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_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_f_f)8 int ostei_h_p_f_f(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_f)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__h_p_f_f);
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__h_s_i_s = work + (SIMINT_NSHELL_SIMD * 966);
33 double * const INT__i_s_f_s = work + (SIMINT_NSHELL_SIMD * 1554);
34 double * const INT__i_s_g_s = work + (SIMINT_NSHELL_SIMD * 1834);
35 double * const INT__i_s_h_s = work + (SIMINT_NSHELL_SIMD * 2254);
36 double * const INT__i_s_i_s = work + (SIMINT_NSHELL_SIMD * 2842);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3626);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 13;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 31;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 67;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 121;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 211;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 277;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 385;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 565;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 805;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 905;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 1085;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 1385;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 1785;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 2235;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 2370;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 2640;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 3090;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 3690;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 4365;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 4995;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 5163;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 5541;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 6171;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 7011;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 7956;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 8838;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 9426;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 9622;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 10126;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 10966;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 12086;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 13346;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_i_s = primwork + 14522;
72 double * const hrrwork = (double *)(primwork + 15306);
73 double * const HRR_INT__h_p_f_s = hrrwork + 0;
74 double * const HRR_INT__h_p_f_p = hrrwork + 630;
75 double * const HRR_INT__h_p_f_d = hrrwork + 2520;
76 double * const HRR_INT__h_p_g_s = hrrwork + 6300;
77 double * const HRR_INT__h_p_g_p = hrrwork + 7245;
78 double * const HRR_INT__h_p_g_d = hrrwork + 10080;
79 double * const HRR_INT__h_p_h_s = hrrwork + 15750;
80 double * const HRR_INT__h_p_h_p = hrrwork + 17073;
81 double * const HRR_INT__h_p_i_s = hrrwork + 21042;
82
83
84 // Create constants
85 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
86 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
87 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
88 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
89 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
90 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
91 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
92
93
94 ////////////////////////////////////////
95 // Loop over shells and primitives
96 ////////////////////////////////////////
97
98 real_abcd = 0;
99 istart = 0;
100 for(ab = 0; ab < P.nshell12_clip; ++ab)
101 {
102 const int iend = istart + P.nprim12[ab];
103
104 cd = 0;
105 jstart = 0;
106
107 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
108 {
109 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
110 int jend = jstart;
111 for(i = 0; i < nshellbatch; i++)
112 jend += Q.nprim12[cd+i];
113
114 // Clear the beginning of the workspace (where we are accumulating integrals)
115 memset(work, 0, SIMINT_NSHELL_SIMD * 3626 * sizeof(double));
116 abcd = 0;
117
118
119 for(i = istart; i < iend; ++i)
120 {
121 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
122
123 if(check_screen)
124 {
125 // Skip this whole thing if always insignificant
126 if((P.screen[i] * Q.screen_max) < screen_tol)
127 continue;
128 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
129 }
130
131 icd = 0;
132 iprimcd = 0;
133 nprim_icd = Q.nprim12[cd];
134 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
135 double * restrict PRIM_PTR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
136 double * restrict PRIM_PTR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
137 double * restrict PRIM_PTR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
138 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
139 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
140 double * restrict PRIM_PTR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
141 double * restrict PRIM_PTR_INT__i_s_i_s = INT__i_s_i_s + abcd * 784;
142
143
144
145 // Load these one per loop over i
146 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
147 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
148 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
149
150 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
151
152 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
153 {
154 // calculate the shell offsets
155 // these are the offset from the shell pointed to by cd
156 // for each element
157 int shelloffsets[SIMINT_SIMD_LEN] = {0};
158 int lastoffset = 0;
159 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
160
161 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
162 {
163 // Handle if the first element of the vector is a new shell
164 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
165 {
166 nprim_icd += Q.nprim12[cd + (++icd)];
167 PRIM_PTR_INT__h_s_f_s += 210;
168 PRIM_PTR_INT__h_s_g_s += 315;
169 PRIM_PTR_INT__h_s_h_s += 441;
170 PRIM_PTR_INT__h_s_i_s += 588;
171 PRIM_PTR_INT__i_s_f_s += 280;
172 PRIM_PTR_INT__i_s_g_s += 420;
173 PRIM_PTR_INT__i_s_h_s += 588;
174 PRIM_PTR_INT__i_s_i_s += 784;
175 }
176 iprimcd++;
177 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
178 {
179 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
180 {
181 shelloffsets[n] = shelloffsets[n-1] + 1;
182 lastoffset++;
183 nprim_icd += Q.nprim12[cd + (++icd)];
184 }
185 else
186 shelloffsets[n] = shelloffsets[n-1];
187 iprimcd++;
188 }
189 }
190 else
191 iprimcd += SIMINT_SIMD_LEN;
192
193 // Do we have to compute this vector (or has it been screened out)?
194 // (not_screened != 0 means we have to do this vector)
195 if(check_screen)
196 {
197 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
198 if(vmax < screen_tol)
199 {
200 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
201 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
202 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
203 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
204 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
205 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
206 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
207 PRIM_PTR_INT__i_s_i_s += lastoffset*784;
208 continue;
209 }
210 }
211
212 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
213 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
214 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
215 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
216
217
218 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
219 SIMINT_DBLTYPE PQ[3];
220 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
221 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
222 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
223 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
224 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
225 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
226
227 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
228 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
229 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
230 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
231 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
232 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
233 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
234
235 // NOTE: Minus sign!
236 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
237 SIMINT_DBLTYPE aop_PQ[3];
238 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
239 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
240 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
241
242 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
243 SIMINT_DBLTYPE aoq_PQ[3];
244 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
245 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
246 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
247 // Put a minus sign here so we don't have to in RR routines
248 a_over_q = SIMINT_NEG(a_over_q);
249
250
251 //////////////////////////////////////////////
252 // Fjt function section
253 // Maximum v value: 12
254 //////////////////////////////////////////////
255 // The parameter to the Fjt function
256 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
257
258
259 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
260
261
262 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
263 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
264 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
265 for(n = 0; n <= 12; n++)
266 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
267
268 //////////////////////////////////////////////
269 // Primitive integrals: Vertical recurrance
270 //////////////////////////////////////////////
271
272 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
273 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
274 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
275 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
276 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
277 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
278 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
279 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
280 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
281 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
282 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
283 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
284 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
285 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
286 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
288
289
290
291 // Forming PRIM_INT__p_s_s_s[12 * 3];
292 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
293 {
294
295 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
296 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]);
297
298 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
299 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]);
300
301 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
302 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]);
303
304 }
305
306
307
308 // Forming PRIM_INT__d_s_s_s[11 * 6];
309 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
310 {
311
312 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
313 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]);
314 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]);
315
316 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
317 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]);
318
319 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
320 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]);
321
322 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
323 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]);
324 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]);
325
326 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
327 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]);
328
329 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
330 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]);
331 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]);
332
333 }
334
335
336
337 // Forming PRIM_INT__f_s_s_s[10 * 10];
338 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
339 {
340
341 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
342 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]);
343 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]);
344
345 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
346 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]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
349 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]);
350
351 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
352 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]);
353
354 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
355 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]);
356
357 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
358 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]);
359
360 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
361 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]);
362 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]);
363
364 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
365 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]);
366
367 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
368 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]);
369
370 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
371 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]);
372 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]);
373
374 }
375
376
377 VRR_I_g_s_s_s(
378 PRIM_INT__g_s_s_s,
379 PRIM_INT__f_s_s_s,
380 PRIM_INT__d_s_s_s,
381 P_PA,
382 a_over_p,
383 aop_PQ,
384 one_over_2p,
385 9);
386
387
388 VRR_I_h_s_s_s(
389 PRIM_INT__h_s_s_s,
390 PRIM_INT__g_s_s_s,
391 PRIM_INT__f_s_s_s,
392 P_PA,
393 a_over_p,
394 aop_PQ,
395 one_over_2p,
396 8);
397
398
399 ostei_general_vrr_K(5, 0, 1, 0, 6,
400 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
401 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
402
403
404 VRR_K_g_s_p_s(
405 PRIM_INT__g_s_p_s,
406 PRIM_INT__g_s_s_s,
407 PRIM_INT__f_s_s_s,
408 Q_PA,
409 aoq_PQ,
410 one_over_2pq,
411 6);
412
413
414 ostei_general_vrr_K(5, 0, 2, 0, 5,
415 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
416 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);
417
418
419 VRR_K_f_s_p_s(
420 PRIM_INT__f_s_p_s,
421 PRIM_INT__f_s_s_s,
422 PRIM_INT__d_s_s_s,
423 Q_PA,
424 aoq_PQ,
425 one_over_2pq,
426 6);
427
428
429 ostei_general_vrr_K(4, 0, 2, 0, 5,
430 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
431 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);
432
433
434 ostei_general_vrr_K(5, 0, 3, 0, 4,
435 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
436 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);
437
438
439
440 // Forming PRIM_INT__d_s_p_s[6 * 18];
441 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
442 {
443
444 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
445 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]);
446 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]);
447
448 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
449 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]);
450
451 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
452 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]);
453
454 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
455 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]);
456 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]);
457
458 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
459 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]);
460 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]);
461
462 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
463 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]);
464
465 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
466 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]);
467 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]);
468
469 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
470 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]);
471
472 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
473 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]);
474 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]);
475
476 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
477 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]);
478
479 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
480 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]);
481 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]);
482
483 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
484 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]);
485
486 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
487 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]);
488
489 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
490 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]);
491 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]);
492
493 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
494 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]);
495 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]);
496
497 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
498 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]);
499
500 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
501 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]);
502
503 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
504 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]);
505 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]);
506
507 }
508
509
510 VRR_K_f_s_d_s(
511 PRIM_INT__f_s_d_s,
512 PRIM_INT__f_s_p_s,
513 PRIM_INT__f_s_s_s,
514 PRIM_INT__d_s_p_s,
515 Q_PA,
516 a_over_q,
517 aoq_PQ,
518 one_over_2pq,
519 one_over_2q,
520 5);
521
522
523 ostei_general_vrr_K(4, 0, 3, 0, 4,
524 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
525 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);
526
527
528 ostei_general_vrr_K(5, 0, 4, 0, 3,
529 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
530 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);
531
532
533 ostei_general_vrr1_I(6, 7,
534 one_over_2p, a_over_p, aop_PQ, P_PA,
535 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
536
537
538 ostei_general_vrr_K(6, 0, 1, 0, 6,
539 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
540 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
541
542
543 ostei_general_vrr_K(6, 0, 2, 0, 5,
544 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
545 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);
546
547
548 ostei_general_vrr_K(6, 0, 3, 0, 4,
549 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
550 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);
551
552
553
554 // Forming PRIM_INT__p_s_p_s[6 * 9];
555 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
556 {
557
558 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
559 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]);
560 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]);
561
562 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
563 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]);
564
565 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
566 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]);
567
568 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
569 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]);
570
571 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
572 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]);
573 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]);
574
575 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
576 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]);
577
578 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
579 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]);
580
581 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
582 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]);
583
584 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
585 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]);
586 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]);
587
588 }
589
590
591 VRR_K_d_s_d_s(
592 PRIM_INT__d_s_d_s,
593 PRIM_INT__d_s_p_s,
594 PRIM_INT__d_s_s_s,
595 PRIM_INT__p_s_p_s,
596 Q_PA,
597 a_over_q,
598 aoq_PQ,
599 one_over_2pq,
600 one_over_2q,
601 5);
602
603
604 ostei_general_vrr_K(3, 0, 3, 0, 4,
605 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
606 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);
607
608
609 ostei_general_vrr_K(4, 0, 4, 0, 3,
610 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
611 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);
612
613
614 ostei_general_vrr_K(5, 0, 5, 0, 2,
615 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
616 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);
617
618
619 ostei_general_vrr_K(6, 0, 4, 0, 3,
620 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
621 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);
622
623
624
625 // Forming PRIM_INT__s_s_p_s[6 * 3];
626 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
627 {
628
629 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
630 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]);
631
632 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
633 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 1]);
634
635 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
636 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 2]);
637
638 }
639
640
641
642 // Forming PRIM_INT__p_s_d_s[5 * 18];
643 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
644 {
645
646 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_p_s[n * 9 + 0]);
647 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_p_s[(n+1) * 9 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
648 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__p_s_d_s[n * 18 + 0]);
649 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
650
651 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_p_s[n * 9 + 1]);
652 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_p_s[(n+1) * 9 + 1], PRIM_INT__p_s_d_s[n * 18 + 3]);
653 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__p_s_d_s[n * 18 + 3]);
654
655 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_p_s[n * 9 + 2]);
656 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_p_s[(n+1) * 9 + 2], PRIM_INT__p_s_d_s[n * 18 + 5]);
657 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 0], PRIM_INT__p_s_s_s[n * 3 + 0]), PRIM_INT__p_s_d_s[n * 18 + 5]);
658
659 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_p_s[n * 9 + 3]);
660 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_p_s[(n+1) * 9 + 3], PRIM_INT__p_s_d_s[n * 18 + 6]);
661 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__p_s_d_s[n * 18 + 6]);
662
663 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_p_s[n * 9 + 4]);
664 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_p_s[(n+1) * 9 + 4], PRIM_INT__p_s_d_s[n * 18 + 9]);
665 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__p_s_d_s[n * 18 + 9]);
666 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 9]);
667
668 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_p_s[n * 9 + 5]);
669 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_p_s[(n+1) * 9 + 5], PRIM_INT__p_s_d_s[n * 18 + 11]);
670 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 1], PRIM_INT__p_s_s_s[n * 3 + 1]), PRIM_INT__p_s_d_s[n * 18 + 11]);
671
672 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_p_s[n * 9 + 6]);
673 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__p_s_p_s[(n+1) * 9 + 6], PRIM_INT__p_s_d_s[n * 18 + 12]);
674 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__p_s_d_s[n * 18 + 12]);
675
676 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_p_s[n * 9 + 7]);
677 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__p_s_p_s[(n+1) * 9 + 7], PRIM_INT__p_s_d_s[n * 18 + 15]);
678 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__p_s_d_s[n * 18 + 15]);
679
680 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_p_s[n * 9 + 8]);
681 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__p_s_p_s[(n+1) * 9 + 8], PRIM_INT__p_s_d_s[n * 18 + 17]);
682 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__p_s_s_s[(n+1) * 3 + 2], PRIM_INT__p_s_s_s[n * 3 + 2]), PRIM_INT__p_s_d_s[n * 18 + 17]);
683 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 17]);
684
685 }
686
687
688 VRR_K_d_s_f_s(
689 PRIM_INT__d_s_f_s,
690 PRIM_INT__d_s_d_s,
691 PRIM_INT__d_s_p_s,
692 PRIM_INT__p_s_d_s,
693 Q_PA,
694 a_over_q,
695 aoq_PQ,
696 one_over_2pq,
697 one_over_2q,
698 4);
699
700
701 ostei_general_vrr_K(3, 0, 4, 0, 3,
702 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
703 PRIM_INT__f_s_f_s, PRIM_INT__f_s_d_s, NULL, PRIM_INT__d_s_f_s, NULL, PRIM_INT__f_s_g_s);
704
705
706 ostei_general_vrr_K(4, 0, 5, 0, 2,
707 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
708 PRIM_INT__g_s_g_s, PRIM_INT__g_s_f_s, NULL, PRIM_INT__f_s_g_s, NULL, PRIM_INT__g_s_h_s);
709
710
711 ostei_general_vrr_K(5, 0, 6, 0, 1,
712 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
713 PRIM_INT__h_s_h_s, PRIM_INT__h_s_g_s, NULL, PRIM_INT__g_s_h_s, NULL, PRIM_INT__h_s_i_s);
714
715
716 ostei_general_vrr_K(6, 0, 5, 0, 2,
717 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
718 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);
719
720
721 ostei_general_vrr_K(6, 0, 6, 0, 1,
722 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
723 PRIM_INT__i_s_h_s, PRIM_INT__i_s_g_s, NULL, PRIM_INT__h_s_h_s, NULL, PRIM_INT__i_s_i_s);
724
725
726
727
728 ////////////////////////////////////
729 // Accumulate contracted integrals
730 ////////////////////////////////////
731 if(lastoffset == 0)
732 {
733 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
734 contract_all(315, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
735 contract_all(441, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
736 contract_all(588, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
737 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
738 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
739 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
740 contract_all(784, PRIM_INT__i_s_i_s, PRIM_PTR_INT__i_s_i_s);
741 }
742 else
743 {
744 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
745 contract(315, shelloffsets, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
746 contract(441, shelloffsets, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
747 contract(588, shelloffsets, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
748 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
749 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
750 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
751 contract(784, shelloffsets, PRIM_INT__i_s_i_s, PRIM_PTR_INT__i_s_i_s);
752 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
753 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
754 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
755 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
756 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
757 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
758 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
759 PRIM_PTR_INT__i_s_i_s += lastoffset*784;
760 }
761
762 } // close loop over j
763 } // close loop over i
764
765 //Advance to the next batch
766 jstart = SIMINT_SIMD_ROUND(jend);
767
768 //////////////////////////////////////////////
769 // Contracted integrals: Horizontal recurrance
770 //////////////////////////////////////////////
771
772
773 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
774
775
776 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
777 {
778 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
779
780 // set up HRR pointers
781 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
782 double const * restrict HRR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
783 double const * restrict HRR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
784 double const * restrict HRR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
785 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
786 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
787 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
788 double const * restrict HRR_INT__i_s_i_s = INT__i_s_i_s + abcd * 784;
789 double * restrict HRR_INT__h_p_f_f = INT__h_p_f_f + real_abcd * 6300;
790
791 // form INT__h_p_f_s
792 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);
793
794 // form INT__h_p_g_s
795 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);
796
797 // form INT__h_p_h_s
798 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);
799
800 // form INT__h_p_i_s
801 ostei_general_hrr_J(5, 1, 6, 0, hAB, HRR_INT__i_s_i_s, HRR_INT__h_s_i_s, HRR_INT__h_p_i_s);
802
803 // form INT__h_p_f_p
804 HRR_L_f_p(
805 HRR_INT__h_p_f_p,
806 HRR_INT__h_p_f_s,
807 HRR_INT__h_p_g_s,
808 hCD, 63);
809
810 // form INT__h_p_g_p
811 HRR_L_g_p(
812 HRR_INT__h_p_g_p,
813 HRR_INT__h_p_g_s,
814 HRR_INT__h_p_h_s,
815 hCD, 63);
816
817 // form INT__h_p_h_p
818 ostei_general_hrr_L(5, 1, 5, 1, hCD, HRR_INT__h_p_i_s, HRR_INT__h_p_h_s, HRR_INT__h_p_h_p);
819
820 // form INT__h_p_f_d
821 HRR_L_f_d(
822 HRR_INT__h_p_f_d,
823 HRR_INT__h_p_f_p,
824 HRR_INT__h_p_g_p,
825 hCD, 63);
826
827 // form INT__h_p_g_d
828 ostei_general_hrr_L(5, 1, 4, 2, hCD, HRR_INT__h_p_h_p, HRR_INT__h_p_g_p, HRR_INT__h_p_g_d);
829
830 // form INT__h_p_f_f
831 ostei_general_hrr_L(5, 1, 3, 3, hCD, HRR_INT__h_p_g_d, HRR_INT__h_p_f_d, HRR_INT__h_p_f_f);
832
833
834 } // close HRR loop
835
836
837 } // close loop cdbatch
838
839 istart = iend;
840 } // close loop over ab
841
842 return P.nshell12_clip * Q.nshell12_clip;
843 }
844
ostei_p_h_f_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_f_f)845 int ostei_p_h_f_f(struct simint_multi_shellpair const P,
846 struct simint_multi_shellpair const Q,
847 double screen_tol,
848 double * const restrict work,
849 double * const restrict INT__p_h_f_f)
850 {
851 double P_AB[3*P.nshell12];
852 struct simint_multi_shellpair P_tmp = P;
853 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
854 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
855 P_tmp.AB_x = P_AB;
856 P_tmp.AB_y = P_AB + P.nshell12;
857 P_tmp.AB_z = P_AB + 2*P.nshell12;
858
859 for(int i = 0; i < P.nshell12; i++)
860 {
861 P_tmp.AB_x[i] = -P.AB_x[i];
862 P_tmp.AB_y[i] = -P.AB_y[i];
863 P_tmp.AB_z[i] = -P.AB_z[i];
864 }
865
866 int ret = ostei_h_p_f_f(P_tmp, Q, screen_tol, work, INT__p_h_f_f);
867 double buffer[6300] SIMINT_ALIGN_ARRAY_DBL;
868
869 for(int q = 0; q < ret; q++)
870 {
871 int idx = 0;
872 for(int a = 0; a < 3; ++a)
873 for(int b = 0; b < 21; ++b)
874 for(int c = 0; c < 10; ++c)
875 for(int d = 0; d < 10; ++d)
876 buffer[idx++] = INT__p_h_f_f[q*6300+b*300+a*100+c*10+d];
877
878 memcpy(INT__p_h_f_f+q*6300, buffer, 6300*sizeof(double));
879 }
880
881 return ret;
882 }
883
884