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_d_d_h_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_d_h_d)8 int ostei_d_d_h_d(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__d_d_h_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__d_d_h_d);
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__d_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__d_s_i_s = work + (SIMINT_NSHELL_SIMD * 126);
31 double * const INT__d_s_k_s = work + (SIMINT_NSHELL_SIMD * 294);
32 double * const INT__f_s_h_s = work + (SIMINT_NSHELL_SIMD * 510);
33 double * const INT__f_s_i_s = work + (SIMINT_NSHELL_SIMD * 720);
34 double * const INT__f_s_k_s = work + (SIMINT_NSHELL_SIMD * 1000);
35 double * const INT__g_s_h_s = work + (SIMINT_NSHELL_SIMD * 1360);
36 double * const INT__g_s_i_s = work + (SIMINT_NSHELL_SIMD * 1675);
37 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 2095);
38 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2635);
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 12;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 45;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 105;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 195;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 315;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 462;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 630;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 810;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 882;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1002;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1182;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1434;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 1770;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 2202;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 2382;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 2652;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 3030;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 3534;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 4182;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 4482;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 4902;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 5462;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 6182;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 6497;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 6917;
65 double * const hrrwork = (double *)(primwork + 7457);
66 double * const HRR_INT__d_p_h_s = hrrwork + 0;
67 double * const HRR_INT__d_p_i_s = hrrwork + 378;
68 double * const HRR_INT__d_p_k_s = hrrwork + 882;
69 double * const HRR_INT__d_d_h_s = hrrwork + 1530;
70 double * const HRR_INT__d_d_h_p = hrrwork + 2286;
71 double * const HRR_INT__d_d_i_s = hrrwork + 4554;
72 double * const HRR_INT__d_d_i_p = hrrwork + 5562;
73 double * const HRR_INT__d_d_k_s = hrrwork + 8586;
74 double * const HRR_INT__f_p_h_s = hrrwork + 9882;
75 double * const HRR_INT__f_p_i_s = hrrwork + 10512;
76 double * const HRR_INT__f_p_k_s = hrrwork + 11352;
77
78
79 // Create constants
80 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
81 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
82 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
83 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
84 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
85 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
86 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
87 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
88
89
90 ////////////////////////////////////////
91 // Loop over shells and primitives
92 ////////////////////////////////////////
93
94 real_abcd = 0;
95 istart = 0;
96 for(ab = 0; ab < P.nshell12_clip; ++ab)
97 {
98 const int iend = istart + P.nprim12[ab];
99
100 cd = 0;
101 jstart = 0;
102
103 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
104 {
105 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
106 int jend = jstart;
107 for(i = 0; i < nshellbatch; i++)
108 jend += Q.nprim12[cd+i];
109
110 // Clear the beginning of the workspace (where we are accumulating integrals)
111 memset(work, 0, SIMINT_NSHELL_SIMD * 2635 * sizeof(double));
112 abcd = 0;
113
114
115 for(i = istart; i < iend; ++i)
116 {
117 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
118
119 if(check_screen)
120 {
121 // Skip this whole thing if always insignificant
122 if((P.screen[i] * Q.screen_max) < screen_tol)
123 continue;
124 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
125 }
126
127 icd = 0;
128 iprimcd = 0;
129 nprim_icd = Q.nprim12[cd];
130 double * restrict PRIM_PTR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
131 double * restrict PRIM_PTR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
132 double * restrict PRIM_PTR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
133 double * restrict PRIM_PTR_INT__f_s_h_s = INT__f_s_h_s + abcd * 210;
134 double * restrict PRIM_PTR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
135 double * restrict PRIM_PTR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
136 double * restrict PRIM_PTR_INT__g_s_h_s = INT__g_s_h_s + abcd * 315;
137 double * restrict PRIM_PTR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
138 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
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__d_s_h_s += 126;
165 PRIM_PTR_INT__d_s_i_s += 168;
166 PRIM_PTR_INT__d_s_k_s += 216;
167 PRIM_PTR_INT__f_s_h_s += 210;
168 PRIM_PTR_INT__f_s_i_s += 280;
169 PRIM_PTR_INT__f_s_k_s += 360;
170 PRIM_PTR_INT__g_s_h_s += 315;
171 PRIM_PTR_INT__g_s_i_s += 420;
172 PRIM_PTR_INT__g_s_k_s += 540;
173 }
174 iprimcd++;
175 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
176 {
177 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
178 {
179 shelloffsets[n] = shelloffsets[n-1] + 1;
180 lastoffset++;
181 nprim_icd += Q.nprim12[cd + (++icd)];
182 }
183 else
184 shelloffsets[n] = shelloffsets[n-1];
185 iprimcd++;
186 }
187 }
188 else
189 iprimcd += SIMINT_SIMD_LEN;
190
191 // Do we have to compute this vector (or has it been screened out)?
192 // (not_screened != 0 means we have to do this vector)
193 if(check_screen)
194 {
195 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
196 if(vmax < screen_tol)
197 {
198 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
199 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
200 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
201 PRIM_PTR_INT__f_s_h_s += lastoffset*210;
202 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
203 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
204 PRIM_PTR_INT__g_s_h_s += lastoffset*315;
205 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
206 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
207 continue;
208 }
209 }
210
211 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
212 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
213 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
214 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
215
216
217 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
218 SIMINT_DBLTYPE PQ[3];
219 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
220 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
221 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
222 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
223 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
224 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
225
226 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
227 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
228 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
229 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
230 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
231 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
232 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
233
234 // NOTE: Minus sign!
235 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
236 SIMINT_DBLTYPE aop_PQ[3];
237 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
238 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
239 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
240
241 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
242 SIMINT_DBLTYPE aoq_PQ[3];
243 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
244 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
245 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
246 // Put a minus sign here so we don't have to in RR routines
247 a_over_q = SIMINT_NEG(a_over_q);
248
249
250 //////////////////////////////////////////////
251 // Fjt function section
252 // Maximum v value: 11
253 //////////////////////////////////////////////
254 // The parameter to the Fjt function
255 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
256
257
258 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
259
260
261 boys_F_split(PRIM_INT__s_s_s_s, F_x, 11);
262 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
263 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
264 for(n = 0; n <= 11; n++)
265 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
266
267 //////////////////////////////////////////////
268 // Primitive integrals: Vertical recurrance
269 //////////////////////////////////////////////
270
271 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
272 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
273 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
274 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
275 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
276 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
277 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
278 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
279 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
280 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
281 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
282 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
283 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
284 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
285 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
287
288
289
290 // Forming PRIM_INT__s_s_p_s[11 * 3];
291 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
292 {
293
294 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
295 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]);
296
297 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
298 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]);
299
300 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
301 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]);
302
303 }
304
305
306
307 // Forming PRIM_INT__s_s_d_s[10 * 6];
308 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
309 {
310
311 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
312 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]);
313 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]);
314
315 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
316 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]);
317
318 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
319 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]);
320
321 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
322 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]);
323 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]);
324
325 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
326 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]);
327
328 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
329 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]);
330 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]);
331
332 }
333
334
335
336 // Forming PRIM_INT__s_s_f_s[9 * 10];
337 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
338 {
339
340 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
341 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]);
342 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]);
343
344 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
345 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]);
346
347 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
348 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]);
349
350 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
351 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]);
352
353 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
354 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]);
355
356 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
357 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]);
358
359 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
360 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]);
361 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]);
362
363 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
364 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]);
365
366 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
367 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]);
368
369 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
370 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]);
371 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]);
372
373 }
374
375
376 VRR_K_s_s_g_s(
377 PRIM_INT__s_s_g_s,
378 PRIM_INT__s_s_f_s,
379 PRIM_INT__s_s_d_s,
380 Q_PA,
381 a_over_q,
382 aoq_PQ,
383 one_over_2q,
384 8);
385
386
387 VRR_K_s_s_h_s(
388 PRIM_INT__s_s_h_s,
389 PRIM_INT__s_s_g_s,
390 PRIM_INT__s_s_f_s,
391 Q_PA,
392 a_over_q,
393 aoq_PQ,
394 one_over_2q,
395 7);
396
397
398 ostei_general_vrr_I(1, 0, 5, 0, 4,
399 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
400 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
401
402
403 VRR_I_p_s_g_s(
404 PRIM_INT__p_s_g_s,
405 PRIM_INT__s_s_g_s,
406 PRIM_INT__s_s_f_s,
407 P_PA,
408 aop_PQ,
409 one_over_2pq,
410 4);
411
412
413 ostei_general_vrr_I(2, 0, 5, 0, 3,
414 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
415 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);
416
417
418 ostei_general_vrr1_K(6, 6,
419 one_over_2q, a_over_q, aoq_PQ, Q_PA,
420 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
421
422
423 ostei_general_vrr_I(1, 0, 6, 0, 4,
424 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
425 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
426
427
428 ostei_general_vrr_I(2, 0, 6, 0, 3,
429 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
430 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);
431
432
433 VRR_I_p_s_f_s(
434 PRIM_INT__p_s_f_s,
435 PRIM_INT__s_s_f_s,
436 PRIM_INT__s_s_d_s,
437 P_PA,
438 aop_PQ,
439 one_over_2pq,
440 4);
441
442
443 ostei_general_vrr_I(2, 0, 4, 0, 3,
444 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
445 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);
446
447
448 ostei_general_vrr_I(3, 0, 5, 0, 2,
449 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
450 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);
451
452
453 ostei_general_vrr1_K(7, 5,
454 one_over_2q, a_over_q, aoq_PQ, Q_PA,
455 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
456
457
458 ostei_general_vrr_I(1, 0, 7, 0, 4,
459 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
460 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
461
462
463 ostei_general_vrr_I(2, 0, 7, 0, 3,
464 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
465 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);
466
467
468 ostei_general_vrr_I(3, 0, 6, 0, 2,
469 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
470 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);
471
472
473
474 // Forming PRIM_INT__p_s_d_s[4 * 18];
475 for(n = 0; n < 4; ++n) // loop over orders of auxiliary function
476 {
477
478 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
479 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]);
480 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]);
481
482 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
483 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]);
484 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]);
485
486 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
487 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]);
488 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]);
489
490 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
491 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]);
492
493 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
494 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]);
495
496 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
497 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]);
498
499 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
500 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]);
501
502 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
503 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]);
504 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]);
505
506 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
507 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]);
508
509 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
510 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]);
511 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]);
512
513 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
514 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]);
515 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]);
516
517 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
518 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]);
519
520 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
521 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]);
522
523 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
524 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]);
525
526 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
527 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]);
528 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]);
529
530 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
531 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]);
532
533 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
534 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]);
535 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]);
536
537 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
538 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]);
539 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]);
540
541 }
542
543
544 VRR_I_d_s_f_s(
545 PRIM_INT__d_s_f_s,
546 PRIM_INT__p_s_f_s,
547 PRIM_INT__s_s_f_s,
548 PRIM_INT__p_s_d_s,
549 P_PA,
550 a_over_p,
551 aop_PQ,
552 one_over_2p,
553 one_over_2pq,
554 3);
555
556
557 ostei_general_vrr_I(3, 0, 4, 0, 2,
558 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
559 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);
560
561
562 ostei_general_vrr_I(4, 0, 5, 0, 1,
563 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
564 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);
565
566
567 ostei_general_vrr_I(3, 0, 7, 0, 2,
568 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
569 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);
570
571
572 ostei_general_vrr_I(4, 0, 6, 0, 1,
573 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
574 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);
575
576
577 ostei_general_vrr_I(4, 0, 7, 0, 1,
578 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
579 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);
580
581
582
583
584 ////////////////////////////////////
585 // Accumulate contracted integrals
586 ////////////////////////////////////
587 if(lastoffset == 0)
588 {
589 contract_all(126, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
590 contract_all(168, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
591 contract_all(216, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
592 contract_all(210, PRIM_INT__f_s_h_s, PRIM_PTR_INT__f_s_h_s);
593 contract_all(280, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
594 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
595 contract_all(315, PRIM_INT__g_s_h_s, PRIM_PTR_INT__g_s_h_s);
596 contract_all(420, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
597 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
598 }
599 else
600 {
601 contract(126, shelloffsets, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
602 contract(168, shelloffsets, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
603 contract(216, shelloffsets, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
604 contract(210, shelloffsets, PRIM_INT__f_s_h_s, PRIM_PTR_INT__f_s_h_s);
605 contract(280, shelloffsets, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
606 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
607 contract(315, shelloffsets, PRIM_INT__g_s_h_s, PRIM_PTR_INT__g_s_h_s);
608 contract(420, shelloffsets, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
609 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
610 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
611 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
612 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
613 PRIM_PTR_INT__f_s_h_s += lastoffset*210;
614 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
615 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
616 PRIM_PTR_INT__g_s_h_s += lastoffset*315;
617 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
618 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
619 }
620
621 } // close loop over j
622 } // close loop over i
623
624 //Advance to the next batch
625 jstart = SIMINT_SIMD_ROUND(jend);
626
627 //////////////////////////////////////////////
628 // Contracted integrals: Horizontal recurrance
629 //////////////////////////////////////////////
630
631
632 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
633
634
635 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
636 {
637 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
638
639 // set up HRR pointers
640 double const * restrict HRR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
641 double const * restrict HRR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
642 double const * restrict HRR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
643 double const * restrict HRR_INT__f_s_h_s = INT__f_s_h_s + abcd * 210;
644 double const * restrict HRR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
645 double const * restrict HRR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
646 double const * restrict HRR_INT__g_s_h_s = INT__g_s_h_s + abcd * 315;
647 double const * restrict HRR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
648 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
649 double * restrict HRR_INT__d_d_h_d = INT__d_d_h_d + real_abcd * 4536;
650
651 // form INT__d_p_h_s
652 for(iket = 0; iket < 21; ++iket)
653 {
654 HRR_INT__d_p_h_s[0 * 21 + iket] = HRR_INT__f_s_h_s[0 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[0 * 21 + iket] );
655
656 HRR_INT__d_p_h_s[1 * 21 + iket] = HRR_INT__f_s_h_s[1 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[0 * 21 + iket] );
657
658 HRR_INT__d_p_h_s[2 * 21 + iket] = HRR_INT__f_s_h_s[2 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[0 * 21 + iket] );
659
660 HRR_INT__d_p_h_s[3 * 21 + iket] = HRR_INT__f_s_h_s[1 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[1 * 21 + iket] );
661
662 HRR_INT__d_p_h_s[4 * 21 + iket] = HRR_INT__f_s_h_s[3 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[1 * 21 + iket] );
663
664 HRR_INT__d_p_h_s[5 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[1 * 21 + iket] );
665
666 HRR_INT__d_p_h_s[6 * 21 + iket] = HRR_INT__f_s_h_s[2 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[2 * 21 + iket] );
667
668 HRR_INT__d_p_h_s[7 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[2 * 21 + iket] );
669
670 HRR_INT__d_p_h_s[8 * 21 + iket] = HRR_INT__f_s_h_s[5 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[2 * 21 + iket] );
671
672 HRR_INT__d_p_h_s[9 * 21 + iket] = HRR_INT__f_s_h_s[3 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[3 * 21 + iket] );
673
674 HRR_INT__d_p_h_s[10 * 21 + iket] = HRR_INT__f_s_h_s[6 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[3 * 21 + iket] );
675
676 HRR_INT__d_p_h_s[11 * 21 + iket] = HRR_INT__f_s_h_s[7 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[3 * 21 + iket] );
677
678 HRR_INT__d_p_h_s[12 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[4 * 21 + iket] );
679
680 HRR_INT__d_p_h_s[13 * 21 + iket] = HRR_INT__f_s_h_s[7 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[4 * 21 + iket] );
681
682 HRR_INT__d_p_h_s[14 * 21 + iket] = HRR_INT__f_s_h_s[8 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[4 * 21 + iket] );
683
684 HRR_INT__d_p_h_s[15 * 21 + iket] = HRR_INT__f_s_h_s[5 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[5 * 21 + iket] );
685
686 HRR_INT__d_p_h_s[16 * 21 + iket] = HRR_INT__f_s_h_s[8 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[5 * 21 + iket] );
687
688 HRR_INT__d_p_h_s[17 * 21 + iket] = HRR_INT__f_s_h_s[9 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[5 * 21 + iket] );
689
690 }
691
692
693 // form INT__d_p_i_s
694 for(iket = 0; iket < 28; ++iket)
695 {
696 HRR_INT__d_p_i_s[0 * 28 + iket] = HRR_INT__f_s_i_s[0 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[0 * 28 + iket] );
697
698 HRR_INT__d_p_i_s[1 * 28 + iket] = HRR_INT__f_s_i_s[1 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[0 * 28 + iket] );
699
700 HRR_INT__d_p_i_s[2 * 28 + iket] = HRR_INT__f_s_i_s[2 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[0 * 28 + iket] );
701
702 HRR_INT__d_p_i_s[3 * 28 + iket] = HRR_INT__f_s_i_s[1 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[1 * 28 + iket] );
703
704 HRR_INT__d_p_i_s[4 * 28 + iket] = HRR_INT__f_s_i_s[3 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[1 * 28 + iket] );
705
706 HRR_INT__d_p_i_s[5 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[1 * 28 + iket] );
707
708 HRR_INT__d_p_i_s[6 * 28 + iket] = HRR_INT__f_s_i_s[2 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[2 * 28 + iket] );
709
710 HRR_INT__d_p_i_s[7 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[2 * 28 + iket] );
711
712 HRR_INT__d_p_i_s[8 * 28 + iket] = HRR_INT__f_s_i_s[5 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[2 * 28 + iket] );
713
714 HRR_INT__d_p_i_s[9 * 28 + iket] = HRR_INT__f_s_i_s[3 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[3 * 28 + iket] );
715
716 HRR_INT__d_p_i_s[10 * 28 + iket] = HRR_INT__f_s_i_s[6 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[3 * 28 + iket] );
717
718 HRR_INT__d_p_i_s[11 * 28 + iket] = HRR_INT__f_s_i_s[7 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[3 * 28 + iket] );
719
720 HRR_INT__d_p_i_s[12 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[4 * 28 + iket] );
721
722 HRR_INT__d_p_i_s[13 * 28 + iket] = HRR_INT__f_s_i_s[7 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[4 * 28 + iket] );
723
724 HRR_INT__d_p_i_s[14 * 28 + iket] = HRR_INT__f_s_i_s[8 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[4 * 28 + iket] );
725
726 HRR_INT__d_p_i_s[15 * 28 + iket] = HRR_INT__f_s_i_s[5 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[5 * 28 + iket] );
727
728 HRR_INT__d_p_i_s[16 * 28 + iket] = HRR_INT__f_s_i_s[8 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[5 * 28 + iket] );
729
730 HRR_INT__d_p_i_s[17 * 28 + iket] = HRR_INT__f_s_i_s[9 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[5 * 28 + iket] );
731
732 }
733
734
735 // form INT__d_p_k_s
736 for(iket = 0; iket < 36; ++iket)
737 {
738 HRR_INT__d_p_k_s[0 * 36 + iket] = HRR_INT__f_s_k_s[0 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[0 * 36 + iket] );
739
740 HRR_INT__d_p_k_s[1 * 36 + iket] = HRR_INT__f_s_k_s[1 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[0 * 36 + iket] );
741
742 HRR_INT__d_p_k_s[2 * 36 + iket] = HRR_INT__f_s_k_s[2 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[0 * 36 + iket] );
743
744 HRR_INT__d_p_k_s[3 * 36 + iket] = HRR_INT__f_s_k_s[1 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[1 * 36 + iket] );
745
746 HRR_INT__d_p_k_s[4 * 36 + iket] = HRR_INT__f_s_k_s[3 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[1 * 36 + iket] );
747
748 HRR_INT__d_p_k_s[5 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[1 * 36 + iket] );
749
750 HRR_INT__d_p_k_s[6 * 36 + iket] = HRR_INT__f_s_k_s[2 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[2 * 36 + iket] );
751
752 HRR_INT__d_p_k_s[7 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[2 * 36 + iket] );
753
754 HRR_INT__d_p_k_s[8 * 36 + iket] = HRR_INT__f_s_k_s[5 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[2 * 36 + iket] );
755
756 HRR_INT__d_p_k_s[9 * 36 + iket] = HRR_INT__f_s_k_s[3 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[3 * 36 + iket] );
757
758 HRR_INT__d_p_k_s[10 * 36 + iket] = HRR_INT__f_s_k_s[6 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[3 * 36 + iket] );
759
760 HRR_INT__d_p_k_s[11 * 36 + iket] = HRR_INT__f_s_k_s[7 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[3 * 36 + iket] );
761
762 HRR_INT__d_p_k_s[12 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[4 * 36 + iket] );
763
764 HRR_INT__d_p_k_s[13 * 36 + iket] = HRR_INT__f_s_k_s[7 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[4 * 36 + iket] );
765
766 HRR_INT__d_p_k_s[14 * 36 + iket] = HRR_INT__f_s_k_s[8 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[4 * 36 + iket] );
767
768 HRR_INT__d_p_k_s[15 * 36 + iket] = HRR_INT__f_s_k_s[5 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[5 * 36 + iket] );
769
770 HRR_INT__d_p_k_s[16 * 36 + iket] = HRR_INT__f_s_k_s[8 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[5 * 36 + iket] );
771
772 HRR_INT__d_p_k_s[17 * 36 + iket] = HRR_INT__f_s_k_s[9 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[5 * 36 + iket] );
773
774 }
775
776
777 // form INT__f_p_h_s
778 HRR_J_f_p(
779 HRR_INT__f_p_h_s,
780 HRR_INT__f_s_h_s,
781 HRR_INT__g_s_h_s,
782 hAB, 21);
783
784 // form INT__f_p_i_s
785 HRR_J_f_p(
786 HRR_INT__f_p_i_s,
787 HRR_INT__f_s_i_s,
788 HRR_INT__g_s_i_s,
789 hAB, 28);
790
791 // form INT__f_p_k_s
792 HRR_J_f_p(
793 HRR_INT__f_p_k_s,
794 HRR_INT__f_s_k_s,
795 HRR_INT__g_s_k_s,
796 hAB, 36);
797
798 // form INT__d_d_h_s
799 HRR_J_d_d(
800 HRR_INT__d_d_h_s,
801 HRR_INT__d_p_h_s,
802 HRR_INT__f_p_h_s,
803 hAB, 21);
804
805 // form INT__d_d_i_s
806 HRR_J_d_d(
807 HRR_INT__d_d_i_s,
808 HRR_INT__d_p_i_s,
809 HRR_INT__f_p_i_s,
810 hAB, 28);
811
812 // form INT__d_d_k_s
813 HRR_J_d_d(
814 HRR_INT__d_d_k_s,
815 HRR_INT__d_p_k_s,
816 HRR_INT__f_p_k_s,
817 hAB, 36);
818
819 // form INT__d_d_h_p
820 ostei_general_hrr_L(2, 2, 5, 1, hCD, HRR_INT__d_d_i_s, HRR_INT__d_d_h_s, HRR_INT__d_d_h_p);
821
822 // form INT__d_d_i_p
823 ostei_general_hrr_L(2, 2, 6, 1, hCD, HRR_INT__d_d_k_s, HRR_INT__d_d_i_s, HRR_INT__d_d_i_p);
824
825 // form INT__d_d_h_d
826 ostei_general_hrr_L(2, 2, 5, 2, hCD, HRR_INT__d_d_i_p, HRR_INT__d_d_h_p, HRR_INT__d_d_h_d);
827
828
829 } // close HRR loop
830
831
832 } // close loop cdbatch
833
834 istart = iend;
835 } // close loop over ab
836
837 return P.nshell12_clip * Q.nshell12_clip;
838 }
839
ostei_d_d_d_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_d_d_h)840 int ostei_d_d_d_h(struct simint_multi_shellpair const P,
841 struct simint_multi_shellpair const Q,
842 double screen_tol,
843 double * const restrict work,
844 double * const restrict INT__d_d_d_h)
845 {
846 double Q_AB[3*Q.nshell12];
847 struct simint_multi_shellpair Q_tmp = Q;
848 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
849 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
850 Q_tmp.AB_x = Q_AB;
851 Q_tmp.AB_y = Q_AB + Q.nshell12;
852 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
853
854 for(int i = 0; i < Q.nshell12; i++)
855 {
856 Q_tmp.AB_x[i] = -Q.AB_x[i];
857 Q_tmp.AB_y[i] = -Q.AB_y[i];
858 Q_tmp.AB_z[i] = -Q.AB_z[i];
859 }
860
861 int ret = ostei_d_d_h_d(P, Q_tmp, screen_tol, work, INT__d_d_d_h);
862 double buffer[4536] SIMINT_ALIGN_ARRAY_DBL;
863
864 for(int q = 0; q < ret; q++)
865 {
866 int idx = 0;
867 for(int a = 0; a < 6; ++a)
868 for(int b = 0; b < 6; ++b)
869 for(int c = 0; c < 6; ++c)
870 for(int d = 0; d < 21; ++d)
871 buffer[idx++] = INT__d_d_d_h[q*4536+a*756+b*126+d*6+c];
872
873 memcpy(INT__d_d_d_h+q*4536, buffer, 4536*sizeof(double));
874 }
875
876 return ret;
877 }
878
879