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_f_f_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__f_f_d_p)8 int ostei_f_f_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__f_f_d_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__f_f_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__f_s_d_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__f_s_f_s = work + (SIMINT_NSHELL_SIMD * 60);
31 double * const INT__g_s_d_s = work + (SIMINT_NSHELL_SIMD * 160);
32 double * const INT__g_s_f_s = work + (SIMINT_NSHELL_SIMD * 250);
33 double * const INT__h_s_d_s = work + (SIMINT_NSHELL_SIMD * 400);
34 double * const INT__h_s_f_s = work + (SIMINT_NSHELL_SIMD * 526);
35 double * const INT__i_s_d_s = work + (SIMINT_NSHELL_SIMD * 736);
36 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 904);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*1184);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 10;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 37;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 64;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 112;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 166;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 238;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 308;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 398;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 518;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 618;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 708;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 843;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1023;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1173;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 1278;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 1467;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 1719;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 1929;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 2041;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 2293;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 2629;
60 double * const hrrwork = (double *)(primwork + 2909);
61 double * const HRR_INT__f_p_d_s = hrrwork + 0;
62 double * const HRR_INT__f_p_f_s = hrrwork + 180;
63 double * const HRR_INT__f_d_d_s = hrrwork + 480;
64 double * const HRR_INT__f_d_f_s = hrrwork + 840;
65 double * const HRR_INT__f_f_d_s = hrrwork + 1440;
66 double * const HRR_INT__f_f_f_s = hrrwork + 2040;
67 double * const HRR_INT__g_p_d_s = hrrwork + 3040;
68 double * const HRR_INT__g_p_f_s = hrrwork + 3310;
69 double * const HRR_INT__g_d_d_s = hrrwork + 3760;
70 double * const HRR_INT__g_d_f_s = hrrwork + 4300;
71 double * const HRR_INT__h_p_d_s = hrrwork + 5200;
72 double * const HRR_INT__h_p_f_s = hrrwork + 5578;
73
74
75 // Create constants
76 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
77 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
78 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
79 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
80 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
81 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
82 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
83
84
85 ////////////////////////////////////////
86 // Loop over shells and primitives
87 ////////////////////////////////////////
88
89 real_abcd = 0;
90 istart = 0;
91 for(ab = 0; ab < P.nshell12_clip; ++ab)
92 {
93 const int iend = istart + P.nprim12[ab];
94
95 cd = 0;
96 jstart = 0;
97
98 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
99 {
100 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
101 int jend = jstart;
102 for(i = 0; i < nshellbatch; i++)
103 jend += Q.nprim12[cd+i];
104
105 // Clear the beginning of the workspace (where we are accumulating integrals)
106 memset(work, 0, SIMINT_NSHELL_SIMD * 1184 * sizeof(double));
107 abcd = 0;
108
109
110 for(i = istart; i < iend; ++i)
111 {
112 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
113
114 if(check_screen)
115 {
116 // Skip this whole thing if always insignificant
117 if((P.screen[i] * Q.screen_max) < screen_tol)
118 continue;
119 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
120 }
121
122 icd = 0;
123 iprimcd = 0;
124 nprim_icd = Q.nprim12[cd];
125 double * restrict PRIM_PTR_INT__f_s_d_s = INT__f_s_d_s + abcd * 60;
126 double * restrict PRIM_PTR_INT__f_s_f_s = INT__f_s_f_s + abcd * 100;
127 double * restrict PRIM_PTR_INT__g_s_d_s = INT__g_s_d_s + abcd * 90;
128 double * restrict PRIM_PTR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
129 double * restrict PRIM_PTR_INT__h_s_d_s = INT__h_s_d_s + abcd * 126;
130 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
131 double * restrict PRIM_PTR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
132 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
133
134
135
136 // Load these one per loop over i
137 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
138 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
139 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
140
141 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
142
143 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
144 {
145 // calculate the shell offsets
146 // these are the offset from the shell pointed to by cd
147 // for each element
148 int shelloffsets[SIMINT_SIMD_LEN] = {0};
149 int lastoffset = 0;
150 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
151
152 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
153 {
154 // Handle if the first element of the vector is a new shell
155 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
156 {
157 nprim_icd += Q.nprim12[cd + (++icd)];
158 PRIM_PTR_INT__f_s_d_s += 60;
159 PRIM_PTR_INT__f_s_f_s += 100;
160 PRIM_PTR_INT__g_s_d_s += 90;
161 PRIM_PTR_INT__g_s_f_s += 150;
162 PRIM_PTR_INT__h_s_d_s += 126;
163 PRIM_PTR_INT__h_s_f_s += 210;
164 PRIM_PTR_INT__i_s_d_s += 168;
165 PRIM_PTR_INT__i_s_f_s += 280;
166 }
167 iprimcd++;
168 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
169 {
170 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
171 {
172 shelloffsets[n] = shelloffsets[n-1] + 1;
173 lastoffset++;
174 nprim_icd += Q.nprim12[cd + (++icd)];
175 }
176 else
177 shelloffsets[n] = shelloffsets[n-1];
178 iprimcd++;
179 }
180 }
181 else
182 iprimcd += SIMINT_SIMD_LEN;
183
184 // Do we have to compute this vector (or has it been screened out)?
185 // (not_screened != 0 means we have to do this vector)
186 if(check_screen)
187 {
188 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
189 if(vmax < screen_tol)
190 {
191 PRIM_PTR_INT__f_s_d_s += lastoffset*60;
192 PRIM_PTR_INT__f_s_f_s += lastoffset*100;
193 PRIM_PTR_INT__g_s_d_s += lastoffset*90;
194 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
195 PRIM_PTR_INT__h_s_d_s += lastoffset*126;
196 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
197 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
198 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
199 continue;
200 }
201 }
202
203 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
204 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
205 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
206 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
207
208
209 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
210 SIMINT_DBLTYPE PQ[3];
211 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
212 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
213 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
214 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
215 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
216 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
217
218 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
219 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
220 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
221 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
222 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
223 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
224 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
225
226 // NOTE: Minus sign!
227 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
228 SIMINT_DBLTYPE aop_PQ[3];
229 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
230 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
231 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
232
233 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
234 SIMINT_DBLTYPE aoq_PQ[3];
235 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
236 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
237 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
238 // Put a minus sign here so we don't have to in RR routines
239 a_over_q = SIMINT_NEG(a_over_q);
240
241
242 //////////////////////////////////////////////
243 // Fjt function section
244 // Maximum v value: 9
245 //////////////////////////////////////////////
246 // The parameter to the Fjt function
247 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
248
249
250 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
251
252
253 boys_F_split(PRIM_INT__s_s_s_s, F_x, 9);
254 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
255 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
256 for(n = 0; n <= 9; n++)
257 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
258
259 //////////////////////////////////////////////
260 // Primitive integrals: Vertical recurrance
261 //////////////////////////////////////////////
262
263 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
264 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
265 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
266 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
267 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
268 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
269 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
270 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
271 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
272 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
273 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
274 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
275 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
276
277
278
279 // Forming PRIM_INT__p_s_s_s[9 * 3];
280 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
281 {
282
283 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
284 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]);
285
286 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
287 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]);
288
289 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
290 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]);
291
292 }
293
294
295
296 // Forming PRIM_INT__d_s_s_s[8 * 6];
297 for(n = 0; n < 8; ++n) // loop over orders of auxiliary function
298 {
299
300 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
301 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]);
302 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]);
303
304 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
305 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]);
306
307 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
308 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]);
309
310 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
311 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]);
312 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]);
313
314 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
315 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]);
316
317 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
318 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]);
319 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]);
320
321 }
322
323
324
325 // Forming PRIM_INT__f_s_s_s[7 * 10];
326 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
327 {
328
329 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
330 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]);
331 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]);
332
333 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
334 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]);
335
336 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
337 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]);
338
339 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
340 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]);
341
342 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
343 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]);
344
345 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
346 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]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
349 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]);
350 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]);
351
352 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
353 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]);
354
355 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
356 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]);
357
358 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
359 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]);
360 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]);
361
362 }
363
364
365 VRR_K_f_s_p_s(
366 PRIM_INT__f_s_p_s,
367 PRIM_INT__f_s_s_s,
368 PRIM_INT__d_s_s_s,
369 Q_PA,
370 aoq_PQ,
371 one_over_2pq,
372 3);
373
374
375
376 // Forming PRIM_INT__d_s_p_s[3 * 18];
377 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
378 {
379
380 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
381 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
382 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 0]);
383
384 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
385 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 1]);
386
387 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
388 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 0], PRIM_INT__d_s_p_s[n * 18 + 2]);
389
390 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
391 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
392 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 3]);
393
394 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
395 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 4]);
396 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 4]);
397
398 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
399 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 1], PRIM_INT__d_s_p_s[n * 18 + 5]);
400
401 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
402 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
403 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 6]);
404
405 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
406 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 7]);
407
408 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
409 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 2], PRIM_INT__d_s_p_s[n * 18 + 8]);
410 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__d_s_p_s[n * 18 + 8]);
411
412 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
413 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 9]);
414
415 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
416 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 10]);
417 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 10]);
418
419 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
420 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 3], PRIM_INT__d_s_p_s[n * 18 + 11]);
421
422 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
423 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 12]);
424
425 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
426 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 13]);
427 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 13]);
428
429 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
430 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 4], PRIM_INT__d_s_p_s[n * 18 + 14]);
431 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__d_s_p_s[n * 18 + 14]);
432
433 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
434 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 15]);
435
436 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
437 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 16]);
438
439 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
440 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__d_s_s_s[(n+1) * 6 + 5], PRIM_INT__d_s_p_s[n * 18 + 17]);
441 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__d_s_p_s[n * 18 + 17]);
442
443 }
444
445
446 VRR_K_f_s_d_s(
447 PRIM_INT__f_s_d_s,
448 PRIM_INT__f_s_p_s,
449 PRIM_INT__f_s_s_s,
450 PRIM_INT__d_s_p_s,
451 Q_PA,
452 a_over_q,
453 aoq_PQ,
454 one_over_2pq,
455 one_over_2q,
456 2);
457
458
459
460 // Forming PRIM_INT__p_s_p_s[3 * 9];
461 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
462 {
463
464 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
465 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
466 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
467
468 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
469 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 1]);
470
471 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
472 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 2]);
473
474 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
475 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 3]);
476
477 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
478 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 4]);
479 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 4]);
480
481 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
482 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 5]);
483
484 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
485 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 6]);
486
487 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
488 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 7]);
489
490 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
491 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 8]);
492 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__p_s_p_s[n * 9 + 8]);
493
494 }
495
496
497 VRR_K_d_s_d_s(
498 PRIM_INT__d_s_d_s,
499 PRIM_INT__d_s_p_s,
500 PRIM_INT__d_s_s_s,
501 PRIM_INT__p_s_p_s,
502 Q_PA,
503 a_over_q,
504 aoq_PQ,
505 one_over_2pq,
506 one_over_2q,
507 2);
508
509
510 ostei_general_vrr_K(3, 0, 3, 0, 1,
511 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
512 PRIM_INT__f_s_d_s, PRIM_INT__f_s_p_s, NULL, PRIM_INT__d_s_d_s, NULL, PRIM_INT__f_s_f_s);
513
514
515 VRR_I_g_s_s_s(
516 PRIM_INT__g_s_s_s,
517 PRIM_INT__f_s_s_s,
518 PRIM_INT__d_s_s_s,
519 P_PA,
520 a_over_p,
521 aop_PQ,
522 one_over_2p,
523 6);
524
525
526 VRR_K_g_s_p_s(
527 PRIM_INT__g_s_p_s,
528 PRIM_INT__g_s_s_s,
529 PRIM_INT__f_s_s_s,
530 Q_PA,
531 aoq_PQ,
532 one_over_2pq,
533 3);
534
535
536 ostei_general_vrr_K(4, 0, 2, 0, 2,
537 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
538 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);
539
540
541 ostei_general_vrr_K(4, 0, 3, 0, 1,
542 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
543 PRIM_INT__g_s_d_s, PRIM_INT__g_s_p_s, NULL, PRIM_INT__f_s_d_s, NULL, PRIM_INT__g_s_f_s);
544
545
546 VRR_I_h_s_s_s(
547 PRIM_INT__h_s_s_s,
548 PRIM_INT__g_s_s_s,
549 PRIM_INT__f_s_s_s,
550 P_PA,
551 a_over_p,
552 aop_PQ,
553 one_over_2p,
554 5);
555
556
557 ostei_general_vrr_K(5, 0, 1, 0, 3,
558 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
559 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
560
561
562 ostei_general_vrr_K(5, 0, 2, 0, 2,
563 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
564 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);
565
566
567 ostei_general_vrr_K(5, 0, 3, 0, 1,
568 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
569 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);
570
571
572 ostei_general_vrr1_I(6, 4,
573 one_over_2p, a_over_p, aop_PQ, P_PA,
574 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
575
576
577 ostei_general_vrr_K(6, 0, 1, 0, 3,
578 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
579 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
580
581
582 ostei_general_vrr_K(6, 0, 2, 0, 2,
583 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
584 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);
585
586
587 ostei_general_vrr_K(6, 0, 3, 0, 1,
588 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
589 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);
590
591
592
593
594 ////////////////////////////////////
595 // Accumulate contracted integrals
596 ////////////////////////////////////
597 if(lastoffset == 0)
598 {
599 contract_all(60, PRIM_INT__f_s_d_s, PRIM_PTR_INT__f_s_d_s);
600 contract_all(100, PRIM_INT__f_s_f_s, PRIM_PTR_INT__f_s_f_s);
601 contract_all(90, PRIM_INT__g_s_d_s, PRIM_PTR_INT__g_s_d_s);
602 contract_all(150, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
603 contract_all(126, PRIM_INT__h_s_d_s, PRIM_PTR_INT__h_s_d_s);
604 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
605 contract_all(168, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
606 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
607 }
608 else
609 {
610 contract(60, shelloffsets, PRIM_INT__f_s_d_s, PRIM_PTR_INT__f_s_d_s);
611 contract(100, shelloffsets, PRIM_INT__f_s_f_s, PRIM_PTR_INT__f_s_f_s);
612 contract(90, shelloffsets, PRIM_INT__g_s_d_s, PRIM_PTR_INT__g_s_d_s);
613 contract(150, shelloffsets, PRIM_INT__g_s_f_s, PRIM_PTR_INT__g_s_f_s);
614 contract(126, shelloffsets, PRIM_INT__h_s_d_s, PRIM_PTR_INT__h_s_d_s);
615 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
616 contract(168, shelloffsets, PRIM_INT__i_s_d_s, PRIM_PTR_INT__i_s_d_s);
617 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
618 PRIM_PTR_INT__f_s_d_s += lastoffset*60;
619 PRIM_PTR_INT__f_s_f_s += lastoffset*100;
620 PRIM_PTR_INT__g_s_d_s += lastoffset*90;
621 PRIM_PTR_INT__g_s_f_s += lastoffset*150;
622 PRIM_PTR_INT__h_s_d_s += lastoffset*126;
623 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
624 PRIM_PTR_INT__i_s_d_s += lastoffset*168;
625 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
626 }
627
628 } // close loop over j
629 } // close loop over i
630
631 //Advance to the next batch
632 jstart = SIMINT_SIMD_ROUND(jend);
633
634 //////////////////////////////////////////////
635 // Contracted integrals: Horizontal recurrance
636 //////////////////////////////////////////////
637
638
639 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
640
641
642 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
643 {
644 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
645
646 // set up HRR pointers
647 double const * restrict HRR_INT__f_s_d_s = INT__f_s_d_s + abcd * 60;
648 double const * restrict HRR_INT__f_s_f_s = INT__f_s_f_s + abcd * 100;
649 double const * restrict HRR_INT__g_s_d_s = INT__g_s_d_s + abcd * 90;
650 double const * restrict HRR_INT__g_s_f_s = INT__g_s_f_s + abcd * 150;
651 double const * restrict HRR_INT__h_s_d_s = INT__h_s_d_s + abcd * 126;
652 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
653 double const * restrict HRR_INT__i_s_d_s = INT__i_s_d_s + abcd * 168;
654 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
655 double * restrict HRR_INT__f_f_d_p = INT__f_f_d_p + real_abcd * 1800;
656
657 // form INT__f_p_d_s
658 HRR_J_f_p(
659 HRR_INT__f_p_d_s,
660 HRR_INT__f_s_d_s,
661 HRR_INT__g_s_d_s,
662 hAB, 6);
663
664 // form INT__f_p_f_s
665 HRR_J_f_p(
666 HRR_INT__f_p_f_s,
667 HRR_INT__f_s_f_s,
668 HRR_INT__g_s_f_s,
669 hAB, 10);
670
671 // form INT__g_p_d_s
672 HRR_J_g_p(
673 HRR_INT__g_p_d_s,
674 HRR_INT__g_s_d_s,
675 HRR_INT__h_s_d_s,
676 hAB, 6);
677
678 // form INT__g_p_f_s
679 HRR_J_g_p(
680 HRR_INT__g_p_f_s,
681 HRR_INT__g_s_f_s,
682 HRR_INT__h_s_f_s,
683 hAB, 10);
684
685 // form INT__h_p_d_s
686 ostei_general_hrr_J(5, 1, 2, 0, hAB, HRR_INT__i_s_d_s, HRR_INT__h_s_d_s, HRR_INT__h_p_d_s);
687
688 // form INT__h_p_f_s
689 ostei_general_hrr_J(5, 1, 3, 0, hAB, HRR_INT__i_s_f_s, HRR_INT__h_s_f_s, HRR_INT__h_p_f_s);
690
691 // form INT__f_d_d_s
692 HRR_J_f_d(
693 HRR_INT__f_d_d_s,
694 HRR_INT__f_p_d_s,
695 HRR_INT__g_p_d_s,
696 hAB, 6);
697
698 // form INT__f_d_f_s
699 HRR_J_f_d(
700 HRR_INT__f_d_f_s,
701 HRR_INT__f_p_f_s,
702 HRR_INT__g_p_f_s,
703 hAB, 10);
704
705 // form INT__g_d_d_s
706 ostei_general_hrr_J(4, 2, 2, 0, hAB, HRR_INT__h_p_d_s, HRR_INT__g_p_d_s, HRR_INT__g_d_d_s);
707
708 // form INT__g_d_f_s
709 ostei_general_hrr_J(4, 2, 3, 0, hAB, HRR_INT__h_p_f_s, HRR_INT__g_p_f_s, HRR_INT__g_d_f_s);
710
711 // form INT__f_f_d_s
712 ostei_general_hrr_J(3, 3, 2, 0, hAB, HRR_INT__g_d_d_s, HRR_INT__f_d_d_s, HRR_INT__f_f_d_s);
713
714 // form INT__f_f_f_s
715 ostei_general_hrr_J(3, 3, 3, 0, hAB, HRR_INT__g_d_f_s, HRR_INT__f_d_f_s, HRR_INT__f_f_f_s);
716
717 // form INT__f_f_d_p
718 for(ibra = 0; ibra < 100; ++ibra)
719 {
720 HRR_INT__f_f_d_p[ibra * 18 + 0] = HRR_INT__f_f_f_s[ibra * 10 + 0] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 0] );
721
722 HRR_INT__f_f_d_p[ibra * 18 + 1] = HRR_INT__f_f_f_s[ibra * 10 + 1] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 0] );
723
724 HRR_INT__f_f_d_p[ibra * 18 + 2] = HRR_INT__f_f_f_s[ibra * 10 + 2] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 0] );
725
726 HRR_INT__f_f_d_p[ibra * 18 + 3] = HRR_INT__f_f_f_s[ibra * 10 + 1] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 1] );
727
728 HRR_INT__f_f_d_p[ibra * 18 + 4] = HRR_INT__f_f_f_s[ibra * 10 + 3] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 1] );
729
730 HRR_INT__f_f_d_p[ibra * 18 + 5] = HRR_INT__f_f_f_s[ibra * 10 + 4] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 1] );
731
732 HRR_INT__f_f_d_p[ibra * 18 + 6] = HRR_INT__f_f_f_s[ibra * 10 + 2] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 2] );
733
734 HRR_INT__f_f_d_p[ibra * 18 + 7] = HRR_INT__f_f_f_s[ibra * 10 + 4] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 2] );
735
736 HRR_INT__f_f_d_p[ibra * 18 + 8] = HRR_INT__f_f_f_s[ibra * 10 + 5] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 2] );
737
738 HRR_INT__f_f_d_p[ibra * 18 + 9] = HRR_INT__f_f_f_s[ibra * 10 + 3] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 3] );
739
740 HRR_INT__f_f_d_p[ibra * 18 + 10] = HRR_INT__f_f_f_s[ibra * 10 + 6] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 3] );
741
742 HRR_INT__f_f_d_p[ibra * 18 + 11] = HRR_INT__f_f_f_s[ibra * 10 + 7] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 3] );
743
744 HRR_INT__f_f_d_p[ibra * 18 + 12] = HRR_INT__f_f_f_s[ibra * 10 + 4] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 4] );
745
746 HRR_INT__f_f_d_p[ibra * 18 + 13] = HRR_INT__f_f_f_s[ibra * 10 + 7] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 4] );
747
748 HRR_INT__f_f_d_p[ibra * 18 + 14] = HRR_INT__f_f_f_s[ibra * 10 + 8] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 4] );
749
750 HRR_INT__f_f_d_p[ibra * 18 + 15] = HRR_INT__f_f_f_s[ibra * 10 + 5] + ( hCD[0] * HRR_INT__f_f_d_s[ibra * 6 + 5] );
751
752 HRR_INT__f_f_d_p[ibra * 18 + 16] = HRR_INT__f_f_f_s[ibra * 10 + 8] + ( hCD[1] * HRR_INT__f_f_d_s[ibra * 6 + 5] );
753
754 HRR_INT__f_f_d_p[ibra * 18 + 17] = HRR_INT__f_f_f_s[ibra * 10 + 9] + ( hCD[2] * HRR_INT__f_f_d_s[ibra * 6 + 5] );
755
756 }
757
758
759 } // close HRR loop
760
761
762 } // close loop cdbatch
763
764 istart = iend;
765 } // close loop over ab
766
767 return P.nshell12_clip * Q.nshell12_clip;
768 }
769
770