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_h_s(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_h_s)8 int ostei_i_g_h_s(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_h_s)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_g_h_s);
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
27 // partition workspace
28 double * const INT__i_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
29 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 588);
30 double * const INT__l_s_h_s = work + (SIMINT_NSHELL_SIMD * 1344);
31 double * const INT__m_s_h_s = work + (SIMINT_NSHELL_SIMD * 2289);
32 double * const INT__n_s_h_s = work + (SIMINT_NSHELL_SIMD * 3444);
33 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*4830);
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 16;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 61;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 145;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 235;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 365;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 515;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 755;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 935;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1160;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1520;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1970;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2201;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 2516;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 3020;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 3650;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 4280;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 4560;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 4980;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 5652;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 6492;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 7332;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 7920;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 8244;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 8784;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 9648;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 10728;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 11808;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 12564;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 12924;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 13599;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 14679;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_g_s = primwork + 16029;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_h_s = primwork + 17379;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_s_s = primwork + 18324;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_p_s = primwork + 18709;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_d_s = primwork + 19534;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_f_s = primwork + 20854;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_g_s = primwork + 22504;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_h_s = primwork + 24154;
74 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_s_s = primwork + 25309;
75 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_p_s = primwork + 25705;
76 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_d_s = primwork + 26695;
77 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_f_s = primwork + 28279;
78 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_g_s = primwork + 30259;
79 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_h_s = primwork + 32239;
80 double * const hrrwork = (double *)(primwork + 33625);
81 double * const HRR_INT__i_p_h_s = hrrwork + 0;
82 double * const HRR_INT__i_d_h_s = hrrwork + 1764;
83 double * const HRR_INT__i_f_h_s = hrrwork + 5292;
84 double * const HRR_INT__k_p_h_s = hrrwork + 11172;
85 double * const HRR_INT__k_d_h_s = hrrwork + 13440;
86 double * const HRR_INT__k_f_h_s = hrrwork + 17976;
87 double * const HRR_INT__l_p_h_s = hrrwork + 25536;
88 double * const HRR_INT__l_d_h_s = hrrwork + 28371;
89 double * const HRR_INT__m_p_h_s = hrrwork + 34041;
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 * 4830 * 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_h_s = INT__i_s_h_s + abcd * 588;
147 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
148 double * restrict PRIM_PTR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
149 double * restrict PRIM_PTR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
150 double * restrict PRIM_PTR_INT__n_s_h_s = INT__n_s_h_s + abcd * 1386;
151
152
153
154 // Load these one per loop over i
155 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
156 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
157 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
158
159 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
160
161 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
162 {
163 // calculate the shell offsets
164 // these are the offset from the shell pointed to by cd
165 // for each element
166 int shelloffsets[SIMINT_SIMD_LEN] = {0};
167 int lastoffset = 0;
168 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
169
170 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
171 {
172 // Handle if the first element of the vector is a new shell
173 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
174 {
175 nprim_icd += Q.nprim12[cd + (++icd)];
176 PRIM_PTR_INT__i_s_h_s += 588;
177 PRIM_PTR_INT__k_s_h_s += 756;
178 PRIM_PTR_INT__l_s_h_s += 945;
179 PRIM_PTR_INT__m_s_h_s += 1155;
180 PRIM_PTR_INT__n_s_h_s += 1386;
181 }
182 iprimcd++;
183 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
184 {
185 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
186 {
187 shelloffsets[n] = shelloffsets[n-1] + 1;
188 lastoffset++;
189 nprim_icd += Q.nprim12[cd + (++icd)];
190 }
191 else
192 shelloffsets[n] = shelloffsets[n-1];
193 iprimcd++;
194 }
195 }
196 else
197 iprimcd += SIMINT_SIMD_LEN;
198
199 // Do we have to compute this vector (or has it been screened out)?
200 // (not_screened != 0 means we have to do this vector)
201 if(check_screen)
202 {
203 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
204 if(vmax < screen_tol)
205 {
206 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
207 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
208 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
209 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
210 PRIM_PTR_INT__n_s_h_s += lastoffset*1386;
211 continue;
212 }
213 }
214
215 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
216 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
217 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
218 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
219
220
221 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
222 SIMINT_DBLTYPE PQ[3];
223 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
224 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
225 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
226 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
227 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
228 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
229
230 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
231 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
232 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
233 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
234 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
235 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
236 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
237
238 // NOTE: Minus sign!
239 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
240 SIMINT_DBLTYPE aop_PQ[3];
241 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
242 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
243 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
244
245 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
246 SIMINT_DBLTYPE aoq_PQ[3];
247 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
248 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
249 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
250 // Put a minus sign here so we don't have to in RR routines
251 a_over_q = SIMINT_NEG(a_over_q);
252
253
254 //////////////////////////////////////////////
255 // Fjt function section
256 // Maximum v value: 15
257 //////////////////////////////////////////////
258 // The parameter to the Fjt function
259 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
260
261
262 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
263
264
265 boys_F_split(PRIM_INT__s_s_s_s, F_x, 15);
266 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
267 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
268 for(n = 0; n <= 15; n++)
269 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
270
271 //////////////////////////////////////////////
272 // Primitive integrals: Vertical recurrance
273 //////////////////////////////////////////////
274
275 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
276 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
277 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
278 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
279 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
280 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
281 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
282 const SIMINT_DBLTYPE vrr_const_8_over_2p = SIMINT_MUL(const_8, one_over_2p);
283 const SIMINT_DBLTYPE vrr_const_9_over_2p = SIMINT_MUL(const_9, one_over_2p);
284 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
285 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
286 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
287 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
288 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
289 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
290 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
291 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
292 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
293 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
294 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
295 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
296 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
297 const SIMINT_DBLTYPE vrr_const_10_over_2pq = SIMINT_MUL(const_10, one_over_2pq);
298
299
300
301 // Forming PRIM_INT__p_s_s_s[15 * 3];
302 for(n = 0; n < 15; ++n) // loop over orders of auxiliary function
303 {
304
305 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
306 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]);
307
308 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
309 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]);
310
311 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
312 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]);
313
314 }
315
316
317
318 // Forming PRIM_INT__d_s_s_s[14 * 6];
319 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
320 {
321
322 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
323 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]);
324 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]);
325
326 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
327 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]);
328
329 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
330 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]);
331
332 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
333 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]);
334 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]);
335
336 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
337 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]);
338
339 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
340 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]);
341 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]);
342
343 }
344
345
346
347 // Forming PRIM_INT__f_s_s_s[13 * 10];
348 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
349 {
350
351 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
352 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]);
353 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]);
354
355 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
356 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]);
357
358 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
359 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]);
360
361 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
362 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]);
363
364 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
365 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]);
366
367 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
368 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]);
369
370 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
371 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]);
372 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]);
373
374 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
375 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]);
376
377 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
378 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]);
379
380 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
381 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]);
382 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]);
383
384 }
385
386
387 VRR_I_g_s_s_s(
388 PRIM_INT__g_s_s_s,
389 PRIM_INT__f_s_s_s,
390 PRIM_INT__d_s_s_s,
391 P_PA,
392 a_over_p,
393 aop_PQ,
394 one_over_2p,
395 12);
396
397
398 VRR_I_h_s_s_s(
399 PRIM_INT__h_s_s_s,
400 PRIM_INT__g_s_s_s,
401 PRIM_INT__f_s_s_s,
402 P_PA,
403 a_over_p,
404 aop_PQ,
405 one_over_2p,
406 11);
407
408
409 ostei_general_vrr1_I(6, 10,
410 one_over_2p, a_over_p, aop_PQ, P_PA,
411 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
412
413
414 ostei_general_vrr_K(6, 0, 1, 0, 5,
415 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
416 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
417
418
419 ostei_general_vrr_K(5, 0, 1, 0, 5,
420 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
421 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
422
423
424 ostei_general_vrr_K(6, 0, 2, 0, 4,
425 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
426 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);
427
428
429 VRR_K_g_s_p_s(
430 PRIM_INT__g_s_p_s,
431 PRIM_INT__g_s_s_s,
432 PRIM_INT__f_s_s_s,
433 Q_PA,
434 aoq_PQ,
435 one_over_2pq,
436 5);
437
438
439 ostei_general_vrr_K(5, 0, 2, 0, 4,
440 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
441 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);
442
443
444 ostei_general_vrr_K(6, 0, 3, 0, 3,
445 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
446 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);
447
448
449 VRR_K_f_s_p_s(
450 PRIM_INT__f_s_p_s,
451 PRIM_INT__f_s_s_s,
452 PRIM_INT__d_s_s_s,
453 Q_PA,
454 aoq_PQ,
455 one_over_2pq,
456 5);
457
458
459 ostei_general_vrr_K(4, 0, 2, 0, 4,
460 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
461 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);
462
463
464 ostei_general_vrr_K(5, 0, 3, 0, 3,
465 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
466 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);
467
468
469 ostei_general_vrr_K(6, 0, 4, 0, 2,
470 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
471 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);
472
473
474
475 // Forming PRIM_INT__d_s_p_s[5 * 18];
476 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
477 {
478
479 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
480 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]);
481 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]);
482
483 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
484 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]);
485
486 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
487 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]);
488
489 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
490 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]);
491 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]);
492
493 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
494 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]);
495 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]);
496
497 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
498 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]);
499
500 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
501 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]);
502 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]);
503
504 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
505 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]);
506
507 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
508 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]);
509 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]);
510
511 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
512 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]);
513
514 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
515 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]);
516 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]);
517
518 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
519 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]);
520
521 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
522 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]);
523
524 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
525 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]);
526 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]);
527
528 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
529 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]);
530 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]);
531
532 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
533 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]);
534
535 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
536 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]);
537
538 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
539 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]);
540 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]);
541
542 }
543
544
545 VRR_K_f_s_d_s(
546 PRIM_INT__f_s_d_s,
547 PRIM_INT__f_s_p_s,
548 PRIM_INT__f_s_s_s,
549 PRIM_INT__d_s_p_s,
550 Q_PA,
551 a_over_q,
552 aoq_PQ,
553 one_over_2pq,
554 one_over_2q,
555 4);
556
557
558 ostei_general_vrr_K(4, 0, 3, 0, 3,
559 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
560 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);
561
562
563 ostei_general_vrr_K(5, 0, 4, 0, 2,
564 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
565 PRIM_INT__h_s_f_s, PRIM_INT__h_s_d_s, NULL, PRIM_INT__g_s_f_s, NULL, PRIM_INT__h_s_g_s);
566
567
568 ostei_general_vrr_K(6, 0, 5, 0, 1,
569 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
570 PRIM_INT__i_s_g_s, PRIM_INT__i_s_f_s, NULL, PRIM_INT__h_s_g_s, NULL, PRIM_INT__i_s_h_s);
571
572
573 ostei_general_vrr1_I(7, 9,
574 one_over_2p, a_over_p, aop_PQ, P_PA,
575 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
576
577
578 ostei_general_vrr_K(7, 0, 1, 0, 5,
579 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
580 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
581
582
583 ostei_general_vrr_K(7, 0, 2, 0, 4,
584 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
585 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);
586
587
588 ostei_general_vrr_K(7, 0, 3, 0, 3,
589 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
590 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);
591
592
593 ostei_general_vrr_K(7, 0, 4, 0, 2,
594 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
595 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);
596
597
598 ostei_general_vrr_K(7, 0, 5, 0, 1,
599 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
600 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);
601
602
603 ostei_general_vrr1_I(8, 8,
604 one_over_2p, a_over_p, aop_PQ, P_PA,
605 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
606
607
608 ostei_general_vrr_K(8, 0, 1, 0, 5,
609 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
610 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
611
612
613 ostei_general_vrr_K(8, 0, 2, 0, 4,
614 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
615 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);
616
617
618 ostei_general_vrr_K(8, 0, 3, 0, 3,
619 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
620 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);
621
622
623 ostei_general_vrr_K(8, 0, 4, 0, 2,
624 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
625 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);
626
627
628 ostei_general_vrr_K(8, 0, 5, 0, 1,
629 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
630 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);
631
632
633 ostei_general_vrr1_I(9, 7,
634 one_over_2p, a_over_p, aop_PQ, P_PA,
635 PRIM_INT__l_s_s_s, PRIM_INT__k_s_s_s, PRIM_INT__m_s_s_s);
636
637
638 ostei_general_vrr_K(9, 0, 1, 0, 5,
639 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
640 PRIM_INT__m_s_s_s, NULL, NULL, PRIM_INT__l_s_s_s, NULL, PRIM_INT__m_s_p_s);
641
642
643 ostei_general_vrr_K(9, 0, 2, 0, 4,
644 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
645 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);
646
647
648 ostei_general_vrr_K(9, 0, 3, 0, 3,
649 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
650 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);
651
652
653 ostei_general_vrr_K(9, 0, 4, 0, 2,
654 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
655 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);
656
657
658 ostei_general_vrr_K(9, 0, 5, 0, 1,
659 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
660 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);
661
662
663 ostei_general_vrr1_I(10, 6,
664 one_over_2p, a_over_p, aop_PQ, P_PA,
665 PRIM_INT__m_s_s_s, PRIM_INT__l_s_s_s, PRIM_INT__n_s_s_s);
666
667
668 ostei_general_vrr_K(10, 0, 1, 0, 5,
669 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
670 PRIM_INT__n_s_s_s, NULL, NULL, PRIM_INT__m_s_s_s, NULL, PRIM_INT__n_s_p_s);
671
672
673 ostei_general_vrr_K(10, 0, 2, 0, 4,
674 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
675 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);
676
677
678 ostei_general_vrr_K(10, 0, 3, 0, 3,
679 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
680 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);
681
682
683 ostei_general_vrr_K(10, 0, 4, 0, 2,
684 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
685 PRIM_INT__n_s_f_s, PRIM_INT__n_s_d_s, NULL, PRIM_INT__m_s_f_s, NULL, PRIM_INT__n_s_g_s);
686
687
688 ostei_general_vrr_K(10, 0, 5, 0, 1,
689 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
690 PRIM_INT__n_s_g_s, PRIM_INT__n_s_f_s, NULL, PRIM_INT__m_s_g_s, NULL, PRIM_INT__n_s_h_s);
691
692
693
694
695 ////////////////////////////////////
696 // Accumulate contracted integrals
697 ////////////////////////////////////
698 if(lastoffset == 0)
699 {
700 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
701 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
702 contract_all(945, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
703 contract_all(1155, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
704 contract_all(1386, PRIM_INT__n_s_h_s, PRIM_PTR_INT__n_s_h_s);
705 }
706 else
707 {
708 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
709 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
710 contract(945, shelloffsets, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
711 contract(1155, shelloffsets, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
712 contract(1386, shelloffsets, PRIM_INT__n_s_h_s, PRIM_PTR_INT__n_s_h_s);
713 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
714 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
715 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
716 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
717 PRIM_PTR_INT__n_s_h_s += lastoffset*1386;
718 }
719
720 } // close loop over j
721 } // close loop over i
722
723 //Advance to the next batch
724 jstart = SIMINT_SIMD_ROUND(jend);
725
726 //////////////////////////////////////////////
727 // Contracted integrals: Horizontal recurrance
728 //////////////////////////////////////////////
729
730
731 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
732
733
734 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
735 {
736
737 // set up HRR pointers
738 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
739 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
740 double const * restrict HRR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
741 double const * restrict HRR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
742 double const * restrict HRR_INT__n_s_h_s = INT__n_s_h_s + abcd * 1386;
743 double * restrict HRR_INT__i_g_h_s = INT__i_g_h_s + real_abcd * 8820;
744
745 // form INT__i_p_h_s
746 ostei_general_hrr_J(6, 1, 5, 0, hAB, HRR_INT__k_s_h_s, HRR_INT__i_s_h_s, HRR_INT__i_p_h_s);
747
748 // form INT__k_p_h_s
749 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);
750
751 // form INT__l_p_h_s
752 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);
753
754 // form INT__m_p_h_s
755 ostei_general_hrr_J(9, 1, 5, 0, hAB, HRR_INT__n_s_h_s, HRR_INT__m_s_h_s, HRR_INT__m_p_h_s);
756
757 // form INT__i_d_h_s
758 ostei_general_hrr_J(6, 2, 5, 0, hAB, HRR_INT__k_p_h_s, HRR_INT__i_p_h_s, HRR_INT__i_d_h_s);
759
760 // form INT__k_d_h_s
761 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);
762
763 // form INT__l_d_h_s
764 ostei_general_hrr_J(8, 2, 5, 0, hAB, HRR_INT__m_p_h_s, HRR_INT__l_p_h_s, HRR_INT__l_d_h_s);
765
766 // form INT__i_f_h_s
767 ostei_general_hrr_J(6, 3, 5, 0, hAB, HRR_INT__k_d_h_s, HRR_INT__i_d_h_s, HRR_INT__i_f_h_s);
768
769 // form INT__k_f_h_s
770 ostei_general_hrr_J(7, 3, 5, 0, hAB, HRR_INT__l_d_h_s, HRR_INT__k_d_h_s, HRR_INT__k_f_h_s);
771
772 // form INT__i_g_h_s
773 ostei_general_hrr_J(6, 4, 5, 0, hAB, HRR_INT__k_f_h_s, HRR_INT__i_f_h_s, HRR_INT__i_g_h_s);
774
775
776 } // close HRR loop
777
778
779 } // close loop cdbatch
780
781 istart = iend;
782 } // close loop over ab
783
784 return P.nshell12_clip * Q.nshell12_clip;
785 }
786
ostei_g_i_h_s(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_h_s)787 int ostei_g_i_h_s(struct simint_multi_shellpair const P,
788 struct simint_multi_shellpair const Q,
789 double screen_tol,
790 double * const restrict work,
791 double * const restrict INT__g_i_h_s)
792 {
793 double P_AB[3*P.nshell12];
794 struct simint_multi_shellpair P_tmp = P;
795 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
796 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
797 P_tmp.AB_x = P_AB;
798 P_tmp.AB_y = P_AB + P.nshell12;
799 P_tmp.AB_z = P_AB + 2*P.nshell12;
800
801 for(int i = 0; i < P.nshell12; i++)
802 {
803 P_tmp.AB_x[i] = -P.AB_x[i];
804 P_tmp.AB_y[i] = -P.AB_y[i];
805 P_tmp.AB_z[i] = -P.AB_z[i];
806 }
807
808 int ret = ostei_i_g_h_s(P_tmp, Q, screen_tol, work, INT__g_i_h_s);
809 double buffer[8820] SIMINT_ALIGN_ARRAY_DBL;
810
811 for(int q = 0; q < ret; q++)
812 {
813 int idx = 0;
814 for(int a = 0; a < 15; ++a)
815 for(int b = 0; b < 28; ++b)
816 for(int c = 0; c < 21; ++c)
817 for(int d = 0; d < 1; ++d)
818 buffer[idx++] = INT__g_i_h_s[q*8820+b*315+a*21+c*1+d];
819
820 memcpy(INT__g_i_h_s+q*8820, buffer, 8820*sizeof(double));
821 }
822
823 return ret;
824 }
825
ostei_i_g_s_h(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_s_h)826 int ostei_i_g_s_h(struct simint_multi_shellpair const P,
827 struct simint_multi_shellpair const Q,
828 double screen_tol,
829 double * const restrict work,
830 double * const restrict INT__i_g_s_h)
831 {
832 double Q_AB[3*Q.nshell12];
833 struct simint_multi_shellpair Q_tmp = Q;
834 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
835 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
836 Q_tmp.AB_x = Q_AB;
837 Q_tmp.AB_y = Q_AB + Q.nshell12;
838 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
839
840 for(int i = 0; i < Q.nshell12; i++)
841 {
842 Q_tmp.AB_x[i] = -Q.AB_x[i];
843 Q_tmp.AB_y[i] = -Q.AB_y[i];
844 Q_tmp.AB_z[i] = -Q.AB_z[i];
845 }
846
847 int ret = ostei_i_g_h_s(P, Q_tmp, screen_tol, work, INT__i_g_s_h);
848 double buffer[8820] SIMINT_ALIGN_ARRAY_DBL;
849
850 for(int q = 0; q < ret; q++)
851 {
852 int idx = 0;
853 for(int a = 0; a < 28; ++a)
854 for(int b = 0; b < 15; ++b)
855 for(int c = 0; c < 1; ++c)
856 for(int d = 0; d < 21; ++d)
857 buffer[idx++] = INT__i_g_s_h[q*8820+a*315+b*21+d*1+c];
858
859 memcpy(INT__i_g_s_h+q*8820, buffer, 8820*sizeof(double));
860 }
861
862 return ret;
863 }
864
ostei_g_i_s_h(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_s_h)865 int ostei_g_i_s_h(struct simint_multi_shellpair const P,
866 struct simint_multi_shellpair const Q,
867 double screen_tol,
868 double * const restrict work,
869 double * const restrict INT__g_i_s_h)
870 {
871 double P_AB[3*P.nshell12];
872 struct simint_multi_shellpair P_tmp = P;
873 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
874 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
875 P_tmp.AB_x = P_AB;
876 P_tmp.AB_y = P_AB + P.nshell12;
877 P_tmp.AB_z = P_AB + 2*P.nshell12;
878
879 for(int i = 0; i < P.nshell12; i++)
880 {
881 P_tmp.AB_x[i] = -P.AB_x[i];
882 P_tmp.AB_y[i] = -P.AB_y[i];
883 P_tmp.AB_z[i] = -P.AB_z[i];
884 }
885
886 double Q_AB[3*Q.nshell12];
887 struct simint_multi_shellpair Q_tmp = Q;
888 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
889 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
890 Q_tmp.AB_x = Q_AB;
891 Q_tmp.AB_y = Q_AB + Q.nshell12;
892 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
893
894 for(int i = 0; i < Q.nshell12; i++)
895 {
896 Q_tmp.AB_x[i] = -Q.AB_x[i];
897 Q_tmp.AB_y[i] = -Q.AB_y[i];
898 Q_tmp.AB_z[i] = -Q.AB_z[i];
899 }
900
901 int ret = ostei_i_g_h_s(P_tmp, Q_tmp, screen_tol, work, INT__g_i_s_h);
902 double buffer[8820] SIMINT_ALIGN_ARRAY_DBL;
903
904 for(int q = 0; q < ret; q++)
905 {
906 int idx = 0;
907 for(int a = 0; a < 15; ++a)
908 for(int b = 0; b < 28; ++b)
909 for(int c = 0; c < 1; ++c)
910 for(int d = 0; d < 21; ++d)
911 buffer[idx++] = INT__g_i_s_h[q*8820+b*315+a*21+d*1+c];
912
913 memcpy(INT__g_i_s_h+q*8820, buffer, 8820*sizeof(double));
914 }
915
916 return ret;
917 }
918
919