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_h_s_i_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_s_i_p)8 int ostei_h_s_i_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__h_s_i_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__h_s_i_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 ibra;
26
27 // partition workspace
28 double * const INT__h_s_i_s = work + (SIMINT_NSHELL_SIMD * 0);
29 double * const INT__h_s_k_s = work + (SIMINT_NSHELL_SIMD * 588);
30 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*1344);
31 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
32 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 13;
33 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 49;
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 115;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 215;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 350;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 518;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 714;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 930;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1020;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1170;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1395;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1710;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2130;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 2670;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 2910;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 3270;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 3774;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 4446;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 5310;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 5760;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 6390;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 7230;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 8310;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 8940;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 9780;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 10860;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 11448;
59 double * const hrrwork = (double *)(primwork + 12204);
60
61
62 // Create constants
63 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
64 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
65 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
66 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
67 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
68 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
69 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
70 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
71
72
73 ////////////////////////////////////////
74 // Loop over shells and primitives
75 ////////////////////////////////////////
76
77 real_abcd = 0;
78 istart = 0;
79 for(ab = 0; ab < P.nshell12_clip; ++ab)
80 {
81 const int iend = istart + P.nprim12[ab];
82
83 cd = 0;
84 jstart = 0;
85
86 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
87 {
88 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
89 int jend = jstart;
90 for(i = 0; i < nshellbatch; i++)
91 jend += Q.nprim12[cd+i];
92
93 // Clear the beginning of the workspace (where we are accumulating integrals)
94 memset(work, 0, SIMINT_NSHELL_SIMD * 1344 * sizeof(double));
95 abcd = 0;
96
97
98 for(i = istart; i < iend; ++i)
99 {
100 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
101
102 if(check_screen)
103 {
104 // Skip this whole thing if always insignificant
105 if((P.screen[i] * Q.screen_max) < screen_tol)
106 continue;
107 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
108 }
109
110 icd = 0;
111 iprimcd = 0;
112 nprim_icd = Q.nprim12[cd];
113 double * restrict PRIM_PTR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
114 double * restrict PRIM_PTR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
115
116
117
118 // Load these one per loop over i
119 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
120 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
121 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
122
123 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
124
125 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
126 {
127 // calculate the shell offsets
128 // these are the offset from the shell pointed to by cd
129 // for each element
130 int shelloffsets[SIMINT_SIMD_LEN] = {0};
131 int lastoffset = 0;
132 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
133
134 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
135 {
136 // Handle if the first element of the vector is a new shell
137 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
138 {
139 nprim_icd += Q.nprim12[cd + (++icd)];
140 PRIM_PTR_INT__h_s_i_s += 588;
141 PRIM_PTR_INT__h_s_k_s += 756;
142 }
143 iprimcd++;
144 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
145 {
146 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
147 {
148 shelloffsets[n] = shelloffsets[n-1] + 1;
149 lastoffset++;
150 nprim_icd += Q.nprim12[cd + (++icd)];
151 }
152 else
153 shelloffsets[n] = shelloffsets[n-1];
154 iprimcd++;
155 }
156 }
157 else
158 iprimcd += SIMINT_SIMD_LEN;
159
160 // Do we have to compute this vector (or has it been screened out)?
161 // (not_screened != 0 means we have to do this vector)
162 if(check_screen)
163 {
164 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
165 if(vmax < screen_tol)
166 {
167 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
168 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
169 continue;
170 }
171 }
172
173 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
174 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
175 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
176 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
177
178
179 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
180 SIMINT_DBLTYPE PQ[3];
181 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
182 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
183 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
184 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
185 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
186 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
187
188 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
189 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
190 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
191 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
192 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
193 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
194 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
195
196 // NOTE: Minus sign!
197 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
198 SIMINT_DBLTYPE aop_PQ[3];
199 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
200 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
201 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
202
203 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
204 SIMINT_DBLTYPE aoq_PQ[3];
205 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
206 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
207 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
208 // Put a minus sign here so we don't have to in RR routines
209 a_over_q = SIMINT_NEG(a_over_q);
210
211
212 //////////////////////////////////////////////
213 // Fjt function section
214 // Maximum v value: 12
215 //////////////////////////////////////////////
216 // The parameter to the Fjt function
217 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
218
219
220 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
221
222
223 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
224 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
225 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
226 for(n = 0; n <= 12; n++)
227 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
228
229 //////////////////////////////////////////////
230 // Primitive integrals: Vertical recurrance
231 //////////////////////////////////////////////
232
233 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
234 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
235 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
236 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
237 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
238 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
239 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
240 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
241 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
242 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
243 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
244 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
245 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
246 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
247 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
248 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
249 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
250
251
252
253 // Forming PRIM_INT__s_s_p_s[12 * 3];
254 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
255 {
256
257 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
258 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]);
259
260 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
261 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]);
262
263 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
264 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]);
265
266 }
267
268
269
270 // Forming PRIM_INT__s_s_d_s[11 * 6];
271 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
272 {
273
274 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
275 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]);
276 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]);
277
278 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
279 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]);
280
281 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
282 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]);
283
284 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
285 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]);
286 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]);
287
288 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
289 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]);
290
291 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
292 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]);
293 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]);
294
295 }
296
297
298
299 // Forming PRIM_INT__s_s_f_s[10 * 10];
300 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
301 {
302
303 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
304 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]);
305 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]);
306
307 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
308 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]);
309
310 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
311 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]);
312
313 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
314 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]);
315
316 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
317 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]);
318
319 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
320 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]);
321
322 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
323 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]);
324 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]);
325
326 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
327 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]);
328
329 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
330 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]);
331
332 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
333 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]);
334 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]);
335
336 }
337
338
339 VRR_K_s_s_g_s(
340 PRIM_INT__s_s_g_s,
341 PRIM_INT__s_s_f_s,
342 PRIM_INT__s_s_d_s,
343 Q_PA,
344 a_over_q,
345 aoq_PQ,
346 one_over_2q,
347 9);
348
349
350 VRR_K_s_s_h_s(
351 PRIM_INT__s_s_h_s,
352 PRIM_INT__s_s_g_s,
353 PRIM_INT__s_s_f_s,
354 Q_PA,
355 a_over_q,
356 aoq_PQ,
357 one_over_2q,
358 8);
359
360
361 ostei_general_vrr1_K(6, 7,
362 one_over_2q, a_over_q, aoq_PQ, Q_PA,
363 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
364
365
366 ostei_general_vrr_I(1, 0, 6, 0, 5,
367 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
368 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
369
370
371 ostei_general_vrr_I(1, 0, 5, 0, 5,
372 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
373 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
374
375
376 ostei_general_vrr_I(2, 0, 6, 0, 4,
377 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
378 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);
379
380
381 VRR_I_p_s_g_s(
382 PRIM_INT__p_s_g_s,
383 PRIM_INT__s_s_g_s,
384 PRIM_INT__s_s_f_s,
385 P_PA,
386 aop_PQ,
387 one_over_2pq,
388 5);
389
390
391 ostei_general_vrr_I(2, 0, 5, 0, 4,
392 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
393 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);
394
395
396 ostei_general_vrr_I(3, 0, 6, 0, 3,
397 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
398 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);
399
400
401 VRR_I_p_s_f_s(
402 PRIM_INT__p_s_f_s,
403 PRIM_INT__s_s_f_s,
404 PRIM_INT__s_s_d_s,
405 P_PA,
406 aop_PQ,
407 one_over_2pq,
408 5);
409
410
411 ostei_general_vrr_I(2, 0, 4, 0, 4,
412 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
413 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);
414
415
416 ostei_general_vrr_I(3, 0, 5, 0, 3,
417 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
418 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);
419
420
421 ostei_general_vrr_I(4, 0, 6, 0, 2,
422 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
423 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);
424
425
426
427 // Forming PRIM_INT__p_s_d_s[5 * 18];
428 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
429 {
430
431 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
432 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]);
433 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]);
434
435 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
436 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]);
437 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]);
438
439 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
440 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]);
441 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]);
442
443 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
444 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]);
445
446 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
447 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]);
448
449 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
450 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]);
451
452 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
453 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]);
454
455 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
456 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]);
457 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]);
458
459 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
460 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]);
461
462 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
463 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]);
464 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]);
465
466 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
467 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]);
468 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]);
469
470 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
471 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]);
472
473 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
474 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]);
475
476 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
477 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]);
478
479 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
480 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]);
481 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]);
482
483 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
484 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]);
485
486 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
487 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]);
488 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]);
489
490 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
491 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]);
492 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]);
493
494 }
495
496
497 VRR_I_d_s_f_s(
498 PRIM_INT__d_s_f_s,
499 PRIM_INT__p_s_f_s,
500 PRIM_INT__s_s_f_s,
501 PRIM_INT__p_s_d_s,
502 P_PA,
503 a_over_p,
504 aop_PQ,
505 one_over_2p,
506 one_over_2pq,
507 4);
508
509
510 ostei_general_vrr_I(3, 0, 4, 0, 3,
511 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
512 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);
513
514
515 ostei_general_vrr_I(4, 0, 5, 0, 2,
516 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
517 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);
518
519
520 ostei_general_vrr_I(5, 0, 6, 0, 1,
521 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
522 PRIM_INT__g_s_i_s, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_h_s, NULL, PRIM_INT__h_s_i_s);
523
524
525 ostei_general_vrr1_K(7, 6,
526 one_over_2q, a_over_q, aoq_PQ, Q_PA,
527 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
528
529
530 ostei_general_vrr_I(1, 0, 7, 0, 5,
531 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
532 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
533
534
535 ostei_general_vrr_I(2, 0, 7, 0, 4,
536 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
537 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);
538
539
540 ostei_general_vrr_I(3, 0, 7, 0, 3,
541 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
542 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);
543
544
545 ostei_general_vrr_I(4, 0, 7, 0, 2,
546 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
547 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);
548
549
550 ostei_general_vrr_I(5, 0, 7, 0, 1,
551 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
552 PRIM_INT__g_s_k_s, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_k_s);
553
554
555
556
557 ////////////////////////////////////
558 // Accumulate contracted integrals
559 ////////////////////////////////////
560 if(lastoffset == 0)
561 {
562 contract_all(588, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
563 contract_all(756, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
564 }
565 else
566 {
567 contract(588, shelloffsets, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
568 contract(756, shelloffsets, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
569 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
570 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
571 }
572
573 } // close loop over j
574 } // close loop over i
575
576 //Advance to the next batch
577 jstart = SIMINT_SIMD_ROUND(jend);
578
579 //////////////////////////////////////////////
580 // Contracted integrals: Horizontal recurrance
581 //////////////////////////////////////////////
582
583
584
585
586 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
587 {
588 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
589
590 // set up HRR pointers
591 double const * restrict HRR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
592 double const * restrict HRR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
593 double * restrict HRR_INT__h_s_i_p = INT__h_s_i_p + real_abcd * 1764;
594
595 // form INT__h_s_i_p
596 ostei_general_hrr_L(5, 0, 6, 1, hCD, HRR_INT__h_s_k_s, HRR_INT__h_s_i_s, HRR_INT__h_s_i_p);
597
598
599 } // close HRR loop
600
601
602 } // close loop cdbatch
603
604 istart = iend;
605 } // close loop over ab
606
607 return P.nshell12_clip * Q.nshell12_clip;
608 }
609
ostei_s_h_i_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_h_i_p)610 int ostei_s_h_i_p(struct simint_multi_shellpair const P,
611 struct simint_multi_shellpair const Q,
612 double screen_tol,
613 double * const restrict work,
614 double * const restrict INT__s_h_i_p)
615 {
616 double P_AB[3*P.nshell12];
617 struct simint_multi_shellpair P_tmp = P;
618 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
619 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
620 P_tmp.AB_x = P_AB;
621 P_tmp.AB_y = P_AB + P.nshell12;
622 P_tmp.AB_z = P_AB + 2*P.nshell12;
623
624 for(int i = 0; i < P.nshell12; i++)
625 {
626 P_tmp.AB_x[i] = -P.AB_x[i];
627 P_tmp.AB_y[i] = -P.AB_y[i];
628 P_tmp.AB_z[i] = -P.AB_z[i];
629 }
630
631 int ret = ostei_h_s_i_p(P_tmp, Q, screen_tol, work, INT__s_h_i_p);
632 double buffer[1764] SIMINT_ALIGN_ARRAY_DBL;
633
634 for(int q = 0; q < ret; q++)
635 {
636 int idx = 0;
637 for(int a = 0; a < 1; ++a)
638 for(int b = 0; b < 21; ++b)
639 for(int c = 0; c < 28; ++c)
640 for(int d = 0; d < 3; ++d)
641 buffer[idx++] = INT__s_h_i_p[q*1764+b*84+a*84+c*3+d];
642
643 memcpy(INT__s_h_i_p+q*1764, buffer, 1764*sizeof(double));
644 }
645
646 return ret;
647 }
648
ostei_h_s_p_i(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_s_p_i)649 int ostei_h_s_p_i(struct simint_multi_shellpair const P,
650 struct simint_multi_shellpair const Q,
651 double screen_tol,
652 double * const restrict work,
653 double * const restrict INT__h_s_p_i)
654 {
655 double Q_AB[3*Q.nshell12];
656 struct simint_multi_shellpair Q_tmp = Q;
657 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
658 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
659 Q_tmp.AB_x = Q_AB;
660 Q_tmp.AB_y = Q_AB + Q.nshell12;
661 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
662
663 for(int i = 0; i < Q.nshell12; i++)
664 {
665 Q_tmp.AB_x[i] = -Q.AB_x[i];
666 Q_tmp.AB_y[i] = -Q.AB_y[i];
667 Q_tmp.AB_z[i] = -Q.AB_z[i];
668 }
669
670 int ret = ostei_h_s_i_p(P, Q_tmp, screen_tol, work, INT__h_s_p_i);
671 double buffer[1764] SIMINT_ALIGN_ARRAY_DBL;
672
673 for(int q = 0; q < ret; q++)
674 {
675 int idx = 0;
676 for(int a = 0; a < 21; ++a)
677 for(int b = 0; b < 1; ++b)
678 for(int c = 0; c < 3; ++c)
679 for(int d = 0; d < 28; ++d)
680 buffer[idx++] = INT__h_s_p_i[q*1764+a*84+b*84+d*3+c];
681
682 memcpy(INT__h_s_p_i+q*1764, buffer, 1764*sizeof(double));
683 }
684
685 return ret;
686 }
687
ostei_s_h_p_i(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_h_p_i)688 int ostei_s_h_p_i(struct simint_multi_shellpair const P,
689 struct simint_multi_shellpair const Q,
690 double screen_tol,
691 double * const restrict work,
692 double * const restrict INT__s_h_p_i)
693 {
694 double P_AB[3*P.nshell12];
695 struct simint_multi_shellpair P_tmp = P;
696 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
697 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
698 P_tmp.AB_x = P_AB;
699 P_tmp.AB_y = P_AB + P.nshell12;
700 P_tmp.AB_z = P_AB + 2*P.nshell12;
701
702 for(int i = 0; i < P.nshell12; i++)
703 {
704 P_tmp.AB_x[i] = -P.AB_x[i];
705 P_tmp.AB_y[i] = -P.AB_y[i];
706 P_tmp.AB_z[i] = -P.AB_z[i];
707 }
708
709 double Q_AB[3*Q.nshell12];
710 struct simint_multi_shellpair Q_tmp = Q;
711 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
712 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
713 Q_tmp.AB_x = Q_AB;
714 Q_tmp.AB_y = Q_AB + Q.nshell12;
715 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
716
717 for(int i = 0; i < Q.nshell12; i++)
718 {
719 Q_tmp.AB_x[i] = -Q.AB_x[i];
720 Q_tmp.AB_y[i] = -Q.AB_y[i];
721 Q_tmp.AB_z[i] = -Q.AB_z[i];
722 }
723
724 int ret = ostei_h_s_i_p(P_tmp, Q_tmp, screen_tol, work, INT__s_h_p_i);
725 double buffer[1764] SIMINT_ALIGN_ARRAY_DBL;
726
727 for(int q = 0; q < ret; q++)
728 {
729 int idx = 0;
730 for(int a = 0; a < 1; ++a)
731 for(int b = 0; b < 21; ++b)
732 for(int c = 0; c < 3; ++c)
733 for(int d = 0; d < 28; ++d)
734 buffer[idx++] = INT__s_h_p_i[q*1764+b*84+a*84+d*3+c];
735
736 memcpy(INT__s_h_p_i+q*1764, buffer, 1764*sizeof(double));
737 }
738
739 return ret;
740 }
741
742