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_g_f_f_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_f_f_p)8 int ostei_g_f_f_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__g_f_f_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__g_f_f_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__g_s_f_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__g_s_g_s = work + (SIMINT_NSHELL_SIMD * 150);
31 double * const INT__h_s_f_s = work + (SIMINT_NSHELL_SIMD * 375);
32 double * const INT__h_s_g_s = work + (SIMINT_NSHELL_SIMD * 585);
33 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 900);
34 double * const INT__i_s_g_s = work + (SIMINT_NSHELL_SIMD * 1180);
35 double * const INT__k_s_f_s = work + (SIMINT_NSHELL_SIMD * 1600);
36 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 1960);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2500);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 12;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 45;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 81;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 141;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 213;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 321;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 411;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 531;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 711;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 911;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1031;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1211;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1481;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 1781;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 2006;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2153;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 2405;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 2783;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 3203;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 3518;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 3686;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 4022;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 4526;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 5086;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 5506;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 5686;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 6118;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 6766;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 7486;
68 double * const hrrwork = (double *)(primwork + 8026);
69 double * const HRR_INT__g_p_f_s = hrrwork + 0;
70 double * const HRR_INT__g_p_g_s = hrrwork + 450;
71 double * const HRR_INT__g_d_f_s = hrrwork + 1125;
72 double * const HRR_INT__g_d_g_s = hrrwork + 2025;
73 double * const HRR_INT__g_f_f_s = hrrwork + 3375;
74 double * const HRR_INT__g_f_g_s = hrrwork + 4875;
75 double * const HRR_INT__h_p_f_s = hrrwork + 7125;
76 double * const HRR_INT__h_p_g_s = hrrwork + 7755;
77 double * const HRR_INT__h_d_f_s = hrrwork + 8700;
78 double * const HRR_INT__h_d_g_s = hrrwork + 9960;
79 double * const HRR_INT__i_p_f_s = hrrwork + 11850;
80 double * const HRR_INT__i_p_g_s = hrrwork + 12690;
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 one_half = SIMINT_DBLSET1(0.5);
92
93
94 ////////////////////////////////////////
95 // Loop over shells and primitives
96 ////////////////////////////////////////
97
98 real_abcd = 0;
99 istart = 0;
100 for(ab = 0; ab < P.nshell12_clip; ++ab)
101 {
102 const int iend = istart + P.nprim12[ab];
103
104 cd = 0;
105 jstart = 0;
106
107 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
108 {
109 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
110 int jend = jstart;
111 for(i = 0; i < nshellbatch; i++)
112 jend += Q.nprim12[cd+i];
113
114 // Clear the beginning of the workspace (where we are accumulating integrals)
115 memset(work, 0, SIMINT_NSHELL_SIMD * 2500 * sizeof(double));
116 abcd = 0;
117
118
119 for(i = istart; i < iend; ++i)
120 {
121 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
122
123 if(check_screen)
124 {
125 // Skip this whole thing if always insignificant
126 if((P.screen[i] * Q.screen_max) < screen_tol)
127 continue;
128 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
129 }
130
131 icd = 0;
132 iprimcd = 0;
133 nprim_icd = Q.nprim12[cd];
134 double * restrict PRIM_PTR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
135 double * restrict PRIM_PTR_INT__g_s_g_s = INT__g_s_g_s + abcd * 225;
136 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
137 double * restrict PRIM_PTR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
138 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
139 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
140 double * restrict PRIM_PTR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
141 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
142
143
144
145 // Load these one per loop over i
146 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
147 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
148 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
149
150 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
151
152 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
153 {
154 // calculate the shell offsets
155 // these are the offset from the shell pointed to by cd
156 // for each element
157 int shelloffsets[SIMINT_SIMD_LEN] = {0};
158 int lastoffset = 0;
159 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
160
161 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
162 {
163 // Handle if the first element of the vector is a new shell
164 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
165 {
166 nprim_icd += Q.nprim12[cd + (++icd)];
167 PRIM_PTR_INT__g_s_f_s += 150;
168 PRIM_PTR_INT__g_s_g_s += 225;
169 PRIM_PTR_INT__h_s_f_s += 210;
170 PRIM_PTR_INT__h_s_g_s += 315;
171 PRIM_PTR_INT__i_s_f_s += 280;
172 PRIM_PTR_INT__i_s_g_s += 420;
173 PRIM_PTR_INT__k_s_f_s += 360;
174 PRIM_PTR_INT__k_s_g_s += 540;
175 }
176 iprimcd++;
177 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
178 {
179 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
180 {
181 shelloffsets[n] = shelloffsets[n-1] + 1;
182 lastoffset++;
183 nprim_icd += Q.nprim12[cd + (++icd)];
184 }
185 else
186 shelloffsets[n] = shelloffsets[n-1];
187 iprimcd++;
188 }
189 }
190 else
191 iprimcd += SIMINT_SIMD_LEN;
192
193 // Do we have to compute this vector (or has it been screened out)?
194 // (not_screened != 0 means we have to do this vector)
195 if(check_screen)
196 {
197 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
198 if(vmax < screen_tol)
199 {
200 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
201 PRIM_PTR_INT__g_s_g_s += lastoffset*225;
202 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
203 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
204 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
205 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
206 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
207 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
208 continue;
209 }
210 }
211
212 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
213 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
214 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
215 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
216
217
218 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
219 SIMINT_DBLTYPE PQ[3];
220 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
221 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
222 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
223 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
224 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
225 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
226
227 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
228 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
229 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
230 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
231 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
232 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
233 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
234
235 // NOTE: Minus sign!
236 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
237 SIMINT_DBLTYPE aop_PQ[3];
238 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
239 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
240 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
241
242 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
243 SIMINT_DBLTYPE aoq_PQ[3];
244 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
245 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
246 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
247 // Put a minus sign here so we don't have to in RR routines
248 a_over_q = SIMINT_NEG(a_over_q);
249
250
251 //////////////////////////////////////////////
252 // Fjt function section
253 // Maximum v value: 11
254 //////////////////////////////////////////////
255 // The parameter to the Fjt function
256 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
257
258
259 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
260
261
262 boys_F_split(PRIM_INT__s_s_s_s, F_x, 11);
263 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
264 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
265 for(n = 0; n <= 11; n++)
266 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
267
268 //////////////////////////////////////////////
269 // Primitive integrals: Vertical recurrance
270 //////////////////////////////////////////////
271
272 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
273 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
274 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
275 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
276 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
277 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
278 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
279 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
280 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
281 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
282 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
283 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
284 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
285 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
288
289
290
291 // Forming PRIM_INT__p_s_s_s[11 * 3];
292 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
293 {
294
295 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
296 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]);
297
298 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
299 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 1]);
300
301 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
302 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 2]);
303
304 }
305
306
307
308 // Forming PRIM_INT__d_s_s_s[10 * 6];
309 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
310 {
311
312 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
313 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 0]);
314 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 0]);
315
316 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
317 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 1]);
318
319 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
320 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 2]);
321
322 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
323 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 3]);
324 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 3]);
325
326 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
327 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 4]);
328
329 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
330 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_s_s[n * 6 + 5]);
331 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 5]);
332
333 }
334
335
336
337 // Forming PRIM_INT__f_s_s_s[9 * 10];
338 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
339 {
340
341 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
342 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 0]);
343 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__f_s_s_s[n * 10 + 0]);
344
345 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
346 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 1]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
349 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 2]);
350
351 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
352 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 3]);
353
354 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
355 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__f_s_s_s[n * 10 + 4]);
356
357 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
358 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 5]);
359
360 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
361 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 6]);
362 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__f_s_s_s[n * 10 + 6]);
363
364 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
365 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 7]);
366
367 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
368 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 8]);
369
370 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
371 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 9]);
372 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__f_s_s_s[n * 10 + 9]);
373
374 }
375
376
377 VRR_I_g_s_s_s(
378 PRIM_INT__g_s_s_s,
379 PRIM_INT__f_s_s_s,
380 PRIM_INT__d_s_s_s,
381 P_PA,
382 a_over_p,
383 aop_PQ,
384 one_over_2p,
385 8);
386
387
388 VRR_K_g_s_p_s(
389 PRIM_INT__g_s_p_s,
390 PRIM_INT__g_s_s_s,
391 PRIM_INT__f_s_s_s,
392 Q_PA,
393 aoq_PQ,
394 one_over_2pq,
395 4);
396
397
398 VRR_K_f_s_p_s(
399 PRIM_INT__f_s_p_s,
400 PRIM_INT__f_s_s_s,
401 PRIM_INT__d_s_s_s,
402 Q_PA,
403 aoq_PQ,
404 one_over_2pq,
405 4);
406
407
408 ostei_general_vrr_K(4, 0, 2, 0, 3,
409 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
410 PRIM_INT__g_s_p_s, PRIM_INT__g_s_s_s, NULL, PRIM_INT__f_s_p_s, NULL, PRIM_INT__g_s_d_s);
411
412
413
414 // Forming PRIM_INT__d_s_p_s[4 * 18];
415 for(n = 0; n < 4; ++n) // loop over orders of auxiliary function
416 {
417
418 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
419 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
420 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
421
422 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
423 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 1]);
424
425 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
426 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 2]);
427
428 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
429 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
430 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
431
432 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
433 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 4]);
434 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 4]);
435
436 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
437 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 5]);
438
439 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
440 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
441 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
442
443 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
444 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 7]);
445
446 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
447 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 8]);
448 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 8]);
449
450 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
451 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 9]);
452
453 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
454 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 10]);
455 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 10]);
456
457 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
458 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 11]);
459
460 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
461 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 12]);
462
463 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
464 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 13]);
465 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 13]);
466
467 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
468 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 14]);
469 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 14]);
470
471 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
472 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 15]);
473
474 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
475 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 16]);
476
477 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
478 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 17]);
479 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 17]);
480
481 }
482
483
484 VRR_K_f_s_d_s(
485 PRIM_INT__f_s_d_s,
486 PRIM_INT__f_s_p_s,
487 PRIM_INT__f_s_s_s,
488 PRIM_INT__d_s_p_s,
489 Q_PA,
490 a_over_q,
491 aoq_PQ,
492 one_over_2pq,
493 one_over_2q,
494 3);
495
496
497 ostei_general_vrr_K(4, 0, 3, 0, 2,
498 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
499 PRIM_INT__g_s_d_s, PRIM_INT__g_s_p_s, NULL, PRIM_INT__f_s_d_s, NULL, PRIM_INT__g_s_f_s);
500
501
502
503 // Forming PRIM_INT__p_s_p_s[4 * 9];
504 for(n = 0; n < 4; ++n) // loop over orders of auxiliary function
505 {
506
507 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
508 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
509 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]);
510
511 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
512 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 1]);
513
514 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
515 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 2]);
516
517 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
518 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 3]);
519
520 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
521 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 4]);
522 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]);
523
524 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
525 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 5]);
526
527 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
528 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 6]);
529
530 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
531 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 7]);
532
533 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
534 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 8]);
535 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]);
536
537 }
538
539
540 VRR_K_d_s_d_s(
541 PRIM_INT__d_s_d_s,
542 PRIM_INT__d_s_p_s,
543 PRIM_INT__d_s_s_s,
544 PRIM_INT__p_s_p_s,
545 Q_PA,
546 a_over_q,
547 aoq_PQ,
548 one_over_2pq,
549 one_over_2q,
550 3);
551
552
553 ostei_general_vrr_K(3, 0, 3, 0, 2,
554 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
555 PRIM_INT__f_s_d_s, PRIM_INT__f_s_p_s, NULL, PRIM_INT__d_s_d_s, NULL, PRIM_INT__f_s_f_s);
556
557
558 ostei_general_vrr_K(4, 0, 4, 0, 1,
559 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
560 PRIM_INT__g_s_f_s, PRIM_INT__g_s_d_s, NULL, PRIM_INT__f_s_f_s, NULL, PRIM_INT__g_s_g_s);
561
562
563 VRR_I_h_s_s_s(
564 PRIM_INT__h_s_s_s,
565 PRIM_INT__g_s_s_s,
566 PRIM_INT__f_s_s_s,
567 P_PA,
568 a_over_p,
569 aop_PQ,
570 one_over_2p,
571 7);
572
573
574 ostei_general_vrr_K(5, 0, 1, 0, 4,
575 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
576 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
577
578
579 ostei_general_vrr_K(5, 0, 2, 0, 3,
580 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
581 PRIM_INT__h_s_p_s, PRIM_INT__h_s_s_s, NULL, PRIM_INT__g_s_p_s, NULL, PRIM_INT__h_s_d_s);
582
583
584 ostei_general_vrr_K(5, 0, 3, 0, 2,
585 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
586 PRIM_INT__h_s_d_s, PRIM_INT__h_s_p_s, NULL, PRIM_INT__g_s_d_s, NULL, PRIM_INT__h_s_f_s);
587
588
589 ostei_general_vrr_K(5, 0, 4, 0, 1,
590 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
591 PRIM_INT__h_s_f_s, PRIM_INT__h_s_d_s, NULL, PRIM_INT__g_s_f_s, NULL, PRIM_INT__h_s_g_s);
592
593
594 ostei_general_vrr1_I(6, 6,
595 one_over_2p, a_over_p, aop_PQ, P_PA,
596 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
597
598
599 ostei_general_vrr_K(6, 0, 1, 0, 4,
600 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
601 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
602
603
604 ostei_general_vrr_K(6, 0, 2, 0, 3,
605 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
606 PRIM_INT__i_s_p_s, PRIM_INT__i_s_s_s, NULL, PRIM_INT__h_s_p_s, NULL, PRIM_INT__i_s_d_s);
607
608
609 ostei_general_vrr_K(6, 0, 3, 0, 2,
610 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
611 PRIM_INT__i_s_d_s, PRIM_INT__i_s_p_s, NULL, PRIM_INT__h_s_d_s, NULL, PRIM_INT__i_s_f_s);
612
613
614 ostei_general_vrr_K(6, 0, 4, 0, 1,
615 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
616 PRIM_INT__i_s_f_s, PRIM_INT__i_s_d_s, NULL, PRIM_INT__h_s_f_s, NULL, PRIM_INT__i_s_g_s);
617
618
619 ostei_general_vrr1_I(7, 5,
620 one_over_2p, a_over_p, aop_PQ, P_PA,
621 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
622
623
624 ostei_general_vrr_K(7, 0, 1, 0, 4,
625 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
626 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
627
628
629 ostei_general_vrr_K(7, 0, 2, 0, 3,
630 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
631 PRIM_INT__k_s_p_s, PRIM_INT__k_s_s_s, NULL, PRIM_INT__i_s_p_s, NULL, PRIM_INT__k_s_d_s);
632
633
634 ostei_general_vrr_K(7, 0, 3, 0, 2,
635 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
636 PRIM_INT__k_s_d_s, PRIM_INT__k_s_p_s, NULL, PRIM_INT__i_s_d_s, NULL, PRIM_INT__k_s_f_s);
637
638
639 ostei_general_vrr_K(7, 0, 4, 0, 1,
640 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
641 PRIM_INT__k_s_f_s, PRIM_INT__k_s_d_s, NULL, PRIM_INT__i_s_f_s, NULL, PRIM_INT__k_s_g_s);
642
643
644
645
646 ////////////////////////////////////
647 // Accumulate contracted integrals
648 ////////////////////////////////////
649 if(lastoffset == 0)
650 {
651 contract_all(150, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
652 contract_all(225, PRIM_INT__g_s_g_s, PRIM_PTR_INT__g_s_g_s);
653 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
654 contract_all(315, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
655 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
656 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
657 contract_all(360, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
658 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
659 }
660 else
661 {
662 contract(150, shelloffsets, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
663 contract(225, shelloffsets, PRIM_INT__g_s_g_s, PRIM_PTR_INT__g_s_g_s);
664 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
665 contract(315, shelloffsets, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
666 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
667 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
668 contract(360, shelloffsets, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
669 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
670 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
671 PRIM_PTR_INT__g_s_g_s += lastoffset*225;
672 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
673 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
674 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
675 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
676 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
677 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
678 }
679
680 } // close loop over j
681 } // close loop over i
682
683 //Advance to the next batch
684 jstart = SIMINT_SIMD_ROUND(jend);
685
686 //////////////////////////////////////////////
687 // Contracted integrals: Horizontal recurrance
688 //////////////////////////////////////////////
689
690
691 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
692
693
694 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
695 {
696 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
697
698 // set up HRR pointers
699 double const * restrict HRR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
700 double const * restrict HRR_INT__g_s_g_s = INT__g_s_g_s + abcd * 225;
701 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
702 double const * restrict HRR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
703 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
704 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
705 double const * restrict HRR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
706 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
707 double * restrict HRR_INT__g_f_f_p = INT__g_f_f_p + real_abcd * 4500;
708
709 // form INT__g_p_f_s
710 HRR_J_g_p(
711 HRR_INT__g_p_f_s,
712 HRR_INT__g_s_f_s,
713 HRR_INT__h_s_f_s,
714 hAB, 10);
715
716 // form INT__g_p_g_s
717 HRR_J_g_p(
718 HRR_INT__g_p_g_s,
719 HRR_INT__g_s_g_s,
720 HRR_INT__h_s_g_s,
721 hAB, 15);
722
723 // form INT__h_p_f_s
724 ostei_general_hrr_J(5, 1, 3, 0, hAB, HRR_INT__i_s_f_s, HRR_INT__h_s_f_s, HRR_INT__h_p_f_s);
725
726 // form INT__h_p_g_s
727 ostei_general_hrr_J(5, 1, 4, 0, hAB, HRR_INT__i_s_g_s, HRR_INT__h_s_g_s, HRR_INT__h_p_g_s);
728
729 // form INT__i_p_f_s
730 ostei_general_hrr_J(6, 1, 3, 0, hAB, HRR_INT__k_s_f_s, HRR_INT__i_s_f_s, HRR_INT__i_p_f_s);
731
732 // form INT__i_p_g_s
733 ostei_general_hrr_J(6, 1, 4, 0, hAB, HRR_INT__k_s_g_s, HRR_INT__i_s_g_s, HRR_INT__i_p_g_s);
734
735 // form INT__g_d_f_s
736 ostei_general_hrr_J(4, 2, 3, 0, hAB, HRR_INT__h_p_f_s, HRR_INT__g_p_f_s, HRR_INT__g_d_f_s);
737
738 // form INT__g_d_g_s
739 ostei_general_hrr_J(4, 2, 4, 0, hAB, HRR_INT__h_p_g_s, HRR_INT__g_p_g_s, HRR_INT__g_d_g_s);
740
741 // form INT__h_d_f_s
742 ostei_general_hrr_J(5, 2, 3, 0, hAB, HRR_INT__i_p_f_s, HRR_INT__h_p_f_s, HRR_INT__h_d_f_s);
743
744 // form INT__h_d_g_s
745 ostei_general_hrr_J(5, 2, 4, 0, hAB, HRR_INT__i_p_g_s, HRR_INT__h_p_g_s, HRR_INT__h_d_g_s);
746
747 // form INT__g_f_f_s
748 ostei_general_hrr_J(4, 3, 3, 0, hAB, HRR_INT__h_d_f_s, HRR_INT__g_d_f_s, HRR_INT__g_f_f_s);
749
750 // form INT__g_f_g_s
751 ostei_general_hrr_J(4, 3, 4, 0, hAB, HRR_INT__h_d_g_s, HRR_INT__g_d_g_s, HRR_INT__g_f_g_s);
752
753 // form INT__g_f_f_p
754 HRR_L_f_p(
755 HRR_INT__g_f_f_p,
756 HRR_INT__g_f_f_s,
757 HRR_INT__g_f_g_s,
758 hCD, 150);
759
760
761 } // close HRR loop
762
763
764 } // close loop cdbatch
765
766 istart = iend;
767 } // close loop over ab
768
769 return P.nshell12_clip * Q.nshell12_clip;
770 }
771
ostei_f_g_f_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_g_f_p)772 int ostei_f_g_f_p(struct simint_multi_shellpair const P,
773 struct simint_multi_shellpair const Q,
774 double screen_tol,
775 double * const restrict work,
776 double * const restrict INT__f_g_f_p)
777 {
778 double P_AB[3*P.nshell12];
779 struct simint_multi_shellpair P_tmp = P;
780 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
781 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
782 P_tmp.AB_x = P_AB;
783 P_tmp.AB_y = P_AB + P.nshell12;
784 P_tmp.AB_z = P_AB + 2*P.nshell12;
785
786 for(int i = 0; i < P.nshell12; i++)
787 {
788 P_tmp.AB_x[i] = -P.AB_x[i];
789 P_tmp.AB_y[i] = -P.AB_y[i];
790 P_tmp.AB_z[i] = -P.AB_z[i];
791 }
792
793 int ret = ostei_g_f_f_p(P_tmp, Q, screen_tol, work, INT__f_g_f_p);
794 double buffer[4500] SIMINT_ALIGN_ARRAY_DBL;
795
796 for(int q = 0; q < ret; q++)
797 {
798 int idx = 0;
799 for(int a = 0; a < 10; ++a)
800 for(int b = 0; b < 15; ++b)
801 for(int c = 0; c < 10; ++c)
802 for(int d = 0; d < 3; ++d)
803 buffer[idx++] = INT__f_g_f_p[q*4500+b*300+a*30+c*3+d];
804
805 memcpy(INT__f_g_f_p+q*4500, buffer, 4500*sizeof(double));
806 }
807
808 return ret;
809 }
810
ostei_g_f_p_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_f_p_f)811 int ostei_g_f_p_f(struct simint_multi_shellpair const P,
812 struct simint_multi_shellpair const Q,
813 double screen_tol,
814 double * const restrict work,
815 double * const restrict INT__g_f_p_f)
816 {
817 double Q_AB[3*Q.nshell12];
818 struct simint_multi_shellpair Q_tmp = Q;
819 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
820 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
821 Q_tmp.AB_x = Q_AB;
822 Q_tmp.AB_y = Q_AB + Q.nshell12;
823 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
824
825 for(int i = 0; i < Q.nshell12; i++)
826 {
827 Q_tmp.AB_x[i] = -Q.AB_x[i];
828 Q_tmp.AB_y[i] = -Q.AB_y[i];
829 Q_tmp.AB_z[i] = -Q.AB_z[i];
830 }
831
832 int ret = ostei_g_f_f_p(P, Q_tmp, screen_tol, work, INT__g_f_p_f);
833 double buffer[4500] SIMINT_ALIGN_ARRAY_DBL;
834
835 for(int q = 0; q < ret; q++)
836 {
837 int idx = 0;
838 for(int a = 0; a < 15; ++a)
839 for(int b = 0; b < 10; ++b)
840 for(int c = 0; c < 3; ++c)
841 for(int d = 0; d < 10; ++d)
842 buffer[idx++] = INT__g_f_p_f[q*4500+a*300+b*30+d*3+c];
843
844 memcpy(INT__g_f_p_f+q*4500, buffer, 4500*sizeof(double));
845 }
846
847 return ret;
848 }
849
ostei_f_g_p_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_g_p_f)850 int ostei_f_g_p_f(struct simint_multi_shellpair const P,
851 struct simint_multi_shellpair const Q,
852 double screen_tol,
853 double * const restrict work,
854 double * const restrict INT__f_g_p_f)
855 {
856 double P_AB[3*P.nshell12];
857 struct simint_multi_shellpair P_tmp = P;
858 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
859 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
860 P_tmp.AB_x = P_AB;
861 P_tmp.AB_y = P_AB + P.nshell12;
862 P_tmp.AB_z = P_AB + 2*P.nshell12;
863
864 for(int i = 0; i < P.nshell12; i++)
865 {
866 P_tmp.AB_x[i] = -P.AB_x[i];
867 P_tmp.AB_y[i] = -P.AB_y[i];
868 P_tmp.AB_z[i] = -P.AB_z[i];
869 }
870
871 double Q_AB[3*Q.nshell12];
872 struct simint_multi_shellpair Q_tmp = Q;
873 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
874 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
875 Q_tmp.AB_x = Q_AB;
876 Q_tmp.AB_y = Q_AB + Q.nshell12;
877 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
878
879 for(int i = 0; i < Q.nshell12; i++)
880 {
881 Q_tmp.AB_x[i] = -Q.AB_x[i];
882 Q_tmp.AB_y[i] = -Q.AB_y[i];
883 Q_tmp.AB_z[i] = -Q.AB_z[i];
884 }
885
886 int ret = ostei_g_f_f_p(P_tmp, Q_tmp, screen_tol, work, INT__f_g_p_f);
887 double buffer[4500] SIMINT_ALIGN_ARRAY_DBL;
888
889 for(int q = 0; q < ret; q++)
890 {
891 int idx = 0;
892 for(int a = 0; a < 10; ++a)
893 for(int b = 0; b < 15; ++b)
894 for(int c = 0; c < 3; ++c)
895 for(int d = 0; d < 10; ++d)
896 buffer[idx++] = INT__f_g_p_f[q*4500+b*300+a*30+d*3+c];
897
898 memcpy(INT__f_g_p_f+q*4500, buffer, 4500*sizeof(double));
899 }
900
901 return ret;
902 }
903
904