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_d_p_h_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_p_h_g)8 int ostei_d_p_h_g(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__d_p_h_g)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__d_p_h_g);
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__d_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__d_s_i_s = work + (SIMINT_NSHELL_SIMD * 126);
31 double * const INT__d_s_k_s = work + (SIMINT_NSHELL_SIMD * 294);
32 double * const INT__d_s_l_s = work + (SIMINT_NSHELL_SIMD * 510);
33 double * const INT__d_s_m_s = work + (SIMINT_NSHELL_SIMD * 780);
34 double * const INT__f_s_h_s = work + (SIMINT_NSHELL_SIMD * 1110);
35 double * const INT__f_s_i_s = work + (SIMINT_NSHELL_SIMD * 1320);
36 double * const INT__f_s_k_s = work + (SIMINT_NSHELL_SIMD * 1600);
37 double * const INT__f_s_l_s = work + (SIMINT_NSHELL_SIMD * 1960);
38 double * const INT__f_s_m_s = work + (SIMINT_NSHELL_SIMD * 2410);
39 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2960);
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 13;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 49;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 115;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 215;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 350;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 518;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 714;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 930;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_m_s = primwork + 1155;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1375;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1465;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1600;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1789;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2041;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 2365;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_m_s = primwork + 2770;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 3265;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 3445;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 3697;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 4033;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 4465;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_m_s = primwork + 5005;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 5665;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 5875;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 6155;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 6515;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_m_s = primwork + 6965;
68 double * const hrrwork = (double *)(primwork + 7515);
69 double * const HRR_INT__d_p_h_s = hrrwork + 0;
70 double * const HRR_INT__d_p_h_p = hrrwork + 378;
71 double * const HRR_INT__d_p_h_d = hrrwork + 1512;
72 double * const HRR_INT__d_p_h_f = hrrwork + 3780;
73 double * const HRR_INT__d_p_i_s = hrrwork + 7560;
74 double * const HRR_INT__d_p_i_p = hrrwork + 8064;
75 double * const HRR_INT__d_p_i_d = hrrwork + 9576;
76 double * const HRR_INT__d_p_i_f = hrrwork + 12600;
77 double * const HRR_INT__d_p_k_s = hrrwork + 17640;
78 double * const HRR_INT__d_p_k_p = hrrwork + 18288;
79 double * const HRR_INT__d_p_k_d = hrrwork + 20232;
80 double * const HRR_INT__d_p_l_s = hrrwork + 24120;
81 double * const HRR_INT__d_p_l_p = hrrwork + 24930;
82 double * const HRR_INT__d_p_m_s = hrrwork + 27360;
83
84
85 // Create constants
86 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
87 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
88 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
89 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
90 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
91 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
92 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
93 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
94 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
95 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
96
97
98 ////////////////////////////////////////
99 // Loop over shells and primitives
100 ////////////////////////////////////////
101
102 real_abcd = 0;
103 istart = 0;
104 for(ab = 0; ab < P.nshell12_clip; ++ab)
105 {
106 const int iend = istart + P.nprim12[ab];
107
108 cd = 0;
109 jstart = 0;
110
111 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
112 {
113 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
114 int jend = jstart;
115 for(i = 0; i < nshellbatch; i++)
116 jend += Q.nprim12[cd+i];
117
118 // Clear the beginning of the workspace (where we are accumulating integrals)
119 memset(work, 0, SIMINT_NSHELL_SIMD * 2960 * sizeof(double));
120 abcd = 0;
121
122
123 for(i = istart; i < iend; ++i)
124 {
125 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
126
127 if(check_screen)
128 {
129 // Skip this whole thing if always insignificant
130 if((P.screen[i] * Q.screen_max) < screen_tol)
131 continue;
132 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
133 }
134
135 icd = 0;
136 iprimcd = 0;
137 nprim_icd = Q.nprim12[cd];
138 double * restrict PRIM_PTR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
139 double * restrict PRIM_PTR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
140 double * restrict PRIM_PTR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
141 double * restrict PRIM_PTR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
142 double * restrict PRIM_PTR_INT__d_s_m_s = INT__d_s_m_s + abcd * 330;
143 double * restrict PRIM_PTR_INT__f_s_h_s = INT__f_s_h_s + abcd * 210;
144 double * restrict PRIM_PTR_INT__f_s_i_s = INT__f_s_i_s + abcd * 280;
145 double * restrict PRIM_PTR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
146 double * restrict PRIM_PTR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
147 double * restrict PRIM_PTR_INT__f_s_m_s = INT__f_s_m_s + abcd * 550;
148
149
150
151 // Load these one per loop over i
152 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
153 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
154 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
155
156 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
157
158 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
159 {
160 // calculate the shell offsets
161 // these are the offset from the shell pointed to by cd
162 // for each element
163 int shelloffsets[SIMINT_SIMD_LEN] = {0};
164 int lastoffset = 0;
165 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
166
167 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
168 {
169 // Handle if the first element of the vector is a new shell
170 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
171 {
172 nprim_icd += Q.nprim12[cd + (++icd)];
173 PRIM_PTR_INT__d_s_h_s += 126;
174 PRIM_PTR_INT__d_s_i_s += 168;
175 PRIM_PTR_INT__d_s_k_s += 216;
176 PRIM_PTR_INT__d_s_l_s += 270;
177 PRIM_PTR_INT__d_s_m_s += 330;
178 PRIM_PTR_INT__f_s_h_s += 210;
179 PRIM_PTR_INT__f_s_i_s += 280;
180 PRIM_PTR_INT__f_s_k_s += 360;
181 PRIM_PTR_INT__f_s_l_s += 450;
182 PRIM_PTR_INT__f_s_m_s += 550;
183 }
184 iprimcd++;
185 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
186 {
187 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
188 {
189 shelloffsets[n] = shelloffsets[n-1] + 1;
190 lastoffset++;
191 nprim_icd += Q.nprim12[cd + (++icd)];
192 }
193 else
194 shelloffsets[n] = shelloffsets[n-1];
195 iprimcd++;
196 }
197 }
198 else
199 iprimcd += SIMINT_SIMD_LEN;
200
201 // Do we have to compute this vector (or has it been screened out)?
202 // (not_screened != 0 means we have to do this vector)
203 if(check_screen)
204 {
205 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
206 if(vmax < screen_tol)
207 {
208 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
209 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
210 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
211 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
212 PRIM_PTR_INT__d_s_m_s += lastoffset*330;
213 PRIM_PTR_INT__f_s_h_s += lastoffset*210;
214 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
215 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
216 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
217 PRIM_PTR_INT__f_s_m_s += lastoffset*550;
218 continue;
219 }
220 }
221
222 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
223 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
224 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
225 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
226
227
228 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
229 SIMINT_DBLTYPE PQ[3];
230 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
231 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
232 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
233 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
234 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
235 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
236
237 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
238 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
239 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
240 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
241 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
242 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
243 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
244
245 // NOTE: Minus sign!
246 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
247 SIMINT_DBLTYPE aop_PQ[3];
248 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
249 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
250 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
251
252 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
253 SIMINT_DBLTYPE aoq_PQ[3];
254 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
255 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
256 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
257 // Put a minus sign here so we don't have to in RR routines
258 a_over_q = SIMINT_NEG(a_over_q);
259
260
261 //////////////////////////////////////////////
262 // Fjt function section
263 // Maximum v value: 12
264 //////////////////////////////////////////////
265 // The parameter to the Fjt function
266 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
267
268
269 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
270
271
272 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
273 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
274 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
275 for(n = 0; n <= 12; n++)
276 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
277
278 //////////////////////////////////////////////
279 // Primitive integrals: Vertical recurrance
280 //////////////////////////////////////////////
281
282 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
283 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
284 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
285 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
286 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
287 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
288 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
289 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
290 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
291 const SIMINT_DBLTYPE vrr_const_8_over_2q = SIMINT_MUL(const_8, one_over_2q);
292 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
293 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
294 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
295 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
296 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
297 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
298 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
299 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
300 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
301
302
303
304 // Forming PRIM_INT__s_s_p_s[12 * 3];
305 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
306 {
307
308 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
309 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]);
310
311 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
312 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]);
313
314 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
315 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]);
316
317 }
318
319
320
321 // Forming PRIM_INT__s_s_d_s[11 * 6];
322 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
323 {
324
325 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
326 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]);
327 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]);
328
329 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
330 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]);
331
332 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
333 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]);
334
335 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
336 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]);
337 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]);
338
339 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
340 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]);
341
342 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
343 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]);
344 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]);
345
346 }
347
348
349
350 // Forming PRIM_INT__s_s_f_s[10 * 10];
351 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
352 {
353
354 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
355 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]);
356 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]);
357
358 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
359 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]);
360
361 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
362 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]);
363
364 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
365 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]);
366
367 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
368 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]);
369
370 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
371 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]);
372
373 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
374 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]);
375 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]);
376
377 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
378 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]);
379
380 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
381 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]);
382
383 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
384 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]);
385 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]);
386
387 }
388
389
390 VRR_K_s_s_g_s(
391 PRIM_INT__s_s_g_s,
392 PRIM_INT__s_s_f_s,
393 PRIM_INT__s_s_d_s,
394 Q_PA,
395 a_over_q,
396 aoq_PQ,
397 one_over_2q,
398 9);
399
400
401 VRR_K_s_s_h_s(
402 PRIM_INT__s_s_h_s,
403 PRIM_INT__s_s_g_s,
404 PRIM_INT__s_s_f_s,
405 Q_PA,
406 a_over_q,
407 aoq_PQ,
408 one_over_2q,
409 8);
410
411
412 ostei_general_vrr_I(1, 0, 5, 0, 3,
413 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
414 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
415
416
417 VRR_I_p_s_g_s(
418 PRIM_INT__p_s_g_s,
419 PRIM_INT__s_s_g_s,
420 PRIM_INT__s_s_f_s,
421 P_PA,
422 aop_PQ,
423 one_over_2pq,
424 3);
425
426
427 ostei_general_vrr_I(2, 0, 5, 0, 2,
428 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
429 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);
430
431
432 ostei_general_vrr1_K(6, 7,
433 one_over_2q, a_over_q, aoq_PQ, Q_PA,
434 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
435
436
437 ostei_general_vrr_I(1, 0, 6, 0, 3,
438 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
439 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
440
441
442 ostei_general_vrr_I(2, 0, 6, 0, 2,
443 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
444 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);
445
446
447 VRR_I_p_s_f_s(
448 PRIM_INT__p_s_f_s,
449 PRIM_INT__s_s_f_s,
450 PRIM_INT__s_s_d_s,
451 P_PA,
452 aop_PQ,
453 one_over_2pq,
454 3);
455
456
457 ostei_general_vrr_I(2, 0, 4, 0, 2,
458 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
459 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);
460
461
462 ostei_general_vrr_I(3, 0, 5, 0, 1,
463 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
464 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);
465
466
467 ostei_general_vrr1_K(7, 6,
468 one_over_2q, a_over_q, aoq_PQ, Q_PA,
469 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
470
471
472 ostei_general_vrr_I(1, 0, 7, 0, 3,
473 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
474 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
475
476
477 ostei_general_vrr_I(2, 0, 7, 0, 2,
478 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
479 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);
480
481
482 ostei_general_vrr_I(3, 0, 6, 0, 1,
483 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
484 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);
485
486
487 ostei_general_vrr1_K(8, 5,
488 one_over_2q, a_over_q, aoq_PQ, Q_PA,
489 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
490
491
492 ostei_general_vrr_I(1, 0, 8, 0, 3,
493 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
494 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
495
496
497 ostei_general_vrr_I(2, 0, 8, 0, 2,
498 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
499 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);
500
501
502 ostei_general_vrr_I(3, 0, 7, 0, 1,
503 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
504 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);
505
506
507 ostei_general_vrr1_K(9, 4,
508 one_over_2q, a_over_q, aoq_PQ, Q_PA,
509 PRIM_INT__s_s_l_s, PRIM_INT__s_s_k_s, PRIM_INT__s_s_m_s);
510
511
512 ostei_general_vrr_I(1, 0, 9, 0, 3,
513 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
514 PRIM_INT__s_s_m_s, NULL, NULL, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_m_s);
515
516
517 ostei_general_vrr_I(2, 0, 9, 0, 2,
518 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
519 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);
520
521
522 ostei_general_vrr_I(3, 0, 8, 0, 1,
523 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
524 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);
525
526
527 ostei_general_vrr_I(3, 0, 9, 0, 1,
528 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
529 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);
530
531
532
533
534 ////////////////////////////////////
535 // Accumulate contracted integrals
536 ////////////////////////////////////
537 if(lastoffset == 0)
538 {
539 contract_all(126, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
540 contract_all(168, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
541 contract_all(216, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
542 contract_all(270, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
543 contract_all(330, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
544 contract_all(210, PRIM_INT__f_s_h_s, PRIM_PTR_INT__f_s_h_s);
545 contract_all(280, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
546 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
547 contract_all(450, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
548 contract_all(550, PRIM_INT__f_s_m_s, PRIM_PTR_INT__f_s_m_s);
549 }
550 else
551 {
552 contract(126, shelloffsets, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
553 contract(168, shelloffsets, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
554 contract(216, shelloffsets, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
555 contract(270, shelloffsets, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
556 contract(330, shelloffsets, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
557 contract(210, shelloffsets, PRIM_INT__f_s_h_s, PRIM_PTR_INT__f_s_h_s);
558 contract(280, shelloffsets, PRIM_INT__f_s_i_s, PRIM_PTR_INT__f_s_i_s);
559 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
560 contract(450, shelloffsets, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
561 contract(550, shelloffsets, PRIM_INT__f_s_m_s, PRIM_PTR_INT__f_s_m_s);
562 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
563 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
564 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
565 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
566 PRIM_PTR_INT__d_s_m_s += lastoffset*330;
567 PRIM_PTR_INT__f_s_h_s += lastoffset*210;
568 PRIM_PTR_INT__f_s_i_s += lastoffset*280;
569 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
570 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
571 PRIM_PTR_INT__f_s_m_s += lastoffset*550;
572 }
573
574 } // close loop over j
575 } // close loop over i
576
577 //Advance to the next batch
578 jstart = SIMINT_SIMD_ROUND(jend);
579
580 //////////////////////////////////////////////
581 // Contracted integrals: Horizontal recurrance
582 //////////////////////////////////////////////
583
584
585 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
586
587
588 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
589 {
590 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
591
592 // set up HRR pointers
593 double const * restrict HRR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
594 double const * restrict HRR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
595 double const * restrict HRR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
596 double const * restrict HRR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
597 double const * restrict HRR_INT__d_s_m_s = INT__d_s_m_s + abcd * 330;
598 double const * restrict HRR_INT__f_s_h_s = INT__f_s_h_s + abcd * 210;
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 * restrict HRR_INT__d_p_h_g = INT__d_p_h_g + real_abcd * 5670;
604
605 // form INT__d_p_h_s
606 for(iket = 0; iket < 21; ++iket)
607 {
608 HRR_INT__d_p_h_s[0 * 21 + iket] = HRR_INT__f_s_h_s[0 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[0 * 21 + iket] );
609
610 HRR_INT__d_p_h_s[1 * 21 + iket] = HRR_INT__f_s_h_s[1 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[0 * 21 + iket] );
611
612 HRR_INT__d_p_h_s[2 * 21 + iket] = HRR_INT__f_s_h_s[2 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[0 * 21 + iket] );
613
614 HRR_INT__d_p_h_s[3 * 21 + iket] = HRR_INT__f_s_h_s[1 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[1 * 21 + iket] );
615
616 HRR_INT__d_p_h_s[4 * 21 + iket] = HRR_INT__f_s_h_s[3 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[1 * 21 + iket] );
617
618 HRR_INT__d_p_h_s[5 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[1 * 21 + iket] );
619
620 HRR_INT__d_p_h_s[6 * 21 + iket] = HRR_INT__f_s_h_s[2 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[2 * 21 + iket] );
621
622 HRR_INT__d_p_h_s[7 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[2 * 21 + iket] );
623
624 HRR_INT__d_p_h_s[8 * 21 + iket] = HRR_INT__f_s_h_s[5 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[2 * 21 + iket] );
625
626 HRR_INT__d_p_h_s[9 * 21 + iket] = HRR_INT__f_s_h_s[3 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[3 * 21 + iket] );
627
628 HRR_INT__d_p_h_s[10 * 21 + iket] = HRR_INT__f_s_h_s[6 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[3 * 21 + iket] );
629
630 HRR_INT__d_p_h_s[11 * 21 + iket] = HRR_INT__f_s_h_s[7 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[3 * 21 + iket] );
631
632 HRR_INT__d_p_h_s[12 * 21 + iket] = HRR_INT__f_s_h_s[4 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[4 * 21 + iket] );
633
634 HRR_INT__d_p_h_s[13 * 21 + iket] = HRR_INT__f_s_h_s[7 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[4 * 21 + iket] );
635
636 HRR_INT__d_p_h_s[14 * 21 + iket] = HRR_INT__f_s_h_s[8 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[4 * 21 + iket] );
637
638 HRR_INT__d_p_h_s[15 * 21 + iket] = HRR_INT__f_s_h_s[5 * 21 + iket] + ( hAB[0] * HRR_INT__d_s_h_s[5 * 21 + iket] );
639
640 HRR_INT__d_p_h_s[16 * 21 + iket] = HRR_INT__f_s_h_s[8 * 21 + iket] + ( hAB[1] * HRR_INT__d_s_h_s[5 * 21 + iket] );
641
642 HRR_INT__d_p_h_s[17 * 21 + iket] = HRR_INT__f_s_h_s[9 * 21 + iket] + ( hAB[2] * HRR_INT__d_s_h_s[5 * 21 + iket] );
643
644 }
645
646
647 // form INT__d_p_i_s
648 for(iket = 0; iket < 28; ++iket)
649 {
650 HRR_INT__d_p_i_s[0 * 28 + iket] = HRR_INT__f_s_i_s[0 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[0 * 28 + iket] );
651
652 HRR_INT__d_p_i_s[1 * 28 + iket] = HRR_INT__f_s_i_s[1 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[0 * 28 + iket] );
653
654 HRR_INT__d_p_i_s[2 * 28 + iket] = HRR_INT__f_s_i_s[2 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[0 * 28 + iket] );
655
656 HRR_INT__d_p_i_s[3 * 28 + iket] = HRR_INT__f_s_i_s[1 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[1 * 28 + iket] );
657
658 HRR_INT__d_p_i_s[4 * 28 + iket] = HRR_INT__f_s_i_s[3 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[1 * 28 + iket] );
659
660 HRR_INT__d_p_i_s[5 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[1 * 28 + iket] );
661
662 HRR_INT__d_p_i_s[6 * 28 + iket] = HRR_INT__f_s_i_s[2 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[2 * 28 + iket] );
663
664 HRR_INT__d_p_i_s[7 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[2 * 28 + iket] );
665
666 HRR_INT__d_p_i_s[8 * 28 + iket] = HRR_INT__f_s_i_s[5 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[2 * 28 + iket] );
667
668 HRR_INT__d_p_i_s[9 * 28 + iket] = HRR_INT__f_s_i_s[3 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[3 * 28 + iket] );
669
670 HRR_INT__d_p_i_s[10 * 28 + iket] = HRR_INT__f_s_i_s[6 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[3 * 28 + iket] );
671
672 HRR_INT__d_p_i_s[11 * 28 + iket] = HRR_INT__f_s_i_s[7 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[3 * 28 + iket] );
673
674 HRR_INT__d_p_i_s[12 * 28 + iket] = HRR_INT__f_s_i_s[4 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[4 * 28 + iket] );
675
676 HRR_INT__d_p_i_s[13 * 28 + iket] = HRR_INT__f_s_i_s[7 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[4 * 28 + iket] );
677
678 HRR_INT__d_p_i_s[14 * 28 + iket] = HRR_INT__f_s_i_s[8 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[4 * 28 + iket] );
679
680 HRR_INT__d_p_i_s[15 * 28 + iket] = HRR_INT__f_s_i_s[5 * 28 + iket] + ( hAB[0] * HRR_INT__d_s_i_s[5 * 28 + iket] );
681
682 HRR_INT__d_p_i_s[16 * 28 + iket] = HRR_INT__f_s_i_s[8 * 28 + iket] + ( hAB[1] * HRR_INT__d_s_i_s[5 * 28 + iket] );
683
684 HRR_INT__d_p_i_s[17 * 28 + iket] = HRR_INT__f_s_i_s[9 * 28 + iket] + ( hAB[2] * HRR_INT__d_s_i_s[5 * 28 + iket] );
685
686 }
687
688
689 // form INT__d_p_k_s
690 for(iket = 0; iket < 36; ++iket)
691 {
692 HRR_INT__d_p_k_s[0 * 36 + iket] = HRR_INT__f_s_k_s[0 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[0 * 36 + iket] );
693
694 HRR_INT__d_p_k_s[1 * 36 + iket] = HRR_INT__f_s_k_s[1 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[0 * 36 + iket] );
695
696 HRR_INT__d_p_k_s[2 * 36 + iket] = HRR_INT__f_s_k_s[2 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[0 * 36 + iket] );
697
698 HRR_INT__d_p_k_s[3 * 36 + iket] = HRR_INT__f_s_k_s[1 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[1 * 36 + iket] );
699
700 HRR_INT__d_p_k_s[4 * 36 + iket] = HRR_INT__f_s_k_s[3 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[1 * 36 + iket] );
701
702 HRR_INT__d_p_k_s[5 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[1 * 36 + iket] );
703
704 HRR_INT__d_p_k_s[6 * 36 + iket] = HRR_INT__f_s_k_s[2 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[2 * 36 + iket] );
705
706 HRR_INT__d_p_k_s[7 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[2 * 36 + iket] );
707
708 HRR_INT__d_p_k_s[8 * 36 + iket] = HRR_INT__f_s_k_s[5 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[2 * 36 + iket] );
709
710 HRR_INT__d_p_k_s[9 * 36 + iket] = HRR_INT__f_s_k_s[3 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[3 * 36 + iket] );
711
712 HRR_INT__d_p_k_s[10 * 36 + iket] = HRR_INT__f_s_k_s[6 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[3 * 36 + iket] );
713
714 HRR_INT__d_p_k_s[11 * 36 + iket] = HRR_INT__f_s_k_s[7 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[3 * 36 + iket] );
715
716 HRR_INT__d_p_k_s[12 * 36 + iket] = HRR_INT__f_s_k_s[4 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[4 * 36 + iket] );
717
718 HRR_INT__d_p_k_s[13 * 36 + iket] = HRR_INT__f_s_k_s[7 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[4 * 36 + iket] );
719
720 HRR_INT__d_p_k_s[14 * 36 + iket] = HRR_INT__f_s_k_s[8 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[4 * 36 + iket] );
721
722 HRR_INT__d_p_k_s[15 * 36 + iket] = HRR_INT__f_s_k_s[5 * 36 + iket] + ( hAB[0] * HRR_INT__d_s_k_s[5 * 36 + iket] );
723
724 HRR_INT__d_p_k_s[16 * 36 + iket] = HRR_INT__f_s_k_s[8 * 36 + iket] + ( hAB[1] * HRR_INT__d_s_k_s[5 * 36 + iket] );
725
726 HRR_INT__d_p_k_s[17 * 36 + iket] = HRR_INT__f_s_k_s[9 * 36 + iket] + ( hAB[2] * HRR_INT__d_s_k_s[5 * 36 + iket] );
727
728 }
729
730
731 // form INT__d_p_l_s
732 for(iket = 0; iket < 45; ++iket)
733 {
734 HRR_INT__d_p_l_s[0 * 45 + iket] = HRR_INT__f_s_l_s[0 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[0 * 45 + iket] );
735
736 HRR_INT__d_p_l_s[1 * 45 + iket] = HRR_INT__f_s_l_s[1 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[0 * 45 + iket] );
737
738 HRR_INT__d_p_l_s[2 * 45 + iket] = HRR_INT__f_s_l_s[2 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[0 * 45 + iket] );
739
740 HRR_INT__d_p_l_s[3 * 45 + iket] = HRR_INT__f_s_l_s[1 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[1 * 45 + iket] );
741
742 HRR_INT__d_p_l_s[4 * 45 + iket] = HRR_INT__f_s_l_s[3 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[1 * 45 + iket] );
743
744 HRR_INT__d_p_l_s[5 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[1 * 45 + iket] );
745
746 HRR_INT__d_p_l_s[6 * 45 + iket] = HRR_INT__f_s_l_s[2 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[2 * 45 + iket] );
747
748 HRR_INT__d_p_l_s[7 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[2 * 45 + iket] );
749
750 HRR_INT__d_p_l_s[8 * 45 + iket] = HRR_INT__f_s_l_s[5 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[2 * 45 + iket] );
751
752 HRR_INT__d_p_l_s[9 * 45 + iket] = HRR_INT__f_s_l_s[3 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[3 * 45 + iket] );
753
754 HRR_INT__d_p_l_s[10 * 45 + iket] = HRR_INT__f_s_l_s[6 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[3 * 45 + iket] );
755
756 HRR_INT__d_p_l_s[11 * 45 + iket] = HRR_INT__f_s_l_s[7 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[3 * 45 + iket] );
757
758 HRR_INT__d_p_l_s[12 * 45 + iket] = HRR_INT__f_s_l_s[4 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[4 * 45 + iket] );
759
760 HRR_INT__d_p_l_s[13 * 45 + iket] = HRR_INT__f_s_l_s[7 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[4 * 45 + iket] );
761
762 HRR_INT__d_p_l_s[14 * 45 + iket] = HRR_INT__f_s_l_s[8 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[4 * 45 + iket] );
763
764 HRR_INT__d_p_l_s[15 * 45 + iket] = HRR_INT__f_s_l_s[5 * 45 + iket] + ( hAB[0] * HRR_INT__d_s_l_s[5 * 45 + iket] );
765
766 HRR_INT__d_p_l_s[16 * 45 + iket] = HRR_INT__f_s_l_s[8 * 45 + iket] + ( hAB[1] * HRR_INT__d_s_l_s[5 * 45 + iket] );
767
768 HRR_INT__d_p_l_s[17 * 45 + iket] = HRR_INT__f_s_l_s[9 * 45 + iket] + ( hAB[2] * HRR_INT__d_s_l_s[5 * 45 + iket] );
769
770 }
771
772
773 // form INT__d_p_m_s
774 for(iket = 0; iket < 55; ++iket)
775 {
776 HRR_INT__d_p_m_s[0 * 55 + iket] = HRR_INT__f_s_m_s[0 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[0 * 55 + iket] );
777
778 HRR_INT__d_p_m_s[1 * 55 + iket] = HRR_INT__f_s_m_s[1 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[0 * 55 + iket] );
779
780 HRR_INT__d_p_m_s[2 * 55 + iket] = HRR_INT__f_s_m_s[2 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[0 * 55 + iket] );
781
782 HRR_INT__d_p_m_s[3 * 55 + iket] = HRR_INT__f_s_m_s[1 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[1 * 55 + iket] );
783
784 HRR_INT__d_p_m_s[4 * 55 + iket] = HRR_INT__f_s_m_s[3 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[1 * 55 + iket] );
785
786 HRR_INT__d_p_m_s[5 * 55 + iket] = HRR_INT__f_s_m_s[4 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[1 * 55 + iket] );
787
788 HRR_INT__d_p_m_s[6 * 55 + iket] = HRR_INT__f_s_m_s[2 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[2 * 55 + iket] );
789
790 HRR_INT__d_p_m_s[7 * 55 + iket] = HRR_INT__f_s_m_s[4 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[2 * 55 + iket] );
791
792 HRR_INT__d_p_m_s[8 * 55 + iket] = HRR_INT__f_s_m_s[5 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[2 * 55 + iket] );
793
794 HRR_INT__d_p_m_s[9 * 55 + iket] = HRR_INT__f_s_m_s[3 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[3 * 55 + iket] );
795
796 HRR_INT__d_p_m_s[10 * 55 + iket] = HRR_INT__f_s_m_s[6 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[3 * 55 + iket] );
797
798 HRR_INT__d_p_m_s[11 * 55 + iket] = HRR_INT__f_s_m_s[7 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[3 * 55 + iket] );
799
800 HRR_INT__d_p_m_s[12 * 55 + iket] = HRR_INT__f_s_m_s[4 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[4 * 55 + iket] );
801
802 HRR_INT__d_p_m_s[13 * 55 + iket] = HRR_INT__f_s_m_s[7 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[4 * 55 + iket] );
803
804 HRR_INT__d_p_m_s[14 * 55 + iket] = HRR_INT__f_s_m_s[8 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[4 * 55 + iket] );
805
806 HRR_INT__d_p_m_s[15 * 55 + iket] = HRR_INT__f_s_m_s[5 * 55 + iket] + ( hAB[0] * HRR_INT__d_s_m_s[5 * 55 + iket] );
807
808 HRR_INT__d_p_m_s[16 * 55 + iket] = HRR_INT__f_s_m_s[8 * 55 + iket] + ( hAB[1] * HRR_INT__d_s_m_s[5 * 55 + iket] );
809
810 HRR_INT__d_p_m_s[17 * 55 + iket] = HRR_INT__f_s_m_s[9 * 55 + iket] + ( hAB[2] * HRR_INT__d_s_m_s[5 * 55 + iket] );
811
812 }
813
814
815 // form INT__d_p_h_p
816 ostei_general_hrr_L(2, 1, 5, 1, hCD, HRR_INT__d_p_i_s, HRR_INT__d_p_h_s, HRR_INT__d_p_h_p);
817
818 // form INT__d_p_i_p
819 ostei_general_hrr_L(2, 1, 6, 1, hCD, HRR_INT__d_p_k_s, HRR_INT__d_p_i_s, HRR_INT__d_p_i_p);
820
821 // form INT__d_p_k_p
822 ostei_general_hrr_L(2, 1, 7, 1, hCD, HRR_INT__d_p_l_s, HRR_INT__d_p_k_s, HRR_INT__d_p_k_p);
823
824 // form INT__d_p_l_p
825 ostei_general_hrr_L(2, 1, 8, 1, hCD, HRR_INT__d_p_m_s, HRR_INT__d_p_l_s, HRR_INT__d_p_l_p);
826
827 // form INT__d_p_h_d
828 ostei_general_hrr_L(2, 1, 5, 2, hCD, HRR_INT__d_p_i_p, HRR_INT__d_p_h_p, HRR_INT__d_p_h_d);
829
830 // form INT__d_p_i_d
831 ostei_general_hrr_L(2, 1, 6, 2, hCD, HRR_INT__d_p_k_p, HRR_INT__d_p_i_p, HRR_INT__d_p_i_d);
832
833 // form INT__d_p_k_d
834 ostei_general_hrr_L(2, 1, 7, 2, hCD, HRR_INT__d_p_l_p, HRR_INT__d_p_k_p, HRR_INT__d_p_k_d);
835
836 // form INT__d_p_h_f
837 ostei_general_hrr_L(2, 1, 5, 3, hCD, HRR_INT__d_p_i_d, HRR_INT__d_p_h_d, HRR_INT__d_p_h_f);
838
839 // form INT__d_p_i_f
840 ostei_general_hrr_L(2, 1, 6, 3, hCD, HRR_INT__d_p_k_d, HRR_INT__d_p_i_d, HRR_INT__d_p_i_f);
841
842 // form INT__d_p_h_g
843 ostei_general_hrr_L(2, 1, 5, 4, hCD, HRR_INT__d_p_i_f, HRR_INT__d_p_h_f, HRR_INT__d_p_h_g);
844
845
846 } // close HRR loop
847
848
849 } // close loop cdbatch
850
851 istart = iend;
852 } // close loop over ab
853
854 return P.nshell12_clip * Q.nshell12_clip;
855 }
856
ostei_p_d_h_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_d_h_g)857 int ostei_p_d_h_g(struct simint_multi_shellpair const P,
858 struct simint_multi_shellpair const Q,
859 double screen_tol,
860 double * const restrict work,
861 double * const restrict INT__p_d_h_g)
862 {
863 double P_AB[3*P.nshell12];
864 struct simint_multi_shellpair P_tmp = P;
865 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
866 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
867 P_tmp.AB_x = P_AB;
868 P_tmp.AB_y = P_AB + P.nshell12;
869 P_tmp.AB_z = P_AB + 2*P.nshell12;
870
871 for(int i = 0; i < P.nshell12; i++)
872 {
873 P_tmp.AB_x[i] = -P.AB_x[i];
874 P_tmp.AB_y[i] = -P.AB_y[i];
875 P_tmp.AB_z[i] = -P.AB_z[i];
876 }
877
878 int ret = ostei_d_p_h_g(P_tmp, Q, screen_tol, work, INT__p_d_h_g);
879 double buffer[5670] SIMINT_ALIGN_ARRAY_DBL;
880
881 for(int q = 0; q < ret; q++)
882 {
883 int idx = 0;
884 for(int a = 0; a < 3; ++a)
885 for(int b = 0; b < 6; ++b)
886 for(int c = 0; c < 21; ++c)
887 for(int d = 0; d < 15; ++d)
888 buffer[idx++] = INT__p_d_h_g[q*5670+b*945+a*315+c*15+d];
889
890 memcpy(INT__p_d_h_g+q*5670, buffer, 5670*sizeof(double));
891 }
892
893 return ret;
894 }
895
ostei_d_p_g_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_p_g_h)896 int ostei_d_p_g_h(struct simint_multi_shellpair const P,
897 struct simint_multi_shellpair const Q,
898 double screen_tol,
899 double * const restrict work,
900 double * const restrict INT__d_p_g_h)
901 {
902 double Q_AB[3*Q.nshell12];
903 struct simint_multi_shellpair Q_tmp = Q;
904 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
905 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
906 Q_tmp.AB_x = Q_AB;
907 Q_tmp.AB_y = Q_AB + Q.nshell12;
908 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
909
910 for(int i = 0; i < Q.nshell12; i++)
911 {
912 Q_tmp.AB_x[i] = -Q.AB_x[i];
913 Q_tmp.AB_y[i] = -Q.AB_y[i];
914 Q_tmp.AB_z[i] = -Q.AB_z[i];
915 }
916
917 int ret = ostei_d_p_h_g(P, Q_tmp, screen_tol, work, INT__d_p_g_h);
918 double buffer[5670] SIMINT_ALIGN_ARRAY_DBL;
919
920 for(int q = 0; q < ret; q++)
921 {
922 int idx = 0;
923 for(int a = 0; a < 6; ++a)
924 for(int b = 0; b < 3; ++b)
925 for(int c = 0; c < 15; ++c)
926 for(int d = 0; d < 21; ++d)
927 buffer[idx++] = INT__d_p_g_h[q*5670+a*945+b*315+d*15+c];
928
929 memcpy(INT__d_p_g_h+q*5670, buffer, 5670*sizeof(double));
930 }
931
932 return ret;
933 }
934
ostei_p_d_g_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_d_g_h)935 int ostei_p_d_g_h(struct simint_multi_shellpair const P,
936 struct simint_multi_shellpair const Q,
937 double screen_tol,
938 double * const restrict work,
939 double * const restrict INT__p_d_g_h)
940 {
941 double P_AB[3*P.nshell12];
942 struct simint_multi_shellpair P_tmp = P;
943 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
944 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
945 P_tmp.AB_x = P_AB;
946 P_tmp.AB_y = P_AB + P.nshell12;
947 P_tmp.AB_z = P_AB + 2*P.nshell12;
948
949 for(int i = 0; i < P.nshell12; i++)
950 {
951 P_tmp.AB_x[i] = -P.AB_x[i];
952 P_tmp.AB_y[i] = -P.AB_y[i];
953 P_tmp.AB_z[i] = -P.AB_z[i];
954 }
955
956 double Q_AB[3*Q.nshell12];
957 struct simint_multi_shellpair Q_tmp = Q;
958 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
959 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
960 Q_tmp.AB_x = Q_AB;
961 Q_tmp.AB_y = Q_AB + Q.nshell12;
962 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
963
964 for(int i = 0; i < Q.nshell12; i++)
965 {
966 Q_tmp.AB_x[i] = -Q.AB_x[i];
967 Q_tmp.AB_y[i] = -Q.AB_y[i];
968 Q_tmp.AB_z[i] = -Q.AB_z[i];
969 }
970
971 int ret = ostei_d_p_h_g(P_tmp, Q_tmp, screen_tol, work, INT__p_d_g_h);
972 double buffer[5670] SIMINT_ALIGN_ARRAY_DBL;
973
974 for(int q = 0; q < ret; q++)
975 {
976 int idx = 0;
977 for(int a = 0; a < 3; ++a)
978 for(int b = 0; b < 6; ++b)
979 for(int c = 0; c < 15; ++c)
980 for(int d = 0; d < 21; ++d)
981 buffer[idx++] = INT__p_d_g_h[q*5670+b*945+a*315+d*15+c];
982
983 memcpy(INT__p_d_g_h+q*5670, buffer, 5670*sizeof(double));
984 }
985
986 return ret;
987 }
988
989