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_p_i_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_p_i_f)8 int ostei_f_p_i_f(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_p_i_f)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__f_p_i_f);
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__f_s_i_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__f_s_k_s = work + (SIMINT_NSHELL_SIMD * 280);
31 double * const INT__f_s_l_s = work + (SIMINT_NSHELL_SIMD * 640);
32 double * const INT__f_s_m_s = work + (SIMINT_NSHELL_SIMD * 1090);
33 double * const INT__g_s_i_s = work + (SIMINT_NSHELL_SIMD * 1640);
34 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 2060);
35 double * const INT__g_s_l_s = work + (SIMINT_NSHELL_SIMD * 2600);
36 double * const INT__g_s_m_s = work + (SIMINT_NSHELL_SIMD * 3275);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*4100);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 14;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 53;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 125;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 235;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 385;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 574;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 798;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 1050;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_m_s = primwork + 1320;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1595;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1715;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1895;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2147;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2483;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 2915;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_m_s = primwork + 3455;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 4115;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 4385;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 4763;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 5267;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 5915;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_m_s = primwork + 6725;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 7715;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 8135;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 8695;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 9415;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_m_s = primwork + 10315;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 11415;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 11835;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_l_s = primwork + 12375;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_m_s = primwork + 13050;
70 double * const hrrwork = (double *)(primwork + 13875);
71 double * const HRR_INT__f_p_i_s = hrrwork + 0;
72 double * const HRR_INT__f_p_i_p = hrrwork + 840;
73 double * const HRR_INT__f_p_i_d = hrrwork + 3360;
74 double * const HRR_INT__f_p_k_s = hrrwork + 8400;
75 double * const HRR_INT__f_p_k_p = hrrwork + 9480;
76 double * const HRR_INT__f_p_k_d = hrrwork + 12720;
77 double * const HRR_INT__f_p_l_s = hrrwork + 19200;
78 double * const HRR_INT__f_p_l_p = hrrwork + 20550;
79 double * const HRR_INT__f_p_m_s = hrrwork + 24600;
80
81
82 // Create constants
83 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
84 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
85 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
86 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
87 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
88 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
89 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
90 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
91 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
92 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
93
94
95 ////////////////////////////////////////
96 // Loop over shells and primitives
97 ////////////////////////////////////////
98
99 real_abcd = 0;
100 istart = 0;
101 for(ab = 0; ab < P.nshell12_clip; ++ab)
102 {
103 const int iend = istart + P.nprim12[ab];
104
105 cd = 0;
106 jstart = 0;
107
108 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
109 {
110 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
111 int jend = jstart;
112 for(i = 0; i < nshellbatch; i++)
113 jend += Q.nprim12[cd+i];
114
115 // Clear the beginning of the workspace (where we are accumulating integrals)
116 memset(work, 0, SIMINT_NSHELL_SIMD * 4100 * sizeof(double));
117 abcd = 0;
118
119
120 for(i = istart; i < iend; ++i)
121 {
122 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
123
124 if(check_screen)
125 {
126 // Skip this whole thing if always insignificant
127 if((P.screen[i] * Q.screen_max) < screen_tol)
128 continue;
129 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
130 }
131
132 icd = 0;
133 iprimcd = 0;
134 nprim_icd = Q.nprim12[cd];
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__f_s_m_s = INT__f_s_m_s + abcd * 550;
139 double * restrict PRIM_PTR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
140 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
141 double * restrict PRIM_PTR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
142 double * restrict PRIM_PTR_INT__g_s_m_s = INT__g_s_m_s + abcd * 825;
143
144
145
146 // Load these one per loop over i
147 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
148 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
149 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
150
151 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
152
153 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
154 {
155 // calculate the shell offsets
156 // these are the offset from the shell pointed to by cd
157 // for each element
158 int shelloffsets[SIMINT_SIMD_LEN] = {0};
159 int lastoffset = 0;
160 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
161
162 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
163 {
164 // Handle if the first element of the vector is a new shell
165 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
166 {
167 nprim_icd += Q.nprim12[cd + (++icd)];
168 PRIM_PTR_INT__f_s_i_s += 280;
169 PRIM_PTR_INT__f_s_k_s += 360;
170 PRIM_PTR_INT__f_s_l_s += 450;
171 PRIM_PTR_INT__f_s_m_s += 550;
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 PRIM_PTR_INT__g_s_m_s += 825;
176 }
177 iprimcd++;
178 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
179 {
180 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
181 {
182 shelloffsets[n] = shelloffsets[n-1] + 1;
183 lastoffset++;
184 nprim_icd += Q.nprim12[cd + (++icd)];
185 }
186 else
187 shelloffsets[n] = shelloffsets[n-1];
188 iprimcd++;
189 }
190 }
191 else
192 iprimcd += SIMINT_SIMD_LEN;
193
194 // Do we have to compute this vector (or has it been screened out)?
195 // (not_screened != 0 means we have to do this vector)
196 if(check_screen)
197 {
198 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
199 if(vmax < screen_tol)
200 {
201 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
202 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
203 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
204 PRIM_PTR_INT__f_s_m_s += lastoffset*550;
205 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
206 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
207 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
208 PRIM_PTR_INT__g_s_m_s += lastoffset*825;
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: 13
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, 13);
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 <= 13; 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_8_over_2q = SIMINT_MUL(const_8, one_over_2q);
284 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
285 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
288 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
289 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
290 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
291 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
292 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
293
294
295
296 // Forming PRIM_INT__s_s_p_s[13 * 3];
297 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
298 {
299
300 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
301 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]);
302
303 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
304 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]);
305
306 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
307 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]);
308
309 }
310
311
312
313 // Forming PRIM_INT__s_s_d_s[12 * 6];
314 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
315 {
316
317 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
318 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]);
319 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]);
320
321 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
322 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]);
323
324 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
325 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]);
326
327 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
328 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]);
329 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]);
330
331 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
332 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]);
333
334 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
335 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]);
336 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]);
337
338 }
339
340
341
342 // Forming PRIM_INT__s_s_f_s[11 * 10];
343 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
344 {
345
346 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
347 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]);
348 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]);
349
350 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
351 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]);
352
353 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
354 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]);
355
356 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
357 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]);
358
359 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
360 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]);
361
362 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
363 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]);
364
365 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
366 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]);
367 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]);
368
369 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
370 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]);
371
372 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
373 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]);
374
375 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
376 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]);
377 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]);
378
379 }
380
381
382 VRR_K_s_s_g_s(
383 PRIM_INT__s_s_g_s,
384 PRIM_INT__s_s_f_s,
385 PRIM_INT__s_s_d_s,
386 Q_PA,
387 a_over_q,
388 aoq_PQ,
389 one_over_2q,
390 10);
391
392
393 VRR_K_s_s_h_s(
394 PRIM_INT__s_s_h_s,
395 PRIM_INT__s_s_g_s,
396 PRIM_INT__s_s_f_s,
397 Q_PA,
398 a_over_q,
399 aoq_PQ,
400 one_over_2q,
401 9);
402
403
404 ostei_general_vrr1_K(6, 8,
405 one_over_2q, a_over_q, aoq_PQ, Q_PA,
406 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
407
408
409 ostei_general_vrr_I(1, 0, 6, 0, 4,
410 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
411 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
412
413
414 ostei_general_vrr_I(1, 0, 5, 0, 4,
415 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
416 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
417
418
419 ostei_general_vrr_I(2, 0, 6, 0, 3,
420 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
421 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);
422
423
424 VRR_I_p_s_g_s(
425 PRIM_INT__p_s_g_s,
426 PRIM_INT__s_s_g_s,
427 PRIM_INT__s_s_f_s,
428 P_PA,
429 aop_PQ,
430 one_over_2pq,
431 4);
432
433
434 ostei_general_vrr_I(2, 0, 5, 0, 3,
435 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
436 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);
437
438
439 ostei_general_vrr_I(3, 0, 6, 0, 2,
440 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
441 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);
442
443
444 ostei_general_vrr1_K(7, 7,
445 one_over_2q, a_over_q, aoq_PQ, Q_PA,
446 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
447
448
449 ostei_general_vrr_I(1, 0, 7, 0, 4,
450 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
451 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
452
453
454 ostei_general_vrr_I(2, 0, 7, 0, 3,
455 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
456 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);
457
458
459 ostei_general_vrr_I(3, 0, 7, 0, 2,
460 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
461 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);
462
463
464 VRR_I_p_s_f_s(
465 PRIM_INT__p_s_f_s,
466 PRIM_INT__s_s_f_s,
467 PRIM_INT__s_s_d_s,
468 P_PA,
469 aop_PQ,
470 one_over_2pq,
471 4);
472
473
474 ostei_general_vrr_I(2, 0, 4, 0, 3,
475 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
476 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);
477
478
479 ostei_general_vrr_I(3, 0, 5, 0, 2,
480 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
481 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);
482
483
484 ostei_general_vrr_I(4, 0, 6, 0, 1,
485 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
486 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);
487
488
489 ostei_general_vrr1_K(8, 6,
490 one_over_2q, a_over_q, aoq_PQ, Q_PA,
491 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
492
493
494 ostei_general_vrr_I(1, 0, 8, 0, 4,
495 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
496 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
497
498
499 ostei_general_vrr_I(2, 0, 8, 0, 3,
500 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
501 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);
502
503
504 ostei_general_vrr_I(3, 0, 8, 0, 2,
505 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
506 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);
507
508
509 ostei_general_vrr_I(4, 0, 7, 0, 1,
510 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
511 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);
512
513
514 ostei_general_vrr1_K(9, 5,
515 one_over_2q, a_over_q, aoq_PQ, Q_PA,
516 PRIM_INT__s_s_l_s, PRIM_INT__s_s_k_s, PRIM_INT__s_s_m_s);
517
518
519 ostei_general_vrr_I(1, 0, 9, 0, 4,
520 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
521 PRIM_INT__s_s_m_s, NULL, NULL, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_m_s);
522
523
524 ostei_general_vrr_I(2, 0, 9, 0, 3,
525 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
526 PRIM_INT__p_s_m_s, PRIM_INT__s_s_m_s, NULL, PRIM_INT__p_s_l_s, NULL, PRIM_INT__d_s_m_s);
527
528
529 ostei_general_vrr_I(3, 0, 9, 0, 2,
530 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
531 PRIM_INT__d_s_m_s, PRIM_INT__p_s_m_s, NULL, PRIM_INT__d_s_l_s, NULL, PRIM_INT__f_s_m_s);
532
533
534 ostei_general_vrr_I(4, 0, 8, 0, 1,
535 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
536 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);
537
538
539 ostei_general_vrr_I(4, 0, 9, 0, 1,
540 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
541 PRIM_INT__f_s_m_s, PRIM_INT__d_s_m_s, NULL, PRIM_INT__f_s_l_s, NULL, PRIM_INT__g_s_m_s);
542
543
544
545
546 ////////////////////////////////////
547 // Accumulate contracted integrals
548 ////////////////////////////////////
549 if(lastoffset == 0)
550 {
551 contract_all(280, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
552 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
553 contract_all(450, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
554 contract_all(550, PRIM_INT__f_s_m_s, PRIM_PTR_INT__f_s_m_s);
555 contract_all(420, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
556 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
557 contract_all(675, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
558 contract_all(825, PRIM_INT__g_s_m_s, PRIM_PTR_INT__g_s_m_s);
559 }
560 else
561 {
562 contract(280, shelloffsets, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
563 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
564 contract(450, shelloffsets, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
565 contract(550, shelloffsets, PRIM_INT__f_s_m_s, PRIM_PTR_INT__f_s_m_s);
566 contract(420, shelloffsets, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
567 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
568 contract(675, shelloffsets, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
569 contract(825, shelloffsets, PRIM_INT__g_s_m_s, PRIM_PTR_INT__g_s_m_s);
570 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
571 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
572 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
573 PRIM_PTR_INT__f_s_m_s += lastoffset*550;
574 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
575 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
576 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
577 PRIM_PTR_INT__g_s_m_s += lastoffset*825;
578 }
579
580 } // close loop over j
581 } // close loop over i
582
583 //Advance to the next batch
584 jstart = SIMINT_SIMD_ROUND(jend);
585
586 //////////////////////////////////////////////
587 // Contracted integrals: Horizontal recurrance
588 //////////////////////////////////////////////
589
590
591 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
592
593
594 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
595 {
596 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
597
598 // set up HRR pointers
599 double const * restrict HRR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
600 double const * restrict HRR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
601 double const * restrict HRR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
602 double const * restrict HRR_INT__f_s_m_s = INT__f_s_m_s + abcd * 550;
603 double const * restrict HRR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
604 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
605 double const * restrict HRR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
606 double const * restrict HRR_INT__g_s_m_s = INT__g_s_m_s + abcd * 825;
607 double * restrict HRR_INT__f_p_i_f = INT__f_p_i_f + real_abcd * 8400;
608
609 // form INT__f_p_i_s
610 HRR_J_f_p(
611 HRR_INT__f_p_i_s,
612 HRR_INT__f_s_i_s,
613 HRR_INT__g_s_i_s,
614 hAB, 28);
615
616 // form INT__f_p_k_s
617 HRR_J_f_p(
618 HRR_INT__f_p_k_s,
619 HRR_INT__f_s_k_s,
620 HRR_INT__g_s_k_s,
621 hAB, 36);
622
623 // form INT__f_p_l_s
624 HRR_J_f_p(
625 HRR_INT__f_p_l_s,
626 HRR_INT__f_s_l_s,
627 HRR_INT__g_s_l_s,
628 hAB, 45);
629
630 // form INT__f_p_m_s
631 HRR_J_f_p(
632 HRR_INT__f_p_m_s,
633 HRR_INT__f_s_m_s,
634 HRR_INT__g_s_m_s,
635 hAB, 55);
636
637 // form INT__f_p_i_p
638 ostei_general_hrr_L(3, 1, 6, 1, hCD, HRR_INT__f_p_k_s, HRR_INT__f_p_i_s, HRR_INT__f_p_i_p);
639
640 // form INT__f_p_k_p
641 ostei_general_hrr_L(3, 1, 7, 1, hCD, HRR_INT__f_p_l_s, HRR_INT__f_p_k_s, HRR_INT__f_p_k_p);
642
643 // form INT__f_p_l_p
644 ostei_general_hrr_L(3, 1, 8, 1, hCD, HRR_INT__f_p_m_s, HRR_INT__f_p_l_s, HRR_INT__f_p_l_p);
645
646 // form INT__f_p_i_d
647 ostei_general_hrr_L(3, 1, 6, 2, hCD, HRR_INT__f_p_k_p, HRR_INT__f_p_i_p, HRR_INT__f_p_i_d);
648
649 // form INT__f_p_k_d
650 ostei_general_hrr_L(3, 1, 7, 2, hCD, HRR_INT__f_p_l_p, HRR_INT__f_p_k_p, HRR_INT__f_p_k_d);
651
652 // form INT__f_p_i_f
653 ostei_general_hrr_L(3, 1, 6, 3, hCD, HRR_INT__f_p_k_d, HRR_INT__f_p_i_d, HRR_INT__f_p_i_f);
654
655
656 } // close HRR loop
657
658
659 } // close loop cdbatch
660
661 istart = iend;
662 } // close loop over ab
663
664 return P.nshell12_clip * Q.nshell12_clip;
665 }
666
ostei_p_f_i_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_f_i_f)667 int ostei_p_f_i_f(struct simint_multi_shellpair const P,
668 struct simint_multi_shellpair const Q,
669 double screen_tol,
670 double * const restrict work,
671 double * const restrict INT__p_f_i_f)
672 {
673 double P_AB[3*P.nshell12];
674 struct simint_multi_shellpair P_tmp = P;
675 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
676 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
677 P_tmp.AB_x = P_AB;
678 P_tmp.AB_y = P_AB + P.nshell12;
679 P_tmp.AB_z = P_AB + 2*P.nshell12;
680
681 for(int i = 0; i < P.nshell12; i++)
682 {
683 P_tmp.AB_x[i] = -P.AB_x[i];
684 P_tmp.AB_y[i] = -P.AB_y[i];
685 P_tmp.AB_z[i] = -P.AB_z[i];
686 }
687
688 int ret = ostei_f_p_i_f(P_tmp, Q, screen_tol, work, INT__p_f_i_f);
689 double buffer[8400] SIMINT_ALIGN_ARRAY_DBL;
690
691 for(int q = 0; q < ret; q++)
692 {
693 int idx = 0;
694 for(int a = 0; a < 3; ++a)
695 for(int b = 0; b < 10; ++b)
696 for(int c = 0; c < 28; ++c)
697 for(int d = 0; d < 10; ++d)
698 buffer[idx++] = INT__p_f_i_f[q*8400+b*840+a*280+c*10+d];
699
700 memcpy(INT__p_f_i_f+q*8400, buffer, 8400*sizeof(double));
701 }
702
703 return ret;
704 }
705
ostei_f_p_f_i(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_p_f_i)706 int ostei_f_p_f_i(struct simint_multi_shellpair const P,
707 struct simint_multi_shellpair const Q,
708 double screen_tol,
709 double * const restrict work,
710 double * const restrict INT__f_p_f_i)
711 {
712 double Q_AB[3*Q.nshell12];
713 struct simint_multi_shellpair Q_tmp = Q;
714 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
715 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
716 Q_tmp.AB_x = Q_AB;
717 Q_tmp.AB_y = Q_AB + Q.nshell12;
718 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
719
720 for(int i = 0; i < Q.nshell12; i++)
721 {
722 Q_tmp.AB_x[i] = -Q.AB_x[i];
723 Q_tmp.AB_y[i] = -Q.AB_y[i];
724 Q_tmp.AB_z[i] = -Q.AB_z[i];
725 }
726
727 int ret = ostei_f_p_i_f(P, Q_tmp, screen_tol, work, INT__f_p_f_i);
728 double buffer[8400] SIMINT_ALIGN_ARRAY_DBL;
729
730 for(int q = 0; q < ret; q++)
731 {
732 int idx = 0;
733 for(int a = 0; a < 10; ++a)
734 for(int b = 0; b < 3; ++b)
735 for(int c = 0; c < 10; ++c)
736 for(int d = 0; d < 28; ++d)
737 buffer[idx++] = INT__f_p_f_i[q*8400+a*840+b*280+d*10+c];
738
739 memcpy(INT__f_p_f_i+q*8400, buffer, 8400*sizeof(double));
740 }
741
742 return ret;
743 }
744
ostei_p_f_f_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_f_f_i)745 int ostei_p_f_f_i(struct simint_multi_shellpair const P,
746 struct simint_multi_shellpair const Q,
747 double screen_tol,
748 double * const restrict work,
749 double * const restrict INT__p_f_f_i)
750 {
751 double P_AB[3*P.nshell12];
752 struct simint_multi_shellpair P_tmp = P;
753 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
754 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
755 P_tmp.AB_x = P_AB;
756 P_tmp.AB_y = P_AB + P.nshell12;
757 P_tmp.AB_z = P_AB + 2*P.nshell12;
758
759 for(int i = 0; i < P.nshell12; i++)
760 {
761 P_tmp.AB_x[i] = -P.AB_x[i];
762 P_tmp.AB_y[i] = -P.AB_y[i];
763 P_tmp.AB_z[i] = -P.AB_z[i];
764 }
765
766 double Q_AB[3*Q.nshell12];
767 struct simint_multi_shellpair Q_tmp = Q;
768 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
769 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
770 Q_tmp.AB_x = Q_AB;
771 Q_tmp.AB_y = Q_AB + Q.nshell12;
772 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
773
774 for(int i = 0; i < Q.nshell12; i++)
775 {
776 Q_tmp.AB_x[i] = -Q.AB_x[i];
777 Q_tmp.AB_y[i] = -Q.AB_y[i];
778 Q_tmp.AB_z[i] = -Q.AB_z[i];
779 }
780
781 int ret = ostei_f_p_i_f(P_tmp, Q_tmp, screen_tol, work, INT__p_f_f_i);
782 double buffer[8400] SIMINT_ALIGN_ARRAY_DBL;
783
784 for(int q = 0; q < ret; q++)
785 {
786 int idx = 0;
787 for(int a = 0; a < 3; ++a)
788 for(int b = 0; b < 10; ++b)
789 for(int c = 0; c < 10; ++c)
790 for(int d = 0; d < 28; ++d)
791 buffer[idx++] = INT__p_f_f_i[q*8400+b*840+a*280+d*10+c];
792
793 memcpy(INT__p_f_f_i+q*8400, buffer, 8400*sizeof(double));
794 }
795
796 return ret;
797 }
798
799