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_h_p_f_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_p_f_d)8 int ostei_h_p_f_d(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__h_p_f_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__h_p_f_d);
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__h_s_f_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__h_s_g_s = work + (SIMINT_NSHELL_SIMD * 210);
31 double * const INT__h_s_h_s = work + (SIMINT_NSHELL_SIMD * 525);
32 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 966);
33 double * const INT__i_s_g_s = work + (SIMINT_NSHELL_SIMD * 1246);
34 double * const INT__i_s_h_s = work + (SIMINT_NSHELL_SIMD * 1666);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*2254);
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 12;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 45;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 90;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 150;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 240;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 384;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 474;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 624;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 864;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 1164;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1284;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1509;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1869;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 2319;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 2769;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2916;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 3231;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 3735;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 4365;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 4995;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 5436;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 5604;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 6024;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 6696;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 7536;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 8376;
63 double * const hrrwork = (double *)(primwork + 8964);
64 double * const HRR_INT__h_p_f_s = hrrwork + 0;
65 double * const HRR_INT__h_p_f_p = hrrwork + 630;
66 double * const HRR_INT__h_p_g_s = hrrwork + 2520;
67 double * const HRR_INT__h_p_g_p = hrrwork + 3465;
68 double * const HRR_INT__h_p_h_s = hrrwork + 6300;
69
70
71 // Create constants
72 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
73 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
74 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
75 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
76 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
77 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
78 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
79
80
81 ////////////////////////////////////////
82 // Loop over shells and primitives
83 ////////////////////////////////////////
84
85 real_abcd = 0;
86 istart = 0;
87 for(ab = 0; ab < P.nshell12_clip; ++ab)
88 {
89 const int iend = istart + P.nprim12[ab];
90
91 cd = 0;
92 jstart = 0;
93
94 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
95 {
96 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
97 int jend = jstart;
98 for(i = 0; i < nshellbatch; i++)
99 jend += Q.nprim12[cd+i];
100
101 // Clear the beginning of the workspace (where we are accumulating integrals)
102 memset(work, 0, SIMINT_NSHELL_SIMD * 2254 * sizeof(double));
103 abcd = 0;
104
105
106 for(i = istart; i < iend; ++i)
107 {
108 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
109
110 if(check_screen)
111 {
112 // Skip this whole thing if always insignificant
113 if((P.screen[i] * Q.screen_max) < screen_tol)
114 continue;
115 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
116 }
117
118 icd = 0;
119 iprimcd = 0;
120 nprim_icd = Q.nprim12[cd];
121 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
122 double * restrict PRIM_PTR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
123 double * restrict PRIM_PTR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
124 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
125 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
126 double * restrict PRIM_PTR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
127
128
129
130 // Load these one per loop over i
131 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
132 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
133 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
134
135 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
136
137 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
138 {
139 // calculate the shell offsets
140 // these are the offset from the shell pointed to by cd
141 // for each element
142 int shelloffsets[SIMINT_SIMD_LEN] = {0};
143 int lastoffset = 0;
144 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
145
146 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
147 {
148 // Handle if the first element of the vector is a new shell
149 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
150 {
151 nprim_icd += Q.nprim12[cd + (++icd)];
152 PRIM_PTR_INT__h_s_f_s += 210;
153 PRIM_PTR_INT__h_s_g_s += 315;
154 PRIM_PTR_INT__h_s_h_s += 441;
155 PRIM_PTR_INT__i_s_f_s += 280;
156 PRIM_PTR_INT__i_s_g_s += 420;
157 PRIM_PTR_INT__i_s_h_s += 588;
158 }
159 iprimcd++;
160 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
161 {
162 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
163 {
164 shelloffsets[n] = shelloffsets[n-1] + 1;
165 lastoffset++;
166 nprim_icd += Q.nprim12[cd + (++icd)];
167 }
168 else
169 shelloffsets[n] = shelloffsets[n-1];
170 iprimcd++;
171 }
172 }
173 else
174 iprimcd += SIMINT_SIMD_LEN;
175
176 // Do we have to compute this vector (or has it been screened out)?
177 // (not_screened != 0 means we have to do this vector)
178 if(check_screen)
179 {
180 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
181 if(vmax < screen_tol)
182 {
183 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
184 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
185 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
186 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
187 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
188 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
189 continue;
190 }
191 }
192
193 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
194 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
195 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
196 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
197
198
199 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
200 SIMINT_DBLTYPE PQ[3];
201 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
202 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
203 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
204 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
205 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
206 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
207
208 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
209 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
210 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
211 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
212 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
213 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
214 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
215
216 // NOTE: Minus sign!
217 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
218 SIMINT_DBLTYPE aop_PQ[3];
219 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
220 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
221 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
222
223 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
224 SIMINT_DBLTYPE aoq_PQ[3];
225 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
226 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
227 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
228 // Put a minus sign here so we don't have to in RR routines
229 a_over_q = SIMINT_NEG(a_over_q);
230
231
232 //////////////////////////////////////////////
233 // Fjt function section
234 // Maximum v value: 11
235 //////////////////////////////////////////////
236 // The parameter to the Fjt function
237 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
238
239
240 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
241
242
243 boys_F_split(PRIM_INT__s_s_s_s, F_x, 11);
244 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
245 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
246 for(n = 0; n <= 11; n++)
247 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
248
249 //////////////////////////////////////////////
250 // Primitive integrals: Vertical recurrance
251 //////////////////////////////////////////////
252
253 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
254 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
255 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
256 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
257 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
258 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
259 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
260 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
261 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
262 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
263 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
264 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
265 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
266 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
267 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
268
269
270
271 // Forming PRIM_INT__p_s_s_s[11 * 3];
272 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
273 {
274
275 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
276 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]);
277
278 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
279 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]);
280
281 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
282 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]);
283
284 }
285
286
287
288 // Forming PRIM_INT__d_s_s_s[10 * 6];
289 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
290 {
291
292 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
293 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]);
294 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]);
295
296 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
297 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]);
298
299 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
300 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]);
301
302 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
303 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]);
304 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]);
305
306 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
307 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]);
308
309 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
310 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]);
311 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]);
312
313 }
314
315
316
317 // Forming PRIM_INT__f_s_s_s[9 * 10];
318 for(n = 0; n < 9; ++n) // loop over orders of auxiliary function
319 {
320
321 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
322 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]);
323 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]);
324
325 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
326 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]);
327
328 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
329 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]);
330
331 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
332 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]);
333
334 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
335 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]);
336
337 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
338 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]);
339
340 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
341 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]);
342 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]);
343
344 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
345 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]);
346
347 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
348 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]);
349
350 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
351 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]);
352 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]);
353
354 }
355
356
357 VRR_I_g_s_s_s(
358 PRIM_INT__g_s_s_s,
359 PRIM_INT__f_s_s_s,
360 PRIM_INT__d_s_s_s,
361 P_PA,
362 a_over_p,
363 aop_PQ,
364 one_over_2p,
365 8);
366
367
368 VRR_I_h_s_s_s(
369 PRIM_INT__h_s_s_s,
370 PRIM_INT__g_s_s_s,
371 PRIM_INT__f_s_s_s,
372 P_PA,
373 a_over_p,
374 aop_PQ,
375 one_over_2p,
376 7);
377
378
379 ostei_general_vrr_K(5, 0, 1, 0, 5,
380 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
381 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
382
383
384 VRR_K_g_s_p_s(
385 PRIM_INT__g_s_p_s,
386 PRIM_INT__g_s_s_s,
387 PRIM_INT__f_s_s_s,
388 Q_PA,
389 aoq_PQ,
390 one_over_2pq,
391 5);
392
393
394 ostei_general_vrr_K(5, 0, 2, 0, 4,
395 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
396 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);
397
398
399 VRR_K_f_s_p_s(
400 PRIM_INT__f_s_p_s,
401 PRIM_INT__f_s_s_s,
402 PRIM_INT__d_s_s_s,
403 Q_PA,
404 aoq_PQ,
405 one_over_2pq,
406 5);
407
408
409 ostei_general_vrr_K(4, 0, 2, 0, 4,
410 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
411 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);
412
413
414 ostei_general_vrr_K(5, 0, 3, 0, 3,
415 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
416 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);
417
418
419
420 // Forming PRIM_INT__d_s_p_s[5 * 18];
421 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
422 {
423
424 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
425 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]);
426 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]);
427
428 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
429 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]);
430
431 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
432 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]);
433
434 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
435 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]);
436 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]);
437
438 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
439 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]);
440 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]);
441
442 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
443 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]);
444
445 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
446 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]);
447 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]);
448
449 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
450 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]);
451
452 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
453 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]);
454 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]);
455
456 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
457 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]);
458
459 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
460 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]);
461 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]);
462
463 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
464 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]);
465
466 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
467 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]);
468
469 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
470 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]);
471 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]);
472
473 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
474 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]);
475 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]);
476
477 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
478 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]);
479
480 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
481 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]);
482
483 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
484 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]);
485 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]);
486
487 }
488
489
490 VRR_K_f_s_d_s(
491 PRIM_INT__f_s_d_s,
492 PRIM_INT__f_s_p_s,
493 PRIM_INT__f_s_s_s,
494 PRIM_INT__d_s_p_s,
495 Q_PA,
496 a_over_q,
497 aoq_PQ,
498 one_over_2pq,
499 one_over_2q,
500 4);
501
502
503 ostei_general_vrr_K(4, 0, 3, 0, 3,
504 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
505 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);
506
507
508 ostei_general_vrr_K(5, 0, 4, 0, 2,
509 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
510 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);
511
512
513 ostei_general_vrr1_I(6, 6,
514 one_over_2p, a_over_p, aop_PQ, P_PA,
515 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
516
517
518 ostei_general_vrr_K(6, 0, 1, 0, 5,
519 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
520 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
521
522
523 ostei_general_vrr_K(6, 0, 2, 0, 4,
524 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
525 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);
526
527
528 ostei_general_vrr_K(6, 0, 3, 0, 3,
529 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
530 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);
531
532
533
534 // Forming PRIM_INT__p_s_p_s[5 * 9];
535 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
536 {
537
538 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
539 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]);
540 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]);
541
542 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
543 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]);
544
545 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
546 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]);
547
548 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
549 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]);
550
551 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
552 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]);
553 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]);
554
555 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
556 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]);
557
558 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
559 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]);
560
561 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
562 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]);
563
564 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
565 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]);
566 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]);
567
568 }
569
570
571 VRR_K_d_s_d_s(
572 PRIM_INT__d_s_d_s,
573 PRIM_INT__d_s_p_s,
574 PRIM_INT__d_s_s_s,
575 PRIM_INT__p_s_p_s,
576 Q_PA,
577 a_over_q,
578 aoq_PQ,
579 one_over_2pq,
580 one_over_2q,
581 4);
582
583
584 ostei_general_vrr_K(3, 0, 3, 0, 3,
585 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
586 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);
587
588
589 ostei_general_vrr_K(4, 0, 4, 0, 2,
590 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
591 PRIM_INT__g_s_f_s, PRIM_INT__g_s_d_s, NULL, PRIM_INT__f_s_f_s, NULL, PRIM_INT__g_s_g_s);
592
593
594 ostei_general_vrr_K(5, 0, 5, 0, 1,
595 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
596 PRIM_INT__h_s_g_s, PRIM_INT__h_s_f_s, NULL, PRIM_INT__g_s_g_s, NULL, PRIM_INT__h_s_h_s);
597
598
599 ostei_general_vrr_K(6, 0, 4, 0, 2,
600 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
601 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);
602
603
604 ostei_general_vrr_K(6, 0, 5, 0, 1,
605 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
606 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);
607
608
609
610
611 ////////////////////////////////////
612 // Accumulate contracted integrals
613 ////////////////////////////////////
614 if(lastoffset == 0)
615 {
616 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
617 contract_all(315, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
618 contract_all(441, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
619 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
620 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
621 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
622 }
623 else
624 {
625 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
626 contract(315, shelloffsets, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
627 contract(441, shelloffsets, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
628 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
629 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
630 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
631 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
632 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
633 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
634 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
635 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
636 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
637 }
638
639 } // close loop over j
640 } // close loop over i
641
642 //Advance to the next batch
643 jstart = SIMINT_SIMD_ROUND(jend);
644
645 //////////////////////////////////////////////
646 // Contracted integrals: Horizontal recurrance
647 //////////////////////////////////////////////
648
649
650 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
651
652
653 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
654 {
655 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
656
657 // set up HRR pointers
658 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
659 double const * restrict HRR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
660 double const * restrict HRR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
661 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
662 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
663 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
664 double * restrict HRR_INT__h_p_f_d = INT__h_p_f_d + real_abcd * 3780;
665
666 // form INT__h_p_f_s
667 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);
668
669 // form INT__h_p_g_s
670 ostei_general_hrr_J(5, 1, 4, 0, hAB, HRR_INT__i_s_g_s, HRR_INT__h_s_g_s, HRR_INT__h_p_g_s);
671
672 // form INT__h_p_h_s
673 ostei_general_hrr_J(5, 1, 5, 0, hAB, HRR_INT__i_s_h_s, HRR_INT__h_s_h_s, HRR_INT__h_p_h_s);
674
675 // form INT__h_p_f_p
676 HRR_L_f_p(
677 HRR_INT__h_p_f_p,
678 HRR_INT__h_p_f_s,
679 HRR_INT__h_p_g_s,
680 hCD, 63);
681
682 // form INT__h_p_g_p
683 HRR_L_g_p(
684 HRR_INT__h_p_g_p,
685 HRR_INT__h_p_g_s,
686 HRR_INT__h_p_h_s,
687 hCD, 63);
688
689 // form INT__h_p_f_d
690 HRR_L_f_d(
691 HRR_INT__h_p_f_d,
692 HRR_INT__h_p_f_p,
693 HRR_INT__h_p_g_p,
694 hCD, 63);
695
696
697 } // close HRR loop
698
699
700 } // close loop cdbatch
701
702 istart = iend;
703 } // close loop over ab
704
705 return P.nshell12_clip * Q.nshell12_clip;
706 }
707
ostei_p_h_f_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_h_f_d)708 int ostei_p_h_f_d(struct simint_multi_shellpair const P,
709 struct simint_multi_shellpair const Q,
710 double screen_tol,
711 double * const restrict work,
712 double * const restrict INT__p_h_f_d)
713 {
714 double P_AB[3*P.nshell12];
715 struct simint_multi_shellpair P_tmp = P;
716 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
717 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
718 P_tmp.AB_x = P_AB;
719 P_tmp.AB_y = P_AB + P.nshell12;
720 P_tmp.AB_z = P_AB + 2*P.nshell12;
721
722 for(int i = 0; i < P.nshell12; i++)
723 {
724 P_tmp.AB_x[i] = -P.AB_x[i];
725 P_tmp.AB_y[i] = -P.AB_y[i];
726 P_tmp.AB_z[i] = -P.AB_z[i];
727 }
728
729 int ret = ostei_h_p_f_d(P_tmp, Q, screen_tol, work, INT__p_h_f_d);
730 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
731
732 for(int q = 0; q < ret; q++)
733 {
734 int idx = 0;
735 for(int a = 0; a < 3; ++a)
736 for(int b = 0; b < 21; ++b)
737 for(int c = 0; c < 10; ++c)
738 for(int d = 0; d < 6; ++d)
739 buffer[idx++] = INT__p_h_f_d[q*3780+b*180+a*60+c*6+d];
740
741 memcpy(INT__p_h_f_d+q*3780, buffer, 3780*sizeof(double));
742 }
743
744 return ret;
745 }
746
ostei_h_p_d_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_p_d_f)747 int ostei_h_p_d_f(struct simint_multi_shellpair const P,
748 struct simint_multi_shellpair const Q,
749 double screen_tol,
750 double * const restrict work,
751 double * const restrict INT__h_p_d_f)
752 {
753 double Q_AB[3*Q.nshell12];
754 struct simint_multi_shellpair Q_tmp = Q;
755 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
756 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
757 Q_tmp.AB_x = Q_AB;
758 Q_tmp.AB_y = Q_AB + Q.nshell12;
759 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
760
761 for(int i = 0; i < Q.nshell12; i++)
762 {
763 Q_tmp.AB_x[i] = -Q.AB_x[i];
764 Q_tmp.AB_y[i] = -Q.AB_y[i];
765 Q_tmp.AB_z[i] = -Q.AB_z[i];
766 }
767
768 int ret = ostei_h_p_f_d(P, Q_tmp, screen_tol, work, INT__h_p_d_f);
769 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
770
771 for(int q = 0; q < ret; q++)
772 {
773 int idx = 0;
774 for(int a = 0; a < 21; ++a)
775 for(int b = 0; b < 3; ++b)
776 for(int c = 0; c < 6; ++c)
777 for(int d = 0; d < 10; ++d)
778 buffer[idx++] = INT__h_p_d_f[q*3780+a*180+b*60+d*6+c];
779
780 memcpy(INT__h_p_d_f+q*3780, buffer, 3780*sizeof(double));
781 }
782
783 return ret;
784 }
785
ostei_p_h_d_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_h_d_f)786 int ostei_p_h_d_f(struct simint_multi_shellpair const P,
787 struct simint_multi_shellpair const Q,
788 double screen_tol,
789 double * const restrict work,
790 double * const restrict INT__p_h_d_f)
791 {
792 double P_AB[3*P.nshell12];
793 struct simint_multi_shellpair P_tmp = P;
794 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
795 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
796 P_tmp.AB_x = P_AB;
797 P_tmp.AB_y = P_AB + P.nshell12;
798 P_tmp.AB_z = P_AB + 2*P.nshell12;
799
800 for(int i = 0; i < P.nshell12; i++)
801 {
802 P_tmp.AB_x[i] = -P.AB_x[i];
803 P_tmp.AB_y[i] = -P.AB_y[i];
804 P_tmp.AB_z[i] = -P.AB_z[i];
805 }
806
807 double Q_AB[3*Q.nshell12];
808 struct simint_multi_shellpair Q_tmp = Q;
809 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
810 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
811 Q_tmp.AB_x = Q_AB;
812 Q_tmp.AB_y = Q_AB + Q.nshell12;
813 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
814
815 for(int i = 0; i < Q.nshell12; i++)
816 {
817 Q_tmp.AB_x[i] = -Q.AB_x[i];
818 Q_tmp.AB_y[i] = -Q.AB_y[i];
819 Q_tmp.AB_z[i] = -Q.AB_z[i];
820 }
821
822 int ret = ostei_h_p_f_d(P_tmp, Q_tmp, screen_tol, work, INT__p_h_d_f);
823 double buffer[3780] SIMINT_ALIGN_ARRAY_DBL;
824
825 for(int q = 0; q < ret; q++)
826 {
827 int idx = 0;
828 for(int a = 0; a < 3; ++a)
829 for(int b = 0; b < 21; ++b)
830 for(int c = 0; c < 6; ++c)
831 for(int d = 0; d < 10; ++d)
832 buffer[idx++] = INT__p_h_d_f[q*3780+b*180+a*60+d*6+c];
833
834 memcpy(INT__p_h_d_f+q*3780, buffer, 3780*sizeof(double));
835 }
836
837 return ret;
838 }
839
840