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_k_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_k_g)8 int ostei_p_p_k_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_k_g)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__p_p_k_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_k_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__p_s_l_s = work + (SIMINT_NSHELL_SIMD * 108);
31 double * const INT__p_s_m_s = work + (SIMINT_NSHELL_SIMD * 243);
32 double * const INT__p_s_n_s = work + (SIMINT_NSHELL_SIMD * 408);
33 double * const INT__p_s_o_s = work + (SIMINT_NSHELL_SIMD * 606);
34 double * const INT__d_s_k_s = work + (SIMINT_NSHELL_SIMD * 840);
35 double * const INT__d_s_l_s = work + (SIMINT_NSHELL_SIMD * 1056);
36 double * const INT__d_s_m_s = work + (SIMINT_NSHELL_SIMD * 1326);
37 double * const INT__d_s_n_s = work + (SIMINT_NSHELL_SIMD * 1656);
38 double * const INT__d_s_o_s = work + (SIMINT_NSHELL_SIMD * 2052);
39 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2520);
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 + 14;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 53;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 125;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 235;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 385;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 574;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 798;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 1050;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_m_s = primwork + 1320;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_n_s = primwork + 1595;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_o_s = primwork + 1859;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2093;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2261;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 2477;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_m_s = primwork + 2747;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_n_s = primwork + 3077;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_o_s = primwork + 3473;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 3941;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 4157;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_m_s = primwork + 4427;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_n_s = primwork + 4757;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_o_s = primwork + 5153;
63 double * const hrrwork = (double *)(primwork + 5621);
64 double * const HRR_INT__p_p_k_s = hrrwork + 0;
65 double * const HRR_INT__p_p_k_p = hrrwork + 324;
66 double * const HRR_INT__p_p_k_d = hrrwork + 1296;
67 double * const HRR_INT__p_p_k_f = hrrwork + 3240;
68 double * const HRR_INT__p_p_l_s = hrrwork + 6480;
69 double * const HRR_INT__p_p_l_p = hrrwork + 6885;
70 double * const HRR_INT__p_p_l_d = hrrwork + 8100;
71 double * const HRR_INT__p_p_l_f = hrrwork + 10530;
72 double * const HRR_INT__p_p_m_s = hrrwork + 14580;
73 double * const HRR_INT__p_p_m_p = hrrwork + 15075;
74 double * const HRR_INT__p_p_m_d = hrrwork + 16560;
75 double * const HRR_INT__p_p_n_s = hrrwork + 19530;
76 double * const HRR_INT__p_p_n_p = hrrwork + 20124;
77 double * const HRR_INT__p_p_o_s = hrrwork + 21906;
78
79
80 // Create constants
81 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
82 const SIMINT_DBLTYPE const_10 = SIMINT_DBLSET1(10);
83 const SIMINT_DBLTYPE const_11 = SIMINT_DBLSET1(11);
84 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
85 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
86 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
87 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
88 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
89 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
90 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
91 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
92 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
93
94
95 ////////////////////////////////////////
96 // Loop over shells and primitives
97 ////////////////////////////////////////
98
99 real_abcd = 0;
100 istart = 0;
101 for(ab = 0; ab < P.nshell12_clip; ++ab)
102 {
103 const int iend = istart + P.nprim12[ab];
104
105 cd = 0;
106 jstart = 0;
107
108 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
109 {
110 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
111 int jend = jstart;
112 for(i = 0; i < nshellbatch; i++)
113 jend += Q.nprim12[cd+i];
114
115 // Clear the beginning of the workspace (where we are accumulating integrals)
116 memset(work, 0, SIMINT_NSHELL_SIMD * 2520 * sizeof(double));
117 abcd = 0;
118
119
120 for(i = istart; i < iend; ++i)
121 {
122 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
123
124 if(check_screen)
125 {
126 // Skip this whole thing if always insignificant
127 if((P.screen[i] * Q.screen_max) < screen_tol)
128 continue;
129 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
130 }
131
132 icd = 0;
133 iprimcd = 0;
134 nprim_icd = Q.nprim12[cd];
135 double * restrict PRIM_PTR_INT__p_s_k_s = INT__p_s_k_s + abcd * 108;
136 double * restrict PRIM_PTR_INT__p_s_l_s = INT__p_s_l_s + abcd * 135;
137 double * restrict PRIM_PTR_INT__p_s_m_s = INT__p_s_m_s + abcd * 165;
138 double * restrict PRIM_PTR_INT__p_s_n_s = INT__p_s_n_s + abcd * 198;
139 double * restrict PRIM_PTR_INT__p_s_o_s = INT__p_s_o_s + abcd * 234;
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__d_s_n_s = INT__d_s_n_s + abcd * 396;
144 double * restrict PRIM_PTR_INT__d_s_o_s = INT__d_s_o_s + abcd * 468;
145
146
147
148 // Load these one per loop over i
149 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
150 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
151 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
152
153 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
154
155 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
156 {
157 // calculate the shell offsets
158 // these are the offset from the shell pointed to by cd
159 // for each element
160 int shelloffsets[SIMINT_SIMD_LEN] = {0};
161 int lastoffset = 0;
162 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
163
164 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
165 {
166 // Handle if the first element of the vector is a new shell
167 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
168 {
169 nprim_icd += Q.nprim12[cd + (++icd)];
170 PRIM_PTR_INT__p_s_k_s += 108;
171 PRIM_PTR_INT__p_s_l_s += 135;
172 PRIM_PTR_INT__p_s_m_s += 165;
173 PRIM_PTR_INT__p_s_n_s += 198;
174 PRIM_PTR_INT__p_s_o_s += 234;
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__d_s_n_s += 396;
179 PRIM_PTR_INT__d_s_o_s += 468;
180 }
181 iprimcd++;
182 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
183 {
184 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
185 {
186 shelloffsets[n] = shelloffsets[n-1] + 1;
187 lastoffset++;
188 nprim_icd += Q.nprim12[cd + (++icd)];
189 }
190 else
191 shelloffsets[n] = shelloffsets[n-1];
192 iprimcd++;
193 }
194 }
195 else
196 iprimcd += SIMINT_SIMD_LEN;
197
198 // Do we have to compute this vector (or has it been screened out)?
199 // (not_screened != 0 means we have to do this vector)
200 if(check_screen)
201 {
202 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
203 if(vmax < screen_tol)
204 {
205 PRIM_PTR_INT__p_s_k_s += lastoffset*108;
206 PRIM_PTR_INT__p_s_l_s += lastoffset*135;
207 PRIM_PTR_INT__p_s_m_s += lastoffset*165;
208 PRIM_PTR_INT__p_s_n_s += lastoffset*198;
209 PRIM_PTR_INT__p_s_o_s += lastoffset*234;
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__d_s_n_s += lastoffset*396;
214 PRIM_PTR_INT__d_s_o_s += lastoffset*468;
215 continue;
216 }
217 }
218
219 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
220 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
221 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
222 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
223
224
225 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
226 SIMINT_DBLTYPE PQ[3];
227 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
228 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
229 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
230 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
231 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
232 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
233
234 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
235 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
236 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
237 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
238 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
239 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
240 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
241
242 // NOTE: Minus sign!
243 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
244 SIMINT_DBLTYPE aop_PQ[3];
245 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
246 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
247 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
248
249 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
250 SIMINT_DBLTYPE aoq_PQ[3];
251 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
252 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
253 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
254 // Put a minus sign here so we don't have to in RR routines
255 a_over_q = SIMINT_NEG(a_over_q);
256
257
258 //////////////////////////////////////////////
259 // Fjt function section
260 // Maximum v value: 13
261 //////////////////////////////////////////////
262 // The parameter to the Fjt function
263 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
264
265
266 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
267
268
269 boys_F_split(PRIM_INT__s_s_s_s, F_x, 13);
270 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
271 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
272 for(n = 0; n <= 13; n++)
273 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
274
275 //////////////////////////////////////////////
276 // Primitive integrals: Vertical recurrance
277 //////////////////////////////////////////////
278
279 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
280 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
281 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
282 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
283 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
284 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
285 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
286 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
287 const SIMINT_DBLTYPE vrr_const_8_over_2q = SIMINT_MUL(const_8, one_over_2q);
288 const SIMINT_DBLTYPE vrr_const_9_over_2q = SIMINT_MUL(const_9, one_over_2q);
289 const SIMINT_DBLTYPE vrr_const_10_over_2q = SIMINT_MUL(const_10, one_over_2q);
290 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
291 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
292 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
293 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
294 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
295 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
296 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
297 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
298 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
299 const SIMINT_DBLTYPE vrr_const_10_over_2pq = SIMINT_MUL(const_10, one_over_2pq);
300 const SIMINT_DBLTYPE vrr_const_11_over_2pq = SIMINT_MUL(const_11, one_over_2pq);
301
302
303
304 // Forming PRIM_INT__s_s_p_s[13 * 3];
305 for(n = 0; n < 13; ++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[12 * 6];
322 for(n = 0; n < 12; ++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 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
330 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]);
331 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]);
332
333 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
334 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]);
335 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]);
336
337 }
338
339
340
341 // Forming PRIM_INT__s_s_f_s[11 * 10];
342 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
343 {
344
345 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
346 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]);
347 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]);
348
349 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
350 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]);
351
352 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
353 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]);
354
355 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
356 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]);
357 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]);
358
359 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
360 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]);
361
362 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
363 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]);
364 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]);
365
366 }
367
368
369 VRR_K_s_s_g_s(
370 PRIM_INT__s_s_g_s,
371 PRIM_INT__s_s_f_s,
372 PRIM_INT__s_s_d_s,
373 Q_PA,
374 a_over_q,
375 aoq_PQ,
376 one_over_2q,
377 10);
378
379
380 VRR_K_s_s_h_s(
381 PRIM_INT__s_s_h_s,
382 PRIM_INT__s_s_g_s,
383 PRIM_INT__s_s_f_s,
384 Q_PA,
385 a_over_q,
386 aoq_PQ,
387 one_over_2q,
388 9);
389
390
391 ostei_general_vrr1_K(6, 8,
392 one_over_2q, a_over_q, aoq_PQ, Q_PA,
393 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
394
395
396 ostei_general_vrr1_K(7, 7,
397 one_over_2q, a_over_q, aoq_PQ, Q_PA,
398 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
399
400
401 ostei_general_vrr_I(1, 0, 7, 0, 2,
402 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
403 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
404
405
406 ostei_general_vrr1_K(8, 6,
407 one_over_2q, a_over_q, aoq_PQ, Q_PA,
408 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
409
410
411 ostei_general_vrr_I(1, 0, 8, 0, 2,
412 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
413 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
414
415
416 ostei_general_vrr_I(1, 0, 6, 0, 2,
417 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
418 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
419
420
421 ostei_general_vrr_I(2, 0, 7, 0, 1,
422 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
423 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);
424
425
426 ostei_general_vrr1_K(9, 5,
427 one_over_2q, a_over_q, aoq_PQ, Q_PA,
428 PRIM_INT__s_s_l_s, PRIM_INT__s_s_k_s, PRIM_INT__s_s_m_s);
429
430
431 ostei_general_vrr_I(1, 0, 9, 0, 2,
432 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
433 PRIM_INT__s_s_m_s, NULL, NULL, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_m_s);
434
435
436 ostei_general_vrr_I(2, 0, 8, 0, 1,
437 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
438 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);
439
440
441 ostei_general_vrr1_K(10, 4,
442 one_over_2q, a_over_q, aoq_PQ, Q_PA,
443 PRIM_INT__s_s_m_s, PRIM_INT__s_s_l_s, PRIM_INT__s_s_n_s);
444
445
446 ostei_general_vrr_I(1, 0, 10, 0, 2,
447 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
448 PRIM_INT__s_s_n_s, NULL, NULL, PRIM_INT__s_s_m_s, NULL, PRIM_INT__p_s_n_s);
449
450
451 ostei_general_vrr_I(2, 0, 9, 0, 1,
452 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
453 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);
454
455
456 ostei_general_vrr1_K(11, 3,
457 one_over_2q, a_over_q, aoq_PQ, Q_PA,
458 PRIM_INT__s_s_n_s, PRIM_INT__s_s_m_s, PRIM_INT__s_s_o_s);
459
460
461 ostei_general_vrr_I(1, 0, 11, 0, 2,
462 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
463 PRIM_INT__s_s_o_s, NULL, NULL, PRIM_INT__s_s_n_s, NULL, PRIM_INT__p_s_o_s);
464
465
466 ostei_general_vrr_I(2, 0, 10, 0, 1,
467 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
468 PRIM_INT__p_s_n_s, PRIM_INT__s_s_n_s, NULL, PRIM_INT__p_s_m_s, NULL, PRIM_INT__d_s_n_s);
469
470
471 ostei_general_vrr_I(2, 0, 11, 0, 1,
472 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
473 PRIM_INT__p_s_o_s, PRIM_INT__s_s_o_s, NULL, PRIM_INT__p_s_n_s, NULL, PRIM_INT__d_s_o_s);
474
475
476
477
478 ////////////////////////////////////
479 // Accumulate contracted integrals
480 ////////////////////////////////////
481 if(lastoffset == 0)
482 {
483 contract_all(108, PRIM_INT__p_s_k_s, PRIM_PTR_INT__p_s_k_s);
484 contract_all(135, PRIM_INT__p_s_l_s, PRIM_PTR_INT__p_s_l_s);
485 contract_all(165, PRIM_INT__p_s_m_s, PRIM_PTR_INT__p_s_m_s);
486 contract_all(198, PRIM_INT__p_s_n_s, PRIM_PTR_INT__p_s_n_s);
487 contract_all(234, PRIM_INT__p_s_o_s, PRIM_PTR_INT__p_s_o_s);
488 contract_all(216, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
489 contract_all(270, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
490 contract_all(330, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
491 contract_all(396, PRIM_INT__d_s_n_s, PRIM_PTR_INT__d_s_n_s);
492 contract_all(468, PRIM_INT__d_s_o_s, PRIM_PTR_INT__d_s_o_s);
493 }
494 else
495 {
496 contract(108, shelloffsets, PRIM_INT__p_s_k_s, PRIM_PTR_INT__p_s_k_s);
497 contract(135, shelloffsets, PRIM_INT__p_s_l_s, PRIM_PTR_INT__p_s_l_s);
498 contract(165, shelloffsets, PRIM_INT__p_s_m_s, PRIM_PTR_INT__p_s_m_s);
499 contract(198, shelloffsets, PRIM_INT__p_s_n_s, PRIM_PTR_INT__p_s_n_s);
500 contract(234, shelloffsets, PRIM_INT__p_s_o_s, PRIM_PTR_INT__p_s_o_s);
501 contract(216, shelloffsets, PRIM_INT__d_s_k_s, PRIM_PTR_INT__d_s_k_s);
502 contract(270, shelloffsets, PRIM_INT__d_s_l_s, PRIM_PTR_INT__d_s_l_s);
503 contract(330, shelloffsets, PRIM_INT__d_s_m_s, PRIM_PTR_INT__d_s_m_s);
504 contract(396, shelloffsets, PRIM_INT__d_s_n_s, PRIM_PTR_INT__d_s_n_s);
505 contract(468, shelloffsets, PRIM_INT__d_s_o_s, PRIM_PTR_INT__d_s_o_s);
506 PRIM_PTR_INT__p_s_k_s += lastoffset*108;
507 PRIM_PTR_INT__p_s_l_s += lastoffset*135;
508 PRIM_PTR_INT__p_s_m_s += lastoffset*165;
509 PRIM_PTR_INT__p_s_n_s += lastoffset*198;
510 PRIM_PTR_INT__p_s_o_s += lastoffset*234;
511 PRIM_PTR_INT__d_s_k_s += lastoffset*216;
512 PRIM_PTR_INT__d_s_l_s += lastoffset*270;
513 PRIM_PTR_INT__d_s_m_s += lastoffset*330;
514 PRIM_PTR_INT__d_s_n_s += lastoffset*396;
515 PRIM_PTR_INT__d_s_o_s += lastoffset*468;
516 }
517
518 } // close loop over j
519 } // close loop over i
520
521 //Advance to the next batch
522 jstart = SIMINT_SIMD_ROUND(jend);
523
524 //////////////////////////////////////////////
525 // Contracted integrals: Horizontal recurrance
526 //////////////////////////////////////////////
527
528
529 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
530
531
532 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
533 {
534 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
535
536 // set up HRR pointers
537 double const * restrict HRR_INT__p_s_k_s = INT__p_s_k_s + abcd * 108;
538 double const * restrict HRR_INT__p_s_l_s = INT__p_s_l_s + abcd * 135;
539 double const * restrict HRR_INT__p_s_m_s = INT__p_s_m_s + abcd * 165;
540 double const * restrict HRR_INT__p_s_n_s = INT__p_s_n_s + abcd * 198;
541 double const * restrict HRR_INT__p_s_o_s = INT__p_s_o_s + abcd * 234;
542 double const * restrict HRR_INT__d_s_k_s = INT__d_s_k_s + abcd * 216;
543 double const * restrict HRR_INT__d_s_l_s = INT__d_s_l_s + abcd * 270;
544 double const * restrict HRR_INT__d_s_m_s = INT__d_s_m_s + abcd * 330;
545 double const * restrict HRR_INT__d_s_n_s = INT__d_s_n_s + abcd * 396;
546 double const * restrict HRR_INT__d_s_o_s = INT__d_s_o_s + abcd * 468;
547 double * restrict HRR_INT__p_p_k_g = INT__p_p_k_g + real_abcd * 4860;
548
549 // form INT__p_p_k_s
550 for(iket = 0; iket < 36; ++iket)
551 {
552 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] );
553
554 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] );
555
556 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] );
557
558 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] );
559
560 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] );
561
562 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] );
563
564 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] );
565
566 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] );
567
568 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] );
569
570 }
571
572
573 // form INT__p_p_l_s
574 for(iket = 0; iket < 45; ++iket)
575 {
576 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] );
577
578 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] );
579
580 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] );
581
582 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] );
583
584 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] );
585
586 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] );
587
588 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] );
589
590 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] );
591
592 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] );
593
594 }
595
596
597 // form INT__p_p_m_s
598 for(iket = 0; iket < 55; ++iket)
599 {
600 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] );
601
602 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] );
603
604 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] );
605
606 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] );
607
608 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] );
609
610 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] );
611
612 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] );
613
614 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] );
615
616 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] );
617
618 }
619
620
621 // form INT__p_p_n_s
622 for(iket = 0; iket < 66; ++iket)
623 {
624 HRR_INT__p_p_n_s[0 * 66 + iket] = HRR_INT__d_s_n_s[0 * 66 + iket] + ( hAB[0] * HRR_INT__p_s_n_s[0 * 66 + iket] );
625
626 HRR_INT__p_p_n_s[1 * 66 + iket] = HRR_INT__d_s_n_s[1 * 66 + iket] + ( hAB[1] * HRR_INT__p_s_n_s[0 * 66 + iket] );
627
628 HRR_INT__p_p_n_s[2 * 66 + iket] = HRR_INT__d_s_n_s[2 * 66 + iket] + ( hAB[2] * HRR_INT__p_s_n_s[0 * 66 + iket] );
629
630 HRR_INT__p_p_n_s[3 * 66 + iket] = HRR_INT__d_s_n_s[1 * 66 + iket] + ( hAB[0] * HRR_INT__p_s_n_s[1 * 66 + iket] );
631
632 HRR_INT__p_p_n_s[4 * 66 + iket] = HRR_INT__d_s_n_s[3 * 66 + iket] + ( hAB[1] * HRR_INT__p_s_n_s[1 * 66 + iket] );
633
634 HRR_INT__p_p_n_s[5 * 66 + iket] = HRR_INT__d_s_n_s[4 * 66 + iket] + ( hAB[2] * HRR_INT__p_s_n_s[1 * 66 + iket] );
635
636 HRR_INT__p_p_n_s[6 * 66 + iket] = HRR_INT__d_s_n_s[2 * 66 + iket] + ( hAB[0] * HRR_INT__p_s_n_s[2 * 66 + iket] );
637
638 HRR_INT__p_p_n_s[7 * 66 + iket] = HRR_INT__d_s_n_s[4 * 66 + iket] + ( hAB[1] * HRR_INT__p_s_n_s[2 * 66 + iket] );
639
640 HRR_INT__p_p_n_s[8 * 66 + iket] = HRR_INT__d_s_n_s[5 * 66 + iket] + ( hAB[2] * HRR_INT__p_s_n_s[2 * 66 + iket] );
641
642 }
643
644
645 // form INT__p_p_o_s
646 for(iket = 0; iket < 78; ++iket)
647 {
648 HRR_INT__p_p_o_s[0 * 78 + iket] = HRR_INT__d_s_o_s[0 * 78 + iket] + ( hAB[0] * HRR_INT__p_s_o_s[0 * 78 + iket] );
649
650 HRR_INT__p_p_o_s[1 * 78 + iket] = HRR_INT__d_s_o_s[1 * 78 + iket] + ( hAB[1] * HRR_INT__p_s_o_s[0 * 78 + iket] );
651
652 HRR_INT__p_p_o_s[2 * 78 + iket] = HRR_INT__d_s_o_s[2 * 78 + iket] + ( hAB[2] * HRR_INT__p_s_o_s[0 * 78 + iket] );
653
654 HRR_INT__p_p_o_s[3 * 78 + iket] = HRR_INT__d_s_o_s[1 * 78 + iket] + ( hAB[0] * HRR_INT__p_s_o_s[1 * 78 + iket] );
655
656 HRR_INT__p_p_o_s[4 * 78 + iket] = HRR_INT__d_s_o_s[3 * 78 + iket] + ( hAB[1] * HRR_INT__p_s_o_s[1 * 78 + iket] );
657
658 HRR_INT__p_p_o_s[5 * 78 + iket] = HRR_INT__d_s_o_s[4 * 78 + iket] + ( hAB[2] * HRR_INT__p_s_o_s[1 * 78 + iket] );
659
660 HRR_INT__p_p_o_s[6 * 78 + iket] = HRR_INT__d_s_o_s[2 * 78 + iket] + ( hAB[0] * HRR_INT__p_s_o_s[2 * 78 + iket] );
661
662 HRR_INT__p_p_o_s[7 * 78 + iket] = HRR_INT__d_s_o_s[4 * 78 + iket] + ( hAB[1] * HRR_INT__p_s_o_s[2 * 78 + iket] );
663
664 HRR_INT__p_p_o_s[8 * 78 + iket] = HRR_INT__d_s_o_s[5 * 78 + iket] + ( hAB[2] * HRR_INT__p_s_o_s[2 * 78 + iket] );
665
666 }
667
668
669 // form INT__p_p_k_p
670 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);
671
672 // form INT__p_p_l_p
673 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);
674
675 // form INT__p_p_m_p
676 ostei_general_hrr_L(1, 1, 9, 1, hCD, HRR_INT__p_p_n_s, HRR_INT__p_p_m_s, HRR_INT__p_p_m_p);
677
678 // form INT__p_p_n_p
679 ostei_general_hrr_L(1, 1, 10, 1, hCD, HRR_INT__p_p_o_s, HRR_INT__p_p_n_s, HRR_INT__p_p_n_p);
680
681 // form INT__p_p_k_d
682 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);
683
684 // form INT__p_p_l_d
685 ostei_general_hrr_L(1, 1, 8, 2, hCD, HRR_INT__p_p_m_p, HRR_INT__p_p_l_p, HRR_INT__p_p_l_d);
686
687 // form INT__p_p_m_d
688 ostei_general_hrr_L(1, 1, 9, 2, hCD, HRR_INT__p_p_n_p, HRR_INT__p_p_m_p, HRR_INT__p_p_m_d);
689
690 // form INT__p_p_k_f
691 ostei_general_hrr_L(1, 1, 7, 3, hCD, HRR_INT__p_p_l_d, HRR_INT__p_p_k_d, HRR_INT__p_p_k_f);
692
693 // form INT__p_p_l_f
694 ostei_general_hrr_L(1, 1, 8, 3, hCD, HRR_INT__p_p_m_d, HRR_INT__p_p_l_d, HRR_INT__p_p_l_f);
695
696 // form INT__p_p_k_g
697 ostei_general_hrr_L(1, 1, 7, 4, hCD, HRR_INT__p_p_l_f, HRR_INT__p_p_k_f, HRR_INT__p_p_k_g);
698
699
700 } // close HRR loop
701
702
703 } // close loop cdbatch
704
705 istart = iend;
706 } // close loop over ab
707
708 return P.nshell12_clip * Q.nshell12_clip;
709 }
710
ostei_p_p_g_k(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_k)711 int ostei_p_p_g_k(struct simint_multi_shellpair const P,
712 struct simint_multi_shellpair const Q,
713 double screen_tol,
714 double * const restrict work,
715 double * const restrict INT__p_p_g_k)
716 {
717 double Q_AB[3*Q.nshell12];
718 struct simint_multi_shellpair Q_tmp = Q;
719 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
720 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
721 Q_tmp.AB_x = Q_AB;
722 Q_tmp.AB_y = Q_AB + Q.nshell12;
723 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
724
725 for(int i = 0; i < Q.nshell12; i++)
726 {
727 Q_tmp.AB_x[i] = -Q.AB_x[i];
728 Q_tmp.AB_y[i] = -Q.AB_y[i];
729 Q_tmp.AB_z[i] = -Q.AB_z[i];
730 }
731
732 int ret = ostei_p_p_k_g(P, Q_tmp, screen_tol, work, INT__p_p_g_k);
733 double buffer[4860] SIMINT_ALIGN_ARRAY_DBL;
734
735 for(int q = 0; q < ret; q++)
736 {
737 int idx = 0;
738 for(int a = 0; a < 3; ++a)
739 for(int b = 0; b < 3; ++b)
740 for(int c = 0; c < 15; ++c)
741 for(int d = 0; d < 36; ++d)
742 buffer[idx++] = INT__p_p_g_k[q*4860+a*1620+b*540+d*15+c];
743
744 memcpy(INT__p_p_g_k+q*4860, buffer, 4860*sizeof(double));
745 }
746
747 return ret;
748 }
749
750