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_k_d_g_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_d_g_p)8 int ostei_k_d_g_p(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__k_d_g_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__k_d_g_p);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 540);
31 double * const INT__l_s_g_s = work + (SIMINT_NSHELL_SIMD * 1296);
32 double * const INT__l_s_h_s = work + (SIMINT_NSHELL_SIMD * 1971);
33 double * const INT__m_s_g_s = work + (SIMINT_NSHELL_SIMD * 2916);
34 double * const INT__m_s_h_s = work + (SIMINT_NSHELL_SIMD * 3741);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*4896);
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 15;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 57;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 135;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 255;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 405;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 570;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 795;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1155;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 1365;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 1680;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 2184;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 2814;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 3066;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 3486;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 4158;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 4998;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 5838;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 6126;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 6666;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 7530;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 8610;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 9690;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 10446;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 10761;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 11436;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 12516;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_g_s = primwork + 13866;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_h_s = primwork + 15216;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_s_s = primwork + 16161;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_p_s = primwork + 16491;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_d_s = primwork + 17316;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_f_s = primwork + 18636;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_g_s = primwork + 20286;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_h_s = primwork + 21936;
71 double * const hrrwork = (double *)(primwork + 23091);
72 double * const HRR_INT__k_p_g_s = hrrwork + 0;
73 double * const HRR_INT__k_p_h_s = hrrwork + 1620;
74 double * const HRR_INT__k_d_g_s = hrrwork + 3888;
75 double * const HRR_INT__k_d_h_s = hrrwork + 7128;
76 double * const HRR_INT__l_p_g_s = hrrwork + 11664;
77 double * const HRR_INT__l_p_h_s = hrrwork + 13689;
78
79
80 // Create constants
81 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
82 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
83 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
84 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
85 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
86 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
87 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
88 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
89 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
90 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
91
92
93 ////////////////////////////////////////
94 // Loop over shells and primitives
95 ////////////////////////////////////////
96
97 real_abcd = 0;
98 istart = 0;
99 for(ab = 0; ab < P.nshell12_clip; ++ab)
100 {
101 const int iend = istart + P.nprim12[ab];
102
103 cd = 0;
104 jstart = 0;
105
106 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
107 {
108 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
109 int jend = jstart;
110 for(i = 0; i < nshellbatch; i++)
111 jend += Q.nprim12[cd+i];
112
113 // Clear the beginning of the workspace (where we are accumulating integrals)
114 memset(work, 0, SIMINT_NSHELL_SIMD * 4896 * sizeof(double));
115 abcd = 0;
116
117
118 for(i = istart; i < iend; ++i)
119 {
120 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
121
122 if(check_screen)
123 {
124 // Skip this whole thing if always insignificant
125 if((P.screen[i] * Q.screen_max) < screen_tol)
126 continue;
127 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
128 }
129
130 icd = 0;
131 iprimcd = 0;
132 nprim_icd = Q.nprim12[cd];
133 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
134 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
135 double * restrict PRIM_PTR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
136 double * restrict PRIM_PTR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
137 double * restrict PRIM_PTR_INT__m_s_g_s = INT__m_s_g_s + abcd * 825;
138 double * restrict PRIM_PTR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
139
140
141
142 // Load these one per loop over i
143 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
144 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
145 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
146
147 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
148
149 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
150 {
151 // calculate the shell offsets
152 // these are the offset from the shell pointed to by cd
153 // for each element
154 int shelloffsets[SIMINT_SIMD_LEN] = {0};
155 int lastoffset = 0;
156 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
157
158 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
159 {
160 // Handle if the first element of the vector is a new shell
161 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
162 {
163 nprim_icd += Q.nprim12[cd + (++icd)];
164 PRIM_PTR_INT__k_s_g_s += 540;
165 PRIM_PTR_INT__k_s_h_s += 756;
166 PRIM_PTR_INT__l_s_g_s += 675;
167 PRIM_PTR_INT__l_s_h_s += 945;
168 PRIM_PTR_INT__m_s_g_s += 825;
169 PRIM_PTR_INT__m_s_h_s += 1155;
170 }
171 iprimcd++;
172 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
173 {
174 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
175 {
176 shelloffsets[n] = shelloffsets[n-1] + 1;
177 lastoffset++;
178 nprim_icd += Q.nprim12[cd + (++icd)];
179 }
180 else
181 shelloffsets[n] = shelloffsets[n-1];
182 iprimcd++;
183 }
184 }
185 else
186 iprimcd += SIMINT_SIMD_LEN;
187
188 // Do we have to compute this vector (or has it been screened out)?
189 // (not_screened != 0 means we have to do this vector)
190 if(check_screen)
191 {
192 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
193 if(vmax < screen_tol)
194 {
195 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
196 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
197 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
198 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
199 PRIM_PTR_INT__m_s_g_s += lastoffset*825;
200 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
201 continue;
202 }
203 }
204
205 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
206 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
207 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
208 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
209
210
211 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
212 SIMINT_DBLTYPE PQ[3];
213 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
214 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
215 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
216 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
217 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
218 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
219
220 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
221 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
222 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
223 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
224 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
225 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
226 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
227
228 // NOTE: Minus sign!
229 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
230 SIMINT_DBLTYPE aop_PQ[3];
231 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
232 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
233 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
234
235 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
236 SIMINT_DBLTYPE aoq_PQ[3];
237 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
238 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
239 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
240 // Put a minus sign here so we don't have to in RR routines
241 a_over_q = SIMINT_NEG(a_over_q);
242
243
244 //////////////////////////////////////////////
245 // Fjt function section
246 // Maximum v value: 14
247 //////////////////////////////////////////////
248 // The parameter to the Fjt function
249 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
250
251
252 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
253
254
255 boys_F_split(PRIM_INT__s_s_s_s, F_x, 14);
256 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
257 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
258 for(n = 0; n <= 14; n++)
259 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
260
261 //////////////////////////////////////////////
262 // Primitive integrals: Vertical recurrance
263 //////////////////////////////////////////////
264
265 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
266 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
267 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
268 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
269 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
270 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
271 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
272 const SIMINT_DBLTYPE vrr_const_8_over_2p = SIMINT_MUL(const_8, one_over_2p);
273 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
274 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
275 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
276 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
277 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
278 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
279 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
280 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
281 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
282 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
283 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
284 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
285 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
286
287
288
289 // Forming PRIM_INT__p_s_s_s[14 * 3];
290 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
291 {
292
293 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
294 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]);
295
296 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
297 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 1]);
298
299 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
300 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_s_s[n * 3 + 2]);
301
302 }
303
304
305
306 // Forming PRIM_INT__d_s_s_s[13 * 6];
307 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
308 {
309
310 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
311 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 0]);
312 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 0]);
313
314 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
315 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 1]);
316
317 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
318 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_s_s[n * 6 + 2]);
319
320 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
321 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 3]);
322 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 3]);
323
324 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
325 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_s_s[n * 6 + 4]);
326
327 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
328 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_s_s[n * 6 + 5]);
329 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__d_s_s_s[n * 6 + 5]);
330
331 }
332
333
334
335 // Forming PRIM_INT__f_s_s_s[12 * 10];
336 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
337 {
338
339 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
340 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 0]);
341 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__f_s_s_s[n * 10 + 0]);
342
343 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
344 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 1]);
345
346 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
347 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__f_s_s_s[n * 10 + 2]);
348
349 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
350 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 3]);
351
352 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
353 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__f_s_s_s[n * 10 + 4]);
354
355 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
356 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 5]);
357
358 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
359 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 6]);
360 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__f_s_s_s[n * 10 + 6]);
361
362 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
363 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__f_s_s_s[n * 10 + 7]);
364
365 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
366 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 8]);
367
368 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
369 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__f_s_s_s[n * 10 + 9]);
370 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__f_s_s_s[n * 10 + 9]);
371
372 }
373
374
375 VRR_I_g_s_s_s(
376 PRIM_INT__g_s_s_s,
377 PRIM_INT__f_s_s_s,
378 PRIM_INT__d_s_s_s,
379 P_PA,
380 a_over_p,
381 aop_PQ,
382 one_over_2p,
383 11);
384
385
386 VRR_I_h_s_s_s(
387 PRIM_INT__h_s_s_s,
388 PRIM_INT__g_s_s_s,
389 PRIM_INT__f_s_s_s,
390 P_PA,
391 a_over_p,
392 aop_PQ,
393 one_over_2p,
394 10);
395
396
397 ostei_general_vrr1_I(6, 9,
398 one_over_2p, a_over_p, aop_PQ, P_PA,
399 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
400
401
402 ostei_general_vrr1_I(7, 8,
403 one_over_2p, a_over_p, aop_PQ, P_PA,
404 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
405
406
407 ostei_general_vrr_K(7, 0, 1, 0, 5,
408 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
409 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
410
411
412 ostei_general_vrr_K(6, 0, 1, 0, 5,
413 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
414 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
415
416
417 ostei_general_vrr_K(7, 0, 2, 0, 4,
418 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
419 PRIM_INT__k_s_p_s, PRIM_INT__k_s_s_s, NULL, PRIM_INT__i_s_p_s, NULL, PRIM_INT__k_s_d_s);
420
421
422 ostei_general_vrr_K(5, 0, 1, 0, 5,
423 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
424 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
425
426
427 ostei_general_vrr_K(6, 0, 2, 0, 4,
428 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
429 PRIM_INT__i_s_p_s, PRIM_INT__i_s_s_s, NULL, PRIM_INT__h_s_p_s, NULL, PRIM_INT__i_s_d_s);
430
431
432 ostei_general_vrr_K(7, 0, 3, 0, 3,
433 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
434 PRIM_INT__k_s_d_s, PRIM_INT__k_s_p_s, NULL, PRIM_INT__i_s_d_s, NULL, PRIM_INT__k_s_f_s);
435
436
437 VRR_K_g_s_p_s(
438 PRIM_INT__g_s_p_s,
439 PRIM_INT__g_s_s_s,
440 PRIM_INT__f_s_s_s,
441 Q_PA,
442 aoq_PQ,
443 one_over_2pq,
444 5);
445
446
447 ostei_general_vrr_K(5, 0, 2, 0, 4,
448 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
449 PRIM_INT__h_s_p_s, PRIM_INT__h_s_s_s, NULL, PRIM_INT__g_s_p_s, NULL, PRIM_INT__h_s_d_s);
450
451
452 ostei_general_vrr_K(6, 0, 3, 0, 3,
453 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
454 PRIM_INT__i_s_d_s, PRIM_INT__i_s_p_s, NULL, PRIM_INT__h_s_d_s, NULL, PRIM_INT__i_s_f_s);
455
456
457 ostei_general_vrr_K(7, 0, 4, 0, 2,
458 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
459 PRIM_INT__k_s_f_s, PRIM_INT__k_s_d_s, NULL, PRIM_INT__i_s_f_s, NULL, PRIM_INT__k_s_g_s);
460
461
462 VRR_K_f_s_p_s(
463 PRIM_INT__f_s_p_s,
464 PRIM_INT__f_s_s_s,
465 PRIM_INT__d_s_s_s,
466 Q_PA,
467 aoq_PQ,
468 one_over_2pq,
469 5);
470
471
472 ostei_general_vrr_K(4, 0, 2, 0, 4,
473 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
474 PRIM_INT__g_s_p_s, PRIM_INT__g_s_s_s, NULL, PRIM_INT__f_s_p_s, NULL, PRIM_INT__g_s_d_s);
475
476
477 ostei_general_vrr_K(5, 0, 3, 0, 3,
478 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
479 PRIM_INT__h_s_d_s, PRIM_INT__h_s_p_s, NULL, PRIM_INT__g_s_d_s, NULL, PRIM_INT__h_s_f_s);
480
481
482 ostei_general_vrr_K(6, 0, 4, 0, 2,
483 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
484 PRIM_INT__i_s_f_s, PRIM_INT__i_s_d_s, NULL, PRIM_INT__h_s_f_s, NULL, PRIM_INT__i_s_g_s);
485
486
487 ostei_general_vrr_K(7, 0, 5, 0, 1,
488 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
489 PRIM_INT__k_s_g_s, PRIM_INT__k_s_f_s, NULL, PRIM_INT__i_s_g_s, NULL, PRIM_INT__k_s_h_s);
490
491
492 ostei_general_vrr1_I(8, 7,
493 one_over_2p, a_over_p, aop_PQ, P_PA,
494 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
495
496
497 ostei_general_vrr_K(8, 0, 1, 0, 5,
498 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
499 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
500
501
502 ostei_general_vrr_K(8, 0, 2, 0, 4,
503 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
504 PRIM_INT__l_s_p_s, PRIM_INT__l_s_s_s, NULL, PRIM_INT__k_s_p_s, NULL, PRIM_INT__l_s_d_s);
505
506
507 ostei_general_vrr_K(8, 0, 3, 0, 3,
508 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
509 PRIM_INT__l_s_d_s, PRIM_INT__l_s_p_s, NULL, PRIM_INT__k_s_d_s, NULL, PRIM_INT__l_s_f_s);
510
511
512 ostei_general_vrr_K(8, 0, 4, 0, 2,
513 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
514 PRIM_INT__l_s_f_s, PRIM_INT__l_s_d_s, NULL, PRIM_INT__k_s_f_s, NULL, PRIM_INT__l_s_g_s);
515
516
517 ostei_general_vrr_K(8, 0, 5, 0, 1,
518 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
519 PRIM_INT__l_s_g_s, PRIM_INT__l_s_f_s, NULL, PRIM_INT__k_s_g_s, NULL, PRIM_INT__l_s_h_s);
520
521
522 ostei_general_vrr1_I(9, 6,
523 one_over_2p, a_over_p, aop_PQ, P_PA,
524 PRIM_INT__l_s_s_s, PRIM_INT__k_s_s_s, PRIM_INT__m_s_s_s);
525
526
527 ostei_general_vrr_K(9, 0, 1, 0, 5,
528 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
529 PRIM_INT__m_s_s_s, NULL, NULL, PRIM_INT__l_s_s_s, NULL, PRIM_INT__m_s_p_s);
530
531
532 ostei_general_vrr_K(9, 0, 2, 0, 4,
533 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
534 PRIM_INT__m_s_p_s, PRIM_INT__m_s_s_s, NULL, PRIM_INT__l_s_p_s, NULL, PRIM_INT__m_s_d_s);
535
536
537 ostei_general_vrr_K(9, 0, 3, 0, 3,
538 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
539 PRIM_INT__m_s_d_s, PRIM_INT__m_s_p_s, NULL, PRIM_INT__l_s_d_s, NULL, PRIM_INT__m_s_f_s);
540
541
542 ostei_general_vrr_K(9, 0, 4, 0, 2,
543 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
544 PRIM_INT__m_s_f_s, PRIM_INT__m_s_d_s, NULL, PRIM_INT__l_s_f_s, NULL, PRIM_INT__m_s_g_s);
545
546
547 ostei_general_vrr_K(9, 0, 5, 0, 1,
548 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
549 PRIM_INT__m_s_g_s, PRIM_INT__m_s_f_s, NULL, PRIM_INT__l_s_g_s, NULL, PRIM_INT__m_s_h_s);
550
551
552
553
554 ////////////////////////////////////
555 // Accumulate contracted integrals
556 ////////////////////////////////////
557 if(lastoffset == 0)
558 {
559 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
560 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
561 contract_all(675, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
562 contract_all(945, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
563 contract_all(825, PRIM_INT__m_s_g_s, PRIM_PTR_INT__m_s_g_s);
564 contract_all(1155, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
565 }
566 else
567 {
568 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
569 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
570 contract(675, shelloffsets, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
571 contract(945, shelloffsets, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
572 contract(825, shelloffsets, PRIM_INT__m_s_g_s, PRIM_PTR_INT__m_s_g_s);
573 contract(1155, shelloffsets, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
574 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
575 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
576 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
577 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
578 PRIM_PTR_INT__m_s_g_s += lastoffset*825;
579 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
580 }
581
582 } // close loop over j
583 } // close loop over i
584
585 //Advance to the next batch
586 jstart = SIMINT_SIMD_ROUND(jend);
587
588 //////////////////////////////////////////////
589 // Contracted integrals: Horizontal recurrance
590 //////////////////////////////////////////////
591
592
593 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
594
595
596 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
597 {
598 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
599
600 // set up HRR pointers
601 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
602 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
603 double const * restrict HRR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
604 double const * restrict HRR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
605 double const * restrict HRR_INT__m_s_g_s = INT__m_s_g_s + abcd * 825;
606 double const * restrict HRR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
607 double * restrict HRR_INT__k_d_g_p = INT__k_d_g_p + real_abcd * 9720;
608
609 // form INT__k_p_g_s
610 ostei_general_hrr_J(7, 1, 4, 0, hAB, HRR_INT__l_s_g_s, HRR_INT__k_s_g_s, HRR_INT__k_p_g_s);
611
612 // form INT__k_p_h_s
613 ostei_general_hrr_J(7, 1, 5, 0, hAB, HRR_INT__l_s_h_s, HRR_INT__k_s_h_s, HRR_INT__k_p_h_s);
614
615 // form INT__l_p_g_s
616 ostei_general_hrr_J(8, 1, 4, 0, hAB, HRR_INT__m_s_g_s, HRR_INT__l_s_g_s, HRR_INT__l_p_g_s);
617
618 // form INT__l_p_h_s
619 ostei_general_hrr_J(8, 1, 5, 0, hAB, HRR_INT__m_s_h_s, HRR_INT__l_s_h_s, HRR_INT__l_p_h_s);
620
621 // form INT__k_d_g_s
622 ostei_general_hrr_J(7, 2, 4, 0, hAB, HRR_INT__l_p_g_s, HRR_INT__k_p_g_s, HRR_INT__k_d_g_s);
623
624 // form INT__k_d_h_s
625 ostei_general_hrr_J(7, 2, 5, 0, hAB, HRR_INT__l_p_h_s, HRR_INT__k_p_h_s, HRR_INT__k_d_h_s);
626
627 // form INT__k_d_g_p
628 HRR_L_g_p(
629 HRR_INT__k_d_g_p,
630 HRR_INT__k_d_g_s,
631 HRR_INT__k_d_h_s,
632 hCD, 216);
633
634
635 } // close HRR loop
636
637
638 } // close loop cdbatch
639
640 istart = iend;
641 } // close loop over ab
642
643 return P.nshell12_clip * Q.nshell12_clip;
644 }
645
ostei_d_k_g_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_k_g_p)646 int ostei_d_k_g_p(struct simint_multi_shellpair const P,
647 struct simint_multi_shellpair const Q,
648 double screen_tol,
649 double * const restrict work,
650 double * const restrict INT__d_k_g_p)
651 {
652 double P_AB[3*P.nshell12];
653 struct simint_multi_shellpair P_tmp = P;
654 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
655 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
656 P_tmp.AB_x = P_AB;
657 P_tmp.AB_y = P_AB + P.nshell12;
658 P_tmp.AB_z = P_AB + 2*P.nshell12;
659
660 for(int i = 0; i < P.nshell12; i++)
661 {
662 P_tmp.AB_x[i] = -P.AB_x[i];
663 P_tmp.AB_y[i] = -P.AB_y[i];
664 P_tmp.AB_z[i] = -P.AB_z[i];
665 }
666
667 int ret = ostei_k_d_g_p(P_tmp, Q, screen_tol, work, INT__d_k_g_p);
668 double buffer[9720] SIMINT_ALIGN_ARRAY_DBL;
669
670 for(int q = 0; q < ret; q++)
671 {
672 int idx = 0;
673 for(int a = 0; a < 6; ++a)
674 for(int b = 0; b < 36; ++b)
675 for(int c = 0; c < 15; ++c)
676 for(int d = 0; d < 3; ++d)
677 buffer[idx++] = INT__d_k_g_p[q*9720+b*270+a*45+c*3+d];
678
679 memcpy(INT__d_k_g_p+q*9720, buffer, 9720*sizeof(double));
680 }
681
682 return ret;
683 }
684
ostei_k_d_p_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_d_p_g)685 int ostei_k_d_p_g(struct simint_multi_shellpair const P,
686 struct simint_multi_shellpair const Q,
687 double screen_tol,
688 double * const restrict work,
689 double * const restrict INT__k_d_p_g)
690 {
691 double Q_AB[3*Q.nshell12];
692 struct simint_multi_shellpair Q_tmp = Q;
693 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
694 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
695 Q_tmp.AB_x = Q_AB;
696 Q_tmp.AB_y = Q_AB + Q.nshell12;
697 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
698
699 for(int i = 0; i < Q.nshell12; i++)
700 {
701 Q_tmp.AB_x[i] = -Q.AB_x[i];
702 Q_tmp.AB_y[i] = -Q.AB_y[i];
703 Q_tmp.AB_z[i] = -Q.AB_z[i];
704 }
705
706 int ret = ostei_k_d_g_p(P, Q_tmp, screen_tol, work, INT__k_d_p_g);
707 double buffer[9720] SIMINT_ALIGN_ARRAY_DBL;
708
709 for(int q = 0; q < ret; q++)
710 {
711 int idx = 0;
712 for(int a = 0; a < 36; ++a)
713 for(int b = 0; b < 6; ++b)
714 for(int c = 0; c < 3; ++c)
715 for(int d = 0; d < 15; ++d)
716 buffer[idx++] = INT__k_d_p_g[q*9720+a*270+b*45+d*3+c];
717
718 memcpy(INT__k_d_p_g+q*9720, buffer, 9720*sizeof(double));
719 }
720
721 return ret;
722 }
723
ostei_d_k_p_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_k_p_g)724 int ostei_d_k_p_g(struct simint_multi_shellpair const P,
725 struct simint_multi_shellpair const Q,
726 double screen_tol,
727 double * const restrict work,
728 double * const restrict INT__d_k_p_g)
729 {
730 double P_AB[3*P.nshell12];
731 struct simint_multi_shellpair P_tmp = P;
732 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
733 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
734 P_tmp.AB_x = P_AB;
735 P_tmp.AB_y = P_AB + P.nshell12;
736 P_tmp.AB_z = P_AB + 2*P.nshell12;
737
738 for(int i = 0; i < P.nshell12; i++)
739 {
740 P_tmp.AB_x[i] = -P.AB_x[i];
741 P_tmp.AB_y[i] = -P.AB_y[i];
742 P_tmp.AB_z[i] = -P.AB_z[i];
743 }
744
745 double Q_AB[3*Q.nshell12];
746 struct simint_multi_shellpair Q_tmp = Q;
747 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
748 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
749 Q_tmp.AB_x = Q_AB;
750 Q_tmp.AB_y = Q_AB + Q.nshell12;
751 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
752
753 for(int i = 0; i < Q.nshell12; i++)
754 {
755 Q_tmp.AB_x[i] = -Q.AB_x[i];
756 Q_tmp.AB_y[i] = -Q.AB_y[i];
757 Q_tmp.AB_z[i] = -Q.AB_z[i];
758 }
759
760 int ret = ostei_k_d_g_p(P_tmp, Q_tmp, screen_tol, work, INT__d_k_p_g);
761 double buffer[9720] SIMINT_ALIGN_ARRAY_DBL;
762
763 for(int q = 0; q < ret; q++)
764 {
765 int idx = 0;
766 for(int a = 0; a < 6; ++a)
767 for(int b = 0; b < 36; ++b)
768 for(int c = 0; c < 3; ++c)
769 for(int d = 0; d < 15; ++d)
770 buffer[idx++] = INT__d_k_p_g[q*9720+b*270+a*45+d*3+c];
771
772 memcpy(INT__d_k_p_g+q*9720, buffer, 9720*sizeof(double));
773 }
774
775 return ret;
776 }
777
778