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_f_s_k_s(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_s_k_s)8 int ostei_f_s_k_s(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__f_s_k_s)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__f_s_k_s);
17 memset(INT__f_s_k_s, 0, P.nshell12_clip * Q.nshell12_clip * 360 * sizeof(double));
18
19 int ab, cd, abcd;
20 int istart, jstart;
21 int iprimcd, nprim_icd, icd;
22 const int check_screen = (screen_tol > 0.0);
23 int i, j;
24 int n;
25 int not_screened;
26
27 // partition workspace
28 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*0);
29 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
30 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 11;
31 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 41;
32 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 95;
33 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 175;
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 280;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 406;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 546;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 690;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 879;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 1131;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 1455;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 1791;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 2223;
43 double * const hrrwork = (double *)(primwork + 2583);
44
45
46 // Create constants
47 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
48 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
49 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
50 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
51 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
52 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
53 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
54 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
55
56
57 ////////////////////////////////////////
58 // Loop over shells and primitives
59 ////////////////////////////////////////
60
61 abcd = 0;
62 istart = 0;
63 for(ab = 0; ab < P.nshell12_clip; ++ab)
64 {
65 const int iend = istart + P.nprim12[ab];
66
67 cd = 0;
68 jstart = 0;
69
70 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
71 {
72 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
73 int jend = jstart;
74 for(i = 0; i < nshellbatch; i++)
75 jend += Q.nprim12[cd+i];
76
77
78 for(i = istart; i < iend; ++i)
79 {
80 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
81
82 if(check_screen)
83 {
84 // Skip this whole thing if always insignificant
85 if((P.screen[i] * Q.screen_max) < screen_tol)
86 continue;
87 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
88 }
89
90 icd = 0;
91 iprimcd = 0;
92 nprim_icd = Q.nprim12[cd];
93 double * restrict PRIM_PTR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
94
95
96
97 // Load these one per loop over i
98 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
99 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
100 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
101
102 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
103
104 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
105 {
106 // calculate the shell offsets
107 // these are the offset from the shell pointed to by cd
108 // for each element
109 int shelloffsets[SIMINT_SIMD_LEN] = {0};
110 int lastoffset = 0;
111 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
112
113 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
114 {
115 // Handle if the first element of the vector is a new shell
116 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
117 {
118 nprim_icd += Q.nprim12[cd + (++icd)];
119 PRIM_PTR_INT__f_s_k_s += 360;
120 }
121 iprimcd++;
122 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
123 {
124 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
125 {
126 shelloffsets[n] = shelloffsets[n-1] + 1;
127 lastoffset++;
128 nprim_icd += Q.nprim12[cd + (++icd)];
129 }
130 else
131 shelloffsets[n] = shelloffsets[n-1];
132 iprimcd++;
133 }
134 }
135 else
136 iprimcd += SIMINT_SIMD_LEN;
137
138 // Do we have to compute this vector (or has it been screened out)?
139 // (not_screened != 0 means we have to do this vector)
140 if(check_screen)
141 {
142 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
143 if(vmax < screen_tol)
144 {
145 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
146 continue;
147 }
148 }
149
150 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
151 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
152 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
153 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
154
155
156 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
157 SIMINT_DBLTYPE PQ[3];
158 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
159 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
160 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
161 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
162 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
163 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
164
165 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
166 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
167 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
168 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
169 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
170 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
171 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
172
173 // NOTE: Minus sign!
174 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
175 SIMINT_DBLTYPE aop_PQ[3];
176 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
177 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
178 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
179
180 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
181 SIMINT_DBLTYPE aoq_PQ[3];
182 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
183 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
184 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
185 // Put a minus sign here so we don't have to in RR routines
186 a_over_q = SIMINT_NEG(a_over_q);
187
188
189 //////////////////////////////////////////////
190 // Fjt function section
191 // Maximum v value: 10
192 //////////////////////////////////////////////
193 // The parameter to the Fjt function
194 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
195
196
197 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
198
199
200 boys_F_split(PRIM_INT__s_s_s_s, F_x, 10);
201 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
202 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
203 for(n = 0; n <= 10; n++)
204 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
205
206 //////////////////////////////////////////////
207 // Primitive integrals: Vertical recurrance
208 //////////////////////////////////////////////
209
210 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
211 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
212 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
213 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
214 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
215 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
216 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
217 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
218 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
219 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
220 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
221 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
222 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
223 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
224 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
225
226
227
228 // Forming PRIM_INT__s_s_p_s[10 * 3];
229 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
230 {
231
232 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
233 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]);
234
235 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
236 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]);
237
238 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
239 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]);
240
241 }
242
243
244
245 // Forming PRIM_INT__s_s_d_s[9 * 6];
246 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
247 {
248
249 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
250 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]);
251 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]);
252
253 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
254 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]);
255 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]);
256
257 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
258 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]);
259 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]);
260
261 }
262
263
264
265 // Forming PRIM_INT__s_s_f_s[8 * 10];
266 for(n = 0; n < 8; ++n) // loop over orders of auxiliary function
267 {
268
269 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
270 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]);
271 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]);
272
273 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
274 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]);
275
276 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
277 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]);
278
279 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
280 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]);
281
282 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
283 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]);
284
285 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
286 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]);
287 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]);
288
289 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
290 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]);
291
292 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
293 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]);
294 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]);
295
296 }
297
298
299 VRR_K_s_s_g_s(
300 PRIM_INT__s_s_g_s,
301 PRIM_INT__s_s_f_s,
302 PRIM_INT__s_s_d_s,
303 Q_PA,
304 a_over_q,
305 aoq_PQ,
306 one_over_2q,
307 7);
308
309
310 VRR_K_s_s_h_s(
311 PRIM_INT__s_s_h_s,
312 PRIM_INT__s_s_g_s,
313 PRIM_INT__s_s_f_s,
314 Q_PA,
315 a_over_q,
316 aoq_PQ,
317 one_over_2q,
318 6);
319
320
321 ostei_general_vrr1_K(6, 5,
322 one_over_2q, a_over_q, aoq_PQ, Q_PA,
323 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
324
325
326 ostei_general_vrr1_K(7, 4,
327 one_over_2q, a_over_q, aoq_PQ, Q_PA,
328 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
329
330
331 ostei_general_vrr_I(1, 0, 7, 0, 3,
332 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
333 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
334
335
336 ostei_general_vrr_I(1, 0, 6, 0, 3,
337 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
338 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
339
340
341 ostei_general_vrr_I(2, 0, 7, 0, 2,
342 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
343 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);
344
345
346 ostei_general_vrr_I(1, 0, 5, 0, 3,
347 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
348 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
349
350
351 ostei_general_vrr_I(2, 0, 6, 0, 2,
352 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
353 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);
354
355
356 ostei_general_vrr_I(3, 0, 7, 0, 1,
357 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
358 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);
359
360
361
362
363 ////////////////////////////////////
364 // Accumulate contracted integrals
365 ////////////////////////////////////
366 if(lastoffset == 0)
367 {
368 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
369 }
370 else
371 {
372 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
373 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
374 }
375
376 } // close loop over j
377 } // close loop over i
378
379 //Advance to the next batch
380 jstart = SIMINT_SIMD_ROUND(jend);
381 abcd += nshellbatch;
382
383 } // close loop cdbatch
384
385 istart = iend;
386 } // close loop over ab
387
388 return P.nshell12_clip * Q.nshell12_clip;
389 }
390
ostei_s_f_k_s(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_f_k_s)391 int ostei_s_f_k_s(struct simint_multi_shellpair const P,
392 struct simint_multi_shellpair const Q,
393 double screen_tol,
394 double * const restrict work,
395 double * const restrict INT__s_f_k_s)
396 {
397 double P_AB[3*P.nshell12];
398 struct simint_multi_shellpair P_tmp = P;
399 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
400 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
401 P_tmp.AB_x = P_AB;
402 P_tmp.AB_y = P_AB + P.nshell12;
403 P_tmp.AB_z = P_AB + 2*P.nshell12;
404
405 for(int i = 0; i < P.nshell12; i++)
406 {
407 P_tmp.AB_x[i] = -P.AB_x[i];
408 P_tmp.AB_y[i] = -P.AB_y[i];
409 P_tmp.AB_z[i] = -P.AB_z[i];
410 }
411
412 int ret = ostei_f_s_k_s(P_tmp, Q, screen_tol, work, INT__s_f_k_s);
413
414 return ret;
415 }
416
ostei_f_s_s_k(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_s_s_k)417 int ostei_f_s_s_k(struct simint_multi_shellpair const P,
418 struct simint_multi_shellpair const Q,
419 double screen_tol,
420 double * const restrict work,
421 double * const restrict INT__f_s_s_k)
422 {
423 double Q_AB[3*Q.nshell12];
424 struct simint_multi_shellpair Q_tmp = Q;
425 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
426 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
427 Q_tmp.AB_x = Q_AB;
428 Q_tmp.AB_y = Q_AB + Q.nshell12;
429 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
430
431 for(int i = 0; i < Q.nshell12; i++)
432 {
433 Q_tmp.AB_x[i] = -Q.AB_x[i];
434 Q_tmp.AB_y[i] = -Q.AB_y[i];
435 Q_tmp.AB_z[i] = -Q.AB_z[i];
436 }
437
438 int ret = ostei_f_s_k_s(P, Q_tmp, screen_tol, work, INT__f_s_s_k);
439
440 return ret;
441 }
442
ostei_s_f_s_k(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_f_s_k)443 int ostei_s_f_s_k(struct simint_multi_shellpair const P,
444 struct simint_multi_shellpair const Q,
445 double screen_tol,
446 double * const restrict work,
447 double * const restrict INT__s_f_s_k)
448 {
449 double P_AB[3*P.nshell12];
450 struct simint_multi_shellpair P_tmp = P;
451 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
452 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
453 P_tmp.AB_x = P_AB;
454 P_tmp.AB_y = P_AB + P.nshell12;
455 P_tmp.AB_z = P_AB + 2*P.nshell12;
456
457 for(int i = 0; i < P.nshell12; i++)
458 {
459 P_tmp.AB_x[i] = -P.AB_x[i];
460 P_tmp.AB_y[i] = -P.AB_y[i];
461 P_tmp.AB_z[i] = -P.AB_z[i];
462 }
463
464 double Q_AB[3*Q.nshell12];
465 struct simint_multi_shellpair Q_tmp = Q;
466 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
467 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
468 Q_tmp.AB_x = Q_AB;
469 Q_tmp.AB_y = Q_AB + Q.nshell12;
470 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
471
472 for(int i = 0; i < Q.nshell12; i++)
473 {
474 Q_tmp.AB_x[i] = -Q.AB_x[i];
475 Q_tmp.AB_y[i] = -Q.AB_y[i];
476 Q_tmp.AB_z[i] = -Q.AB_z[i];
477 }
478
479 int ret = ostei_f_s_k_s(P_tmp, Q_tmp, screen_tol, work, INT__s_f_s_k);
480
481 return ret;
482 }
483
484