1 #include "simint/boys/boys.h"
2 #include "simint/ostei/gen/ostei_generated.h"
3 #include "simint/vectorization/vectorization.h"
4 #include <math.h>
5 #include <string.h>
6
7
ostei_f_f_k_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_f_k_p)8 int ostei_f_f_k_p(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__f_f_k_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__f_f_k_p);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__f_s_k_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__f_s_l_s = work + (SIMINT_NSHELL_SIMD * 360);
31 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 810);
32 double * const INT__g_s_l_s = work + (SIMINT_NSHELL_SIMD * 1350);
33 double * const INT__h_s_k_s = work + (SIMINT_NSHELL_SIMD * 2025);
34 double * const INT__h_s_l_s = work + (SIMINT_NSHELL_SIMD * 2781);
35 double * const INT__i_s_k_s = work + (SIMINT_NSHELL_SIMD * 3726);
36 double * const INT__i_s_l_s = work + (SIMINT_NSHELL_SIMD * 4734);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*5994);
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 + 15;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 57;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 135;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 255;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 420;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 630;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 882;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 1170;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 1485;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1593;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1773;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 2043;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2421;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2925;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 3573;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 4383;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 4683;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 5133;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 5763;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 6603;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 7683;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 9033;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 9633;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 10473;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 11593;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 13033;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 14833;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 15778;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 17038;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_l_s = primwork + 18658;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 20683;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 21859;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_l_s = primwork + 23371;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_k_s = primwork + 25261;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_l_s = primwork + 26269;
74 double * const hrrwork = (double *)(primwork + 27529);
75 double * const HRR_INT__f_p_k_s = hrrwork + 0;
76 double * const HRR_INT__f_p_l_s = hrrwork + 1080;
77 double * const HRR_INT__f_d_k_s = hrrwork + 2430;
78 double * const HRR_INT__f_d_l_s = hrrwork + 4590;
79 double * const HRR_INT__f_f_k_s = hrrwork + 7290;
80 double * const HRR_INT__f_f_l_s = hrrwork + 10890;
81 double * const HRR_INT__g_p_k_s = hrrwork + 15390;
82 double * const HRR_INT__g_p_l_s = hrrwork + 17010;
83 double * const HRR_INT__g_d_k_s = hrrwork + 19035;
84 double * const HRR_INT__g_d_l_s = hrrwork + 22275;
85 double * const HRR_INT__h_p_k_s = hrrwork + 26325;
86 double * const HRR_INT__h_p_l_s = hrrwork + 28593;
87
88
89 // Create constants
90 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
91 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
92 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
93 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
94 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
95 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
96 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
97 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
98 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
99
100
101 ////////////////////////////////////////
102 // Loop over shells and primitives
103 ////////////////////////////////////////
104
105 real_abcd = 0;
106 istart = 0;
107 for(ab = 0; ab < P.nshell12_clip; ++ab)
108 {
109 const int iend = istart + P.nprim12[ab];
110
111 cd = 0;
112 jstart = 0;
113
114 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
115 {
116 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
117 int jend = jstart;
118 for(i = 0; i < nshellbatch; i++)
119 jend += Q.nprim12[cd+i];
120
121 // Clear the beginning of the workspace (where we are accumulating integrals)
122 memset(work, 0, SIMINT_NSHELL_SIMD * 5994 * sizeof(double));
123 abcd = 0;
124
125
126 for(i = istart; i < iend; ++i)
127 {
128 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
129
130 if(check_screen)
131 {
132 // Skip this whole thing if always insignificant
133 if((P.screen[i] * Q.screen_max) < screen_tol)
134 continue;
135 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
136 }
137
138 icd = 0;
139 iprimcd = 0;
140 nprim_icd = Q.nprim12[cd];
141 double * restrict PRIM_PTR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
142 double * restrict PRIM_PTR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
143 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
144 double * restrict PRIM_PTR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
145 double * restrict PRIM_PTR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
146 double * restrict PRIM_PTR_INT__h_s_l_s = INT__h_s_l_s + abcd * 945;
147 double * restrict PRIM_PTR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
148 double * restrict PRIM_PTR_INT__i_s_l_s = INT__i_s_l_s + abcd * 1260;
149
150
151
152 // Load these one per loop over i
153 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
154 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
155 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
156
157 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
158
159 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
160 {
161 // calculate the shell offsets
162 // these are the offset from the shell pointed to by cd
163 // for each element
164 int shelloffsets[SIMINT_SIMD_LEN] = {0};
165 int lastoffset = 0;
166 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
167
168 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
169 {
170 // Handle if the first element of the vector is a new shell
171 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
172 {
173 nprim_icd += Q.nprim12[cd + (++icd)];
174 PRIM_PTR_INT__f_s_k_s += 360;
175 PRIM_PTR_INT__f_s_l_s += 450;
176 PRIM_PTR_INT__g_s_k_s += 540;
177 PRIM_PTR_INT__g_s_l_s += 675;
178 PRIM_PTR_INT__h_s_k_s += 756;
179 PRIM_PTR_INT__h_s_l_s += 945;
180 PRIM_PTR_INT__i_s_k_s += 1008;
181 PRIM_PTR_INT__i_s_l_s += 1260;
182 }
183 iprimcd++;
184 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
185 {
186 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
187 {
188 shelloffsets[n] = shelloffsets[n-1] + 1;
189 lastoffset++;
190 nprim_icd += Q.nprim12[cd + (++icd)];
191 }
192 else
193 shelloffsets[n] = shelloffsets[n-1];
194 iprimcd++;
195 }
196 }
197 else
198 iprimcd += SIMINT_SIMD_LEN;
199
200 // Do we have to compute this vector (or has it been screened out)?
201 // (not_screened != 0 means we have to do this vector)
202 if(check_screen)
203 {
204 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
205 if(vmax < screen_tol)
206 {
207 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
208 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
209 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
210 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
211 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
212 PRIM_PTR_INT__h_s_l_s += lastoffset*945;
213 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
214 PRIM_PTR_INT__i_s_l_s += lastoffset*1260;
215 continue;
216 }
217 }
218
219 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
220 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
221 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
222 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
223
224
225 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
226 SIMINT_DBLTYPE PQ[3];
227 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
228 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
229 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
230 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
231 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
232 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
233
234 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
235 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
236 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
237 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
238 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
239 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
240 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
241
242 // NOTE: Minus sign!
243 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
244 SIMINT_DBLTYPE aop_PQ[3];
245 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
246 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
247 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
248
249 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
250 SIMINT_DBLTYPE aoq_PQ[3];
251 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
252 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
253 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
254 // Put a minus sign here so we don't have to in RR routines
255 a_over_q = SIMINT_NEG(a_over_q);
256
257
258 //////////////////////////////////////////////
259 // Fjt function section
260 // Maximum v value: 14
261 //////////////////////////////////////////////
262 // The parameter to the Fjt function
263 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
264
265
266 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
267
268
269 boys_F_split(PRIM_INT__s_s_s_s, F_x, 14);
270 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
271 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
272 for(n = 0; n <= 14; n++)
273 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
274
275 //////////////////////////////////////////////
276 // Primitive integrals: Vertical recurrance
277 //////////////////////////////////////////////
278
279 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
280 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
281 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
282 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
283 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, 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_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
289 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
290 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
291 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
292 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
293 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
294 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
295 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
296 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
297 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
298 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
299
300
301
302 // Forming PRIM_INT__s_s_p_s[14 * 3];
303 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
304 {
305
306 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
307 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]);
308
309 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
310 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]);
311
312 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
313 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]);
314
315 }
316
317
318
319 // Forming PRIM_INT__s_s_d_s[13 * 6];
320 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
321 {
322
323 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
324 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 0]);
325 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 0]);
326
327 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
328 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 1]);
329
330 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
331 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 2]);
332
333 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
334 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 3]);
335 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 3]);
336
337 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
338 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 4]);
339
340 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
341 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_d_s[n * 6 + 5]);
342 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 5]);
343
344 }
345
346
347
348 // Forming PRIM_INT__s_s_f_s[12 * 10];
349 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
350 {
351
352 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
353 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 0]);
354 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]), PRIM_INT__s_s_f_s[n * 10 + 0]);
355
356 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
357 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 1]);
358
359 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
360 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 2]);
361
362 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
363 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 3]);
364
365 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
366 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__s_s_f_s[n * 10 + 4]);
367
368 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
369 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 5]);
370
371 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
372 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 6]);
373 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_p_s[n * 3 + 1]), PRIM_INT__s_s_f_s[n * 10 + 6]);
374
375 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
376 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 7]);
377
378 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
379 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 8]);
380
381 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
382 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 9]);
383 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_p_s[n * 3 + 2]), PRIM_INT__s_s_f_s[n * 10 + 9]);
384
385 }
386
387
388 VRR_K_s_s_g_s(
389 PRIM_INT__s_s_g_s,
390 PRIM_INT__s_s_f_s,
391 PRIM_INT__s_s_d_s,
392 Q_PA,
393 a_over_q,
394 aoq_PQ,
395 one_over_2q,
396 11);
397
398
399 VRR_K_s_s_h_s(
400 PRIM_INT__s_s_h_s,
401 PRIM_INT__s_s_g_s,
402 PRIM_INT__s_s_f_s,
403 Q_PA,
404 a_over_q,
405 aoq_PQ,
406 one_over_2q,
407 10);
408
409
410 ostei_general_vrr1_K(6, 9,
411 one_over_2q, a_over_q, aoq_PQ, Q_PA,
412 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
413
414
415 ostei_general_vrr1_K(7, 8,
416 one_over_2q, a_over_q, aoq_PQ, Q_PA,
417 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
418
419
420 ostei_general_vrr_I(1, 0, 7, 0, 6,
421 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
422 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
423
424
425 ostei_general_vrr_I(1, 0, 6, 0, 6,
426 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
427 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
428
429
430 ostei_general_vrr_I(2, 0, 7, 0, 5,
431 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
432 PRIM_INT__p_s_k_s, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_k_s);
433
434
435 ostei_general_vrr_I(1, 0, 5, 0, 6,
436 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
437 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
438
439
440 ostei_general_vrr_I(2, 0, 6, 0, 5,
441 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
442 PRIM_INT__p_s_i_s, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_i_s);
443
444
445 ostei_general_vrr_I(3, 0, 7, 0, 4,
446 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
447 PRIM_INT__d_s_k_s, PRIM_INT__p_s_k_s, NULL, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_k_s);
448
449
450 ostei_general_vrr1_K(8, 7,
451 one_over_2q, a_over_q, aoq_PQ, Q_PA,
452 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
453
454
455 ostei_general_vrr_I(1, 0, 8, 0, 6,
456 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
457 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
458
459
460 ostei_general_vrr_I(2, 0, 8, 0, 5,
461 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
462 PRIM_INT__p_s_l_s, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_k_s, NULL, PRIM_INT__d_s_l_s);
463
464
465 ostei_general_vrr_I(3, 0, 8, 0, 4,
466 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
467 PRIM_INT__d_s_l_s, PRIM_INT__p_s_l_s, NULL, PRIM_INT__d_s_k_s, NULL, PRIM_INT__f_s_l_s);
468
469
470 VRR_I_p_s_g_s(
471 PRIM_INT__p_s_g_s,
472 PRIM_INT__s_s_g_s,
473 PRIM_INT__s_s_f_s,
474 P_PA,
475 aop_PQ,
476 one_over_2pq,
477 6);
478
479
480 ostei_general_vrr_I(2, 0, 5, 0, 5,
481 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
482 PRIM_INT__p_s_h_s, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_h_s);
483
484
485 ostei_general_vrr_I(3, 0, 6, 0, 4,
486 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
487 PRIM_INT__d_s_i_s, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_i_s);
488
489
490 ostei_general_vrr_I(4, 0, 7, 0, 3,
491 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
492 PRIM_INT__f_s_k_s, PRIM_INT__d_s_k_s, NULL, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_k_s);
493
494
495 ostei_general_vrr_I(4, 0, 8, 0, 3,
496 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
497 PRIM_INT__f_s_l_s, PRIM_INT__d_s_l_s, NULL, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_l_s);
498
499
500 VRR_I_p_s_f_s(
501 PRIM_INT__p_s_f_s,
502 PRIM_INT__s_s_f_s,
503 PRIM_INT__s_s_d_s,
504 P_PA,
505 aop_PQ,
506 one_over_2pq,
507 6);
508
509
510 ostei_general_vrr_I(2, 0, 4, 0, 5,
511 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
512 PRIM_INT__p_s_g_s, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_g_s);
513
514
515 ostei_general_vrr_I(3, 0, 5, 0, 4,
516 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
517 PRIM_INT__d_s_h_s, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_h_s);
518
519
520 ostei_general_vrr_I(4, 0, 6, 0, 3,
521 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
522 PRIM_INT__f_s_i_s, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_i_s);
523
524
525 ostei_general_vrr_I(5, 0, 7, 0, 2,
526 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
527 PRIM_INT__g_s_k_s, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_k_s);
528
529
530 ostei_general_vrr_I(5, 0, 8, 0, 2,
531 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
532 PRIM_INT__g_s_l_s, PRIM_INT__f_s_l_s, NULL, PRIM_INT__g_s_k_s, NULL, PRIM_INT__h_s_l_s);
533
534
535
536 // Forming PRIM_INT__p_s_d_s[6 * 18];
537 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
538 {
539
540 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
541 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
542 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
543
544 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
545 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 1]);
546 PRIM_INT__p_s_d_s[n * 18 + 1] = 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 + 1]);
547
548 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
549 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 2]);
550 PRIM_INT__p_s_d_s[n * 18 + 2] = 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 + 2]);
551
552 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
553 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 3]);
554
555 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
556 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 4]);
557
558 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
559 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 5]);
560
561 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
562 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 6]);
563
564 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
565 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 7]);
566 PRIM_INT__p_s_d_s[n * 18 + 7] = 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 + 7]);
567
568 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
569 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 8]);
570
571 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
572 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 9]);
573 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 9]);
574
575 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
576 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 10]);
577 PRIM_INT__p_s_d_s[n * 18 + 10] = 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 + 10]);
578
579 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
580 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 11]);
581
582 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
583 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 12]);
584
585 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
586 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 13]);
587
588 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
589 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 14]);
590 PRIM_INT__p_s_d_s[n * 18 + 14] = 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 + 14]);
591
592 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
593 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 15]);
594
595 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
596 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 16]);
597 PRIM_INT__p_s_d_s[n * 18 + 16] = 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 + 16]);
598
599 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
600 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 17]);
601 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 17]);
602
603 }
604
605
606 VRR_I_d_s_f_s(
607 PRIM_INT__d_s_f_s,
608 PRIM_INT__p_s_f_s,
609 PRIM_INT__s_s_f_s,
610 PRIM_INT__p_s_d_s,
611 P_PA,
612 a_over_p,
613 aop_PQ,
614 one_over_2p,
615 one_over_2pq,
616 5);
617
618
619 ostei_general_vrr_I(3, 0, 4, 0, 4,
620 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
621 PRIM_INT__d_s_g_s, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_f_s, NULL, PRIM_INT__f_s_g_s);
622
623
624 ostei_general_vrr_I(4, 0, 5, 0, 3,
625 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
626 PRIM_INT__f_s_h_s, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_g_s, NULL, PRIM_INT__g_s_h_s);
627
628
629 ostei_general_vrr_I(5, 0, 6, 0, 2,
630 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
631 PRIM_INT__g_s_i_s, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_h_s, NULL, PRIM_INT__h_s_i_s);
632
633
634 ostei_general_vrr_I(6, 0, 7, 0, 1,
635 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
636 PRIM_INT__h_s_k_s, PRIM_INT__g_s_k_s, NULL, PRIM_INT__h_s_i_s, NULL, PRIM_INT__i_s_k_s);
637
638
639 ostei_general_vrr_I(6, 0, 8, 0, 1,
640 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
641 PRIM_INT__h_s_l_s, PRIM_INT__g_s_l_s, NULL, PRIM_INT__h_s_k_s, NULL, PRIM_INT__i_s_l_s);
642
643
644
645
646 ////////////////////////////////////
647 // Accumulate contracted integrals
648 ////////////////////////////////////
649 if(lastoffset == 0)
650 {
651 contract_all(360, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
652 contract_all(450, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
653 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
654 contract_all(675, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
655 contract_all(756, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
656 contract_all(945, PRIM_INT__h_s_l_s, PRIM_PTR_INT__h_s_l_s);
657 contract_all(1008, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
658 contract_all(1260, PRIM_INT__i_s_l_s, PRIM_PTR_INT__i_s_l_s);
659 }
660 else
661 {
662 contract(360, shelloffsets, PRIM_INT__f_s_k_s, PRIM_PTR_INT__f_s_k_s);
663 contract(450, shelloffsets, PRIM_INT__f_s_l_s, PRIM_PTR_INT__f_s_l_s);
664 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
665 contract(675, shelloffsets, PRIM_INT__g_s_l_s, PRIM_PTR_INT__g_s_l_s);
666 contract(756, shelloffsets, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
667 contract(945, shelloffsets, PRIM_INT__h_s_l_s, PRIM_PTR_INT__h_s_l_s);
668 contract(1008, shelloffsets, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
669 contract(1260, shelloffsets, PRIM_INT__i_s_l_s, PRIM_PTR_INT__i_s_l_s);
670 PRIM_PTR_INT__f_s_k_s += lastoffset*360;
671 PRIM_PTR_INT__f_s_l_s += lastoffset*450;
672 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
673 PRIM_PTR_INT__g_s_l_s += lastoffset*675;
674 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
675 PRIM_PTR_INT__h_s_l_s += lastoffset*945;
676 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
677 PRIM_PTR_INT__i_s_l_s += lastoffset*1260;
678 }
679
680 } // close loop over j
681 } // close loop over i
682
683 //Advance to the next batch
684 jstart = SIMINT_SIMD_ROUND(jend);
685
686 //////////////////////////////////////////////
687 // Contracted integrals: Horizontal recurrance
688 //////////////////////////////////////////////
689
690
691 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
692
693
694 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
695 {
696 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
697
698 // set up HRR pointers
699 double const * restrict HRR_INT__f_s_k_s = INT__f_s_k_s + abcd * 360;
700 double const * restrict HRR_INT__f_s_l_s = INT__f_s_l_s + abcd * 450;
701 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
702 double const * restrict HRR_INT__g_s_l_s = INT__g_s_l_s + abcd * 675;
703 double const * restrict HRR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
704 double const * restrict HRR_INT__h_s_l_s = INT__h_s_l_s + abcd * 945;
705 double const * restrict HRR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
706 double const * restrict HRR_INT__i_s_l_s = INT__i_s_l_s + abcd * 1260;
707 double * restrict HRR_INT__f_f_k_p = INT__f_f_k_p + real_abcd * 10800;
708
709 // form INT__f_p_k_s
710 HRR_J_f_p(
711 HRR_INT__f_p_k_s,
712 HRR_INT__f_s_k_s,
713 HRR_INT__g_s_k_s,
714 hAB, 36);
715
716 // form INT__f_p_l_s
717 HRR_J_f_p(
718 HRR_INT__f_p_l_s,
719 HRR_INT__f_s_l_s,
720 HRR_INT__g_s_l_s,
721 hAB, 45);
722
723 // form INT__g_p_k_s
724 HRR_J_g_p(
725 HRR_INT__g_p_k_s,
726 HRR_INT__g_s_k_s,
727 HRR_INT__h_s_k_s,
728 hAB, 36);
729
730 // form INT__g_p_l_s
731 HRR_J_g_p(
732 HRR_INT__g_p_l_s,
733 HRR_INT__g_s_l_s,
734 HRR_INT__h_s_l_s,
735 hAB, 45);
736
737 // form INT__h_p_k_s
738 ostei_general_hrr_J(5, 1, 7, 0, hAB, HRR_INT__i_s_k_s, HRR_INT__h_s_k_s, HRR_INT__h_p_k_s);
739
740 // form INT__h_p_l_s
741 ostei_general_hrr_J(5, 1, 8, 0, hAB, HRR_INT__i_s_l_s, HRR_INT__h_s_l_s, HRR_INT__h_p_l_s);
742
743 // form INT__f_d_k_s
744 HRR_J_f_d(
745 HRR_INT__f_d_k_s,
746 HRR_INT__f_p_k_s,
747 HRR_INT__g_p_k_s,
748 hAB, 36);
749
750 // form INT__f_d_l_s
751 HRR_J_f_d(
752 HRR_INT__f_d_l_s,
753 HRR_INT__f_p_l_s,
754 HRR_INT__g_p_l_s,
755 hAB, 45);
756
757 // form INT__g_d_k_s
758 ostei_general_hrr_J(4, 2, 7, 0, hAB, HRR_INT__h_p_k_s, HRR_INT__g_p_k_s, HRR_INT__g_d_k_s);
759
760 // form INT__g_d_l_s
761 ostei_general_hrr_J(4, 2, 8, 0, hAB, HRR_INT__h_p_l_s, HRR_INT__g_p_l_s, HRR_INT__g_d_l_s);
762
763 // form INT__f_f_k_s
764 ostei_general_hrr_J(3, 3, 7, 0, hAB, HRR_INT__g_d_k_s, HRR_INT__f_d_k_s, HRR_INT__f_f_k_s);
765
766 // form INT__f_f_l_s
767 ostei_general_hrr_J(3, 3, 8, 0, hAB, HRR_INT__g_d_l_s, HRR_INT__f_d_l_s, HRR_INT__f_f_l_s);
768
769 // form INT__f_f_k_p
770 ostei_general_hrr_L(3, 3, 7, 1, hCD, HRR_INT__f_f_l_s, HRR_INT__f_f_k_s, HRR_INT__f_f_k_p);
771
772
773 } // close HRR loop
774
775
776 } // close loop cdbatch
777
778 istart = iend;
779 } // close loop over ab
780
781 return P.nshell12_clip * Q.nshell12_clip;
782 }
783
ostei_f_f_p_k(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__f_f_p_k)784 int ostei_f_f_p_k(struct simint_multi_shellpair const P,
785 struct simint_multi_shellpair const Q,
786 double screen_tol,
787 double * const restrict work,
788 double * const restrict INT__f_f_p_k)
789 {
790 double Q_AB[3*Q.nshell12];
791 struct simint_multi_shellpair Q_tmp = Q;
792 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
793 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
794 Q_tmp.AB_x = Q_AB;
795 Q_tmp.AB_y = Q_AB + Q.nshell12;
796 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
797
798 for(int i = 0; i < Q.nshell12; i++)
799 {
800 Q_tmp.AB_x[i] = -Q.AB_x[i];
801 Q_tmp.AB_y[i] = -Q.AB_y[i];
802 Q_tmp.AB_z[i] = -Q.AB_z[i];
803 }
804
805 int ret = ostei_f_f_k_p(P, Q_tmp, screen_tol, work, INT__f_f_p_k);
806 double buffer[10800] SIMINT_ALIGN_ARRAY_DBL;
807
808 for(int q = 0; q < ret; q++)
809 {
810 int idx = 0;
811 for(int a = 0; a < 10; ++a)
812 for(int b = 0; b < 10; ++b)
813 for(int c = 0; c < 3; ++c)
814 for(int d = 0; d < 36; ++d)
815 buffer[idx++] = INT__f_f_p_k[q*10800+a*1080+b*108+d*3+c];
816
817 memcpy(INT__f_f_p_k+q*10800, buffer, 10800*sizeof(double));
818 }
819
820 return ret;
821 }
822
823