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_i_p_f_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_p_f_d)8 int ostei_i_p_f_d(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__i_p_f_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_p_f_d);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__i_s_g_s = work + (SIMINT_NSHELL_SIMD * 280);
31 double * const INT__i_s_h_s = work + (SIMINT_NSHELL_SIMD * 700);
32 double * const INT__k_s_f_s = work + (SIMINT_NSHELL_SIMD * 1288);
33 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 1648);
34 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 2188);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2944);
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 13;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 49;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 115;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 205;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 305;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 455;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 695;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 830;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1055;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1415;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1865;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2033;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 2348;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 2852;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 3482;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 4112;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 4308;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 4728;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 5400;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 6240;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 7080;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 7668;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 7884;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 8424;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 9288;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 10368;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 11448;
64 double * const hrrwork = (double *)(primwork + 12204);
65 double * const HRR_INT__i_p_f_s = hrrwork + 0;
66 double * const HRR_INT__i_p_f_p = hrrwork + 840;
67 double * const HRR_INT__i_p_g_s = hrrwork + 3360;
68 double * const HRR_INT__i_p_g_p = hrrwork + 4620;
69 double * const HRR_INT__i_p_h_s = hrrwork + 8400;
70
71
72 // Create constants
73 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
74 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
75 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
76 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
77 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
78 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
79 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
80 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
81
82
83 ////////////////////////////////////////
84 // Loop over shells and primitives
85 ////////////////////////////////////////
86
87 real_abcd = 0;
88 istart = 0;
89 for(ab = 0; ab < P.nshell12_clip; ++ab)
90 {
91 const int iend = istart + P.nprim12[ab];
92
93 cd = 0;
94 jstart = 0;
95
96 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
97 {
98 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
99 int jend = jstart;
100 for(i = 0; i < nshellbatch; i++)
101 jend += Q.nprim12[cd+i];
102
103 // Clear the beginning of the workspace (where we are accumulating integrals)
104 memset(work, 0, SIMINT_NSHELL_SIMD * 2944 * sizeof(double));
105 abcd = 0;
106
107
108 for(i = istart; i < iend; ++i)
109 {
110 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
111
112 if(check_screen)
113 {
114 // Skip this whole thing if always insignificant
115 if((P.screen[i] * Q.screen_max) < screen_tol)
116 continue;
117 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
118 }
119
120 icd = 0;
121 iprimcd = 0;
122 nprim_icd = Q.nprim12[cd];
123 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
124 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
125 double * restrict PRIM_PTR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
126 double * restrict PRIM_PTR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
127 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
128 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
129
130
131
132 // Load these one per loop over i
133 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
134 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
135 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
136
137 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
138
139 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
140 {
141 // calculate the shell offsets
142 // these are the offset from the shell pointed to by cd
143 // for each element
144 int shelloffsets[SIMINT_SIMD_LEN] = {0};
145 int lastoffset = 0;
146 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
147
148 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
149 {
150 // Handle if the first element of the vector is a new shell
151 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
152 {
153 nprim_icd += Q.nprim12[cd + (++icd)];
154 PRIM_PTR_INT__i_s_f_s += 280;
155 PRIM_PTR_INT__i_s_g_s += 420;
156 PRIM_PTR_INT__i_s_h_s += 588;
157 PRIM_PTR_INT__k_s_f_s += 360;
158 PRIM_PTR_INT__k_s_g_s += 540;
159 PRIM_PTR_INT__k_s_h_s += 756;
160 }
161 iprimcd++;
162 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
163 {
164 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
165 {
166 shelloffsets[n] = shelloffsets[n-1] + 1;
167 lastoffset++;
168 nprim_icd += Q.nprim12[cd + (++icd)];
169 }
170 else
171 shelloffsets[n] = shelloffsets[n-1];
172 iprimcd++;
173 }
174 }
175 else
176 iprimcd += SIMINT_SIMD_LEN;
177
178 // Do we have to compute this vector (or has it been screened out)?
179 // (not_screened != 0 means we have to do this vector)
180 if(check_screen)
181 {
182 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
183 if(vmax < screen_tol)
184 {
185 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
186 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
187 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
188 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
189 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
190 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
191 continue;
192 }
193 }
194
195 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
196 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
197 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
198 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
199
200
201 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
202 SIMINT_DBLTYPE PQ[3];
203 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
204 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
205 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
206 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
207 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
208 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
209
210 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
211 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
212 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
213 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
214 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
215 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
216 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
217
218 // NOTE: Minus sign!
219 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
220 SIMINT_DBLTYPE aop_PQ[3];
221 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
222 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
223 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
224
225 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
226 SIMINT_DBLTYPE aoq_PQ[3];
227 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
228 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
229 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
230 // Put a minus sign here so we don't have to in RR routines
231 a_over_q = SIMINT_NEG(a_over_q);
232
233
234 //////////////////////////////////////////////
235 // Fjt function section
236 // Maximum v value: 12
237 //////////////////////////////////////////////
238 // The parameter to the Fjt function
239 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
240
241
242 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
243
244
245 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
246 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
247 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
248 for(n = 0; n <= 12; n++)
249 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
250
251 //////////////////////////////////////////////
252 // Primitive integrals: Vertical recurrance
253 //////////////////////////////////////////////
254
255 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
256 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
257 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
258 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
259 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
260 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
261 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
262 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
263 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
264 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
265 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
266 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
267 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
268 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
269 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
270 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
271 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
272
273
274
275 // Forming PRIM_INT__p_s_s_s[12 * 3];
276 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
277 {
278
279 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
280 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]);
281
282 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
283 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 1]);
284
285 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
286 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 2]);
287
288 }
289
290
291
292 // Forming PRIM_INT__d_s_s_s[11 * 6];
293 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
294 {
295
296 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
297 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 0]);
298 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 0]);
299
300 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
301 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 1]);
302
303 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
304 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 2]);
305
306 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
307 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 3]);
308 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 3]);
309
310 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
311 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 4]);
312
313 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
314 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_s_s[n * 6 + 5]);
315 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 5]);
316
317 }
318
319
320
321 // Forming PRIM_INT__f_s_s_s[10 * 10];
322 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
323 {
324
325 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
326 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 0]);
327 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__f_s_s_s[n * 10 + 0]);
328
329 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
330 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 1]);
331
332 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
333 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 2]);
334
335 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
336 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 3]);
337
338 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
339 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__f_s_s_s[n * 10 + 4]);
340
341 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
342 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 5]);
343
344 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
345 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 6]);
346 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__f_s_s_s[n * 10 + 6]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
349 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 7]);
350
351 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
352 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 8]);
353
354 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
355 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 9]);
356 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__f_s_s_s[n * 10 + 9]);
357
358 }
359
360
361 VRR_I_g_s_s_s(
362 PRIM_INT__g_s_s_s,
363 PRIM_INT__f_s_s_s,
364 PRIM_INT__d_s_s_s,
365 P_PA,
366 a_over_p,
367 aop_PQ,
368 one_over_2p,
369 9);
370
371
372 VRR_I_h_s_s_s(
373 PRIM_INT__h_s_s_s,
374 PRIM_INT__g_s_s_s,
375 PRIM_INT__f_s_s_s,
376 P_PA,
377 a_over_p,
378 aop_PQ,
379 one_over_2p,
380 8);
381
382
383 ostei_general_vrr1_I(6, 7,
384 one_over_2p, a_over_p, aop_PQ, P_PA,
385 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
386
387
388 ostei_general_vrr_K(6, 0, 1, 0, 5,
389 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
390 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
391
392
393 ostei_general_vrr_K(5, 0, 1, 0, 5,
394 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
395 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
396
397
398 ostei_general_vrr_K(6, 0, 2, 0, 4,
399 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
400 PRIM_INT__i_s_p_s, PRIM_INT__i_s_s_s, NULL, PRIM_INT__h_s_p_s, NULL, PRIM_INT__i_s_d_s);
401
402
403 VRR_K_g_s_p_s(
404 PRIM_INT__g_s_p_s,
405 PRIM_INT__g_s_s_s,
406 PRIM_INT__f_s_s_s,
407 Q_PA,
408 aoq_PQ,
409 one_over_2pq,
410 5);
411
412
413 ostei_general_vrr_K(5, 0, 2, 0, 4,
414 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
415 PRIM_INT__h_s_p_s, PRIM_INT__h_s_s_s, NULL, PRIM_INT__g_s_p_s, NULL, PRIM_INT__h_s_d_s);
416
417
418 ostei_general_vrr_K(6, 0, 3, 0, 3,
419 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
420 PRIM_INT__i_s_d_s, PRIM_INT__i_s_p_s, NULL, PRIM_INT__h_s_d_s, NULL, PRIM_INT__i_s_f_s);
421
422
423 VRR_K_f_s_p_s(
424 PRIM_INT__f_s_p_s,
425 PRIM_INT__f_s_s_s,
426 PRIM_INT__d_s_s_s,
427 Q_PA,
428 aoq_PQ,
429 one_over_2pq,
430 5);
431
432
433 ostei_general_vrr_K(4, 0, 2, 0, 4,
434 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
435 PRIM_INT__g_s_p_s, PRIM_INT__g_s_s_s, NULL, PRIM_INT__f_s_p_s, NULL, PRIM_INT__g_s_d_s);
436
437
438 ostei_general_vrr_K(5, 0, 3, 0, 3,
439 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
440 PRIM_INT__h_s_d_s, PRIM_INT__h_s_p_s, NULL, PRIM_INT__g_s_d_s, NULL, PRIM_INT__h_s_f_s);
441
442
443 ostei_general_vrr_K(6, 0, 4, 0, 2,
444 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
445 PRIM_INT__i_s_f_s, PRIM_INT__i_s_d_s, NULL, PRIM_INT__h_s_f_s, NULL, PRIM_INT__i_s_g_s);
446
447
448 ostei_general_vrr1_I(7, 6,
449 one_over_2p, a_over_p, aop_PQ, P_PA,
450 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
451
452
453 ostei_general_vrr_K(7, 0, 1, 0, 5,
454 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
455 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
456
457
458 ostei_general_vrr_K(7, 0, 2, 0, 4,
459 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
460 PRIM_INT__k_s_p_s, PRIM_INT__k_s_s_s, NULL, PRIM_INT__i_s_p_s, NULL, PRIM_INT__k_s_d_s);
461
462
463 ostei_general_vrr_K(7, 0, 3, 0, 3,
464 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
465 PRIM_INT__k_s_d_s, PRIM_INT__k_s_p_s, NULL, PRIM_INT__i_s_d_s, NULL, PRIM_INT__k_s_f_s);
466
467
468
469 // Forming PRIM_INT__d_s_p_s[5 * 18];
470 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
471 {
472
473 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
474 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
475 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
476
477 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
478 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 1]);
479
480 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
481 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 2]);
482
483 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
484 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
485 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
486
487 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
488 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 4]);
489 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 4]);
490
491 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
492 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 5]);
493
494 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
495 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
496 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
497
498 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
499 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 7]);
500
501 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
502 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 8]);
503 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 8]);
504
505 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
506 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 9]);
507
508 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
509 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 10]);
510 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 10]);
511
512 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
513 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 11]);
514
515 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
516 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 12]);
517
518 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
519 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 13]);
520 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 13]);
521
522 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
523 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 14]);
524 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 14]);
525
526 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
527 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 15]);
528
529 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
530 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 16]);
531
532 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
533 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 17]);
534 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 17]);
535
536 }
537
538
539 VRR_K_f_s_d_s(
540 PRIM_INT__f_s_d_s,
541 PRIM_INT__f_s_p_s,
542 PRIM_INT__f_s_s_s,
543 PRIM_INT__d_s_p_s,
544 Q_PA,
545 a_over_q,
546 aoq_PQ,
547 one_over_2pq,
548 one_over_2q,
549 4);
550
551
552 ostei_general_vrr_K(4, 0, 3, 0, 3,
553 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
554 PRIM_INT__g_s_d_s, PRIM_INT__g_s_p_s, NULL, PRIM_INT__f_s_d_s, NULL, PRIM_INT__g_s_f_s);
555
556
557 ostei_general_vrr_K(5, 0, 4, 0, 2,
558 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
559 PRIM_INT__h_s_f_s, PRIM_INT__h_s_d_s, NULL, PRIM_INT__g_s_f_s, NULL, PRIM_INT__h_s_g_s);
560
561
562 ostei_general_vrr_K(6, 0, 5, 0, 1,
563 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
564 PRIM_INT__i_s_g_s, PRIM_INT__i_s_f_s, NULL, PRIM_INT__h_s_g_s, NULL, PRIM_INT__i_s_h_s);
565
566
567 ostei_general_vrr_K(7, 0, 4, 0, 2,
568 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
569 PRIM_INT__k_s_f_s, PRIM_INT__k_s_d_s, NULL, PRIM_INT__i_s_f_s, NULL, PRIM_INT__k_s_g_s);
570
571
572 ostei_general_vrr_K(7, 0, 5, 0, 1,
573 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
574 PRIM_INT__k_s_g_s, PRIM_INT__k_s_f_s, NULL, PRIM_INT__i_s_g_s, NULL, PRIM_INT__k_s_h_s);
575
576
577
578
579 ////////////////////////////////////
580 // Accumulate contracted integrals
581 ////////////////////////////////////
582 if(lastoffset == 0)
583 {
584 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
585 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
586 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
587 contract_all(360, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
588 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
589 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
590 }
591 else
592 {
593 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
594 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
595 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
596 contract(360, shelloffsets, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
597 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
598 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
599 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
600 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
601 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
602 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
603 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
604 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
605 }
606
607 } // close loop over j
608 } // close loop over i
609
610 //Advance to the next batch
611 jstart = SIMINT_SIMD_ROUND(jend);
612
613 //////////////////////////////////////////////
614 // Contracted integrals: Horizontal recurrance
615 //////////////////////////////////////////////
616
617
618 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
619
620
621 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
622 {
623 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
624
625 // set up HRR pointers
626 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
627 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
628 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
629 double const * restrict HRR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
630 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
631 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
632 double * restrict HRR_INT__i_p_f_d = INT__i_p_f_d + real_abcd * 5040;
633
634 // form INT__i_p_f_s
635 ostei_general_hrr_J(6, 1, 3, 0, hAB, HRR_INT__k_s_f_s, HRR_INT__i_s_f_s, HRR_INT__i_p_f_s);
636
637 // form INT__i_p_g_s
638 ostei_general_hrr_J(6, 1, 4, 0, hAB, HRR_INT__k_s_g_s, HRR_INT__i_s_g_s, HRR_INT__i_p_g_s);
639
640 // form INT__i_p_h_s
641 ostei_general_hrr_J(6, 1, 5, 0, hAB, HRR_INT__k_s_h_s, HRR_INT__i_s_h_s, HRR_INT__i_p_h_s);
642
643 // form INT__i_p_f_p
644 HRR_L_f_p(
645 HRR_INT__i_p_f_p,
646 HRR_INT__i_p_f_s,
647 HRR_INT__i_p_g_s,
648 hCD, 84);
649
650 // form INT__i_p_g_p
651 HRR_L_g_p(
652 HRR_INT__i_p_g_p,
653 HRR_INT__i_p_g_s,
654 HRR_INT__i_p_h_s,
655 hCD, 84);
656
657 // form INT__i_p_f_d
658 HRR_L_f_d(
659 HRR_INT__i_p_f_d,
660 HRR_INT__i_p_f_p,
661 HRR_INT__i_p_g_p,
662 hCD, 84);
663
664
665 } // close HRR loop
666
667
668 } // close loop cdbatch
669
670 istart = iend;
671 } // close loop over ab
672
673 return P.nshell12_clip * Q.nshell12_clip;
674 }
675
ostei_p_i_f_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_i_f_d)676 int ostei_p_i_f_d(struct simint_multi_shellpair const P,
677 struct simint_multi_shellpair const Q,
678 double screen_tol,
679 double * const restrict work,
680 double * const restrict INT__p_i_f_d)
681 {
682 double P_AB[3*P.nshell12];
683 struct simint_multi_shellpair P_tmp = P;
684 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
685 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
686 P_tmp.AB_x = P_AB;
687 P_tmp.AB_y = P_AB + P.nshell12;
688 P_tmp.AB_z = P_AB + 2*P.nshell12;
689
690 for(int i = 0; i < P.nshell12; i++)
691 {
692 P_tmp.AB_x[i] = -P.AB_x[i];
693 P_tmp.AB_y[i] = -P.AB_y[i];
694 P_tmp.AB_z[i] = -P.AB_z[i];
695 }
696
697 int ret = ostei_i_p_f_d(P_tmp, Q, screen_tol, work, INT__p_i_f_d);
698 double buffer[5040] SIMINT_ALIGN_ARRAY_DBL;
699
700 for(int q = 0; q < ret; q++)
701 {
702 int idx = 0;
703 for(int a = 0; a < 3; ++a)
704 for(int b = 0; b < 28; ++b)
705 for(int c = 0; c < 10; ++c)
706 for(int d = 0; d < 6; ++d)
707 buffer[idx++] = INT__p_i_f_d[q*5040+b*180+a*60+c*6+d];
708
709 memcpy(INT__p_i_f_d+q*5040, buffer, 5040*sizeof(double));
710 }
711
712 return ret;
713 }
714
ostei_i_p_d_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_p_d_f)715 int ostei_i_p_d_f(struct simint_multi_shellpair const P,
716 struct simint_multi_shellpair const Q,
717 double screen_tol,
718 double * const restrict work,
719 double * const restrict INT__i_p_d_f)
720 {
721 double Q_AB[3*Q.nshell12];
722 struct simint_multi_shellpair Q_tmp = Q;
723 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
724 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
725 Q_tmp.AB_x = Q_AB;
726 Q_tmp.AB_y = Q_AB + Q.nshell12;
727 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
728
729 for(int i = 0; i < Q.nshell12; i++)
730 {
731 Q_tmp.AB_x[i] = -Q.AB_x[i];
732 Q_tmp.AB_y[i] = -Q.AB_y[i];
733 Q_tmp.AB_z[i] = -Q.AB_z[i];
734 }
735
736 int ret = ostei_i_p_f_d(P, Q_tmp, screen_tol, work, INT__i_p_d_f);
737 double buffer[5040] SIMINT_ALIGN_ARRAY_DBL;
738
739 for(int q = 0; q < ret; q++)
740 {
741 int idx = 0;
742 for(int a = 0; a < 28; ++a)
743 for(int b = 0; b < 3; ++b)
744 for(int c = 0; c < 6; ++c)
745 for(int d = 0; d < 10; ++d)
746 buffer[idx++] = INT__i_p_d_f[q*5040+a*180+b*60+d*6+c];
747
748 memcpy(INT__i_p_d_f+q*5040, buffer, 5040*sizeof(double));
749 }
750
751 return ret;
752 }
753
ostei_p_i_d_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_i_d_f)754 int ostei_p_i_d_f(struct simint_multi_shellpair const P,
755 struct simint_multi_shellpair const Q,
756 double screen_tol,
757 double * const restrict work,
758 double * const restrict INT__p_i_d_f)
759 {
760 double P_AB[3*P.nshell12];
761 struct simint_multi_shellpair P_tmp = P;
762 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
763 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
764 P_tmp.AB_x = P_AB;
765 P_tmp.AB_y = P_AB + P.nshell12;
766 P_tmp.AB_z = P_AB + 2*P.nshell12;
767
768 for(int i = 0; i < P.nshell12; i++)
769 {
770 P_tmp.AB_x[i] = -P.AB_x[i];
771 P_tmp.AB_y[i] = -P.AB_y[i];
772 P_tmp.AB_z[i] = -P.AB_z[i];
773 }
774
775 double Q_AB[3*Q.nshell12];
776 struct simint_multi_shellpair Q_tmp = Q;
777 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
778 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
779 Q_tmp.AB_x = Q_AB;
780 Q_tmp.AB_y = Q_AB + Q.nshell12;
781 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
782
783 for(int i = 0; i < Q.nshell12; i++)
784 {
785 Q_tmp.AB_x[i] = -Q.AB_x[i];
786 Q_tmp.AB_y[i] = -Q.AB_y[i];
787 Q_tmp.AB_z[i] = -Q.AB_z[i];
788 }
789
790 int ret = ostei_i_p_f_d(P_tmp, Q_tmp, screen_tol, work, INT__p_i_d_f);
791 double buffer[5040] SIMINT_ALIGN_ARRAY_DBL;
792
793 for(int q = 0; q < ret; q++)
794 {
795 int idx = 0;
796 for(int a = 0; a < 3; ++a)
797 for(int b = 0; b < 28; ++b)
798 for(int c = 0; c < 6; ++c)
799 for(int d = 0; d < 10; ++d)
800 buffer[idx++] = INT__p_i_d_f[q*5040+b*180+a*60+d*6+c];
801
802 memcpy(INT__p_i_d_f+q*5040, buffer, 5040*sizeof(double));
803 }
804
805 return ret;
806 }
807
808