s_h_s,
384 PRIM_INT__s_s_g_s,
385 PRIM_INT__s_s_f_s,
386 Q_PA,
387 a_over_q,
388 aoq_PQ,
389 one_over_2q,
390 11);
391
392
393 ostei_general_vrr1_K(6, 10,
394 one_over_2q, a_over_q, aoq_PQ, Q_PA,
395 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
396
397
398 ostei_general_vrr1_K(7, 9,
399 one_over_2q, a_over_q, aoq_PQ, Q_PA,
400 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
401
402
403 ostei_general_vrr_I(1, 0, 7, 0, 7,
404 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
405 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
406
407
408 ostei_general_vrr_I(1, 0, 6, 0, 7,
409 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
410 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
411
412
413 ostei_general_vrr_I(2, 0, 7, 0, 6,
414 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
415 PRIM_INT__p_s_k_s, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_k_s);
416
417
418 ostei_general_vrr_I(1, 0, 5, 0, 7,
419 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
420 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
421
422
423 ostei_general_vrr_I(2, 0, 6, 0, 6,
424 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
425 PRIM_INT__p_s_i_s, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_i_s);
426
427
428 ostei_general_vrr_I(3, 0, 7, 0, 5,
429 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
430 PRIM_INT__d_s_k_s, PRIM_INT__p_s_k_s, NULL, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_k_s);
431
432
433 VRR_I_p_s_g_s(
434 PRIM_INT__p_s_g_s,
435 PRIM_INT__s_s_g_s,
436 PRIM_INT__s_s_f_s,
437 P_PA,
438 aop_PQ,
439 one_over_2pq,
440 7);
441
442
443 ostei_general_vrr_I(2, 0, 5, 0, 6,
444 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
445 PRIM_INT__p_s_h_s, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_h_s);
446
447
448 ostei_general_vrr_I(3, 0, 6, 0, 5,
449 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
450 PRIM_INT__d_s_i_s, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_i_s);
451
452
453 ostei_general_vrr_I(4, 0, 7, 0, 4,
454 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
455 PRIM_INT__f_s_k_s, PRIM_INT__d_s_k_s, NULL, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_k_s);
456
457
458 VRR_I_p_s_f_s(
459 PRIM_INT__p_s_f_s,
460 PRIM_INT__s_s_f_s,
461 PRIM_INT__s_s_d_s,
462 P_PA,
463 aop_PQ,
464 one_over_2pq,
465 7);
466
467
468 ostei_general_vrr_I(2, 0, 4, 0, 6,
469 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
470 PRIM_INT__p_s_g_s, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_g_s);
471
472
473 ostei_general_vrr_I(3, 0, 5, 0, 5,
474 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
475 PRIM_INT__d_s_h_s, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_h_s);
476
477
478 ostei_general_vrr_I(4, 0, 6, 0, 4,
479 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
480 PRIM_INT__f_s_i_s, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_i_s);
481
482
483 ostei_general_vrr_I(5, 0, 7, 0, 3,
484 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
485 PRIM_INT__g_s_k_s, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_k_s);
486
487
488
489 // Forming PRIM_INT__p_s_d_s[7 * 18];
490 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
491 {
492
493 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
494 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
495 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
496
497 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
498 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 1]);
499 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 1]);
500
501 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
502 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 2]);
503 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 2]);
504
505 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
506 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 3]);
507
508 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
509 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 4]);
510
511 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
512 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 5]);
513
514 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
515 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 6]);
516
517 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
518 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 7]);
519 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 7]);
520
521 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
522 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 8]);
523
524 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
525 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 9]);
526 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 9]);
527
528 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
529 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 10]);
530 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 10]);
531
532 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
533 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 11]);
534
535 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
536 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 12]);
537
538 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
539 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 13]);
540
541 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
542 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 14]);
543 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 14]);
544
545 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
546 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 15]);
547
548 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
549 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 16]);
550 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 16]);
551
552 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
553 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 17]);
554 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 17]);
555
556 }
557
558
559 VRR_I_d_s_f_s(
560 PRIM_INT__d_s_f_s,
561 PRIM_INT__p_s_f_s,
562 PRIM_INT__s_s_f_s,
563 PRIM_INT__p_s_d_s,
564 P_PA,
565 a_over_p,
566 aop_PQ,
567 one_over_2p,
568 one_over_2pq,
569 6);
570
571
572 ostei_general_vrr_I(3, 0, 4, 0, 5,
573 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
574 PRIM_INT__d_s_g_s, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_f_s, NULL, PRIM_INT__f_s_g_s);
575
576
577 ostei_general_vrr_I(4, 0, 5, 0, 4,
578 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
579 PRIM_INT__f_s_h_s, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_g_s, NULL, PRIM_INT__g_s_h_s);
580
581
582 ostei_general_vrr_I(5, 0, 6, 0, 3,
583 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
584 PRIM_INT__g_s_i_s, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_h_s, NULL, PRIM_INT__h_s_i_s);
585
586
587 ostei_general_vrr_I(6, 0, 7, 0, 2,
588 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
589 PRIM_INT__h_s_k_s, PRIM_INT__g_s_k_s, NULL, PRIM_INT__h_s_i_s, NULL, PRIM_INT__i_s_k_s);
590
591
592 ostei_general_vrr1_K(8, 8,
593 one_over_2q, a_over_q, aoq_PQ, Q_PA,
594 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
595
596
597 ostei_general_vrr_I(1, 0, 8, 0, 7,
598 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
599 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
600
601
602 ostei_general_vrr_I(2, 0, 8, 0, 6,
603 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
604 PRIM_INT__p_s_l_s, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_k_s, NULL, PRIM_INT__d_s_l_s);
605
606
607 ostei_general_vrr_I(3, 0, 8, 0, 5,
608 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
609 PRIM_INT__d_s_l_s, PRIM_INT__p_s_l_s, NULL, PRIM_INT__d_s_k_s, NULL, PRIM_INT__f_s_l_s);
610
611
612 ostei_general_vrr_I(4, 0, 8, 0, 4,
613 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
614 PRIM_INT__f_s_l_s, PRIM_INT__d_s_l_s, NULL, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_l_s);
615
616
617 ostei_general_vrr_I(5, 0, 8, 0, 3,
618 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
619 PRIM_INT__g_s_l_s, PRIM_INT__f_s_l_s, NULL, PRIM_INT__g_s_k_s, NULL, PRIM_INT__h_s_l_s);
620
621
622 ostei_general_vrr_I(6, 0, 8, 0, 2,
623 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
624 PRIM_INT__h_s_l_s, PRIM_INT__g_s_l_s, NULL, PRIM_INT__h_s_k_s, NULL, PRIM_INT__i_s_l_s);
625
626
627
628 // Forming PRIM_INT__p_s_p_s[7 * 9];
629 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
630 {
631
632 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
633 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
634 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
635
636 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 1]);
637 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 1]);
638
639 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 2]);
640 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 2]);
641
642 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
643 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 3]);
644
645 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
646 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 4]);
647 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 4]);
648
649 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 2]);
650 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 5]);
651
652 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
653 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 6]);
654
655 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
656 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 7]);
657
658 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
659 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 8]);
660 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 8]);
661
662 }
663
664
665 VRR_I_d_s_d_s(
666 PRIM_INT__d_s_d_s,
667 PRIM_INT__p_s_d_s,
668 PRIM_INT__s_s_d_s,
669 PRIM_INT__p_s_p_s,
670 P_PA,
671 a_over_p,
672 aop_PQ,
673 one_over_2p,
674 one_over_2pq,
675 6);
676
677
678 ostei_general_vrr_I(3, 0, 3, 0, 5,
679 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
680 PRIM_INT__d_s_f_s, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_d_s, NULL, PRIM_INT__f_s_f_s);
681
682
683 ostei_general_vrr_I(4, 0, 4, 0, 4,
684 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
685 PRIM_INT__f_s_g_s, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_f_s, NULL, PRIM_INT__g_s_g_s);
686
687
688 ostei_general_vrr_I(5, 0, 5, 0, 3,
689 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
690 PRIM_INT__g_s_h_s, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_g_s, NULL, PRIM_INT__h_s_h_s);
691
692
693 ostei_general_vrr_I(6, 0, 6, 0, 2,
694 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
695 PRIM_INT__h_s_i_s, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_h_s, NULL, PRIM_INT__i_s_i_s);
696
697
698 ostei_general_vrr_I(7, 0, 7, 0, 1,
699 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
700 PRIM_INT__i_s_k_s, PRIM_INT__h_s_k_s, NULL, PRIM_INT__i_s_i_s, NULL, PRIM_INT__k_s_k_s);
701
702
703 ostei_general_vrr_I(7, 0, 8, 0, 1,
704 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
705 PRIM_INT__i_s_l_s, PRIM_INT__h_s_l_s, NULL, PRIM_INT__i_s_k_s, NULL, PRIM_INT__k_s_l_s);
706
707
708
709
710 ////////////////////////////////////
711 // Accumulate contracted integrals
712 ////////////////////////////////////
713 if(lastoffset == 0)
714 {
715 contract_all(1008, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
716 contract_all(1260, PRIM_INT__i_s_l_s, PRIM_PTR_INT__i_s_l_s);
717 contract_all(1296, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
718 contract_all(1620, PRIM_INT__k_s_l_s, PRIM_PTR_INT__k_s_l_s);
719 }
720 else
721 {
722 contract(1008, shelloffsets, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
723 contract(1260, shelloffsets, PRIM_INT__i_s_l_s, PRIM_PTR_INT__i_s_l_s);
724 contract(1296, shelloffsets, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
725 contract(1620, shelloffsets, PRIM_INT__k_s_l_s, PRIM_PTR_INT__k_s_l_s);
726 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
727 PRIM_PTR_INT__i_s_l_s += lastoffset*1260;
728 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
729 PRIM_PTR_INT__k_s_l_s += lastoffset*1620;
730 }
731
732 } // close loop over j
733 } // close loop over i
734
735 //Advance to the next batch
736 jstart = SIMINT_SIMD_ROUND(jend);
737
738 //////////////////////////////////////////////
739 // Contracted integrals: Horizontal recurrance
740 //////////////////////////////////////////////
741
742
743 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
744
745
746 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
747 {
748 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
749
750 // set up HRR pointers
751 double const * restrict HRR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
752 double const * restrict HRR_INT__i_s_l_s = INT__i_s_l_s + abcd * 1260;
753 double const * restrict HRR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
754 double const * restrict HRR_INT__k_s_l_s = INT__k_s_l_s + abcd * 1620;
755 double * restrict HRR_INT__i_p_k_p = INT__i_p_k_p + real_abcd * 9072;
756
757 // form INT__i_p_k_s
758 ostei_general_hrr_J(6, 1, 7, 0, hAB, HRR_INT__k_s_k_s, HRR_INT__i_s_k_s, HRR_INT__i_p_k_s);
759
760 // form INT__i_p_l_s
761 ostei_general_hrr_J(6, 1, 8, 0, hAB, HRR_INT__k_s_l_s, HRR_INT__i_s_l_s, HRR_INT__i_p_l_s);
762
763 // form INT__i_p_k_p
764 ostei_general_hrr_L(6, 1, 7, 1, hCD, HRR_INT__i_p_l_s, HRR_INT__i_p_k_s, HRR_INT__i_p_k_p);
765
766
767 } // close HRR loop
768
769
770 } // close loop cdbatch
771
772 istart = iend;
773 } // close loop over ab
774
775 return P.nshell12_clip * Q.nshell12_clip;
776 }
777
ostei_p_i_k_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_i_k_p)778 int ostei_p_i_k_p(struct simint_multi_shellpair const P,
779 struct simint_multi_shellpair const Q,
780 double screen_tol,
781 double * const restrict work,
782 double * const restrict INT__p_i_k_p)
783 {
784 double P_AB[3*P.nshell12];
785 struct simint_multi_shellpair P_tmp = P;
786 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
787 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
788 P_tmp.AB_x = P_AB;
789 P_tmp.AB_y = P_AB + P.nshell12;
790 P_tmp.AB_z = P_AB + 2*P.nshell12;
791
792 for(int i = 0; i < P.nshell12; i++)
793 {
794 P_tmp.AB_x[i] = -P.AB_x[i];
795 P_tmp.AB_y[i] = -P.AB_y[i];
796 P_tmp.AB_z[i] = -P.AB_z[i];
797 }
798
799 int ret = ostei_i_p_k_p(P_tmp, Q, screen_tol, work, INT__p_i_k_p);
800 double buffer[9072] SIMINT_ALIGN_ARRAY_DBL;
801
802 for(int q = 0; q < ret; q++)
803 {
804 int idx = 0;
805 for(int a = 0; a < 3; ++a)
806 for(int b = 0; b < 28; ++b)
807 for(int c = 0; c < 36; ++c)
808 for(int d = 0; d < 3; ++d)
809 buffer[idx++] = INT__p_i_k_p[q*9072+b*324+a*108+c*3+d];
810
811 memcpy(INT__p_i_k_p+q*9072, buffer, 9072*sizeof(double));
812 }
813
814 return ret;
815 }
816
ostei_i_p_p_k(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_p_p_k)817 int ostei_i_p_p_k(struct simint_multi_shellpair const P,
818 struct simint_multi_shellpair const Q,
819 double screen_tol,
820 double * const restrict work,
821 double * const restrict INT__i_p_p_k)
822 {
823 double Q_AB[3*Q.nshell12];
824 struct simint_multi_shellpair Q_tmp = Q;
825 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
826 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
827 Q_tmp.AB_x = Q_AB;
828 Q_tmp.AB_y = Q_AB + Q.nshell12;
829 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
830
831 for(int i = 0; i < Q.nshell12; i++)
832 {
833 Q_tmp.AB_x[i] = -Q.AB_x[i];
834 Q_tmp.AB_y[i] = -Q.AB_y[i];
835 Q_tmp.AB_z[i] = -Q.AB_z[i];
836 }
837
838 int ret = ostei_i_p_k_p(P, Q_tmp, screen_tol, work, INT__i_p_p_k);
839 double buffer[9072] SIMINT_ALIGN_ARRAY_DBL;
840
841 for(int q = 0; q < ret; q++)
842 {
843 int idx = 0;
844 for(int a = 0; a < 28; ++a)
845 for(int b = 0; b < 3; ++b)
846 for(int c = 0; c < 3; ++c)
847 for(int d = 0; d < 36; ++d)
848 buffer[idx++] = INT__i_p_p_k[q*9072+a*324+b*108+d*3+c];
849
850 memcpy(INT__i_p_p_k+q*9072, buffer, 9072*sizeof(double));
851 }
852
853 return ret;
854 }
855
ostei_p_i_p_k(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_i_p_k)856 int ostei_p_i_p_k(struct simint_multi_shellpair const P,
857 struct simint_multi_shellpair const Q,
858 double screen_tol,
859 double * const restrict work,
860 double * const restrict INT__p_i_p_k)
861 {
862 double P_AB[3*P.nshell12];
863 struct simint_multi_shellpair P_tmp = P;
864 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
865 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
866 P_tmp.AB_x = P_AB;
867 P_tmp.AB_y = P_AB + P.nshell12;
868 P_tmp.AB_z = P_AB + 2*P.nshell12;
869
870 for(int i = 0; i < P.nshell12; i++)
871 {
872 P_tmp.AB_x[i] = -P.AB_x[i];
873 P_tmp.AB_y[i] = -P.AB_y[i];
874 P_tmp.AB_z[i] = -P.AB_z[i];
875 }
876
877 double Q_AB[3*Q.nshell12];
878 struct simint_multi_shellpair Q_tmp = Q;
879 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
880 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
881 Q_tmp.AB_x = Q_AB;
882 Q_tmp.AB_y = Q_AB + Q.nshell12;
883 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
884
885 for(int i = 0; i < Q.nshell12; i++)
886 {
887 Q_tmp.AB_x[i] = -Q.AB_x[i];
888 Q_tmp.AB_y[i] = -Q.AB_y[i];
889 Q_tmp.AB_z[i] = -Q.AB_z[i];
890 }
891
892 int ret = ostei_i_p_k_p(P_tmp, Q_tmp, screen_tol, work, INT__p_i_p_k);
893 double buffer[9072] SIMINT_ALIGN_ARRAY_DBL;
894
895 for(int q = 0; q < ret; q++)
896 {
897 int idx = 0;
898 for(int a = 0; a < 3; ++a)
899 for(int b = 0; b < 28; ++b)
900 for(int c = 0; c < 3; ++c)
901 for(int d = 0; d < 36; ++d)
902 buffer[idx++] = INT__p_i_p_k[q*9072+b*324+a*108+d*3+c];
903
904 memcpy(INT__p_i_p_k+q*9072, buffer, 9072*sizeof(double));
905 }
906
907 return ret;
908 }
909
910