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_i_g_d_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_g_d_p)8 int ostei_i_g_d_p(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__i_g_d_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_g_d_p);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__i_s_d_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 168);
31 double * const INT__k_s_d_s = work + (SIMINT_NSHELL_SIMD * 448);
32 double * const INT__k_s_f_s = work + (SIMINT_NSHELL_SIMD * 664);
33 double * const INT__l_s_d_s = work + (SIMINT_NSHELL_SIMD * 1024);
34 double * const INT__l_s_f_s = work + (SIMINT_NSHELL_SIMD * 1294);
35 double * const INT__m_s_d_s = work + (SIMINT_NSHELL_SIMD * 1744);
36 double * const INT__m_s_f_s = work + (SIMINT_NSHELL_SIMD * 2074);
37 double * const INT__n_s_d_s = work + (SIMINT_NSHELL_SIMD * 2624);
38 double * const INT__n_s_f_s = work + (SIMINT_NSHELL_SIMD * 3020);
39 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3680);
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 14;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 53;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 125;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 235;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 385;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 520;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 709;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 898;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 1150;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 1374;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 1626;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 1962;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 2242;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 2494;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 2818;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 3250;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 3610;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 3880;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 4285;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 4825;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_s_s = primwork + 5275;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_p_s = primwork + 5550;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_d_s = primwork + 6045;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_f_s = primwork + 6705;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_s_s = primwork + 7255;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_p_s = primwork + 7519;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_d_s = primwork + 8113;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_f_s = primwork + 8905;
69 double * const hrrwork = (double *)(primwork + 9565);
70 double * const HRR_INT__i_p_d_s = hrrwork + 0;
71 double * const HRR_INT__i_p_f_s = hrrwork + 504;
72 double * const HRR_INT__i_d_d_s = hrrwork + 1344;
73 double * const HRR_INT__i_d_f_s = hrrwork + 2352;
74 double * const HRR_INT__i_f_d_s = hrrwork + 4032;
75 double * const HRR_INT__i_f_f_s = hrrwork + 5712;
76 double * const HRR_INT__i_g_d_s = hrrwork + 8512;
77 double * const HRR_INT__i_g_f_s = hrrwork + 11032;
78 double * const HRR_INT__k_p_d_s = hrrwork + 15232;
79 double * const HRR_INT__k_p_f_s = hrrwork + 15880;
80 double * const HRR_INT__k_d_d_s = hrrwork + 16960;
81 double * const HRR_INT__k_d_f_s = hrrwork + 18256;
82 double * const HRR_INT__k_f_d_s = hrrwork + 20416;
83 double * const HRR_INT__k_f_f_s = hrrwork + 22576;
84 double * const HRR_INT__l_p_d_s = hrrwork + 26176;
85 double * const HRR_INT__l_p_f_s = hrrwork + 26986;
86 double * const HRR_INT__l_d_d_s = hrrwork + 28336;
87 double * const HRR_INT__l_d_f_s = hrrwork + 29956;
88 double * const HRR_INT__m_p_d_s = hrrwork + 32656;
89 double * const HRR_INT__m_p_f_s = hrrwork + 33646;
90
91
92 // Create constants
93 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
94 const SIMINT_DBLTYPE const_10 = SIMINT_DBLSET1(10);
95 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
96 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
97 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
98 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
99 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
100 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
101 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
102 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
103 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
104
105
106 ////////////////////////////////////////
107 // Loop over shells and primitives
108 ////////////////////////////////////////
109
110 real_abcd = 0;
111 istart = 0;
112 for(ab = 0; ab < P.nshell12_clip; ++ab)
113 {
114 const int iend = istart + P.nprim12[ab];
115
116 cd = 0;
117 jstart = 0;
118
119 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
120 {
121 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
122 int jend = jstart;
123 for(i = 0; i < nshellbatch; i++)
124 jend += Q.nprim12[cd+i];
125
126 // Clear the beginning of the workspace (where we are accumulating integrals)
127 memset(work, 0, SIMINT_NSHELL_SIMD * 3680 * sizeof(double));
128 abcd = 0;
129
130
131 for(i = istart; i < iend; ++i)
132 {
133 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
134
135 if(check_screen)
136 {
137 // Skip this whole thing if always insignificant
138 if((P.screen[i] * Q.screen_max) < screen_tol)
139 continue;
140 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
141 }
142
143 icd = 0;
144 iprimcd = 0;
145 nprim_icd = Q.nprim12[cd];
146 double * restrict PRIM_PTR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
147 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
148 double * restrict PRIM_PTR_INT__k_s_d_s = INT__k_s_d_s + abcd * 216;
149 double * restrict PRIM_PTR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
150 double * restrict PRIM_PTR_INT__l_s_d_s = INT__l_s_d_s + abcd * 270;
151 double * restrict PRIM_PTR_INT__l_s_f_s = INT__l_s_f_s + abcd * 450;
152 double * restrict PRIM_PTR_INT__m_s_d_s = INT__m_s_d_s + abcd * 330;
153 double * restrict PRIM_PTR_INT__m_s_f_s = INT__m_s_f_s + abcd * 550;
154 double * restrict PRIM_PTR_INT__n_s_d_s = INT__n_s_d_s + abcd * 396;
155 double * restrict PRIM_PTR_INT__n_s_f_s = INT__n_s_f_s + abcd * 660;
156
157
158
159 // Load these one per loop over i
160 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
161 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
162 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
163
164 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
165
166 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
167 {
168 // calculate the shell offsets
169 // these are the offset from the shell pointed to by cd
170 // for each element
171 int shelloffsets[SIMINT_SIMD_LEN] = {0};
172 int lastoffset = 0;
173 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
174
175 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
176 {
177 // Handle if the first element of the vector is a new shell
178 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
179 {
180 nprim_icd += Q.nprim12[cd + (++icd)];
181 PRIM_PTR_INT__i_s_d_s += 168;
182 PRIM_PTR_INT__i_s_f_s += 280;
183 PRIM_PTR_INT__k_s_d_s += 216;
184 PRIM_PTR_INT__k_s_f_s += 360;
185 PRIM_PTR_INT__l_s_d_s += 270;
186 PRIM_PTR_INT__l_s_f_s += 450;
187 PRIM_PTR_INT__m_s_d_s += 330;
188 PRIM_PTR_INT__m_s_f_s += 550;
189 PRIM_PTR_INT__n_s_d_s += 396;
190 PRIM_PTR_INT__n_s_f_s += 660;
191 }
192 iprimcd++;
193 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
194 {
195 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
196 {
197 shelloffsets[n] = shelloffsets[n-1] + 1;
198 lastoffset++;
199 nprim_icd += Q.nprim12[cd + (++icd)];
200 }
201 else
202 shelloffsets[n] = shelloffsets[n-1];
203 iprimcd++;
204 }
205 }
206 else
207 iprimcd += SIMINT_SIMD_LEN;
208
209 // Do we have to compute this vector (or has it been screened out)?
210 // (not_screened != 0 means we have to do this vector)
211 if(check_screen)
212 {
213 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
214 if(vmax < screen_tol)
215 {
216 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
217 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
218 PRIM_PTR_INT__k_s_d_s += lastoffset*216;
219 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
220 PRIM_PTR_INT__l_s_d_s += lastoffset*270;
221 PRIM_PTR_INT__l_s_f_s += lastoffset*450;
222 PRIM_PTR_INT__m_s_d_s += lastoffset*330;
223 PRIM_PTR_INT__m_s_f_s += lastoffset*550;
224 PRIM_PTR_INT__n_s_d_s += lastoffset*396;
225 PRIM_PTR_INT__n_s_f_s += lastoffset*660;
226 continue;
227 }
228 }
229
230 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
231 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
232 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
233 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
234
235
236 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
237 SIMINT_DBLTYPE PQ[3];
238 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
239 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
240 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
241 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
242 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
243 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
244
245 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
246 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
247 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
248 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
249 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
250 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
251 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
252
253 // NOTE: Minus sign!
254 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
255 SIMINT_DBLTYPE aop_PQ[3];
256 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
257 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
258 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
259
260 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
261 SIMINT_DBLTYPE aoq_PQ[3];
262 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
263 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
264 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
265 // Put a minus sign here so we don't have to in RR routines
266 a_over_q = SIMINT_NEG(a_over_q);
267
268
269 //////////////////////////////////////////////
270 // Fjt function section
271 // Maximum v value: 13
272 //////////////////////////////////////////////
273 // The parameter to the Fjt function
274 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
275
276
277 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
278
279
280 boys_F_split(PRIM_INT__s_s_s_s, F_x, 13);
281 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
282 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
283 for(n = 0; n <= 13; n++)
284 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
285
286 //////////////////////////////////////////////
287 // Primitive integrals: Vertical recurrance
288 //////////////////////////////////////////////
289
290 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
291 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
292 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
293 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
294 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
295 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
296 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
297 const SIMINT_DBLTYPE vrr_const_8_over_2p = SIMINT_MUL(const_8, one_over_2p);
298 const SIMINT_DBLTYPE vrr_const_9_over_2p = SIMINT_MUL(const_9, one_over_2p);
299 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
300 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
301 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
302 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
303 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
304 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
305 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
306 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
307 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
308 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
309 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
310 const SIMINT_DBLTYPE vrr_const_10_over_2pq = SIMINT_MUL(const_10, one_over_2pq);
311
312
313
314 // Forming PRIM_INT__p_s_s_s[13 * 3];
315 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
316 {
317
318 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
319 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]);
320
321 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
322 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]);
323
324 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
325 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]);
326
327 }
328
329
330
331 // Forming PRIM_INT__d_s_s_s[12 * 6];
332 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
333 {
334
335 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
336 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]);
337 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]);
338
339 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
340 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]);
341
342 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
343 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]);
344 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]);
345
346 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
347 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]);
348 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]);
349
350 }
351
352
353
354 // Forming PRIM_INT__f_s_s_s[11 * 10];
355 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
356 {
357
358 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
359 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]);
360 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]);
361
362 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
363 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]);
364
365 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
366 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]);
367
368 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
369 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]);
370
371 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
372 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]);
373
374 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
375 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]);
376
377 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
378 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]);
379 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]);
380
381 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
382 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]);
383
384 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
385 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]);
386
387 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
388 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]);
389 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]);
390
391 }
392
393
394 VRR_I_g_s_s_s(
395 PRIM_INT__g_s_s_s,
396 PRIM_INT__f_s_s_s,
397 PRIM_INT__d_s_s_s,
398 P_PA,
399 a_over_p,
400 aop_PQ,
401 one_over_2p,
402 10);
403
404
405 VRR_I_h_s_s_s(
406 PRIM_INT__h_s_s_s,
407 PRIM_INT__g_s_s_s,
408 PRIM_INT__f_s_s_s,
409 P_PA,
410 a_over_p,
411 aop_PQ,
412 one_over_2p,
413 9);
414
415
416 ostei_general_vrr1_I(6, 8,
417 one_over_2p, a_over_p, aop_PQ, P_PA,
418 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
419
420
421 ostei_general_vrr_K(6, 0, 1, 0, 3,
422 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
423 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
424
425
426 ostei_general_vrr_K(5, 0, 1, 0, 3,
427 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
428 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
429
430
431 ostei_general_vrr_K(6, 0, 2, 0, 2,
432 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
433 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);
434
435
436 VRR_K_g_s_p_s(
437 PRIM_INT__g_s_p_s,
438 PRIM_INT__g_s_s_s,
439 PRIM_INT__f_s_s_s,
440 Q_PA,
441 aoq_PQ,
442 one_over_2pq,
443 3);
444
445
446 ostei_general_vrr_K(5, 0, 2, 0, 2,
447 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
448 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);
449
450
451 ostei_general_vrr_K(6, 0, 3, 0, 1,
452 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
453 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);
454
455
456 ostei_general_vrr1_I(7, 7,
457 one_over_2p, a_over_p, aop_PQ, P_PA,
458 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
459
460
461 ostei_general_vrr_K(7, 0, 1, 0, 3,
462 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
463 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
464
465
466 ostei_general_vrr_K(7, 0, 2, 0, 2,
467 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
468 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);
469
470
471 ostei_general_vrr_K(7, 0, 3, 0, 1,
472 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
473 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);
474
475
476 ostei_general_vrr1_I(8, 6,
477 one_over_2p, a_over_p, aop_PQ, P_PA,
478 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
479
480
481 ostei_general_vrr_K(8, 0, 1, 0, 3,
482 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
483 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
484
485
486 ostei_general_vrr_K(8, 0, 2, 0, 2,
487 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
488 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);
489
490
491 ostei_general_vrr_K(8, 0, 3, 0, 1,
492 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
493 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);
494
495
496 ostei_general_vrr1_I(9, 5,
497 one_over_2p, a_over_p, aop_PQ, P_PA,
498 PRIM_INT__l_s_s_s, PRIM_INT__k_s_s_s, PRIM_INT__m_s_s_s);
499
500
501 ostei_general_vrr_K(9, 0, 1, 0, 3,
502 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
503 PRIM_INT__m_s_s_s, NULL, NULL, PRIM_INT__l_s_s_s, NULL, PRIM_INT__m_s_p_s);
504
505
506 ostei_general_vrr_K(9, 0, 2, 0, 2,
507 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
508 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);
509
510
511 ostei_general_vrr_K(9, 0, 3, 0, 1,
512 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
513 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);
514
515
516 ostei_general_vrr1_I(10, 4,
517 one_over_2p, a_over_p, aop_PQ, P_PA,
518 PRIM_INT__m_s_s_s, PRIM_INT__l_s_s_s, PRIM_INT__n_s_s_s);
519
520
521 ostei_general_vrr_K(10, 0, 1, 0, 3,
522 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
523 PRIM_INT__n_s_s_s, NULL, NULL, PRIM_INT__m_s_s_s, NULL, PRIM_INT__n_s_p_s);
524
525
526 ostei_general_vrr_K(10, 0, 2, 0, 2,
527 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
528 PRIM_INT__n_s_p_s, PRIM_INT__n_s_s_s, NULL, PRIM_INT__m_s_p_s, NULL, PRIM_INT__n_s_d_s);
529
530
531 ostei_general_vrr_K(10, 0, 3, 0, 1,
532 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
533 PRIM_INT__n_s_d_s, PRIM_INT__n_s_p_s, NULL, PRIM_INT__m_s_d_s, NULL, PRIM_INT__n_s_f_s);
534
535
536
537
538 ////////////////////////////////////
539 // Accumulate contracted integrals
540 ////////////////////////////////////
541 if(lastoffset == 0)
542 {
543 contract_all(168, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
544 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
545 contract_all(216, PRIM_INT__k_s_d_s, PRIM_PTR_INT__k_s_d_s);
546 contract_all(360, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
547 contract_all(270, PRIM_INT__l_s_d_s, PRIM_PTR_INT__l_s_d_s);
548 contract_all(450, PRIM_INT__l_s_f_s, PRIM_PTR_INT__l_s_f_s);
549 contract_all(330, PRIM_INT__m_s_d_s, PRIM_PTR_INT__m_s_d_s);
550 contract_all(550, PRIM_INT__m_s_f_s, PRIM_PTR_INT__m_s_f_s);
551 contract_all(396, PRIM_INT__n_s_d_s, PRIM_PTR_INT__n_s_d_s);
552 contract_all(660, PRIM_INT__n_s_f_s, PRIM_PTR_INT__n_s_f_s);
553 }
554 else
555 {
556 contract(168, shelloffsets, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
557 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
558 contract(216, shelloffsets, PRIM_INT__k_s_d_s, PRIM_PTR_INT__k_s_d_s);
559 contract(360, shelloffsets, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
560 contract(270, shelloffsets, PRIM_INT__l_s_d_s, PRIM_PTR_INT__l_s_d_s);
561 contract(450, shelloffsets, PRIM_INT__l_s_f_s, PRIM_PTR_INT__l_s_f_s);
562 contract(330, shelloffsets, PRIM_INT__m_s_d_s, PRIM_PTR_INT__m_s_d_s);
563 contract(550, shelloffsets, PRIM_INT__m_s_f_s, PRIM_PTR_INT__m_s_f_s);
564 contract(396, shelloffsets, PRIM_INT__n_s_d_s, PRIM_PTR_INT__n_s_d_s);
565 contract(660, shelloffsets, PRIM_INT__n_s_f_s, PRIM_PTR_INT__n_s_f_s);
566 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
567 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
568 PRIM_PTR_INT__k_s_d_s += lastoffset*216;
569 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
570 PRIM_PTR_INT__l_s_d_s += lastoffset*270;
571 PRIM_PTR_INT__l_s_f_s += lastoffset*450;
572 PRIM_PTR_INT__m_s_d_s += lastoffset*330;
573 PRIM_PTR_INT__m_s_f_s += lastoffset*550;
574 PRIM_PTR_INT__n_s_d_s += lastoffset*396;
575 PRIM_PTR_INT__n_s_f_s += lastoffset*660;
576 }
577
578 } // close loop over j
579 } // close loop over i
580
581 //Advance to the next batch
582 jstart = SIMINT_SIMD_ROUND(jend);
583
584 //////////////////////////////////////////////
585 // Contracted integrals: Horizontal recurrance
586 //////////////////////////////////////////////
587
588
589 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
590
591
592 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
593 {
594 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
595
596 // set up HRR pointers
597 double const * restrict HRR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
598 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
599 double const * restrict HRR_INT__k_s_d_s = INT__k_s_d_s + abcd * 216;
600 double const * restrict HRR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
601 double const * restrict HRR_INT__l_s_d_s = INT__l_s_d_s + abcd * 270;
602 double const * restrict HRR_INT__l_s_f_s = INT__l_s_f_s + abcd * 450;
603 double const * restrict HRR_INT__m_s_d_s = INT__m_s_d_s + abcd * 330;
604 double const * restrict HRR_INT__m_s_f_s = INT__m_s_f_s + abcd * 550;
605 double const * restrict HRR_INT__n_s_d_s = INT__n_s_d_s + abcd * 396;
606 double const * restrict HRR_INT__n_s_f_s = INT__n_s_f_s + abcd * 660;
607 double * restrict HRR_INT__i_g_d_p = INT__i_g_d_p + real_abcd * 7560;
608
609 // form INT__i_p_d_s
610 ostei_general_hrr_J(6, 1, 2, 0, hAB, HRR_INT__k_s_d_s, HRR_INT__i_s_d_s, HRR_INT__i_p_d_s);
611
612 // form INT__i_p_f_s
613 ostei_general_hrr_J(6, 1, 3, 0, hAB, HRR_INT__k_s_f_s, HRR_INT__i_s_f_s, HRR_INT__i_p_f_s);
614
615 // form INT__k_p_d_s
616 ostei_general_hrr_J(7, 1, 2, 0, hAB, HRR_INT__l_s_d_s, HRR_INT__k_s_d_s, HRR_INT__k_p_d_s);
617
618 // form INT__k_p_f_s
619 ostei_general_hrr_J(7, 1, 3, 0, hAB, HRR_INT__l_s_f_s, HRR_INT__k_s_f_s, HRR_INT__k_p_f_s);
620
621 // form INT__l_p_d_s
622 ostei_general_hrr_J(8, 1, 2, 0, hAB, HRR_INT__m_s_d_s, HRR_INT__l_s_d_s, HRR_INT__l_p_d_s);
623
624 // form INT__l_p_f_s
625 ostei_general_hrr_J(8, 1, 3, 0, hAB, HRR_INT__m_s_f_s, HRR_INT__l_s_f_s, HRR_INT__l_p_f_s);
626
627 // form INT__m_p_d_s
628 ostei_general_hrr_J(9, 1, 2, 0, hAB, HRR_INT__n_s_d_s, HRR_INT__m_s_d_s, HRR_INT__m_p_d_s);
629
630 // form INT__m_p_f_s
631 ostei_general_hrr_J(9, 1, 3, 0, hAB, HRR_INT__n_s_f_s, HRR_INT__m_s_f_s, HRR_INT__m_p_f_s);
632
633 // form INT__i_d_d_s
634 ostei_general_hrr_J(6, 2, 2, 0, hAB, HRR_INT__k_p_d_s, HRR_INT__i_p_d_s, HRR_INT__i_d_d_s);
635
636 // form INT__i_d_f_s
637 ostei_general_hrr_J(6, 2, 3, 0, hAB, HRR_INT__k_p_f_s, HRR_INT__i_p_f_s, HRR_INT__i_d_f_s);
638
639 // form INT__k_d_d_s
640 ostei_general_hrr_J(7, 2, 2, 0, hAB, HRR_INT__l_p_d_s, HRR_INT__k_p_d_s, HRR_INT__k_d_d_s);
641
642 // form INT__k_d_f_s
643 ostei_general_hrr_J(7, 2, 3, 0, hAB, HRR_INT__l_p_f_s, HRR_INT__k_p_f_s, HRR_INT__k_d_f_s);
644
645 // form INT__l_d_d_s
646 ostei_general_hrr_J(8, 2, 2, 0, hAB, HRR_INT__m_p_d_s, HRR_INT__l_p_d_s, HRR_INT__l_d_d_s);
647
648 // form INT__l_d_f_s
649 ostei_general_hrr_J(8, 2, 3, 0, hAB, HRR_INT__m_p_f_s, HRR_INT__l_p_f_s, HRR_INT__l_d_f_s);
650
651 // form INT__i_f_d_s
652 ostei_general_hrr_J(6, 3, 2, 0, hAB, HRR_INT__k_d_d_s, HRR_INT__i_d_d_s, HRR_INT__i_f_d_s);
653
654 // form INT__i_f_f_s
655 ostei_general_hrr_J(6, 3, 3, 0, hAB, HRR_INT__k_d_f_s, HRR_INT__i_d_f_s, HRR_INT__i_f_f_s);
656
657 // form INT__k_f_d_s
658 ostei_general_hrr_J(7, 3, 2, 0, hAB, HRR_INT__l_d_d_s, HRR_INT__k_d_d_s, HRR_INT__k_f_d_s);
659
660 // form INT__k_f_f_s
661 ostei_general_hrr_J(7, 3, 3, 0, hAB, HRR_INT__l_d_f_s, HRR_INT__k_d_f_s, HRR_INT__k_f_f_s);
662
663 // form INT__i_g_d_s
664 ostei_general_hrr_J(6, 4, 2, 0, hAB, HRR_INT__k_f_d_s, HRR_INT__i_f_d_s, HRR_INT__i_g_d_s);
665
666 // form INT__i_g_f_s
667 ostei_general_hrr_J(6, 4, 3, 0, hAB, HRR_INT__k_f_f_s, HRR_INT__i_f_f_s, HRR_INT__i_g_f_s);
668
669 // form INT__i_g_d_p
670 for(ibra = 0; ibra < 420; ++ibra)
671 {
672 HRR_INT__i_g_d_p[ibra * 18 + 0] = HRR_INT__i_g_f_s[ibra * 10 + 0] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 0] );
673
674 HRR_INT__i_g_d_p[ibra * 18 + 1] = HRR_INT__i_g_f_s[ibra * 10 + 1] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 0] );
675
676 HRR_INT__i_g_d_p[ibra * 18 + 2] = HRR_INT__i_g_f_s[ibra * 10 + 2] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 0] );
677
678 HRR_INT__i_g_d_p[ibra * 18 + 3] = HRR_INT__i_g_f_s[ibra * 10 + 1] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 1] );
679
680 HRR_INT__i_g_d_p[ibra * 18 + 4] = HRR_INT__i_g_f_s[ibra * 10 + 3] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 1] );
681
682 HRR_INT__i_g_d_p[ibra * 18 + 5] = HRR_INT__i_g_f_s[ibra * 10 + 4] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 1] );
683
684 HRR_INT__i_g_d_p[ibra * 18 + 6] = HRR_INT__i_g_f_s[ibra * 10 + 2] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 2] );
685
686 HRR_INT__i_g_d_p[ibra * 18 + 7] = HRR_INT__i_g_f_s[ibra * 10 + 4] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 2] );
687
688 HRR_INT__i_g_d_p[ibra * 18 + 8] = HRR_INT__i_g_f_s[ibra * 10 + 5] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 2] );
689
690 HRR_INT__i_g_d_p[ibra * 18 + 9] = HRR_INT__i_g_f_s[ibra * 10 + 3] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 3] );
691
692 HRR_INT__i_g_d_p[ibra * 18 + 10] = HRR_INT__i_g_f_s[ibra * 10 + 6] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 3] );
693
694 HRR_INT__i_g_d_p[ibra * 18 + 11] = HRR_INT__i_g_f_s[ibra * 10 + 7] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 3] );
695
696 HRR_INT__i_g_d_p[ibra * 18 + 12] = HRR_INT__i_g_f_s[ibra * 10 + 4] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 4] );
697
698 HRR_INT__i_g_d_p[ibra * 18 + 13] = HRR_INT__i_g_f_s[ibra * 10 + 7] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 4] );
699
700 HRR_INT__i_g_d_p[ibra * 18 + 14] = HRR_INT__i_g_f_s[ibra * 10 + 8] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 4] );
701
702 HRR_INT__i_g_d_p[ibra * 18 + 15] = HRR_INT__i_g_f_s[ibra * 10 + 5] + ( hCD[0] * HRR_INT__i_g_d_s[ibra * 6 + 5] );
703
704 HRR_INT__i_g_d_p[ibra * 18 + 16] = HRR_INT__i_g_f_s[ibra * 10 + 8] + ( hCD[1] * HRR_INT__i_g_d_s[ibra * 6 + 5] );
705
706 HRR_INT__i_g_d_p[ibra * 18 + 17] = HRR_INT__i_g_f_s[ibra * 10 + 9] + ( hCD[2] * HRR_INT__i_g_d_s[ibra * 6 + 5] );
707
708 }
709
710
711 } // close HRR loop
712
713
714 } // close loop cdbatch
715
716 istart = iend;
717 } // close loop over ab
718
719 return P.nshell12_clip * Q.nshell12_clip;
720 }
721
ostei_g_i_d_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_i_d_p)722 int ostei_g_i_d_p(struct simint_multi_shellpair const P,
723 struct simint_multi_shellpair const Q,
724 double screen_tol,
725 double * const restrict work,
726 double * const restrict INT__g_i_d_p)
727 {
728 double P_AB[3*P.nshell12];
729 struct simint_multi_shellpair P_tmp = P;
730 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
731 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
732 P_tmp.AB_x = P_AB;
733 P_tmp.AB_y = P_AB + P.nshell12;
734 P_tmp.AB_z = P_AB + 2*P.nshell12;
735
736 for(int i = 0; i < P.nshell12; i++)
737 {
738 P_tmp.AB_x[i] = -P.AB_x[i];
739 P_tmp.AB_y[i] = -P.AB_y[i];
740 P_tmp.AB_z[i] = -P.AB_z[i];
741 }
742
743 int ret = ostei_i_g_d_p(P_tmp, Q, screen_tol, work, INT__g_i_d_p);
744 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
745
746 for(int q = 0; q < ret; q++)
747 {
748 int idx = 0;
749 for(int a = 0; a < 15; ++a)
750 for(int b = 0; b < 28; ++b)
751 for(int c = 0; c < 6; ++c)
752 for(int d = 0; d < 3; ++d)
753 buffer[idx++] = INT__g_i_d_p[q*7560+b*270+a*18+c*3+d];
754
755 memcpy(INT__g_i_d_p+q*7560, buffer, 7560*sizeof(double));
756 }
757
758 return ret;
759 }
760
ostei_i_g_p_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_g_p_d)761 int ostei_i_g_p_d(struct simint_multi_shellpair const P,
762 struct simint_multi_shellpair const Q,
763 double screen_tol,
764 double * const restrict work,
765 double * const restrict INT__i_g_p_d)
766 {
767 double Q_AB[3*Q.nshell12];
768 struct simint_multi_shellpair Q_tmp = Q;
769 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
770 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
771 Q_tmp.AB_x = Q_AB;
772 Q_tmp.AB_y = Q_AB + Q.nshell12;
773 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
774
775 for(int i = 0; i < Q.nshell12; i++)
776 {
777 Q_tmp.AB_x[i] = -Q.AB_x[i];
778 Q_tmp.AB_y[i] = -Q.AB_y[i];
779 Q_tmp.AB_z[i] = -Q.AB_z[i];
780 }
781
782 int ret = ostei_i_g_d_p(P, Q_tmp, screen_tol, work, INT__i_g_p_d);
783 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
784
785 for(int q = 0; q < ret; q++)
786 {
787 int idx = 0;
788 for(int a = 0; a < 28; ++a)
789 for(int b = 0; b < 15; ++b)
790 for(int c = 0; c < 3; ++c)
791 for(int d = 0; d < 6; ++d)
792 buffer[idx++] = INT__i_g_p_d[q*7560+a*270+b*18+d*3+c];
793
794 memcpy(INT__i_g_p_d+q*7560, buffer, 7560*sizeof(double));
795 }
796
797 return ret;
798 }
799
ostei_g_i_p_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_i_p_d)800 int ostei_g_i_p_d(struct simint_multi_shellpair const P,
801 struct simint_multi_shellpair const Q,
802 double screen_tol,
803 double * const restrict work,
804 double * const restrict INT__g_i_p_d)
805 {
806 double P_AB[3*P.nshell12];
807 struct simint_multi_shellpair P_tmp = P;
808 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
809 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
810 P_tmp.AB_x = P_AB;
811 P_tmp.AB_y = P_AB + P.nshell12;
812 P_tmp.AB_z = P_AB + 2*P.nshell12;
813
814 for(int i = 0; i < P.nshell12; i++)
815 {
816 P_tmp.AB_x[i] = -P.AB_x[i];
817 P_tmp.AB_y[i] = -P.AB_y[i];
818 P_tmp.AB_z[i] = -P.AB_z[i];
819 }
820
821 double Q_AB[3*Q.nshell12];
822 struct simint_multi_shellpair Q_tmp = Q;
823 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
824 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
825 Q_tmp.AB_x = Q_AB;
826 Q_tmp.AB_y = Q_AB + Q.nshell12;
827 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
828
829 for(int i = 0; i < Q.nshell12; i++)
830 {
831 Q_tmp.AB_x[i] = -Q.AB_x[i];
832 Q_tmp.AB_y[i] = -Q.AB_y[i];
833 Q_tmp.AB_z[i] = -Q.AB_z[i];
834 }
835
836 int ret = ostei_i_g_d_p(P_tmp, Q_tmp, screen_tol, work, INT__g_i_p_d);
837 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
838
839 for(int q = 0; q < ret; q++)
840 {
841 int idx = 0;
842 for(int a = 0; a < 15; ++a)
843 for(int b = 0; b < 28; ++b)
844 for(int c = 0; c < 3; ++c)
845 for(int d = 0; d < 6; ++d)
846 buffer[idx++] = INT__g_i_p_d[q*7560+b*270+a*18+d*3+c];
847
848 memcpy(INT__g_i_p_d+q*7560, buffer, 7560*sizeof(double));
849 }
850
851 return ret;
852 }
853
854