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_f_d_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_f_d_p)8 int ostei_g_f_d_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_f_d_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__g_f_d_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_d_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__g_s_f_s = work + (SIMINT_NSHELL_SIMD * 90);
31 double * const INT__h_s_d_s = work + (SIMINT_NSHELL_SIMD * 240);
32 double * const INT__h_s_f_s = work + (SIMINT_NSHELL_SIMD * 366);
33 double * const INT__i_s_d_s = work + (SIMINT_NSHELL_SIMD * 576);
34 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 744);
35 double * const INT__k_s_d_s = work + (SIMINT_NSHELL_SIMD * 1024);
36 double * const INT__k_s_f_s = work + (SIMINT_NSHELL_SIMD * 1240);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*1600);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 11;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 41;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 95;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 149;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 229;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 319;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 439;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 544;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 679;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 859;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1009;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 1135;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 1324;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 1576;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 1786;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 1926;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 2178;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 2514;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 2794;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 2938;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 3262;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 3694;
61 double * const hrrwork = (double *)(primwork + 4054);
62 double * const HRR_INT__g_p_d_s = hrrwork + 0;
63 double * const HRR_INT__g_p_f_s = hrrwork + 270;
64 double * const HRR_INT__g_d_d_s = hrrwork + 720;
65 double * const HRR_INT__g_d_f_s = hrrwork + 1260;
66 double * const HRR_INT__g_f_d_s = hrrwork + 2160;
67 double * const HRR_INT__g_f_f_s = hrrwork + 3060;
68 double * const HRR_INT__h_p_d_s = hrrwork + 4560;
69 double * const HRR_INT__h_p_f_s = hrrwork + 4938;
70 double * const HRR_INT__h_d_d_s = hrrwork + 5568;
71 double * const HRR_INT__h_d_f_s = hrrwork + 6324;
72 double * const HRR_INT__i_p_d_s = hrrwork + 7584;
73 double * const HRR_INT__i_p_f_s = hrrwork + 8088;
74
75
76 // Create constants
77 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
78 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
79 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
80 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
81 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
82 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
83 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
84 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
85
86
87 ////////////////////////////////////////
88 // Loop over shells and primitives
89 ////////////////////////////////////////
90
91 real_abcd = 0;
92 istart = 0;
93 for(ab = 0; ab < P.nshell12_clip; ++ab)
94 {
95 const int iend = istart + P.nprim12[ab];
96
97 cd = 0;
98 jstart = 0;
99
100 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
101 {
102 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
103 int jend = jstart;
104 for(i = 0; i < nshellbatch; i++)
105 jend += Q.nprim12[cd+i];
106
107 // Clear the beginning of the workspace (where we are accumulating integrals)
108 memset(work, 0, SIMINT_NSHELL_SIMD * 1600 * sizeof(double));
109 abcd = 0;
110
111
112 for(i = istart; i < iend; ++i)
113 {
114 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
115
116 if(check_screen)
117 {
118 // Skip this whole thing if always insignificant
119 if((P.screen[i] * Q.screen_max) < screen_tol)
120 continue;
121 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
122 }
123
124 icd = 0;
125 iprimcd = 0;
126 nprim_icd = Q.nprim12[cd];
127 double * restrict PRIM_PTR_INT__g_s_d_s = INT__g_s_d_s + abcd * 90;
128 double * restrict PRIM_PTR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
129 double * restrict PRIM_PTR_INT__h_s_d_s = INT__h_s_d_s + abcd * 126;
130 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
131 double * restrict PRIM_PTR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
132 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
133 double * restrict PRIM_PTR_INT__k_s_d_s = INT__k_s_d_s + abcd * 216;
134 double * restrict PRIM_PTR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
135
136
137
138 // Load these one per loop over i
139 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
140 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
141 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
142
143 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
144
145 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
146 {
147 // calculate the shell offsets
148 // these are the offset from the shell pointed to by cd
149 // for each element
150 int shelloffsets[SIMINT_SIMD_LEN] = {0};
151 int lastoffset = 0;
152 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
153
154 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
155 {
156 // Handle if the first element of the vector is a new shell
157 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
158 {
159 nprim_icd += Q.nprim12[cd + (++icd)];
160 PRIM_PTR_INT__g_s_d_s += 90;
161 PRIM_PTR_INT__g_s_f_s += 150;
162 PRIM_PTR_INT__h_s_d_s += 126;
163 PRIM_PTR_INT__h_s_f_s += 210;
164 PRIM_PTR_INT__i_s_d_s += 168;
165 PRIM_PTR_INT__i_s_f_s += 280;
166 PRIM_PTR_INT__k_s_d_s += 216;
167 PRIM_PTR_INT__k_s_f_s += 360;
168 }
169 iprimcd++;
170 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
171 {
172 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
173 {
174 shelloffsets[n] = shelloffsets[n-1] + 1;
175 lastoffset++;
176 nprim_icd += Q.nprim12[cd + (++icd)];
177 }
178 else
179 shelloffsets[n] = shelloffsets[n-1];
180 iprimcd++;
181 }
182 }
183 else
184 iprimcd += SIMINT_SIMD_LEN;
185
186 // Do we have to compute this vector (or has it been screened out)?
187 // (not_screened != 0 means we have to do this vector)
188 if(check_screen)
189 {
190 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
191 if(vmax < screen_tol)
192 {
193 PRIM_PTR_INT__g_s_d_s += lastoffset*90;
194 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
195 PRIM_PTR_INT__h_s_d_s += lastoffset*126;
196 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
197 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
198 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
199 PRIM_PTR_INT__k_s_d_s += lastoffset*216;
200 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
201 continue;
202 }
203 }
204
205 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
206 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
207 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
208 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
209
210
211 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
212 SIMINT_DBLTYPE PQ[3];
213 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
214 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
215 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
216 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
217 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
218 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
219
220 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
221 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
222 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
223 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
224 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
225 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
226 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
227
228 // NOTE: Minus sign!
229 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
230 SIMINT_DBLTYPE aop_PQ[3];
231 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
232 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
233 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
234
235 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
236 SIMINT_DBLTYPE aoq_PQ[3];
237 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
238 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
239 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
240 // Put a minus sign here so we don't have to in RR routines
241 a_over_q = SIMINT_NEG(a_over_q);
242
243
244 //////////////////////////////////////////////
245 // Fjt function section
246 // Maximum v value: 10
247 //////////////////////////////////////////////
248 // The parameter to the Fjt function
249 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
250
251
252 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
253
254
255 boys_F_split(PRIM_INT__s_s_s_s, F_x, 10);
256 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
257 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
258 for(n = 0; n <= 10; n++)
259 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
260
261 //////////////////////////////////////////////
262 // Primitive integrals: Vertical recurrance
263 //////////////////////////////////////////////
264
265 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
266 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
267 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
268 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
269 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
270 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
271 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
272 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
273 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
274 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
275 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
276 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
277 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
278 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
279 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
280
281
282
283 // Forming PRIM_INT__p_s_s_s[10 * 3];
284 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
285 {
286
287 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
288 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]);
289
290 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
291 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]);
292
293 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
294 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]);
295
296 }
297
298
299
300 // Forming PRIM_INT__d_s_s_s[9 * 6];
301 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
302 {
303
304 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
305 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]);
306 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]);
307
308 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
309 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]);
310
311 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
312 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]);
313
314 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
315 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]);
316 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]);
317
318 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
319 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]);
320
321 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
322 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]);
323 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]);
324
325 }
326
327
328
329 // Forming PRIM_INT__f_s_s_s[8 * 10];
330 for(n = 0; n < 8; ++n) // loop over orders of auxiliary function
331 {
332
333 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
334 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]);
335 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]);
336
337 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
338 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]);
339
340 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
341 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]);
342
343 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
344 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]);
345
346 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
347 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]);
348
349 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
350 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]);
351
352 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
353 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]);
354 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]);
355
356 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
357 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]);
358
359 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
360 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]);
361
362 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
363 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]);
364 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]);
365
366 }
367
368
369 VRR_I_g_s_s_s(
370 PRIM_INT__g_s_s_s,
371 PRIM_INT__f_s_s_s,
372 PRIM_INT__d_s_s_s,
373 P_PA,
374 a_over_p,
375 aop_PQ,
376 one_over_2p,
377 7);
378
379
380 VRR_K_g_s_p_s(
381 PRIM_INT__g_s_p_s,
382 PRIM_INT__g_s_s_s,
383 PRIM_INT__f_s_s_s,
384 Q_PA,
385 aoq_PQ,
386 one_over_2pq,
387 3);
388
389
390 VRR_K_f_s_p_s(
391 PRIM_INT__f_s_p_s,
392 PRIM_INT__f_s_s_s,
393 PRIM_INT__d_s_s_s,
394 Q_PA,
395 aoq_PQ,
396 one_over_2pq,
397 3);
398
399
400 ostei_general_vrr_K(4, 0, 2, 0, 2,
401 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
402 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);
403
404
405
406 // Forming PRIM_INT__d_s_p_s[3 * 18];
407 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
408 {
409
410 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
411 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]);
412 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]);
413
414 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
415 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]);
416
417 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
418 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]);
419
420 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
421 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]);
422 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]);
423
424 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
425 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]);
426 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]);
427
428 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
429 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]);
430
431 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
432 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]);
433 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]);
434
435 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
436 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]);
437
438 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
439 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]);
440 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]);
441
442 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
443 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]);
444
445 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
446 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]);
447 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]);
448
449 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
450 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]);
451
452 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
453 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]);
454
455 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
456 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]);
457 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]);
458
459 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
460 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]);
461 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]);
462
463 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
464 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]);
465
466 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
467 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]);
468
469 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
470 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]);
471 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]);
472
473 }
474
475
476 VRR_K_f_s_d_s(
477 PRIM_INT__f_s_d_s,
478 PRIM_INT__f_s_p_s,
479 PRIM_INT__f_s_s_s,
480 PRIM_INT__d_s_p_s,
481 Q_PA,
482 a_over_q,
483 aoq_PQ,
484 one_over_2pq,
485 one_over_2q,
486 2);
487
488
489 ostei_general_vrr_K(4, 0, 3, 0, 1,
490 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
491 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);
492
493
494 VRR_I_h_s_s_s(
495 PRIM_INT__h_s_s_s,
496 PRIM_INT__g_s_s_s,
497 PRIM_INT__f_s_s_s,
498 P_PA,
499 a_over_p,
500 aop_PQ,
501 one_over_2p,
502 6);
503
504
505 ostei_general_vrr_K(5, 0, 1, 0, 3,
506 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
507 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
508
509
510 ostei_general_vrr_K(5, 0, 2, 0, 2,
511 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
512 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);
513
514
515 ostei_general_vrr_K(5, 0, 3, 0, 1,
516 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
517 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);
518
519
520 ostei_general_vrr1_I(6, 5,
521 one_over_2p, a_over_p, aop_PQ, P_PA,
522 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
523
524
525 ostei_general_vrr_K(6, 0, 1, 0, 3,
526 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
527 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
528
529
530 ostei_general_vrr_K(6, 0, 2, 0, 2,
531 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
532 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);
533
534
535 ostei_general_vrr_K(6, 0, 3, 0, 1,
536 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
537 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);
538
539
540 ostei_general_vrr1_I(7, 4,
541 one_over_2p, a_over_p, aop_PQ, P_PA,
542 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
543
544
545 ostei_general_vrr_K(7, 0, 1, 0, 3,
546 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
547 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
548
549
550 ostei_general_vrr_K(7, 0, 2, 0, 2,
551 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
552 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);
553
554
555 ostei_general_vrr_K(7, 0, 3, 0, 1,
556 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
557 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);
558
559
560
561
562 ////////////////////////////////////
563 // Accumulate contracted integrals
564 ////////////////////////////////////
565 if(lastoffset == 0)
566 {
567 contract_all(90, PRIM_INT__g_s_d_s, PRIM_PTR_INT__g_s_d_s);
568 contract_all(150, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
569 contract_all(126, PRIM_INT__h_s_d_s, PRIM_PTR_INT__h_s_d_s);
570 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
571 contract_all(168, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
572 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
573 contract_all(216, PRIM_INT__k_s_d_s, PRIM_PTR_INT__k_s_d_s);
574 contract_all(360, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
575 }
576 else
577 {
578 contract(90, shelloffsets, PRIM_INT__g_s_d_s, PRIM_PTR_INT__g_s_d_s);
579 contract(150, shelloffsets, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
580 contract(126, shelloffsets, PRIM_INT__h_s_d_s, PRIM_PTR_INT__h_s_d_s);
581 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
582 contract(168, shelloffsets, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
583 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
584 contract(216, shelloffsets, PRIM_INT__k_s_d_s, PRIM_PTR_INT__k_s_d_s);
585 contract(360, shelloffsets, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
586 PRIM_PTR_INT__g_s_d_s += lastoffset*90;
587 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
588 PRIM_PTR_INT__h_s_d_s += lastoffset*126;
589 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
590 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
591 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
592 PRIM_PTR_INT__k_s_d_s += lastoffset*216;
593 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
594 }
595
596 } // close loop over j
597 } // close loop over i
598
599 //Advance to the next batch
600 jstart = SIMINT_SIMD_ROUND(jend);
601
602 //////////////////////////////////////////////
603 // Contracted integrals: Horizontal recurrance
604 //////////////////////////////////////////////
605
606
607 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
608
609
610 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
611 {
612 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
613
614 // set up HRR pointers
615 double const * restrict HRR_INT__g_s_d_s = INT__g_s_d_s + abcd * 90;
616 double const * restrict HRR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
617 double const * restrict HRR_INT__h_s_d_s = INT__h_s_d_s + abcd * 126;
618 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
619 double const * restrict HRR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
620 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
621 double const * restrict HRR_INT__k_s_d_s = INT__k_s_d_s + abcd * 216;
622 double const * restrict HRR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
623 double * restrict HRR_INT__g_f_d_p = INT__g_f_d_p + real_abcd * 2700;
624
625 // form INT__g_p_d_s
626 HRR_J_g_p(
627 HRR_INT__g_p_d_s,
628 HRR_INT__g_s_d_s,
629 HRR_INT__h_s_d_s,
630 hAB, 6);
631
632 // form INT__g_p_f_s
633 HRR_J_g_p(
634 HRR_INT__g_p_f_s,
635 HRR_INT__g_s_f_s,
636 HRR_INT__h_s_f_s,
637 hAB, 10);
638
639 // form INT__h_p_d_s
640 ostei_general_hrr_J(5, 1, 2, 0, hAB, HRR_INT__i_s_d_s, HRR_INT__h_s_d_s, HRR_INT__h_p_d_s);
641
642 // form INT__h_p_f_s
643 ostei_general_hrr_J(5, 1, 3, 0, hAB, HRR_INT__i_s_f_s, HRR_INT__h_s_f_s, HRR_INT__h_p_f_s);
644
645 // form INT__i_p_d_s
646 ostei_general_hrr_J(6, 1, 2, 0, hAB, HRR_INT__k_s_d_s, HRR_INT__i_s_d_s, HRR_INT__i_p_d_s);
647
648 // form INT__i_p_f_s
649 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);
650
651 // form INT__g_d_d_s
652 ostei_general_hrr_J(4, 2, 2, 0, hAB, HRR_INT__h_p_d_s, HRR_INT__g_p_d_s, HRR_INT__g_d_d_s);
653
654 // form INT__g_d_f_s
655 ostei_general_hrr_J(4, 2, 3, 0, hAB, HRR_INT__h_p_f_s, HRR_INT__g_p_f_s, HRR_INT__g_d_f_s);
656
657 // form INT__h_d_d_s
658 ostei_general_hrr_J(5, 2, 2, 0, hAB, HRR_INT__i_p_d_s, HRR_INT__h_p_d_s, HRR_INT__h_d_d_s);
659
660 // form INT__h_d_f_s
661 ostei_general_hrr_J(5, 2, 3, 0, hAB, HRR_INT__i_p_f_s, HRR_INT__h_p_f_s, HRR_INT__h_d_f_s);
662
663 // form INT__g_f_d_s
664 ostei_general_hrr_J(4, 3, 2, 0, hAB, HRR_INT__h_d_d_s, HRR_INT__g_d_d_s, HRR_INT__g_f_d_s);
665
666 // form INT__g_f_f_s
667 ostei_general_hrr_J(4, 3, 3, 0, hAB, HRR_INT__h_d_f_s, HRR_INT__g_d_f_s, HRR_INT__g_f_f_s);
668
669 // form INT__g_f_d_p
670 for(ibra = 0; ibra < 150; ++ibra)
671 {
672 HRR_INT__g_f_d_p[ibra * 18 + 0] = HRR_INT__g_f_f_s[ibra * 10 + 0] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 0] );
673
674 HRR_INT__g_f_d_p[ibra * 18 + 1] = HRR_INT__g_f_f_s[ibra * 10 + 1] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 0] );
675
676 HRR_INT__g_f_d_p[ibra * 18 + 2] = HRR_INT__g_f_f_s[ibra * 10 + 2] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 0] );
677
678 HRR_INT__g_f_d_p[ibra * 18 + 3] = HRR_INT__g_f_f_s[ibra * 10 + 1] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 1] );
679
680 HRR_INT__g_f_d_p[ibra * 18 + 4] = HRR_INT__g_f_f_s[ibra * 10 + 3] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 1] );
681
682 HRR_INT__g_f_d_p[ibra * 18 + 5] = HRR_INT__g_f_f_s[ibra * 10 + 4] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 1] );
683
684 HRR_INT__g_f_d_p[ibra * 18 + 6] = HRR_INT__g_f_f_s[ibra * 10 + 2] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 2] );
685
686 HRR_INT__g_f_d_p[ibra * 18 + 7] = HRR_INT__g_f_f_s[ibra * 10 + 4] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 2] );
687
688 HRR_INT__g_f_d_p[ibra * 18 + 8] = HRR_INT__g_f_f_s[ibra * 10 + 5] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 2] );
689
690 HRR_INT__g_f_d_p[ibra * 18 + 9] = HRR_INT__g_f_f_s[ibra * 10 + 3] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 3] );
691
692 HRR_INT__g_f_d_p[ibra * 18 + 10] = HRR_INT__g_f_f_s[ibra * 10 + 6] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 3] );
693
694 HRR_INT__g_f_d_p[ibra * 18 + 11] = HRR_INT__g_f_f_s[ibra * 10 + 7] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 3] );
695
696 HRR_INT__g_f_d_p[ibra * 18 + 12] = HRR_INT__g_f_f_s[ibra * 10 + 4] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 4] );
697
698 HRR_INT__g_f_d_p[ibra * 18 + 13] = HRR_INT__g_f_f_s[ibra * 10 + 7] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 4] );
699
700 HRR_INT__g_f_d_p[ibra * 18 + 14] = HRR_INT__g_f_f_s[ibra * 10 + 8] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 4] );
701
702 HRR_INT__g_f_d_p[ibra * 18 + 15] = HRR_INT__g_f_f_s[ibra * 10 + 5] + ( hCD[0] * HRR_INT__g_f_d_s[ibra * 6 + 5] );
703
704 HRR_INT__g_f_d_p[ibra * 18 + 16] = HRR_INT__g_f_f_s[ibra * 10 + 8] + ( hCD[1] * HRR_INT__g_f_d_s[ibra * 6 + 5] );
705
706 HRR_INT__g_f_d_p[ibra * 18 + 17] = HRR_INT__g_f_f_s[ibra * 10 + 9] + ( hCD[2] * HRR_INT__g_f_d_s[ibra * 6 + 5] );
707
708 }
709
710
711 } // close HRR loop
712
713
714 } // close loop cdbatch
715
716 istart = iend;
717 } // close loop over ab
718
719 return P.nshell12_clip * Q.nshell12_clip;
720 }
721
ostei_f_g_d_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_g_d_p)722 int ostei_f_g_d_p(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__f_g_d_p)
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 int ret = ostei_g_f_d_p(P_tmp, Q, screen_tol, work, INT__f_g_d_p);
744 double buffer[2700] SIMINT_ALIGN_ARRAY_DBL;
745
746 for(int q = 0; q < ret; q++)
747 {
748 int idx = 0;
749 for(int a = 0; a < 10; ++a)
750 for(int b = 0; b < 15; ++b)
751 for(int c = 0; c < 6; ++c)
752 for(int d = 0; d < 3; ++d)
753 buffer[idx++] = INT__f_g_d_p[q*2700+b*180+a*18+c*3+d];
754
755 memcpy(INT__f_g_d_p+q*2700, buffer, 2700*sizeof(double));
756 }
757
758 return ret;
759 }
760
ostei_g_f_p_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_f_p_d)761 int ostei_g_f_p_d(struct simint_multi_shellpair const P,
762 struct simint_multi_shellpair const Q,
763 double screen_tol,
764 double * const restrict work,
765 double * const restrict INT__g_f_p_d)
766 {
767 double Q_AB[3*Q.nshell12];
768 struct simint_multi_shellpair Q_tmp = Q;
769 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
770 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
771 Q_tmp.AB_x = Q_AB;
772 Q_tmp.AB_y = Q_AB + Q.nshell12;
773 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
774
775 for(int i = 0; i < Q.nshell12; i++)
776 {
777 Q_tmp.AB_x[i] = -Q.AB_x[i];
778 Q_tmp.AB_y[i] = -Q.AB_y[i];
779 Q_tmp.AB_z[i] = -Q.AB_z[i];
780 }
781
782 int ret = ostei_g_f_d_p(P, Q_tmp, screen_tol, work, INT__g_f_p_d);
783 double buffer[2700] SIMINT_ALIGN_ARRAY_DBL;
784
785 for(int q = 0; q < ret; q++)
786 {
787 int idx = 0;
788 for(int a = 0; a < 15; ++a)
789 for(int b = 0; b < 10; ++b)
790 for(int c = 0; c < 3; ++c)
791 for(int d = 0; d < 6; ++d)
792 buffer[idx++] = INT__g_f_p_d[q*2700+a*180+b*18+d*3+c];
793
794 memcpy(INT__g_f_p_d+q*2700, buffer, 2700*sizeof(double));
795 }
796
797 return ret;
798 }
799
ostei_f_g_p_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_g_p_d)800 int ostei_f_g_p_d(struct simint_multi_shellpair const P,
801 struct simint_multi_shellpair const Q,
802 double screen_tol,
803 double * const restrict work,
804 double * const restrict INT__f_g_p_d)
805 {
806 double P_AB[3*P.nshell12];
807 struct simint_multi_shellpair P_tmp = P;
808 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
809 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
810 P_tmp.AB_x = P_AB;
811 P_tmp.AB_y = P_AB + P.nshell12;
812 P_tmp.AB_z = P_AB + 2*P.nshell12;
813
814 for(int i = 0; i < P.nshell12; i++)
815 {
816 P_tmp.AB_x[i] = -P.AB_x[i];
817 P_tmp.AB_y[i] = -P.AB_y[i];
818 P_tmp.AB_z[i] = -P.AB_z[i];
819 }
820
821 double Q_AB[3*Q.nshell12];
822 struct simint_multi_shellpair Q_tmp = Q;
823 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
824 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
825 Q_tmp.AB_x = Q_AB;
826 Q_tmp.AB_y = Q_AB + Q.nshell12;
827 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
828
829 for(int i = 0; i < Q.nshell12; i++)
830 {
831 Q_tmp.AB_x[i] = -Q.AB_x[i];
832 Q_tmp.AB_y[i] = -Q.AB_y[i];
833 Q_tmp.AB_z[i] = -Q.AB_z[i];
834 }
835
836 int ret = ostei_g_f_d_p(P_tmp, Q_tmp, screen_tol, work, INT__f_g_p_d);
837 double buffer[2700] SIMINT_ALIGN_ARRAY_DBL;
838
839 for(int q = 0; q < ret; q++)
840 {
841 int idx = 0;
842 for(int a = 0; a < 10; ++a)
843 for(int b = 0; b < 15; ++b)
844 for(int c = 0; c < 3; ++c)
845 for(int d = 0; d < 6; ++d)
846 buffer[idx++] = INT__f_g_p_d[q*2700+b*180+a*18+d*3+c];
847
848 memcpy(INT__f_g_p_d+q*2700, buffer, 2700*sizeof(double));
849 }
850
851 return ret;
852 }
853
854