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_p_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__p_p_h_g)8 int ostei_p_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__p_p_h_g)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__p_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__p_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__p_s_i_s = work + (SIMINT_NSHELL_SIMD * 63);
31 double * const INT__p_s_k_s = work + (SIMINT_NSHELL_SIMD * 147);
32 double * const INT__p_s_l_s = work + (SIMINT_NSHELL_SIMD * 255);
33 double * const INT__p_s_m_s = work + (SIMINT_NSHELL_SIMD * 390);
34 double * const INT__d_s_h_s = work + (SIMINT_NSHELL_SIMD * 555);
35 double * const INT__d_s_i_s = work + (SIMINT_NSHELL_SIMD * 681);
36 double * const INT__d_s_k_s = work + (SIMINT_NSHELL_SIMD * 849);
37 double * const INT__d_s_l_s = work + (SIMINT_NSHELL_SIMD * 1065);
38 double * const INT__d_s_m_s = work + (SIMINT_NSHELL_SIMD * 1335);
39 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*1665);
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 + 12;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 45;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 105;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 195;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 315;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 462;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 630;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 810;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_m_s = primwork + 990;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1155;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1245;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 1371;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 1539;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 1755;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_m_s = primwork + 2025;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 2355;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 2481;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 2649;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 2865;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_m_s = primwork + 3135;
61 double * const hrrwork = (double *)(primwork + 3465);
62 double * const HRR_INT__p_p_h_s = hrrwork + 0;
63 double * const HRR_INT__p_p_h_p = hrrwork + 189;
64 double * const HRR_INT__p_p_h_d = hrrwork + 756;
65 double * const HRR_INT__p_p_h_f = hrrwork + 1890;
66 double * const HRR_INT__p_p_i_s = hrrwork + 3780;
67 double * const HRR_INT__p_p_i_p = hrrwork + 4032;
68 double * const HRR_INT__p_p_i_d = hrrwork + 4788;
69 double * const HRR_INT__p_p_i_f = hrrwork + 6300;
70 double * const HRR_INT__p_p_k_s = hrrwork + 8820;
71 double * const HRR_INT__p_p_k_p = hrrwork + 9144;
72 double * const HRR_INT__p_p_k_d = hrrwork + 10116;
73 double * const HRR_INT__p_p_l_s = hrrwork + 12060;
74 double * const HRR_INT__p_p_l_p = hrrwork + 12465;
75 double * const HRR_INT__p_p_m_s = hrrwork + 13680;
76
77
78 // Create constants
79 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
80 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
81 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
82 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
83 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
84 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
85 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
86 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
87 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
88 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
89
90
91 ////////////////////////////////////////
92 // Loop over shells and primitives
93 ////////////////////////////////////////
94
95 real_abcd = 0;
96 istart = 0;
97 for(ab = 0; ab < P.nshell12_clip; ++ab)
98 {
99 const int iend = istart + P.nprim12[ab];
100
101 cd = 0;
102 jstart = 0;
103
104 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
105 {
106 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
107 int jend = jstart;
108 for(i = 0; i < nshellbatch; i++)
109 jend += Q.nprim12[cd+i];
110
111 // Clear the beginning of the workspace (where we are accumulating integrals)
112 memset(work, 0, SIMINT_NSHELL_SIMD * 1665 * sizeof(double));
113 abcd = 0;
114
115
116 for(i = istart; i < iend; ++i)
117 {
118 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
119
120 if(check_screen)
121 {
122 // Skip this whole thing if always insignificant
123 if((P.screen[i] * Q.screen_max) < screen_tol)
124 continue;
125 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
126 }
127
128 icd = 0;
129 iprimcd = 0;
130 nprim_icd = Q.nprim12[cd];
131 double * restrict PRIM_PTR_INT__p_s_h_s = INT__p_s_h_s + abcd * 63;
132 double * restrict PRIM_PTR_INT__p_s_i_s = INT__p_s_i_s + abcd * 84;
133 double * restrict PRIM_PTR_INT__p_s_k_s = INT__p_s_k_s + abcd * 108;
134 double * restrict PRIM_PTR_INT__p_s_l_s = INT__p_s_l_s + abcd * 135;
135 double * restrict PRIM_PTR_INT__p_s_m_s = INT__p_s_m_s + abcd * 165;
136 double * restrict PRIM_PTR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
137 double * restrict PRIM_PTR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
138 double * restrict PRIM_PTR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
139 double * restrict PRIM_PTR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
140 double * restrict PRIM_PTR_INT__d_s_m_s = INT__d_s_m_s + abcd * 330;
141
142
143
144 // Load these one per loop over i
145 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
146 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
147 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
148
149 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
150
151 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
152 {
153 // calculate the shell offsets
154 // these are the offset from the shell pointed to by cd
155 // for each element
156 int shelloffsets[SIMINT_SIMD_LEN] = {0};
157 int lastoffset = 0;
158 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
159
160 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
161 {
162 // Handle if the first element of the vector is a new shell
163 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
164 {
165 nprim_icd += Q.nprim12[cd + (++icd)];
166 PRIM_PTR_INT__p_s_h_s += 63;
167 PRIM_PTR_INT__p_s_i_s += 84;
168 PRIM_PTR_INT__p_s_k_s += 108;
169 PRIM_PTR_INT__p_s_l_s += 135;
170 PRIM_PTR_INT__p_s_m_s += 165;
171 PRIM_PTR_INT__d_s_h_s += 126;
172 PRIM_PTR_INT__d_s_i_s += 168;
173 PRIM_PTR_INT__d_s_k_s += 216;
174 PRIM_PTR_INT__d_s_l_s += 270;
175 PRIM_PTR_INT__d_s_m_s += 330;
176 }
177 iprimcd++;
178 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
179 {
180 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
181 {
182 shelloffsets[n] = shelloffsets[n-1] + 1;
183 lastoffset++;
184 nprim_icd += Q.nprim12[cd + (++icd)];
185 }
186 else
187 shelloffsets[n] = shelloffsets[n-1];
188 iprimcd++;
189 }
190 }
191 else
192 iprimcd += SIMINT_SIMD_LEN;
193
194 // Do we have to compute this vector (or has it been screened out)?
195 // (not_screened != 0 means we have to do this vector)
196 if(check_screen)
197 {
198 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
199 if(vmax < screen_tol)
200 {
201 PRIM_PTR_INT__p_s_h_s += lastoffset*63;
202 PRIM_PTR_INT__p_s_i_s += lastoffset*84;
203 PRIM_PTR_INT__p_s_k_s += lastoffset*108;
204 PRIM_PTR_INT__p_s_l_s += lastoffset*135;
205 PRIM_PTR_INT__p_s_m_s += lastoffset*165;
206 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
207 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
208 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
209 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
210 PRIM_PTR_INT__d_s_m_s += lastoffset*330;
211 continue;
212 }
213 }
214
215 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
216 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
217 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
218 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
219
220
221 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
222 SIMINT_DBLTYPE PQ[3];
223 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
224 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
225 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
226 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
227 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
228 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
229
230 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
231 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
232 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
233 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
234 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
235 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
236 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
237
238 // NOTE: Minus sign!
239 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
240 SIMINT_DBLTYPE aop_PQ[3];
241 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
242 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
243 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
244
245 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
246 SIMINT_DBLTYPE aoq_PQ[3];
247 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
248 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
249 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
250 // Put a minus sign here so we don't have to in RR routines
251 a_over_q = SIMINT_NEG(a_over_q);
252
253
254 //////////////////////////////////////////////
255 // Fjt function section
256 // Maximum v value: 11
257 //////////////////////////////////////////////
258 // The parameter to the Fjt function
259 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
260
261
262 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
263
264
265 boys_F_split(PRIM_INT__s_s_s_s, F_x, 11);
266 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
267 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
268 for(n = 0; n <= 11; n++)
269 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
270
271 //////////////////////////////////////////////
272 // Primitive integrals: Vertical recurrance
273 //////////////////////////////////////////////
274
275 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
276 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
277 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
278 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
279 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
280 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
281 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
282 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
283 const SIMINT_DBLTYPE vrr_const_8_over_2q = SIMINT_MUL(const_8, one_over_2q);
284 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
285 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
288 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
289 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
290 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
291 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
292 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
293
294
295
296 // Forming PRIM_INT__s_s_p_s[11 * 3];
297 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
298 {
299
300 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
301 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]);
302
303 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
304 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 1]);
305
306 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
307 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 2]);
308
309 }
310
311
312
313 // Forming PRIM_INT__s_s_d_s[10 * 6];
314 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
315 {
316
317 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
318 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 0]);
319 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 0]);
320
321 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
322 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 1]);
323
324 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
325 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]);
326 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]);
327
328 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
329 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]);
330 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]);
331
332 }
333
334
335
336 // Forming PRIM_INT__s_s_f_s[9 * 10];
337 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
338 {
339
340 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
341 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]);
342 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]);
343
344 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
345 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]);
346
347 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
348 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]);
349
350 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
351 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]);
352
353 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
354 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]);
355
356 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
357 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]);
358
359 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
360 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]);
361 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]);
362
363 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
364 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]);
365
366 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
367 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]);
368
369 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
370 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]);
371 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]);
372
373 }
374
375
376 VRR_K_s_s_g_s(
377 PRIM_INT__s_s_g_s,
378 PRIM_INT__s_s_f_s,
379 PRIM_INT__s_s_d_s,
380 Q_PA,
381 a_over_q,
382 aoq_PQ,
383 one_over_2q,
384 8);
385
386
387 VRR_K_s_s_h_s(
388 PRIM_INT__s_s_h_s,
389 PRIM_INT__s_s_g_s,
390 PRIM_INT__s_s_f_s,
391 Q_PA,
392 a_over_q,
393 aoq_PQ,
394 one_over_2q,
395 7);
396
397
398 ostei_general_vrr_I(1, 0, 5, 0, 2,
399 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
400 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
401
402
403 ostei_general_vrr1_K(6, 6,
404 one_over_2q, a_over_q, aoq_PQ, Q_PA,
405 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
406
407
408 ostei_general_vrr_I(1, 0, 6, 0, 2,
409 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
410 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
411
412
413 VRR_I_p_s_g_s(
414 PRIM_INT__p_s_g_s,
415 PRIM_INT__s_s_g_s,
416 PRIM_INT__s_s_f_s,
417 P_PA,
418 aop_PQ,
419 one_over_2pq,
420 2);
421
422
423 ostei_general_vrr_I(2, 0, 5, 0, 1,
424 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
425 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);
426
427
428 ostei_general_vrr1_K(7, 5,
429 one_over_2q, a_over_q, aoq_PQ, Q_PA,
430 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
431
432
433 ostei_general_vrr_I(1, 0, 7, 0, 2,
434 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
435 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
436
437
438 ostei_general_vrr_I(2, 0, 6, 0, 1,
439 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
440 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);
441
442
443 ostei_general_vrr1_K(8, 4,
444 one_over_2q, a_over_q, aoq_PQ, Q_PA,
445 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
446
447
448 ostei_general_vrr_I(1, 0, 8, 0, 2,
449 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
450 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
451
452
453 ostei_general_vrr_I(2, 0, 7, 0, 1,
454 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
455 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);
456
457
458 ostei_general_vrr1_K(9, 3,
459 one_over_2q, a_over_q, aoq_PQ, Q_PA,
460 PRIM_INT__s_s_l_s, PRIM_INT__s_s_k_s, PRIM_INT__s_s_m_s);
461
462
463 ostei_general_vrr_I(1, 0, 9, 0, 2,
464 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
465 PRIM_INT__s_s_m_s, NULL, NULL, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_m_s);
466
467
468 ostei_general_vrr_I(2, 0, 8, 0, 1,
469 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
470 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);
471
472
473 ostei_general_vrr_I(2, 0, 9, 0, 1,
474 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
475 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);
476
477
478
479
480 ////////////////////////////////////
481 // Accumulate contracted integrals
482 ////////////////////////////////////
483 if(lastoffset == 0)
484 {
485 contract_all(63, PRIM_INT__p_s_h_s, PRIM_PTR_INT__p_s_h_s);
486 contract_all(84, PRIM_INT__p_s_i_s, PRIM_PTR_INT__p_s_i_s);
487 contract_all(108, PRIM_INT__p_s_k_s, PRIM_PTR_INT__p_s_k_s);
488 contract_all(135, PRIM_INT__p_s_l_s, PRIM_PTR_INT__p_s_l_s);
489 contract_all(165, PRIM_INT__p_s_m_s, PRIM_PTR_INT__p_s_m_s);
490 contract_all(126, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
491 contract_all(168, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
492 contract_all(216, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
493 contract_all(270, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
494 contract_all(330, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
495 }
496 else
497 {
498 contract(63, shelloffsets, PRIM_INT__p_s_h_s, PRIM_PTR_INT__p_s_h_s);
499 contract(84, shelloffsets, PRIM_INT__p_s_i_s, PRIM_PTR_INT__p_s_i_s);
500 contract(108, shelloffsets, PRIM_INT__p_s_k_s, PRIM_PTR_INT__p_s_k_s);
501 contract(135, shelloffsets, PRIM_INT__p_s_l_s, PRIM_PTR_INT__p_s_l_s);
502 contract(165, shelloffsets, PRIM_INT__p_s_m_s, PRIM_PTR_INT__p_s_m_s);
503 contract(126, shelloffsets, PRIM_INT__d_s_h_s, PRIM_PTR_INT__d_s_h_s);
504 contract(168, shelloffsets, PRIM_INT__d_s_i_s, PRIM_PTR_INT__d_s_i_s);
505 contract(216, shelloffsets, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
506 contract(270, shelloffsets, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
507 contract(330, shelloffsets, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
508 PRIM_PTR_INT__p_s_h_s += lastoffset*63;
509 PRIM_PTR_INT__p_s_i_s += lastoffset*84;
510 PRIM_PTR_INT__p_s_k_s += lastoffset*108;
511 PRIM_PTR_INT__p_s_l_s += lastoffset*135;
512 PRIM_PTR_INT__p_s_m_s += lastoffset*165;
513 PRIM_PTR_INT__d_s_h_s += lastoffset*126;
514 PRIM_PTR_INT__d_s_i_s += lastoffset*168;
515 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
516 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
517 PRIM_PTR_INT__d_s_m_s += lastoffset*330;
518 }
519
520 } // close loop over j
521 } // close loop over i
522
523 //Advance to the next batch
524 jstart = SIMINT_SIMD_ROUND(jend);
525
526 //////////////////////////////////////////////
527 // Contracted integrals: Horizontal recurrance
528 //////////////////////////////////////////////
529
530
531 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
532
533
534 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
535 {
536 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
537
538 // set up HRR pointers
539 double const * restrict HRR_INT__p_s_h_s = INT__p_s_h_s + abcd * 63;
540 double const * restrict HRR_INT__p_s_i_s = INT__p_s_i_s + abcd * 84;
541 double const * restrict HRR_INT__p_s_k_s = INT__p_s_k_s + abcd * 108;
542 double const * restrict HRR_INT__p_s_l_s = INT__p_s_l_s + abcd * 135;
543 double const * restrict HRR_INT__p_s_m_s = INT__p_s_m_s + abcd * 165;
544 double const * restrict HRR_INT__d_s_h_s = INT__d_s_h_s + abcd * 126;
545 double const * restrict HRR_INT__d_s_i_s = INT__d_s_i_s + abcd * 168;
546 double const * restrict HRR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
547 double const * restrict HRR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
548 double const * restrict HRR_INT__d_s_m_s = INT__d_s_m_s + abcd * 330;
549 double * restrict HRR_INT__p_p_h_g = INT__p_p_h_g + real_abcd * 2835;
550
551 // form INT__p_p_h_s
552 for(iket = 0; iket < 21; ++iket)
553 {
554 HRR_INT__p_p_h_s[0 * 21 + iket] = HRR_INT__d_s_h_s[0 * 21 + iket] + ( hAB[0] * HRR_INT__p_s_h_s[0 * 21 + iket] );
555
556 HRR_INT__p_p_h_s[1 * 21 + iket] = HRR_INT__d_s_h_s[1 * 21 + iket] + ( hAB[1] * HRR_INT__p_s_h_s[0 * 21 + iket] );
557
558 HRR_INT__p_p_h_s[2 * 21 + iket] = HRR_INT__d_s_h_s[2 * 21 + iket] + ( hAB[2] * HRR_INT__p_s_h_s[0 * 21 + iket] );
559
560 HRR_INT__p_p_h_s[3 * 21 + iket] = HRR_INT__d_s_h_s[1 * 21 + iket] + ( hAB[0] * HRR_INT__p_s_h_s[1 * 21 + iket] );
561
562 HRR_INT__p_p_h_s[4 * 21 + iket] = HRR_INT__d_s_h_s[3 * 21 + iket] + ( hAB[1] * HRR_INT__p_s_h_s[1 * 21 + iket] );
563
564 HRR_INT__p_p_h_s[5 * 21 + iket] = HRR_INT__d_s_h_s[4 * 21 + iket] + ( hAB[2] * HRR_INT__p_s_h_s[1 * 21 + iket] );
565
566 HRR_INT__p_p_h_s[6 * 21 + iket] = HRR_INT__d_s_h_s[2 * 21 + iket] + ( hAB[0] * HRR_INT__p_s_h_s[2 * 21 + iket] );
567
568 HRR_INT__p_p_h_s[7 * 21 + iket] = HRR_INT__d_s_h_s[4 * 21 + iket] + ( hAB[1] * HRR_INT__p_s_h_s[2 * 21 + iket] );
569
570 HRR_INT__p_p_h_s[8 * 21 + iket] = HRR_INT__d_s_h_s[5 * 21 + iket] + ( hAB[2] * HRR_INT__p_s_h_s[2 * 21 + iket] );
571
572 }
573
574
575 // form INT__p_p_i_s
576 for(iket = 0; iket < 28; ++iket)
577 {
578 HRR_INT__p_p_i_s[0 * 28 + iket] = HRR_INT__d_s_i_s[0 * 28 + iket] + ( hAB[0] * HRR_INT__p_s_i_s[0 * 28 + iket] );
579
580 HRR_INT__p_p_i_s[1 * 28 + iket] = HRR_INT__d_s_i_s[1 * 28 + iket] + ( hAB[1] * HRR_INT__p_s_i_s[0 * 28 + iket] );
581
582 HRR_INT__p_p_i_s[2 * 28 + iket] = HRR_INT__d_s_i_s[2 * 28 + iket] + ( hAB[2] * HRR_INT__p_s_i_s[0 * 28 + iket] );
583
584 HRR_INT__p_p_i_s[3 * 28 + iket] = HRR_INT__d_s_i_s[1 * 28 + iket] + ( hAB[0] * HRR_INT__p_s_i_s[1 * 28 + iket] );
585
586 HRR_INT__p_p_i_s[4 * 28 + iket] = HRR_INT__d_s_i_s[3 * 28 + iket] + ( hAB[1] * HRR_INT__p_s_i_s[1 * 28 + iket] );
587
588 HRR_INT__p_p_i_s[5 * 28 + iket] = HRR_INT__d_s_i_s[4 * 28 + iket] + ( hAB[2] * HRR_INT__p_s_i_s[1 * 28 + iket] );
589
590 HRR_INT__p_p_i_s[6 * 28 + iket] = HRR_INT__d_s_i_s[2 * 28 + iket] + ( hAB[0] * HRR_INT__p_s_i_s[2 * 28 + iket] );
591
592 HRR_INT__p_p_i_s[7 * 28 + iket] = HRR_INT__d_s_i_s[4 * 28 + iket] + ( hAB[1] * HRR_INT__p_s_i_s[2 * 28 + iket] );
593
594 HRR_INT__p_p_i_s[8 * 28 + iket] = HRR_INT__d_s_i_s[5 * 28 + iket] + ( hAB[2] * HRR_INT__p_s_i_s[2 * 28 + iket] );
595
596 }
597
598
599 // form INT__p_p_k_s
600 for(iket = 0; iket < 36; ++iket)
601 {
602 HRR_INT__p_p_k_s[0 * 36 + iket] = HRR_INT__d_s_k_s[0 * 36 + iket] + ( hAB[0] * HRR_INT__p_s_k_s[0 * 36 + iket] );
603
604 HRR_INT__p_p_k_s[1 * 36 + iket] = HRR_INT__d_s_k_s[1 * 36 + iket] + ( hAB[1] * HRR_INT__p_s_k_s[0 * 36 + iket] );
605
606 HRR_INT__p_p_k_s[2 * 36 + iket] = HRR_INT__d_s_k_s[2 * 36 + iket] + ( hAB[2] * HRR_INT__p_s_k_s[0 * 36 + iket] );
607
608 HRR_INT__p_p_k_s[3 * 36 + iket] = HRR_INT__d_s_k_s[1 * 36 + iket] + ( hAB[0] * HRR_INT__p_s_k_s[1 * 36 + iket] );
609
610 HRR_INT__p_p_k_s[4 * 36 + iket] = HRR_INT__d_s_k_s[3 * 36 + iket] + ( hAB[1] * HRR_INT__p_s_k_s[1 * 36 + iket] );
611
612 HRR_INT__p_p_k_s[5 * 36 + iket] = HRR_INT__d_s_k_s[4 * 36 + iket] + ( hAB[2] * HRR_INT__p_s_k_s[1 * 36 + iket] );
613
614 HRR_INT__p_p_k_s[6 * 36 + iket] = HRR_INT__d_s_k_s[2 * 36 + iket] + ( hAB[0] * HRR_INT__p_s_k_s[2 * 36 + iket] );
615
616 HRR_INT__p_p_k_s[7 * 36 + iket] = HRR_INT__d_s_k_s[4 * 36 + iket] + ( hAB[1] * HRR_INT__p_s_k_s[2 * 36 + iket] );
617
618 HRR_INT__p_p_k_s[8 * 36 + iket] = HRR_INT__d_s_k_s[5 * 36 + iket] + ( hAB[2] * HRR_INT__p_s_k_s[2 * 36 + iket] );
619
620 }
621
622
623 // form INT__p_p_l_s
624 for(iket = 0; iket < 45; ++iket)
625 {
626 HRR_INT__p_p_l_s[0 * 45 + iket] = HRR_INT__d_s_l_s[0 * 45 + iket] + ( hAB[0] * HRR_INT__p_s_l_s[0 * 45 + iket] );
627
628 HRR_INT__p_p_l_s[1 * 45 + iket] = HRR_INT__d_s_l_s[1 * 45 + iket] + ( hAB[1] * HRR_INT__p_s_l_s[0 * 45 + iket] );
629
630 HRR_INT__p_p_l_s[2 * 45 + iket] = HRR_INT__d_s_l_s[2 * 45 + iket] + ( hAB[2] * HRR_INT__p_s_l_s[0 * 45 + iket] );
631
632 HRR_INT__p_p_l_s[3 * 45 + iket] = HRR_INT__d_s_l_s[1 * 45 + iket] + ( hAB[0] * HRR_INT__p_s_l_s[1 * 45 + iket] );
633
634 HRR_INT__p_p_l_s[4 * 45 + iket] = HRR_INT__d_s_l_s[3 * 45 + iket] + ( hAB[1] * HRR_INT__p_s_l_s[1 * 45 + iket] );
635
636 HRR_INT__p_p_l_s[5 * 45 + iket] = HRR_INT__d_s_l_s[4 * 45 + iket] + ( hAB[2] * HRR_INT__p_s_l_s[1 * 45 + iket] );
637
638 HRR_INT__p_p_l_s[6 * 45 + iket] = HRR_INT__d_s_l_s[2 * 45 + iket] + ( hAB[0] * HRR_INT__p_s_l_s[2 * 45 + iket] );
639
640 HRR_INT__p_p_l_s[7 * 45 + iket] = HRR_INT__d_s_l_s[4 * 45 + iket] + ( hAB[1] * HRR_INT__p_s_l_s[2 * 45 + iket] );
641
642 HRR_INT__p_p_l_s[8 * 45 + iket] = HRR_INT__d_s_l_s[5 * 45 + iket] + ( hAB[2] * HRR_INT__p_s_l_s[2 * 45 + iket] );
643
644 }
645
646
647 // form INT__p_p_m_s
648 for(iket = 0; iket < 55; ++iket)
649 {
650 HRR_INT__p_p_m_s[0 * 55 + iket] = HRR_INT__d_s_m_s[0 * 55 + iket] + ( hAB[0] * HRR_INT__p_s_m_s[0 * 55 + iket] );
651
652 HRR_INT__p_p_m_s[1 * 55 + iket] = HRR_INT__d_s_m_s[1 * 55 + iket] + ( hAB[1] * HRR_INT__p_s_m_s[0 * 55 + iket] );
653
654 HRR_INT__p_p_m_s[2 * 55 + iket] = HRR_INT__d_s_m_s[2 * 55 + iket] + ( hAB[2] * HRR_INT__p_s_m_s[0 * 55 + iket] );
655
656 HRR_INT__p_p_m_s[3 * 55 + iket] = HRR_INT__d_s_m_s[1 * 55 + iket] + ( hAB[0] * HRR_INT__p_s_m_s[1 * 55 + iket] );
657
658 HRR_INT__p_p_m_s[4 * 55 + iket] = HRR_INT__d_s_m_s[3 * 55 + iket] + ( hAB[1] * HRR_INT__p_s_m_s[1 * 55 + iket] );
659
660 HRR_INT__p_p_m_s[5 * 55 + iket] = HRR_INT__d_s_m_s[4 * 55 + iket] + ( hAB[2] * HRR_INT__p_s_m_s[1 * 55 + iket] );
661
662 HRR_INT__p_p_m_s[6 * 55 + iket] = HRR_INT__d_s_m_s[2 * 55 + iket] + ( hAB[0] * HRR_INT__p_s_m_s[2 * 55 + iket] );
663
664 HRR_INT__p_p_m_s[7 * 55 + iket] = HRR_INT__d_s_m_s[4 * 55 + iket] + ( hAB[1] * HRR_INT__p_s_m_s[2 * 55 + iket] );
665
666 HRR_INT__p_p_m_s[8 * 55 + iket] = HRR_INT__d_s_m_s[5 * 55 + iket] + ( hAB[2] * HRR_INT__p_s_m_s[2 * 55 + iket] );
667
668 }
669
670
671 // form INT__p_p_h_p
672 ostei_general_hrr_L(1, 1, 5, 1, hCD, HRR_INT__p_p_i_s, HRR_INT__p_p_h_s, HRR_INT__p_p_h_p);
673
674 // form INT__p_p_i_p
675 ostei_general_hrr_L(1, 1, 6, 1, hCD, HRR_INT__p_p_k_s, HRR_INT__p_p_i_s, HRR_INT__p_p_i_p);
676
677 // form INT__p_p_k_p
678 ostei_general_hrr_L(1, 1, 7, 1, hCD, HRR_INT__p_p_l_s, HRR_INT__p_p_k_s, HRR_INT__p_p_k_p);
679
680 // form INT__p_p_l_p
681 ostei_general_hrr_L(1, 1, 8, 1, hCD, HRR_INT__p_p_m_s, HRR_INT__p_p_l_s, HRR_INT__p_p_l_p);
682
683 // form INT__p_p_h_d
684 ostei_general_hrr_L(1, 1, 5, 2, hCD, HRR_INT__p_p_i_p, HRR_INT__p_p_h_p, HRR_INT__p_p_h_d);
685
686 // form INT__p_p_i_d
687 ostei_general_hrr_L(1, 1, 6, 2, hCD, HRR_INT__p_p_k_p, HRR_INT__p_p_i_p, HRR_INT__p_p_i_d);
688
689 // form INT__p_p_k_d
690 ostei_general_hrr_L(1, 1, 7, 2, hCD, HRR_INT__p_p_l_p, HRR_INT__p_p_k_p, HRR_INT__p_p_k_d);
691
692 // form INT__p_p_h_f
693 ostei_general_hrr_L(1, 1, 5, 3, hCD, HRR_INT__p_p_i_d, HRR_INT__p_p_h_d, HRR_INT__p_p_h_f);
694
695 // form INT__p_p_i_f
696 ostei_general_hrr_L(1, 1, 6, 3, hCD, HRR_INT__p_p_k_d, HRR_INT__p_p_i_d, HRR_INT__p_p_i_f);
697
698 // form INT__p_p_h_g
699 ostei_general_hrr_L(1, 1, 5, 4, hCD, HRR_INT__p_p_i_f, HRR_INT__p_p_h_f, HRR_INT__p_p_h_g);
700
701
702 } // close HRR loop
703
704
705 } // close loop cdbatch
706
707 istart = iend;
708 } // close loop over ab
709
710 return P.nshell12_clip * Q.nshell12_clip;
711 }
712
ostei_p_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__p_p_g_h)713 int ostei_p_p_g_h(struct simint_multi_shellpair const P,
714 struct simint_multi_shellpair const Q,
715 double screen_tol,
716 double * const restrict work,
717 double * const restrict INT__p_p_g_h)
718 {
719 double Q_AB[3*Q.nshell12];
720 struct simint_multi_shellpair Q_tmp = Q;
721 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
722 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
723 Q_tmp.AB_x = Q_AB;
724 Q_tmp.AB_y = Q_AB + Q.nshell12;
725 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
726
727 for(int i = 0; i < Q.nshell12; i++)
728 {
729 Q_tmp.AB_x[i] = -Q.AB_x[i];
730 Q_tmp.AB_y[i] = -Q.AB_y[i];
731 Q_tmp.AB_z[i] = -Q.AB_z[i];
732 }
733
734 int ret = ostei_p_p_h_g(P, Q_tmp, screen_tol, work, INT__p_p_g_h);
735 double buffer[2835] SIMINT_ALIGN_ARRAY_DBL;
736
737 for(int q = 0; q < ret; q++)
738 {
739 int idx = 0;
740 for(int a = 0; a < 3; ++a)
741 for(int b = 0; b < 3; ++b)
742 for(int c = 0; c < 15; ++c)
743 for(int d = 0; d < 21; ++d)
744 buffer[idx++] = INT__p_p_g_h[q*2835+a*945+b*315+d*15+c];
745
746 memcpy(INT__p_p_g_h+q*2835, buffer, 2835*sizeof(double));
747 }
748
749 return ret;
750 }
751
752