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_d_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_d_f_d)8 int ostei_h_d_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_d_f_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__h_d_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 double * const INT__k_s_f_s = work + (SIMINT_NSHELL_SIMD * 2254);
36 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 2614);
37 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 3154);
38 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3910);
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 13;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 49;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 94;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 160;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 250;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 394;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 494;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 644;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 884;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 1184;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1319;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1544;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1904;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 2354;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 2804;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2972;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 3287;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 3791;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 4421;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 5051;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 5492;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 5688;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 6108;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 6780;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 7620;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 8460;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 9048;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 9264;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 9804;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 10668;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 11748;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 12828;
72 double * const hrrwork = (double *)(primwork + 13584);
73 double * const HRR_INT__h_p_f_s = hrrwork + 0;
74 double * const HRR_INT__h_p_g_s = hrrwork + 630;
75 double * const HRR_INT__h_p_h_s = hrrwork + 1575;
76 double * const HRR_INT__h_d_f_s = hrrwork + 2898;
77 double * const HRR_INT__h_d_f_p = hrrwork + 4158;
78 double * const HRR_INT__h_d_g_s = hrrwork + 7938;
79 double * const HRR_INT__h_d_g_p = hrrwork + 9828;
80 double * const HRR_INT__h_d_h_s = hrrwork + 15498;
81 double * const HRR_INT__i_p_f_s = hrrwork + 18144;
82 double * const HRR_INT__i_p_g_s = hrrwork + 18984;
83 double * const HRR_INT__i_p_h_s = hrrwork + 20244;
84
85
86 // Create constants
87 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
88 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
89 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
90 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
91 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
92 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
93 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
94 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
95
96
97 ////////////////////////////////////////
98 // Loop over shells and primitives
99 ////////////////////////////////////////
100
101 real_abcd = 0;
102 istart = 0;
103 for(ab = 0; ab < P.nshell12_clip; ++ab)
104 {
105 const int iend = istart + P.nprim12[ab];
106
107 cd = 0;
108 jstart = 0;
109
110 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
111 {
112 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
113 int jend = jstart;
114 for(i = 0; i < nshellbatch; i++)
115 jend += Q.nprim12[cd+i];
116
117 // Clear the beginning of the workspace (where we are accumulating integrals)
118 memset(work, 0, SIMINT_NSHELL_SIMD * 3910 * sizeof(double));
119 abcd = 0;
120
121
122 for(i = istart; i < iend; ++i)
123 {
124 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
125
126 if(check_screen)
127 {
128 // Skip this whole thing if always insignificant
129 if((P.screen[i] * Q.screen_max) < screen_tol)
130 continue;
131 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
132 }
133
134 icd = 0;
135 iprimcd = 0;
136 nprim_icd = Q.nprim12[cd];
137 double * restrict PRIM_PTR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
138 double * restrict PRIM_PTR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
139 double * restrict PRIM_PTR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
140 double * restrict PRIM_PTR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
141 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
142 double * restrict PRIM_PTR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
143 double * restrict PRIM_PTR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
144 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
145 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
146
147
148
149 // Load these one per loop over i
150 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
151 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
152 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
153
154 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
155
156 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
157 {
158 // calculate the shell offsets
159 // these are the offset from the shell pointed to by cd
160 // for each element
161 int shelloffsets[SIMINT_SIMD_LEN] = {0};
162 int lastoffset = 0;
163 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
164
165 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
166 {
167 // Handle if the first element of the vector is a new shell
168 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
169 {
170 nprim_icd += Q.nprim12[cd + (++icd)];
171 PRIM_PTR_INT__h_s_f_s += 210;
172 PRIM_PTR_INT__h_s_g_s += 315;
173 PRIM_PTR_INT__h_s_h_s += 441;
174 PRIM_PTR_INT__i_s_f_s += 280;
175 PRIM_PTR_INT__i_s_g_s += 420;
176 PRIM_PTR_INT__i_s_h_s += 588;
177 PRIM_PTR_INT__k_s_f_s += 360;
178 PRIM_PTR_INT__k_s_g_s += 540;
179 PRIM_PTR_INT__k_s_h_s += 756;
180 }
181 iprimcd++;
182 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
183 {
184 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
185 {
186 shelloffsets[n] = shelloffsets[n-1] + 1;
187 lastoffset++;
188 nprim_icd += Q.nprim12[cd + (++icd)];
189 }
190 else
191 shelloffsets[n] = shelloffsets[n-1];
192 iprimcd++;
193 }
194 }
195 else
196 iprimcd += SIMINT_SIMD_LEN;
197
198 // Do we have to compute this vector (or has it been screened out)?
199 // (not_screened != 0 means we have to do this vector)
200 if(check_screen)
201 {
202 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
203 if(vmax < screen_tol)
204 {
205 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
206 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
207 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
208 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
209 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
210 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
211 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
212 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
213 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
214 continue;
215 }
216 }
217
218 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
219 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
220 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
221 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
222
223
224 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
225 SIMINT_DBLTYPE PQ[3];
226 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
227 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
228 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
229 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
230 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
231 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
232
233 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
234 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
235 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
236 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
237 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
238 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
239 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
240
241 // NOTE: Minus sign!
242 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
243 SIMINT_DBLTYPE aop_PQ[3];
244 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
245 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
246 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
247
248 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
249 SIMINT_DBLTYPE aoq_PQ[3];
250 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
251 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
252 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
253 // Put a minus sign here so we don't have to in RR routines
254 a_over_q = SIMINT_NEG(a_over_q);
255
256
257 //////////////////////////////////////////////
258 // Fjt function section
259 // Maximum v value: 12
260 //////////////////////////////////////////////
261 // The parameter to the Fjt function
262 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
263
264
265 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
266
267
268 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
269 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
270 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
271 for(n = 0; n <= 12; n++)
272 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
273
274 //////////////////////////////////////////////
275 // Primitive integrals: Vertical recurrance
276 //////////////////////////////////////////////
277
278 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
279 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
280 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
281 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
282 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
283 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, 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
296
297
298 // Forming PRIM_INT__p_s_s_s[12 * 3];
299 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
300 {
301
302 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
303 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]);
304
305 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
306 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]);
307
308 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
309 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]);
310
311 }
312
313
314
315 // Forming PRIM_INT__d_s_s_s[11 * 6];
316 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
317 {
318
319 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
320 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]);
321 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]);
322
323 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
324 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]);
325
326 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
327 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]);
328
329 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
330 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]);
331 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]);
332
333 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
334 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]);
335
336 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
337 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]);
338 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]);
339
340 }
341
342
343
344 // Forming PRIM_INT__f_s_s_s[10 * 10];
345 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
346 {
347
348 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
349 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]);
350 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]);
351
352 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
353 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]);
354
355 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
356 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]);
357
358 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
359 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]);
360
361 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
362 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]);
363
364 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
365 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]);
366
367 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
368 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]);
369 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]);
370
371 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
372 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]);
373
374 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
375 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]);
376
377 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
378 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]);
379 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]);
380
381 }
382
383
384 VRR_I_g_s_s_s(
385 PRIM_INT__g_s_s_s,
386 PRIM_INT__f_s_s_s,
387 PRIM_INT__d_s_s_s,
388 P_PA,
389 a_over_p,
390 aop_PQ,
391 one_over_2p,
392 9);
393
394
395 VRR_I_h_s_s_s(
396 PRIM_INT__h_s_s_s,
397 PRIM_INT__g_s_s_s,
398 PRIM_INT__f_s_s_s,
399 P_PA,
400 a_over_p,
401 aop_PQ,
402 one_over_2p,
403 8);
404
405
406 ostei_general_vrr_K(5, 0, 1, 0, 5,
407 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
408 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
409
410
411 VRR_K_g_s_p_s(
412 PRIM_INT__g_s_p_s,
413 PRIM_INT__g_s_s_s,
414 PRIM_INT__f_s_s_s,
415 Q_PA,
416 aoq_PQ,
417 one_over_2pq,
418 5);
419
420
421 ostei_general_vrr_K(5, 0, 2, 0, 4,
422 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
423 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);
424
425
426 VRR_K_f_s_p_s(
427 PRIM_INT__f_s_p_s,
428 PRIM_INT__f_s_s_s,
429 PRIM_INT__d_s_s_s,
430 Q_PA,
431 aoq_PQ,
432 one_over_2pq,
433 5);
434
435
436 ostei_general_vrr_K(4, 0, 2, 0, 4,
437 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
438 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);
439
440
441 ostei_general_vrr_K(5, 0, 3, 0, 3,
442 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
443 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);
444
445
446
447 // Forming PRIM_INT__d_s_p_s[5 * 18];
448 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
449 {
450
451 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
452 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]);
453 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]);
454
455 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
456 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]);
457
458 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
459 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]);
460
461 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
462 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]);
463 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]);
464
465 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
466 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]);
467 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]);
468
469 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
470 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]);
471
472 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
473 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]);
474 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]);
475
476 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
477 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]);
478
479 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
480 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]);
481 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]);
482
483 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
484 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]);
485
486 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
487 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]);
488 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]);
489
490 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
491 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]);
492
493 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
494 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]);
495
496 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
497 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]);
498 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]);
499
500 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
501 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]);
502 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]);
503
504 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
505 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]);
506
507 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
508 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]);
509
510 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
511 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]);
512 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]);
513
514 }
515
516
517 VRR_K_f_s_d_s(
518 PRIM_INT__f_s_d_s,
519 PRIM_INT__f_s_p_s,
520 PRIM_INT__f_s_s_s,
521 PRIM_INT__d_s_p_s,
522 Q_PA,
523 a_over_q,
524 aoq_PQ,
525 one_over_2pq,
526 one_over_2q,
527 4);
528
529
530 ostei_general_vrr_K(4, 0, 3, 0, 3,
531 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
532 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);
533
534
535 ostei_general_vrr_K(5, 0, 4, 0, 2,
536 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
537 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);
538
539
540 ostei_general_vrr1_I(6, 7,
541 one_over_2p, a_over_p, aop_PQ, P_PA,
542 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
543
544
545 ostei_general_vrr_K(6, 0, 1, 0, 5,
546 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
547 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
548
549
550 ostei_general_vrr_K(6, 0, 2, 0, 4,
551 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
552 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);
553
554
555 ostei_general_vrr_K(6, 0, 3, 0, 3,
556 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
557 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);
558
559
560
561 // Forming PRIM_INT__p_s_p_s[5 * 9];
562 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
563 {
564
565 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
566 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]);
567 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]);
568
569 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
570 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]);
571
572 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
573 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]);
574
575 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
576 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]);
577
578 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
579 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]);
580 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]);
581
582 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
583 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]);
584
585 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
586 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]);
587
588 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
589 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]);
590
591 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
592 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]);
593 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]);
594
595 }
596
597
598 VRR_K_d_s_d_s(
599 PRIM_INT__d_s_d_s,
600 PRIM_INT__d_s_p_s,
601 PRIM_INT__d_s_s_s,
602 PRIM_INT__p_s_p_s,
603 Q_PA,
604 a_over_q,
605 aoq_PQ,
606 one_over_2pq,
607 one_over_2q,
608 4);
609
610
611 ostei_general_vrr_K(3, 0, 3, 0, 3,
612 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
613 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);
614
615
616 ostei_general_vrr_K(4, 0, 4, 0, 2,
617 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
618 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);
619
620
621 ostei_general_vrr_K(5, 0, 5, 0, 1,
622 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
623 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);
624
625
626 ostei_general_vrr_K(6, 0, 4, 0, 2,
627 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
628 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);
629
630
631 ostei_general_vrr1_I(7, 6,
632 one_over_2p, a_over_p, aop_PQ, P_PA,
633 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
634
635
636 ostei_general_vrr_K(7, 0, 1, 0, 5,
637 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
638 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
639
640
641 ostei_general_vrr_K(7, 0, 2, 0, 4,
642 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
643 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);
644
645
646 ostei_general_vrr_K(7, 0, 3, 0, 3,
647 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
648 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);
649
650
651 ostei_general_vrr_K(6, 0, 5, 0, 1,
652 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
653 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);
654
655
656 ostei_general_vrr_K(7, 0, 4, 0, 2,
657 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
658 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);
659
660
661 ostei_general_vrr_K(7, 0, 5, 0, 1,
662 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
663 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);
664
665
666
667
668 ////////////////////////////////////
669 // Accumulate contracted integrals
670 ////////////////////////////////////
671 if(lastoffset == 0)
672 {
673 contract_all(210, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
674 contract_all(315, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
675 contract_all(441, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
676 contract_all(280, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
677 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
678 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
679 contract_all(360, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
680 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
681 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
682 }
683 else
684 {
685 contract(210, shelloffsets, PRIM_INT__h_s_f_s, PRIM_PTR_INT__h_s_f_s);
686 contract(315, shelloffsets, PRIM_INT__h_s_g_s, PRIM_PTR_INT__h_s_g_s);
687 contract(441, shelloffsets, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
688 contract(280, shelloffsets, PRIM_INT__i_s_f_s, PRIM_PTR_INT__i_s_f_s);
689 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
690 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
691 contract(360, shelloffsets, PRIM_INT__k_s_f_s, PRIM_PTR_INT__k_s_f_s);
692 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
693 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
694 PRIM_PTR_INT__h_s_f_s += lastoffset*210;
695 PRIM_PTR_INT__h_s_g_s += lastoffset*315;
696 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
697 PRIM_PTR_INT__i_s_f_s += lastoffset*280;
698 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
699 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
700 PRIM_PTR_INT__k_s_f_s += lastoffset*360;
701 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
702 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
703 }
704
705 } // close loop over j
706 } // close loop over i
707
708 //Advance to the next batch
709 jstart = SIMINT_SIMD_ROUND(jend);
710
711 //////////////////////////////////////////////
712 // Contracted integrals: Horizontal recurrance
713 //////////////////////////////////////////////
714
715
716 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
717
718
719 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
720 {
721 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
722
723 // set up HRR pointers
724 double const * restrict HRR_INT__h_s_f_s = INT__h_s_f_s + abcd * 210;
725 double const * restrict HRR_INT__h_s_g_s = INT__h_s_g_s + abcd * 315;
726 double const * restrict HRR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
727 double const * restrict HRR_INT__i_s_f_s = INT__i_s_f_s + abcd * 280;
728 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
729 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
730 double const * restrict HRR_INT__k_s_f_s = INT__k_s_f_s + abcd * 360;
731 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
732 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
733 double * restrict HRR_INT__h_d_f_d = INT__h_d_f_d + real_abcd * 7560;
734
735 // form INT__h_p_f_s
736 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);
737
738 // form INT__h_p_g_s
739 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);
740
741 // form INT__h_p_h_s
742 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);
743
744 // form INT__i_p_f_s
745 ostei_general_hrr_J(6, 1, 3, 0, hAB, HRR_INT__k_s_f_s, HRR_INT__i_s_f_s, HRR_INT__i_p_f_s);
746
747 // form INT__i_p_g_s
748 ostei_general_hrr_J(6, 1, 4, 0, hAB, HRR_INT__k_s_g_s, HRR_INT__i_s_g_s, HRR_INT__i_p_g_s);
749
750 // form INT__i_p_h_s
751 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);
752
753 // form INT__h_d_f_s
754 ostei_general_hrr_J(5, 2, 3, 0, hAB, HRR_INT__i_p_f_s, HRR_INT__h_p_f_s, HRR_INT__h_d_f_s);
755
756 // form INT__h_d_g_s
757 ostei_general_hrr_J(5, 2, 4, 0, hAB, HRR_INT__i_p_g_s, HRR_INT__h_p_g_s, HRR_INT__h_d_g_s);
758
759 // form INT__h_d_h_s
760 ostei_general_hrr_J(5, 2, 5, 0, hAB, HRR_INT__i_p_h_s, HRR_INT__h_p_h_s, HRR_INT__h_d_h_s);
761
762 // form INT__h_d_f_p
763 HRR_L_f_p(
764 HRR_INT__h_d_f_p,
765 HRR_INT__h_d_f_s,
766 HRR_INT__h_d_g_s,
767 hCD, 126);
768
769 // form INT__h_d_g_p
770 HRR_L_g_p(
771 HRR_INT__h_d_g_p,
772 HRR_INT__h_d_g_s,
773 HRR_INT__h_d_h_s,
774 hCD, 126);
775
776 // form INT__h_d_f_d
777 HRR_L_f_d(
778 HRR_INT__h_d_f_d,
779 HRR_INT__h_d_f_p,
780 HRR_INT__h_d_g_p,
781 hCD, 126);
782
783
784 } // close HRR loop
785
786
787 } // close loop cdbatch
788
789 istart = iend;
790 } // close loop over ab
791
792 return P.nshell12_clip * Q.nshell12_clip;
793 }
794
ostei_d_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__d_h_f_d)795 int ostei_d_h_f_d(struct simint_multi_shellpair const P,
796 struct simint_multi_shellpair const Q,
797 double screen_tol,
798 double * const restrict work,
799 double * const restrict INT__d_h_f_d)
800 {
801 double P_AB[3*P.nshell12];
802 struct simint_multi_shellpair P_tmp = P;
803 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
804 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
805 P_tmp.AB_x = P_AB;
806 P_tmp.AB_y = P_AB + P.nshell12;
807 P_tmp.AB_z = P_AB + 2*P.nshell12;
808
809 for(int i = 0; i < P.nshell12; i++)
810 {
811 P_tmp.AB_x[i] = -P.AB_x[i];
812 P_tmp.AB_y[i] = -P.AB_y[i];
813 P_tmp.AB_z[i] = -P.AB_z[i];
814 }
815
816 int ret = ostei_h_d_f_d(P_tmp, Q, screen_tol, work, INT__d_h_f_d);
817 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
818
819 for(int q = 0; q < ret; q++)
820 {
821 int idx = 0;
822 for(int a = 0; a < 6; ++a)
823 for(int b = 0; b < 21; ++b)
824 for(int c = 0; c < 10; ++c)
825 for(int d = 0; d < 6; ++d)
826 buffer[idx++] = INT__d_h_f_d[q*7560+b*360+a*60+c*6+d];
827
828 memcpy(INT__d_h_f_d+q*7560, buffer, 7560*sizeof(double));
829 }
830
831 return ret;
832 }
833
ostei_h_d_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_d_d_f)834 int ostei_h_d_d_f(struct simint_multi_shellpair const P,
835 struct simint_multi_shellpair const Q,
836 double screen_tol,
837 double * const restrict work,
838 double * const restrict INT__h_d_d_f)
839 {
840 double Q_AB[3*Q.nshell12];
841 struct simint_multi_shellpair Q_tmp = Q;
842 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
843 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
844 Q_tmp.AB_x = Q_AB;
845 Q_tmp.AB_y = Q_AB + Q.nshell12;
846 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
847
848 for(int i = 0; i < Q.nshell12; i++)
849 {
850 Q_tmp.AB_x[i] = -Q.AB_x[i];
851 Q_tmp.AB_y[i] = -Q.AB_y[i];
852 Q_tmp.AB_z[i] = -Q.AB_z[i];
853 }
854
855 int ret = ostei_h_d_f_d(P, Q_tmp, screen_tol, work, INT__h_d_d_f);
856 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
857
858 for(int q = 0; q < ret; q++)
859 {
860 int idx = 0;
861 for(int a = 0; a < 21; ++a)
862 for(int b = 0; b < 6; ++b)
863 for(int c = 0; c < 6; ++c)
864 for(int d = 0; d < 10; ++d)
865 buffer[idx++] = INT__h_d_d_f[q*7560+a*360+b*60+d*6+c];
866
867 memcpy(INT__h_d_d_f+q*7560, buffer, 7560*sizeof(double));
868 }
869
870 return ret;
871 }
872
ostei_d_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__d_h_d_f)873 int ostei_d_h_d_f(struct simint_multi_shellpair const P,
874 struct simint_multi_shellpair const Q,
875 double screen_tol,
876 double * const restrict work,
877 double * const restrict INT__d_h_d_f)
878 {
879 double P_AB[3*P.nshell12];
880 struct simint_multi_shellpair P_tmp = P;
881 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
882 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
883 P_tmp.AB_x = P_AB;
884 P_tmp.AB_y = P_AB + P.nshell12;
885 P_tmp.AB_z = P_AB + 2*P.nshell12;
886
887 for(int i = 0; i < P.nshell12; i++)
888 {
889 P_tmp.AB_x[i] = -P.AB_x[i];
890 P_tmp.AB_y[i] = -P.AB_y[i];
891 P_tmp.AB_z[i] = -P.AB_z[i];
892 }
893
894 double Q_AB[3*Q.nshell12];
895 struct simint_multi_shellpair Q_tmp = Q;
896 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
897 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
898 Q_tmp.AB_x = Q_AB;
899 Q_tmp.AB_y = Q_AB + Q.nshell12;
900 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
901
902 for(int i = 0; i < Q.nshell12; i++)
903 {
904 Q_tmp.AB_x[i] = -Q.AB_x[i];
905 Q_tmp.AB_y[i] = -Q.AB_y[i];
906 Q_tmp.AB_z[i] = -Q.AB_z[i];
907 }
908
909 int ret = ostei_h_d_f_d(P_tmp, Q_tmp, screen_tol, work, INT__d_h_d_f);
910 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
911
912 for(int q = 0; q < ret; q++)
913 {
914 int idx = 0;
915 for(int a = 0; a < 6; ++a)
916 for(int b = 0; b < 21; ++b)
917 for(int c = 0; c < 6; ++c)
918 for(int d = 0; d < 10; ++d)
919 buffer[idx++] = INT__d_h_d_f[q*7560+b*360+a*60+d*6+c];
920
921 memcpy(INT__d_h_d_f+q*7560, buffer, 7560*sizeof(double));
922 }
923
924 return ret;
925 }
926
927