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_i_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_i_d)8 int ostei_d_d_i_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_i_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__d_d_i_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_i_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__d_s_k_s = work + (SIMINT_NSHELL_SIMD * 168);
31 double * const INT__d_s_l_s = work + (SIMINT_NSHELL_SIMD * 384);
32 double * const INT__f_s_i_s = work + (SIMINT_NSHELL_SIMD * 654);
33 double * const INT__f_s_k_s = work + (SIMINT_NSHELL_SIMD * 934);
34 double * const INT__f_s_l_s = work + (SIMINT_NSHELL_SIMD * 1294);
35 double * const INT__g_s_i_s = work + (SIMINT_NSHELL_SIMD * 1744);
36 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 2164);
37 double * const INT__g_s_l_s = work + (SIMINT_NSHELL_SIMD * 2704);
38 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3379);
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 + 13;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 49;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 115;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 215;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 350;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 518;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 714;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 930;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1155;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1275;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1455;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1707;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2043;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 2475;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 3015;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 3285;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 3663;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 4167;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 4815;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 5625;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 6045;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 6605;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 7325;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 8225;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 8645;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_l_s = primwork + 9185;
66 double * const hrrwork = (double *)(primwork + 9860);
67 double * const HRR_INT__d_p_i_s = hrrwork + 0;
68 double * const HRR_INT__d_p_k_s = hrrwork + 504;
69 double * const HRR_INT__d_p_l_s = hrrwork + 1152;
70 double * const HRR_INT__d_d_i_s = hrrwork + 1962;
71 double * const HRR_INT__d_d_i_p = hrrwork + 2970;
72 double * const HRR_INT__d_d_k_s = hrrwork + 5994;
73 double * const HRR_INT__d_d_k_p = hrrwork + 7290;
74 double * const HRR_INT__d_d_l_s = hrrwork + 11178;
75 double * const HRR_INT__f_p_i_s = hrrwork + 12798;
76 double * const HRR_INT__f_p_k_s = hrrwork + 13638;
77 double * const HRR_INT__f_p_l_s = hrrwork + 14718;
78
79
80 // Create constants
81 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
82 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
83 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
84 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
85 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
86 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
87 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
88 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
89 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
90
91
92 ////////////////////////////////////////
93 // Loop over shells and primitives
94 ////////////////////////////////////////
95
96 real_abcd = 0;
97 istart = 0;
98 for(ab = 0; ab < P.nshell12_clip; ++ab)
99 {
100 const int iend = istart + P.nprim12[ab];
101
102 cd = 0;
103 jstart = 0;
104
105 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
106 {
107 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
108 int jend = jstart;
109 for(i = 0; i < nshellbatch; i++)
110 jend += Q.nprim12[cd+i];
111
112 // Clear the beginning of the workspace (where we are accumulating integrals)
113 memset(work, 0, SIMINT_NSHELL_SIMD * 3379 * sizeof(double));
114 abcd = 0;
115
116
117 for(i = istart; i < iend; ++i)
118 {
119 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
120
121 if(check_screen)
122 {
123 // Skip this whole thing if always insignificant
124 if((P.screen[i] * Q.screen_max) < screen_tol)
125 continue;
126 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
127 }
128
129 icd = 0;
130 iprimcd = 0;
131 nprim_icd = Q.nprim12[cd];
132 double * restrict PRIM_PTR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
133 double * restrict PRIM_PTR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
134 double * restrict PRIM_PTR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
135 double * restrict PRIM_PTR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
136 double * restrict PRIM_PTR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
137 double * restrict PRIM_PTR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
138 double * restrict PRIM_PTR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
139 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
140 double * restrict PRIM_PTR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
141
142
143
144 // Load these one per loop over i
145 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
146 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
147 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
148
149 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
150
151 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
152 {
153 // calculate the shell offsets
154 // these are the offset from the shell pointed to by cd
155 // for each element
156 int shelloffsets[SIMINT_SIMD_LEN] = {0};
157 int lastoffset = 0;
158 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
159
160 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
161 {
162 // Handle if the first element of the vector is a new shell
163 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
164 {
165 nprim_icd += Q.nprim12[cd + (++icd)];
166 PRIM_PTR_INT__d_s_i_s += 168;
167 PRIM_PTR_INT__d_s_k_s += 216;
168 PRIM_PTR_INT__d_s_l_s += 270;
169 PRIM_PTR_INT__f_s_i_s += 280;
170 PRIM_PTR_INT__f_s_k_s += 360;
171 PRIM_PTR_INT__f_s_l_s += 450;
172 PRIM_PTR_INT__g_s_i_s += 420;
173 PRIM_PTR_INT__g_s_k_s += 540;
174 PRIM_PTR_INT__g_s_l_s += 675;
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__d_s_i_s += lastoffset*168;
201 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
202 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
203 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
204 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
205 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
206 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
207 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
208 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
209 continue;
210 }
211 }
212
213 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
214 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
215 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
216 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
217
218
219 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
220 SIMINT_DBLTYPE PQ[3];
221 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
222 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
223 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
224 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
225 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
226 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
227
228 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
229 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
230 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
231 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
232 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
233 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
234 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
235
236 // NOTE: Minus sign!
237 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
238 SIMINT_DBLTYPE aop_PQ[3];
239 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
240 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
241 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
242
243 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
244 SIMINT_DBLTYPE aoq_PQ[3];
245 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
246 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
247 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
248 // Put a minus sign here so we don't have to in RR routines
249 a_over_q = SIMINT_NEG(a_over_q);
250
251
252 //////////////////////////////////////////////
253 // Fjt function section
254 // Maximum v value: 12
255 //////////////////////////////////////////////
256 // The parameter to the Fjt function
257 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
258
259
260 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
261
262
263 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
264 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
265 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
266 for(n = 0; n <= 12; n++)
267 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
268
269 //////////////////////////////////////////////
270 // Primitive integrals: Vertical recurrance
271 //////////////////////////////////////////////
272
273 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
274 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
275 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
276 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
277 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
278 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
279 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
280 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
281 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
282 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
283 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
284 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
285 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
288 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
289 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
290 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
291
292
293
294 // Forming PRIM_INT__s_s_p_s[12 * 3];
295 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
296 {
297
298 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
299 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]);
300
301 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
302 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]);
303
304 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
305 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]);
306
307 }
308
309
310
311 // Forming PRIM_INT__s_s_d_s[11 * 6];
312 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
313 {
314
315 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
316 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]);
317 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]);
318
319 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
320 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]);
321
322 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
323 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]);
324
325 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
326 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]);
327 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]);
328
329 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
330 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]);
331
332 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
333 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]);
334 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]);
335
336 }
337
338
339
340 // Forming PRIM_INT__s_s_f_s[10 * 10];
341 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
342 {
343
344 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
345 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]);
346 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]);
347
348 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
349 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]);
350
351 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
352 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]);
353
354 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
355 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]);
356
357 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
358 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]);
359
360 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
361 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]);
362
363 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
364 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]);
365 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]);
366
367 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
368 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]);
369
370 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
371 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]);
372
373 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
374 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]);
375 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]);
376
377 }
378
379
380 VRR_K_s_s_g_s(
381 PRIM_INT__s_s_g_s,
382 PRIM_INT__s_s_f_s,
383 PRIM_INT__s_s_d_s,
384 Q_PA,
385 a_over_q,
386 aoq_PQ,
387 one_over_2q,
388 9);
389
390
391 VRR_K_s_s_h_s(
392 PRIM_INT__s_s_h_s,
393 PRIM_INT__s_s_g_s,
394 PRIM_INT__s_s_f_s,
395 Q_PA,
396 a_over_q,
397 aoq_PQ,
398 one_over_2q,
399 8);
400
401
402 ostei_general_vrr1_K(6, 7,
403 one_over_2q, a_over_q, aoq_PQ, Q_PA,
404 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
405
406
407 ostei_general_vrr_I(1, 0, 6, 0, 4,
408 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
409 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
410
411
412 ostei_general_vrr_I(1, 0, 5, 0, 4,
413 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
414 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
415
416
417 ostei_general_vrr_I(2, 0, 6, 0, 3,
418 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
419 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);
420
421
422 ostei_general_vrr1_K(7, 6,
423 one_over_2q, a_over_q, aoq_PQ, Q_PA,
424 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
425
426
427 ostei_general_vrr_I(1, 0, 7, 0, 4,
428 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
429 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
430
431
432 ostei_general_vrr_I(2, 0, 7, 0, 3,
433 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
434 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);
435
436
437 VRR_I_p_s_g_s(
438 PRIM_INT__p_s_g_s,
439 PRIM_INT__s_s_g_s,
440 PRIM_INT__s_s_f_s,
441 P_PA,
442 aop_PQ,
443 one_over_2pq,
444 4);
445
446
447 ostei_general_vrr_I(2, 0, 5, 0, 3,
448 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
449 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);
450
451
452 ostei_general_vrr_I(3, 0, 6, 0, 2,
453 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
454 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);
455
456
457 ostei_general_vrr1_K(8, 5,
458 one_over_2q, a_over_q, aoq_PQ, Q_PA,
459 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
460
461
462 ostei_general_vrr_I(1, 0, 8, 0, 4,
463 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
464 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
465
466
467 ostei_general_vrr_I(2, 0, 8, 0, 3,
468 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
469 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);
470
471
472 ostei_general_vrr_I(3, 0, 7, 0, 2,
473 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
474 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);
475
476
477 VRR_I_p_s_f_s(
478 PRIM_INT__p_s_f_s,
479 PRIM_INT__s_s_f_s,
480 PRIM_INT__s_s_d_s,
481 P_PA,
482 aop_PQ,
483 one_over_2pq,
484 4);
485
486
487 ostei_general_vrr_I(2, 0, 4, 0, 3,
488 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
489 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);
490
491
492 ostei_general_vrr_I(3, 0, 5, 0, 2,
493 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
494 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);
495
496
497 ostei_general_vrr_I(4, 0, 6, 0, 1,
498 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
499 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);
500
501
502 ostei_general_vrr_I(3, 0, 8, 0, 2,
503 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
504 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);
505
506
507 ostei_general_vrr_I(4, 0, 7, 0, 1,
508 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
509 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);
510
511
512 ostei_general_vrr_I(4, 0, 8, 0, 1,
513 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
514 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);
515
516
517
518
519 ////////////////////////////////////
520 // Accumulate contracted integrals
521 ////////////////////////////////////
522 if(lastoffset == 0)
523 {
524 contract_all(168, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
525 contract_all(216, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
526 contract_all(270, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
527 contract_all(280, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
528 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
529 contract_all(450, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
530 contract_all(420, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
531 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
532 contract_all(675, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
533 }
534 else
535 {
536 contract(168, shelloffsets, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
537 contract(216, shelloffsets, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
538 contract(270, shelloffsets, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
539 contract(280, shelloffsets, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
540 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
541 contract(450, shelloffsets, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
542 contract(420, shelloffsets, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
543 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
544 contract(675, shelloffsets, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
545 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
546 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
547 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
548 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
549 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
550 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
551 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
552 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
553 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
554 }
555
556 } // close loop over j
557 } // close loop over i
558
559 //Advance to the next batch
560 jstart = SIMINT_SIMD_ROUND(jend);
561
562 //////////////////////////////////////////////
563 // Contracted integrals: Horizontal recurrance
564 //////////////////////////////////////////////
565
566
567 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
568
569
570 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
571 {
572 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
573
574 // set up HRR pointers
575 double const * restrict HRR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
576 double const * restrict HRR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
577 double const * restrict HRR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
578 double const * restrict HRR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
579 double const * restrict HRR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
580 double const * restrict HRR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
581 double const * restrict HRR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
582 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
583 double const * restrict HRR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
584 double * restrict HRR_INT__d_d_i_d = INT__d_d_i_d + real_abcd * 6048;
585
586 // form INT__d_p_i_s
587 for(iket = 0; iket < 28; ++iket)
588 {
589 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] );
590
591 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] );
592
593 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] );
594
595 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] );
596
597 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] );
598
599 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] );
600
601 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] );
602
603 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] );
604
605 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] );
606
607 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] );
608
609 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] );
610
611 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] );
612
613 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] );
614
615 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] );
616
617 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] );
618
619 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] );
620
621 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] );
622
623 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] );
624
625 }
626
627
628 // form INT__d_p_k_s
629 for(iket = 0; iket < 36; ++iket)
630 {
631 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] );
632
633 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] );
634
635 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] );
636
637 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] );
638
639 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] );
640
641 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] );
642
643 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] );
644
645 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] );
646
647 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] );
648
649 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] );
650
651 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] );
652
653 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] );
654
655 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] );
656
657 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] );
658
659 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] );
660
661 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] );
662
663 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] );
664
665 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] );
666
667 }
668
669
670 // form INT__d_p_l_s
671 for(iket = 0; iket < 45; ++iket)
672 {
673 HRR_INT__d_p_l_s[0 * 45 + iket] = HRR_INT__f_s_l_s[0 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[0 * 45 + iket] );
674
675 HRR_INT__d_p_l_s[1 * 45 + iket] = HRR_INT__f_s_l_s[1 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[0 * 45 + iket] );
676
677 HRR_INT__d_p_l_s[2 * 45 + iket] = HRR_INT__f_s_l_s[2 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[0 * 45 + iket] );
678
679 HRR_INT__d_p_l_s[3 * 45 + iket] = HRR_INT__f_s_l_s[1 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[1 * 45 + iket] );
680
681 HRR_INT__d_p_l_s[4 * 45 + iket] = HRR_INT__f_s_l_s[3 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[1 * 45 + iket] );
682
683 HRR_INT__d_p_l_s[5 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[1 * 45 + iket] );
684
685 HRR_INT__d_p_l_s[6 * 45 + iket] = HRR_INT__f_s_l_s[2 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[2 * 45 + iket] );
686
687 HRR_INT__d_p_l_s[7 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[2 * 45 + iket] );
688
689 HRR_INT__d_p_l_s[8 * 45 + iket] = HRR_INT__f_s_l_s[5 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[2 * 45 + iket] );
690
691 HRR_INT__d_p_l_s[9 * 45 + iket] = HRR_INT__f_s_l_s[3 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[3 * 45 + iket] );
692
693 HRR_INT__d_p_l_s[10 * 45 + iket] = HRR_INT__f_s_l_s[6 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[3 * 45 + iket] );
694
695 HRR_INT__d_p_l_s[11 * 45 + iket] = HRR_INT__f_s_l_s[7 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[3 * 45 + iket] );
696
697 HRR_INT__d_p_l_s[12 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[4 * 45 + iket] );
698
699 HRR_INT__d_p_l_s[13 * 45 + iket] = HRR_INT__f_s_l_s[7 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[4 * 45 + iket] );
700
701 HRR_INT__d_p_l_s[14 * 45 + iket] = HRR_INT__f_s_l_s[8 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[4 * 45 + iket] );
702
703 HRR_INT__d_p_l_s[15 * 45 + iket] = HRR_INT__f_s_l_s[5 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[5 * 45 + iket] );
704
705 HRR_INT__d_p_l_s[16 * 45 + iket] = HRR_INT__f_s_l_s[8 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[5 * 45 + iket] );
706
707 HRR_INT__d_p_l_s[17 * 45 + iket] = HRR_INT__f_s_l_s[9 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[5 * 45 + iket] );
708
709 }
710
711
712 // form INT__f_p_i_s
713 HRR_J_f_p(
714 HRR_INT__f_p_i_s,
715 HRR_INT__f_s_i_s,
716 HRR_INT__g_s_i_s,
717 hAB, 28);
718
719 // form INT__f_p_k_s
720 HRR_J_f_p(
721 HRR_INT__f_p_k_s,
722 HRR_INT__f_s_k_s,
723 HRR_INT__g_s_k_s,
724 hAB, 36);
725
726 // form INT__f_p_l_s
727 HRR_J_f_p(
728 HRR_INT__f_p_l_s,
729 HRR_INT__f_s_l_s,
730 HRR_INT__g_s_l_s,
731 hAB, 45);
732
733 // form INT__d_d_i_s
734 HRR_J_d_d(
735 HRR_INT__d_d_i_s,
736 HRR_INT__d_p_i_s,
737 HRR_INT__f_p_i_s,
738 hAB, 28);
739
740 // form INT__d_d_k_s
741 HRR_J_d_d(
742 HRR_INT__d_d_k_s,
743 HRR_INT__d_p_k_s,
744 HRR_INT__f_p_k_s,
745 hAB, 36);
746
747 // form INT__d_d_l_s
748 HRR_J_d_d(
749 HRR_INT__d_d_l_s,
750 HRR_INT__d_p_l_s,
751 HRR_INT__f_p_l_s,
752 hAB, 45);
753
754 // form INT__d_d_i_p
755 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);
756
757 // form INT__d_d_k_p
758 ostei_general_hrr_L(2, 2, 7, 1, hCD, HRR_INT__d_d_l_s, HRR_INT__d_d_k_s, HRR_INT__d_d_k_p);
759
760 // form INT__d_d_i_d
761 ostei_general_hrr_L(2, 2, 6, 2, hCD, HRR_INT__d_d_k_p, HRR_INT__d_d_i_p, HRR_INT__d_d_i_d);
762
763
764 } // close HRR loop
765
766
767 } // close loop cdbatch
768
769 istart = iend;
770 } // close loop over ab
771
772 return P.nshell12_clip * Q.nshell12_clip;
773 }
774
ostei_d_d_d_i(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_i)775 int ostei_d_d_d_i(struct simint_multi_shellpair const P,
776 struct simint_multi_shellpair const Q,
777 double screen_tol,
778 double * const restrict work,
779 double * const restrict INT__d_d_d_i)
780 {
781 double Q_AB[3*Q.nshell12];
782 struct simint_multi_shellpair Q_tmp = Q;
783 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
784 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
785 Q_tmp.AB_x = Q_AB;
786 Q_tmp.AB_y = Q_AB + Q.nshell12;
787 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
788
789 for(int i = 0; i < Q.nshell12; i++)
790 {
791 Q_tmp.AB_x[i] = -Q.AB_x[i];
792 Q_tmp.AB_y[i] = -Q.AB_y[i];
793 Q_tmp.AB_z[i] = -Q.AB_z[i];
794 }
795
796 int ret = ostei_d_d_i_d(P, Q_tmp, screen_tol, work, INT__d_d_d_i);
797 double buffer[6048] SIMINT_ALIGN_ARRAY_DBL;
798
799 for(int q = 0; q < ret; q++)
800 {
801 int idx = 0;
802 for(int a = 0; a < 6; ++a)
803 for(int b = 0; b < 6; ++b)
804 for(int c = 0; c < 6; ++c)
805 for(int d = 0; d < 28; ++d)
806 buffer[idx++] = INT__d_d_d_i[q*6048+a*1008+b*168+d*6+c];
807
808 memcpy(INT__d_d_d_i+q*6048, buffer, 6048*sizeof(double));
809 }
810
811 return ret;
812 }
813
814