1 #include "simint/boys/boys.h"
2 #include "simint/ostei/gen/ostei_generated.h"
3 #include "simint/vectorization/vectorization.h"
4 #include <math.h>
5 #include <string.h>
6
7
ostei_g_p_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__g_p_i_p)8 int ostei_g_p_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__g_p_i_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__g_p_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 iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__g_s_i_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 420);
31 double * const INT__h_s_i_s = work + (SIMINT_NSHELL_SIMD * 960);
32 double * const INT__h_s_k_s = work + (SIMINT_NSHELL_SIMD * 1548);
33 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2304);
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 13;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 49;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 115;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 215;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 350;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 518;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 714;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 930;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1020;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1170;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1395;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1710;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2130;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 2670;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 2910;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 3270;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 3774;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 4446;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 5310;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 5760;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 6390;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 7230;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 8310;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 8940;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 9780;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 10860;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 11448;
62 double * const hrrwork = (double *)(primwork + 12204);
63 double * const HRR_INT__g_p_i_s = hrrwork + 0;
64 double * const HRR_INT__g_p_k_s = hrrwork + 1260;
65
66
67 // Create constants
68 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
69 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
70 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
71 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
72 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
73 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
74 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
75 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
76
77
78 ////////////////////////////////////////
79 // Loop over shells and primitives
80 ////////////////////////////////////////
81
82 real_abcd = 0;
83 istart = 0;
84 for(ab = 0; ab < P.nshell12_clip; ++ab)
85 {
86 const int iend = istart + P.nprim12[ab];
87
88 cd = 0;
89 jstart = 0;
90
91 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
92 {
93 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
94 int jend = jstart;
95 for(i = 0; i < nshellbatch; i++)
96 jend += Q.nprim12[cd+i];
97
98 // Clear the beginning of the workspace (where we are accumulating integrals)
99 memset(work, 0, SIMINT_NSHELL_SIMD * 2304 * sizeof(double));
100 abcd = 0;
101
102
103 for(i = istart; i < iend; ++i)
104 {
105 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
106
107 if(check_screen)
108 {
109 // Skip this whole thing if always insignificant
110 if((P.screen[i] * Q.screen_max) < screen_tol)
111 continue;
112 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
113 }
114
115 icd = 0;
116 iprimcd = 0;
117 nprim_icd = Q.nprim12[cd];
118 double * restrict PRIM_PTR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
119 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
120 double * restrict PRIM_PTR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
121 double * restrict PRIM_PTR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
122
123
124
125 // Load these one per loop over i
126 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
127 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
128 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
129
130 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
131
132 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
133 {
134 // calculate the shell offsets
135 // these are the offset from the shell pointed to by cd
136 // for each element
137 int shelloffsets[SIMINT_SIMD_LEN] = {0};
138 int lastoffset = 0;
139 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
140
141 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
142 {
143 // Handle if the first element of the vector is a new shell
144 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
145 {
146 nprim_icd += Q.nprim12[cd + (++icd)];
147 PRIM_PTR_INT__g_s_i_s += 420;
148 PRIM_PTR_INT__g_s_k_s += 540;
149 PRIM_PTR_INT__h_s_i_s += 588;
150 PRIM_PTR_INT__h_s_k_s += 756;
151 }
152 iprimcd++;
153 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
154 {
155 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
156 {
157 shelloffsets[n] = shelloffsets[n-1] + 1;
158 lastoffset++;
159 nprim_icd += Q.nprim12[cd + (++icd)];
160 }
161 else
162 shelloffsets[n] = shelloffsets[n-1];
163 iprimcd++;
164 }
165 }
166 else
167 iprimcd += SIMINT_SIMD_LEN;
168
169 // Do we have to compute this vector (or has it been screened out)?
170 // (not_screened != 0 means we have to do this vector)
171 if(check_screen)
172 {
173 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
174 if(vmax < screen_tol)
175 {
176 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
177 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
178 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
179 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
180 continue;
181 }
182 }
183
184 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
185 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
186 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
187 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
188
189
190 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
191 SIMINT_DBLTYPE PQ[3];
192 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
193 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
194 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
195 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
196 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
197 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
198
199 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
200 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
201 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
202 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
203 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
204 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
205 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
206
207 // NOTE: Minus sign!
208 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
209 SIMINT_DBLTYPE aop_PQ[3];
210 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
211 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
212 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
213
214 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
215 SIMINT_DBLTYPE aoq_PQ[3];
216 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
217 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
218 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
219 // Put a minus sign here so we don't have to in RR routines
220 a_over_q = SIMINT_NEG(a_over_q);
221
222
223 //////////////////////////////////////////////
224 // Fjt function section
225 // Maximum v value: 12
226 //////////////////////////////////////////////
227 // The parameter to the Fjt function
228 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
229
230
231 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
232
233
234 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
235 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
236 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
237 for(n = 0; n <= 12; n++)
238 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
239
240 //////////////////////////////////////////////
241 // Primitive integrals: Vertical recurrance
242 //////////////////////////////////////////////
243
244 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
245 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
246 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
247 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
248 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
249 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
250 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
251 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
252 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
253 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
254 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
255 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
256 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
257 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
258 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
259 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
260 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
261
262
263
264 // Forming PRIM_INT__s_s_p_s[12 * 3];
265 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
266 {
267
268 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
269 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]);
270
271 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
272 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]);
273
274 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
275 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]);
276
277 }
278
279
280
281 // Forming PRIM_INT__s_s_d_s[11 * 6];
282 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
283 {
284
285 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
286 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]);
287 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]);
288
289 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
290 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]);
291
292 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
293 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]);
294
295 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
296 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]);
297 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]);
298
299 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
300 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]);
301
302 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
303 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]);
304 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]);
305
306 }
307
308
309
310 // Forming PRIM_INT__s_s_f_s[10 * 10];
311 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
312 {
313
314 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
315 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]);
316 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]);
317
318 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
319 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]);
320
321 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
322 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]);
323
324 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
325 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]);
326
327 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
328 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]);
329
330 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
331 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]);
332
333 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
334 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]);
335 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]);
336
337 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
338 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]);
339
340 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
341 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]);
342
343 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
344 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]);
345 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]);
346
347 }
348
349
350 VRR_K_s_s_g_s(
351 PRIM_INT__s_s_g_s,
352 PRIM_INT__s_s_f_s,
353 PRIM_INT__s_s_d_s,
354 Q_PA,
355 a_over_q,
356 aoq_PQ,
357 one_over_2q,
358 9);
359
360
361 VRR_K_s_s_h_s(
362 PRIM_INT__s_s_h_s,
363 PRIM_INT__s_s_g_s,
364 PRIM_INT__s_s_f_s,
365 Q_PA,
366 a_over_q,
367 aoq_PQ,
368 one_over_2q,
369 8);
370
371
372 ostei_general_vrr1_K(6, 7,
373 one_over_2q, a_over_q, aoq_PQ, Q_PA,
374 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
375
376
377 ostei_general_vrr_I(1, 0, 6, 0, 5,
378 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
379 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
380
381
382 ostei_general_vrr_I(1, 0, 5, 0, 5,
383 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
384 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
385
386
387 ostei_general_vrr_I(2, 0, 6, 0, 4,
388 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
389 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);
390
391
392 VRR_I_p_s_g_s(
393 PRIM_INT__p_s_g_s,
394 PRIM_INT__s_s_g_s,
395 PRIM_INT__s_s_f_s,
396 P_PA,
397 aop_PQ,
398 one_over_2pq,
399 5);
400
401
402 ostei_general_vrr_I(2, 0, 5, 0, 4,
403 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
404 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);
405
406
407 ostei_general_vrr_I(3, 0, 6, 0, 3,
408 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
409 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);
410
411
412 VRR_I_p_s_f_s(
413 PRIM_INT__p_s_f_s,
414 PRIM_INT__s_s_f_s,
415 PRIM_INT__s_s_d_s,
416 P_PA,
417 aop_PQ,
418 one_over_2pq,
419 5);
420
421
422 ostei_general_vrr_I(2, 0, 4, 0, 4,
423 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
424 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);
425
426
427 ostei_general_vrr_I(3, 0, 5, 0, 3,
428 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
429 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);
430
431
432 ostei_general_vrr_I(4, 0, 6, 0, 2,
433 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
434 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);
435
436
437 ostei_general_vrr1_K(7, 6,
438 one_over_2q, a_over_q, aoq_PQ, Q_PA,
439 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
440
441
442 ostei_general_vrr_I(1, 0, 7, 0, 5,
443 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
444 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
445
446
447 ostei_general_vrr_I(2, 0, 7, 0, 4,
448 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
449 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);
450
451
452 ostei_general_vrr_I(3, 0, 7, 0, 3,
453 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
454 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);
455
456
457 ostei_general_vrr_I(4, 0, 7, 0, 2,
458 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
459 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);
460
461
462
463 // Forming PRIM_INT__p_s_d_s[5 * 18];
464 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
465 {
466
467 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
468 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]);
469 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]);
470
471 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
472 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]);
473 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]);
474
475 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
476 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]);
477 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]);
478
479 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
480 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]);
481
482 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
483 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]);
484
485 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
486 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]);
487
488 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
489 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]);
490
491 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
492 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]);
493 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]);
494
495 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
496 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]);
497
498 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
499 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]);
500 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]);
501
502 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
503 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]);
504 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]);
505
506 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
507 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]);
508
509 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
510 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]);
511
512 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
513 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]);
514
515 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
516 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]);
517 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]);
518
519 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
520 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]);
521
522 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
523 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]);
524 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]);
525
526 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
527 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]);
528 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]);
529
530 }
531
532
533 VRR_I_d_s_f_s(
534 PRIM_INT__d_s_f_s,
535 PRIM_INT__p_s_f_s,
536 PRIM_INT__s_s_f_s,
537 PRIM_INT__p_s_d_s,
538 P_PA,
539 a_over_p,
540 aop_PQ,
541 one_over_2p,
542 one_over_2pq,
543 4);
544
545
546 ostei_general_vrr_I(3, 0, 4, 0, 3,
547 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
548 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);
549
550
551 ostei_general_vrr_I(4, 0, 5, 0, 2,
552 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
553 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);
554
555
556 ostei_general_vrr_I(5, 0, 6, 0, 1,
557 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
558 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);
559
560
561 ostei_general_vrr_I(5, 0, 7, 0, 1,
562 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
563 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);
564
565
566
567
568 ////////////////////////////////////
569 // Accumulate contracted integrals
570 ////////////////////////////////////
571 if(lastoffset == 0)
572 {
573 contract_all(420, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
574 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
575 contract_all(588, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
576 contract_all(756, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
577 }
578 else
579 {
580 contract(420, shelloffsets, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
581 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
582 contract(588, shelloffsets, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
583 contract(756, shelloffsets, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
584 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
585 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
586 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
587 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
588 }
589
590 } // close loop over j
591 } // close loop over i
592
593 //Advance to the next batch
594 jstart = SIMINT_SIMD_ROUND(jend);
595
596 //////////////////////////////////////////////
597 // Contracted integrals: Horizontal recurrance
598 //////////////////////////////////////////////
599
600
601 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
602
603
604 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
605 {
606 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
607
608 // set up HRR pointers
609 double const * restrict HRR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
610 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
611 double const * restrict HRR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
612 double const * restrict HRR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
613 double * restrict HRR_INT__g_p_i_p = INT__g_p_i_p + real_abcd * 3780;
614
615 // form INT__g_p_i_s
616 HRR_J_g_p(
617 HRR_INT__g_p_i_s,
618 HRR_INT__g_s_i_s,
619 HRR_INT__h_s_i_s,
620 hAB, 28);
621
622 // form INT__g_p_k_s
623 HRR_J_g_p(
624 HRR_INT__g_p_k_s,
625 HRR_INT__g_s_k_s,
626 HRR_INT__h_s_k_s,
627 hAB, 36);
628
629 // form INT__g_p_i_p
630 ostei_general_hrr_L(4, 1, 6, 1, hCD, HRR_INT__g_p_k_s, HRR_INT__g_p_i_s, HRR_INT__g_p_i_p);
631
632
633 } // close HRR loop
634
635
636 } // close loop cdbatch
637
638 istart = iend;
639 } // close loop over ab
640
641 return P.nshell12_clip * Q.nshell12_clip;
642 }
643
ostei_p_g_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__p_g_i_p)644 int ostei_p_g_i_p(struct simint_multi_shellpair const P,
645 struct simint_multi_shellpair const Q,
646 double screen_tol,
647 double * const restrict work,
648 double * const restrict INT__p_g_i_p)
649 {
650 double P_AB[3*P.nshell12];
651 struct simint_multi_shellpair P_tmp = P;
652 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
653 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
654 P_tmp.AB_x = P_AB;
655 P_tmp.AB_y = P_AB + P.nshell12;
656 P_tmp.AB_z = P_AB + 2*P.nshell12;
657
658 for(int i = 0; i < P.nshell12; i++)
659 {
660 P_tmp.AB_x[i] = -P.AB_x[i];
661 P_tmp.AB_y[i] = -P.AB_y[i];
662 P_tmp.AB_z[i] = -P.AB_z[i];
663 }
664
665 int ret = ostei_g_p_i_p(P_tmp, Q, screen_tol, work, INT__p_g_i_p);
666 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
667
668 for(int q = 0; q < ret; q++)
669 {
670 int idx = 0;
671 for(int a = 0; a < 3; ++a)
672 for(int b = 0; b < 15; ++b)
673 for(int c = 0; c < 28; ++c)
674 for(int d = 0; d < 3; ++d)
675 buffer[idx++] = INT__p_g_i_p[q*3780+b*252+a*84+c*3+d];
676
677 memcpy(INT__p_g_i_p+q*3780, buffer, 3780*sizeof(double));
678 }
679
680 return ret;
681 }
682
ostei_g_p_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__g_p_p_i)683 int ostei_g_p_p_i(struct simint_multi_shellpair const P,
684 struct simint_multi_shellpair const Q,
685 double screen_tol,
686 double * const restrict work,
687 double * const restrict INT__g_p_p_i)
688 {
689 double Q_AB[3*Q.nshell12];
690 struct simint_multi_shellpair Q_tmp = Q;
691 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
692 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
693 Q_tmp.AB_x = Q_AB;
694 Q_tmp.AB_y = Q_AB + Q.nshell12;
695 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
696
697 for(int i = 0; i < Q.nshell12; i++)
698 {
699 Q_tmp.AB_x[i] = -Q.AB_x[i];
700 Q_tmp.AB_y[i] = -Q.AB_y[i];
701 Q_tmp.AB_z[i] = -Q.AB_z[i];
702 }
703
704 int ret = ostei_g_p_i_p(P, Q_tmp, screen_tol, work, INT__g_p_p_i);
705 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
706
707 for(int q = 0; q < ret; q++)
708 {
709 int idx = 0;
710 for(int a = 0; a < 15; ++a)
711 for(int b = 0; b < 3; ++b)
712 for(int c = 0; c < 3; ++c)
713 for(int d = 0; d < 28; ++d)
714 buffer[idx++] = INT__g_p_p_i[q*3780+a*252+b*84+d*3+c];
715
716 memcpy(INT__g_p_p_i+q*3780, buffer, 3780*sizeof(double));
717 }
718
719 return ret;
720 }
721
ostei_p_g_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__p_g_p_i)722 int ostei_p_g_p_i(struct simint_multi_shellpair const P,
723 struct simint_multi_shellpair const Q,
724 double screen_tol,
725 double * const restrict work,
726 double * const restrict INT__p_g_p_i)
727 {
728 double P_AB[3*P.nshell12];
729 struct simint_multi_shellpair P_tmp = P;
730 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
731 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
732 P_tmp.AB_x = P_AB;
733 P_tmp.AB_y = P_AB + P.nshell12;
734 P_tmp.AB_z = P_AB + 2*P.nshell12;
735
736 for(int i = 0; i < P.nshell12; i++)
737 {
738 P_tmp.AB_x[i] = -P.AB_x[i];
739 P_tmp.AB_y[i] = -P.AB_y[i];
740 P_tmp.AB_z[i] = -P.AB_z[i];
741 }
742
743 double Q_AB[3*Q.nshell12];
744 struct simint_multi_shellpair Q_tmp = Q;
745 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
746 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
747 Q_tmp.AB_x = Q_AB;
748 Q_tmp.AB_y = Q_AB + Q.nshell12;
749 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
750
751 for(int i = 0; i < Q.nshell12; i++)
752 {
753 Q_tmp.AB_x[i] = -Q.AB_x[i];
754 Q_tmp.AB_y[i] = -Q.AB_y[i];
755 Q_tmp.AB_z[i] = -Q.AB_z[i];
756 }
757
758 int ret = ostei_g_p_i_p(P_tmp, Q_tmp, screen_tol, work, INT__p_g_p_i);
759 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
760
761 for(int q = 0; q < ret; q++)
762 {
763 int idx = 0;
764 for(int a = 0; a < 3; ++a)
765 for(int b = 0; b < 15; ++b)
766 for(int c = 0; c < 3; ++c)
767 for(int d = 0; d < 28; ++d)
768 buffer[idx++] = INT__p_g_p_i[q*3780+b*252+a*84+d*3+c];
769
770 memcpy(INT__p_g_p_i+q*3780, buffer, 3780*sizeof(double));
771 }
772
773 return ret;
774 }
775
776