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_s_h_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_s_h_g)8 int ostei_h_s_h_g(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_s_h_g)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__h_s_h_g);
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 ibra;
26
27 // partition workspace
28 double * const INT__h_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
29 double * const INT__h_s_i_s = work + (SIMINT_NSHELL_SIMD * 441);
30 double * const INT__h_s_k_s = work + (SIMINT_NSHELL_SIMD * 1029);
31 double * const INT__h_s_l_s = work + (SIMINT_NSHELL_SIMD * 1785);
32 double * const INT__h_s_m_s = work + (SIMINT_NSHELL_SIMD * 2730);
33 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3885);
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 15;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 57;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 135;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 255;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 420;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 630;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 882;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_l_s = primwork + 1170;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_m_s = primwork + 1485;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 1815;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 1860;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1950;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 2100;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 2325;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2640;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 3060;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_l_s = primwork + 3600;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_m_s = primwork + 4275;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 5100;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 5244;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 5484;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 5844;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 6348;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 7020;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_l_s = primwork + 7884;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_m_s = primwork + 8964;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 10284;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 10584;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 11034;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 11664;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 12504;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_l_s = primwork + 13584;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_m_s = primwork + 14934;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 16584;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 17034;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 17664;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 18504;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_l_s = primwork + 19584;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_m_s = primwork + 20934;
74 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 22584;
75 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 23025;
76 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 23613;
77 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_l_s = primwork + 24369;
78 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_m_s = primwork + 25314;
79 double * const hrrwork = (double *)(primwork + 26469);
80 double * const HRR_INT__h_s_h_p = hrrwork + 0;
81 double * const HRR_INT__h_s_h_d = hrrwork + 1323;
82 double * const HRR_INT__h_s_h_f = hrrwork + 3969;
83 double * const HRR_INT__h_s_i_p = hrrwork + 8379;
84 double * const HRR_INT__h_s_i_d = hrrwork + 10143;
85 double * const HRR_INT__h_s_i_f = hrrwork + 13671;
86 double * const HRR_INT__h_s_k_p = hrrwork + 19551;
87 double * const HRR_INT__h_s_k_d = hrrwork + 21819;
88 double * const HRR_INT__h_s_l_p = hrrwork + 26355;
89
90
91 // Create constants
92 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
93 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
94 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
95 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
96 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
97 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
98 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
99 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
100 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
101 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
102
103
104 ////////////////////////////////////////
105 // Loop over shells and primitives
106 ////////////////////////////////////////
107
108 real_abcd = 0;
109 istart = 0;
110 for(ab = 0; ab < P.nshell12_clip; ++ab)
111 {
112 const int iend = istart + P.nprim12[ab];
113
114 cd = 0;
115 jstart = 0;
116
117 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
118 {
119 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
120 int jend = jstart;
121 for(i = 0; i < nshellbatch; i++)
122 jend += Q.nprim12[cd+i];
123
124 // Clear the beginning of the workspace (where we are accumulating integrals)
125 memset(work, 0, SIMINT_NSHELL_SIMD * 3885 * sizeof(double));
126 abcd = 0;
127
128
129 for(i = istart; i < iend; ++i)
130 {
131 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
132
133 if(check_screen)
134 {
135 // Skip this whole thing if always insignificant
136 if((P.screen[i] * Q.screen_max) < screen_tol)
137 continue;
138 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
139 }
140
141 icd = 0;
142 iprimcd = 0;
143 nprim_icd = Q.nprim12[cd];
144 double * restrict PRIM_PTR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
145 double * restrict PRIM_PTR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
146 double * restrict PRIM_PTR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
147 double * restrict PRIM_PTR_INT__h_s_l_s = INT__h_s_l_s + abcd * 945;
148 double * restrict PRIM_PTR_INT__h_s_m_s = INT__h_s_m_s + abcd * 1155;
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__h_s_h_s += 441;
175 PRIM_PTR_INT__h_s_i_s += 588;
176 PRIM_PTR_INT__h_s_k_s += 756;
177 PRIM_PTR_INT__h_s_l_s += 945;
178 PRIM_PTR_INT__h_s_m_s += 1155;
179 }
180 iprimcd++;
181 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
182 {
183 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
184 {
185 shelloffsets[n] = shelloffsets[n-1] + 1;
186 lastoffset++;
187 nprim_icd += Q.nprim12[cd + (++icd)];
188 }
189 else
190 shelloffsets[n] = shelloffsets[n-1];
191 iprimcd++;
192 }
193 }
194 else
195 iprimcd += SIMINT_SIMD_LEN;
196
197 // Do we have to compute this vector (or has it been screened out)?
198 // (not_screened != 0 means we have to do this vector)
199 if(check_screen)
200 {
201 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
202 if(vmax < screen_tol)
203 {
204 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
205 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
206 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
207 PRIM_PTR_INT__h_s_l_s += lastoffset*945;
208 PRIM_PTR_INT__h_s_m_s += lastoffset*1155;
209 continue;
210 }
211 }
212
213 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
214 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
215 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
216 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
217
218
219 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
220 SIMINT_DBLTYPE PQ[3];
221 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
222 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
223 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
224 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
225 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
226 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
227
228 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
229 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
230 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
231 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
232 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
233 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
234 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
235
236 // NOTE: Minus sign!
237 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
238 SIMINT_DBLTYPE aop_PQ[3];
239 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
240 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
241 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
242
243 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
244 SIMINT_DBLTYPE aoq_PQ[3];
245 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
246 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
247 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
248 // Put a minus sign here so we don't have to in RR routines
249 a_over_q = SIMINT_NEG(a_over_q);
250
251
252 //////////////////////////////////////////////
253 // Fjt function section
254 // Maximum v value: 14
255 //////////////////////////////////////////////
256 // The parameter to the Fjt function
257 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
258
259
260 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
261
262
263 boys_F_split(PRIM_INT__s_s_s_s, F_x, 14);
264 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
265 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
266 for(n = 0; n <= 14; n++)
267 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
268
269 //////////////////////////////////////////////
270 // Primitive integrals: Vertical recurrance
271 //////////////////////////////////////////////
272
273 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
274 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
275 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
276 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, 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_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
283 const SIMINT_DBLTYPE vrr_const_7_over_2q = SIMINT_MUL(const_7, one_over_2q);
284 const SIMINT_DBLTYPE vrr_const_8_over_2q = SIMINT_MUL(const_8, one_over_2q);
285 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
286 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
287 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
288 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
289 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
290 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
291 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
292 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
293 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
294
295
296
297 // Forming PRIM_INT__s_s_p_s[14 * 3];
298 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
299 {
300
301 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
302 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]);
303
304 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
305 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]);
306
307 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
308 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]);
309
310 }
311
312
313
314 // Forming PRIM_INT__s_s_d_s[13 * 6];
315 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
316 {
317
318 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
319 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]);
320 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]);
321
322 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
323 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]);
324
325 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
326 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]);
327
328 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
329 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]);
330 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]);
331
332 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
333 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]);
334
335 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
336 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]);
337 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]);
338
339 }
340
341
342
343 // Forming PRIM_INT__s_s_f_s[12 * 10];
344 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
345 {
346
347 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
348 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]);
349 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]);
350
351 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
352 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]);
353
354 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
355 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]);
356
357 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
358 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]);
359
360 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
361 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]);
362
363 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
364 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]);
365
366 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
367 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]);
368 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]);
369
370 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
371 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]);
372
373 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
374 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]);
375
376 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
377 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]);
378 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]);
379
380 }
381
382
383 VRR_K_s_s_g_s(
384 PRIM_INT__s_s_g_s,
385 PRIM_INT__s_s_f_s,
386 PRIM_INT__s_s_d_s,
387 Q_PA,
388 a_over_q,
389 aoq_PQ,
390 one_over_2q,
391 11);
392
393
394 VRR_K_s_s_h_s(
395 PRIM_INT__s_s_h_s,
396 PRIM_INT__s_s_g_s,
397 PRIM_INT__s_s_f_s,
398 Q_PA,
399 a_over_q,
400 aoq_PQ,
401 one_over_2q,
402 10);
403
404
405 ostei_general_vrr_I(1, 0, 5, 0, 5,
406 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
407 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
408
409
410 VRR_I_p_s_g_s(
411 PRIM_INT__p_s_g_s,
412 PRIM_INT__s_s_g_s,
413 PRIM_INT__s_s_f_s,
414 P_PA,
415 aop_PQ,
416 one_over_2pq,
417 5);
418
419
420 ostei_general_vrr_I(2, 0, 5, 0, 4,
421 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
422 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);
423
424
425 VRR_I_p_s_f_s(
426 PRIM_INT__p_s_f_s,
427 PRIM_INT__s_s_f_s,
428 PRIM_INT__s_s_d_s,
429 P_PA,
430 aop_PQ,
431 one_over_2pq,
432 5);
433
434
435 ostei_general_vrr_I(2, 0, 4, 0, 4,
436 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
437 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);
438
439
440 ostei_general_vrr_I(3, 0, 5, 0, 3,
441 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
442 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);
443
444
445
446 // Forming PRIM_INT__p_s_d_s[5 * 18];
447 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
448 {
449
450 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
451 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]);
452 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]);
453
454 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
455 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]);
456 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]);
457
458 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
459 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]);
460 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]);
461
462 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
463 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]);
464
465 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
466 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]);
467
468 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
469 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]);
470
471 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
472 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]);
473
474 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
475 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]);
476 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]);
477
478 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
479 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]);
480
481 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
482 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]);
483 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]);
484
485 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
486 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]);
487 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]);
488
489 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
490 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]);
491
492 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
493 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]);
494
495 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
496 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]);
497
498 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
499 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]);
500 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]);
501
502 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
503 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]);
504
505 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
506 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]);
507 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]);
508
509 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
510 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]);
511 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]);
512
513 }
514
515
516 VRR_I_d_s_f_s(
517 PRIM_INT__d_s_f_s,
518 PRIM_INT__p_s_f_s,
519 PRIM_INT__s_s_f_s,
520 PRIM_INT__p_s_d_s,
521 P_PA,
522 a_over_p,
523 aop_PQ,
524 one_over_2p,
525 one_over_2pq,
526 4);
527
528
529 ostei_general_vrr_I(3, 0, 4, 0, 3,
530 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
531 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);
532
533
534 ostei_general_vrr_I(4, 0, 5, 0, 2,
535 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
536 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);
537
538
539
540 // Forming PRIM_INT__p_s_p_s[5 * 9];
541 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
542 {
543
544 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
545 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
546 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]);
547
548 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 1]);
549 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 1]);
550
551 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 2]);
552 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 2]);
553
554 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
555 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 3]);
556
557 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
558 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 4]);
559 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]);
560
561 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 2]);
562 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 5]);
563
564 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
565 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 6]);
566
567 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
568 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 7]);
569
570 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
571 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 8]);
572 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]);
573
574 }
575
576
577 VRR_I_d_s_d_s(
578 PRIM_INT__d_s_d_s,
579 PRIM_INT__p_s_d_s,
580 PRIM_INT__s_s_d_s,
581 PRIM_INT__p_s_p_s,
582 P_PA,
583 a_over_p,
584 aop_PQ,
585 one_over_2p,
586 one_over_2pq,
587 4);
588
589
590 ostei_general_vrr_I(3, 0, 3, 0, 3,
591 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
592 PRIM_INT__d_s_f_s, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_d_s, NULL, PRIM_INT__f_s_f_s);
593
594
595 ostei_general_vrr_I(4, 0, 4, 0, 2,
596 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
597 PRIM_INT__f_s_g_s, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_f_s, NULL, PRIM_INT__g_s_g_s);
598
599
600 ostei_general_vrr_I(5, 0, 5, 0, 1,
601 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
602 PRIM_INT__g_s_h_s, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_g_s, NULL, PRIM_INT__h_s_h_s);
603
604
605 ostei_general_vrr1_K(6, 9,
606 one_over_2q, a_over_q, aoq_PQ, Q_PA,
607 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
608
609
610 ostei_general_vrr_I(1, 0, 6, 0, 5,
611 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
612 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
613
614
615 ostei_general_vrr_I(2, 0, 6, 0, 4,
616 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
617 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);
618
619
620 ostei_general_vrr_I(3, 0, 6, 0, 3,
621 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
622 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);
623
624
625 ostei_general_vrr_I(4, 0, 6, 0, 2,
626 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
627 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);
628
629
630 ostei_general_vrr_I(5, 0, 6, 0, 1,
631 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
632 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);
633
634
635 ostei_general_vrr1_K(7, 8,
636 one_over_2q, a_over_q, aoq_PQ, Q_PA,
637 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
638
639
640 ostei_general_vrr_I(1, 0, 7, 0, 5,
641 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
642 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
643
644
645 ostei_general_vrr_I(2, 0, 7, 0, 4,
646 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
647 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);
648
649
650 ostei_general_vrr_I(3, 0, 7, 0, 3,
651 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
652 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);
653
654
655 ostei_general_vrr_I(4, 0, 7, 0, 2,
656 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
657 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);
658
659
660 ostei_general_vrr_I(5, 0, 7, 0, 1,
661 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
662 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);
663
664
665 ostei_general_vrr1_K(8, 7,
666 one_over_2q, a_over_q, aoq_PQ, Q_PA,
667 PRIM_INT__s_s_k_s, PRIM_INT__s_s_i_s, PRIM_INT__s_s_l_s);
668
669
670 ostei_general_vrr_I(1, 0, 8, 0, 5,
671 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
672 PRIM_INT__s_s_l_s, NULL, NULL, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_l_s);
673
674
675 ostei_general_vrr_I(2, 0, 8, 0, 4,
676 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
677 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);
678
679
680 ostei_general_vrr_I(3, 0, 8, 0, 3,
681 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
682 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);
683
684
685 ostei_general_vrr_I(4, 0, 8, 0, 2,
686 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
687 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);
688
689
690 ostei_general_vrr_I(5, 0, 8, 0, 1,
691 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
692 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);
693
694
695 ostei_general_vrr1_K(9, 6,
696 one_over_2q, a_over_q, aoq_PQ, Q_PA,
697 PRIM_INT__s_s_l_s, PRIM_INT__s_s_k_s, PRIM_INT__s_s_m_s);
698
699
700 ostei_general_vrr_I(1, 0, 9, 0, 5,
701 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
702 PRIM_INT__s_s_m_s, NULL, NULL, PRIM_INT__s_s_l_s, NULL, PRIM_INT__p_s_m_s);
703
704
705 ostei_general_vrr_I(2, 0, 9, 0, 4,
706 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
707 PRIM_INT__p_s_m_s, PRIM_INT__s_s_m_s, NULL, PRIM_INT__p_s_l_s, NULL, PRIM_INT__d_s_m_s);
708
709
710 ostei_general_vrr_I(3, 0, 9, 0, 3,
711 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
712 PRIM_INT__d_s_m_s, PRIM_INT__p_s_m_s, NULL, PRIM_INT__d_s_l_s, NULL, PRIM_INT__f_s_m_s);
713
714
715 ostei_general_vrr_I(4, 0, 9, 0, 2,
716 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
717 PRIM_INT__f_s_m_s, PRIM_INT__d_s_m_s, NULL, PRIM_INT__f_s_l_s, NULL, PRIM_INT__g_s_m_s);
718
719
720 ostei_general_vrr_I(5, 0, 9, 0, 1,
721 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
722 PRIM_INT__g_s_m_s, PRIM_INT__f_s_m_s, NULL, PRIM_INT__g_s_l_s, NULL, PRIM_INT__h_s_m_s);
723
724
725
726
727 ////////////////////////////////////
728 // Accumulate contracted integrals
729 ////////////////////////////////////
730 if(lastoffset == 0)
731 {
732 contract_all(441, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
733 contract_all(588, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
734 contract_all(756, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
735 contract_all(945, PRIM_INT__h_s_l_s, PRIM_PTR_INT__h_s_l_s);
736 contract_all(1155, PRIM_INT__h_s_m_s, PRIM_PTR_INT__h_s_m_s);
737 }
738 else
739 {
740 contract(441, shelloffsets, PRIM_INT__h_s_h_s, PRIM_PTR_INT__h_s_h_s);
741 contract(588, shelloffsets, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
742 contract(756, shelloffsets, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
743 contract(945, shelloffsets, PRIM_INT__h_s_l_s, PRIM_PTR_INT__h_s_l_s);
744 contract(1155, shelloffsets, PRIM_INT__h_s_m_s, PRIM_PTR_INT__h_s_m_s);
745 PRIM_PTR_INT__h_s_h_s += lastoffset*441;
746 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
747 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
748 PRIM_PTR_INT__h_s_l_s += lastoffset*945;
749 PRIM_PTR_INT__h_s_m_s += lastoffset*1155;
750 }
751
752 } // close loop over j
753 } // close loop over i
754
755 //Advance to the next batch
756 jstart = SIMINT_SIMD_ROUND(jend);
757
758 //////////////////////////////////////////////
759 // Contracted integrals: Horizontal recurrance
760 //////////////////////////////////////////////
761
762
763
764
765 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
766 {
767 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
768
769 // set up HRR pointers
770 double const * restrict HRR_INT__h_s_h_s = INT__h_s_h_s + abcd * 441;
771 double const * restrict HRR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
772 double const * restrict HRR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
773 double const * restrict HRR_INT__h_s_l_s = INT__h_s_l_s + abcd * 945;
774 double const * restrict HRR_INT__h_s_m_s = INT__h_s_m_s + abcd * 1155;
775 double * restrict HRR_INT__h_s_h_g = INT__h_s_h_g + real_abcd * 6615;
776
777 // form INT__h_s_h_p
778 ostei_general_hrr_L(5, 0, 5, 1, hCD, HRR_INT__h_s_i_s, HRR_INT__h_s_h_s, HRR_INT__h_s_h_p);
779
780 // form INT__h_s_i_p
781 ostei_general_hrr_L(5, 0, 6, 1, hCD, HRR_INT__h_s_k_s, HRR_INT__h_s_i_s, HRR_INT__h_s_i_p);
782
783 // form INT__h_s_k_p
784 ostei_general_hrr_L(5, 0, 7, 1, hCD, HRR_INT__h_s_l_s, HRR_INT__h_s_k_s, HRR_INT__h_s_k_p);
785
786 // form INT__h_s_l_p
787 ostei_general_hrr_L(5, 0, 8, 1, hCD, HRR_INT__h_s_m_s, HRR_INT__h_s_l_s, HRR_INT__h_s_l_p);
788
789 // form INT__h_s_h_d
790 ostei_general_hrr_L(5, 0, 5, 2, hCD, HRR_INT__h_s_i_p, HRR_INT__h_s_h_p, HRR_INT__h_s_h_d);
791
792 // form INT__h_s_i_d
793 ostei_general_hrr_L(5, 0, 6, 2, hCD, HRR_INT__h_s_k_p, HRR_INT__h_s_i_p, HRR_INT__h_s_i_d);
794
795 // form INT__h_s_k_d
796 ostei_general_hrr_L(5, 0, 7, 2, hCD, HRR_INT__h_s_l_p, HRR_INT__h_s_k_p, HRR_INT__h_s_k_d);
797
798 // form INT__h_s_h_f
799 ostei_general_hrr_L(5, 0, 5, 3, hCD, HRR_INT__h_s_i_d, HRR_INT__h_s_h_d, HRR_INT__h_s_h_f);
800
801 // form INT__h_s_i_f
802 ostei_general_hrr_L(5, 0, 6, 3, hCD, HRR_INT__h_s_k_d, HRR_INT__h_s_i_d, HRR_INT__h_s_i_f);
803
804 // form INT__h_s_h_g
805 ostei_general_hrr_L(5, 0, 5, 4, hCD, HRR_INT__h_s_i_f, HRR_INT__h_s_h_f, HRR_INT__h_s_h_g);
806
807
808 } // close HRR loop
809
810
811 } // close loop cdbatch
812
813 istart = iend;
814 } // close loop over ab
815
816 return P.nshell12_clip * Q.nshell12_clip;
817 }
818
ostei_s_h_h_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_h_h_g)819 int ostei_s_h_h_g(struct simint_multi_shellpair const P,
820 struct simint_multi_shellpair const Q,
821 double screen_tol,
822 double * const restrict work,
823 double * const restrict INT__s_h_h_g)
824 {
825 double P_AB[3*P.nshell12];
826 struct simint_multi_shellpair P_tmp = P;
827 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
828 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
829 P_tmp.AB_x = P_AB;
830 P_tmp.AB_y = P_AB + P.nshell12;
831 P_tmp.AB_z = P_AB + 2*P.nshell12;
832
833 for(int i = 0; i < P.nshell12; i++)
834 {
835 P_tmp.AB_x[i] = -P.AB_x[i];
836 P_tmp.AB_y[i] = -P.AB_y[i];
837 P_tmp.AB_z[i] = -P.AB_z[i];
838 }
839
840 int ret = ostei_h_s_h_g(P_tmp, Q, screen_tol, work, INT__s_h_h_g);
841 double buffer[6615] SIMINT_ALIGN_ARRAY_DBL;
842
843 for(int q = 0; q < ret; q++)
844 {
845 int idx = 0;
846 for(int a = 0; a < 1; ++a)
847 for(int b = 0; b < 21; ++b)
848 for(int c = 0; c < 21; ++c)
849 for(int d = 0; d < 15; ++d)
850 buffer[idx++] = INT__s_h_h_g[q*6615+b*315+a*315+c*15+d];
851
852 memcpy(INT__s_h_h_g+q*6615, buffer, 6615*sizeof(double));
853 }
854
855 return ret;
856 }
857
ostei_h_s_g_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__h_s_g_h)858 int ostei_h_s_g_h(struct simint_multi_shellpair const P,
859 struct simint_multi_shellpair const Q,
860 double screen_tol,
861 double * const restrict work,
862 double * const restrict INT__h_s_g_h)
863 {
864 double Q_AB[3*Q.nshell12];
865 struct simint_multi_shellpair Q_tmp = Q;
866 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
867 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
868 Q_tmp.AB_x = Q_AB;
869 Q_tmp.AB_y = Q_AB + Q.nshell12;
870 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
871
872 for(int i = 0; i < Q.nshell12; i++)
873 {
874 Q_tmp.AB_x[i] = -Q.AB_x[i];
875 Q_tmp.AB_y[i] = -Q.AB_y[i];
876 Q_tmp.AB_z[i] = -Q.AB_z[i];
877 }
878
879 int ret = ostei_h_s_h_g(P, Q_tmp, screen_tol, work, INT__h_s_g_h);
880 double buffer[6615] SIMINT_ALIGN_ARRAY_DBL;
881
882 for(int q = 0; q < ret; q++)
883 {
884 int idx = 0;
885 for(int a = 0; a < 21; ++a)
886 for(int b = 0; b < 1; ++b)
887 for(int c = 0; c < 15; ++c)
888 for(int d = 0; d < 21; ++d)
889 buffer[idx++] = INT__h_s_g_h[q*6615+a*315+b*315+d*15+c];
890
891 memcpy(INT__h_s_g_h+q*6615, buffer, 6615*sizeof(double));
892 }
893
894 return ret;
895 }
896
ostei_s_h_g_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_h_g_h)897 int ostei_s_h_g_h(struct simint_multi_shellpair const P,
898 struct simint_multi_shellpair const Q,
899 double screen_tol,
900 double * const restrict work,
901 double * const restrict INT__s_h_g_h)
902 {
903 double P_AB[3*P.nshell12];
904 struct simint_multi_shellpair P_tmp = P;
905 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
906 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
907 P_tmp.AB_x = P_AB;
908 P_tmp.AB_y = P_AB + P.nshell12;
909 P_tmp.AB_z = P_AB + 2*P.nshell12;
910
911 for(int i = 0; i < P.nshell12; i++)
912 {
913 P_tmp.AB_x[i] = -P.AB_x[i];
914 P_tmp.AB_y[i] = -P.AB_y[i];
915 P_tmp.AB_z[i] = -P.AB_z[i];
916 }
917
918 double Q_AB[3*Q.nshell12];
919 struct simint_multi_shellpair Q_tmp = Q;
920 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
921 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
922 Q_tmp.AB_x = Q_AB;
923 Q_tmp.AB_y = Q_AB + Q.nshell12;
924 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
925
926 for(int i = 0; i < Q.nshell12; i++)
927 {
928 Q_tmp.AB_x[i] = -Q.AB_x[i];
929 Q_tmp.AB_y[i] = -Q.AB_y[i];
930 Q_tmp.AB_z[i] = -Q.AB_z[i];
931 }
932
933 int ret = ostei_h_s_h_g(P_tmp, Q_tmp, screen_tol, work, INT__s_h_g_h);
934 double buffer[6615] SIMINT_ALIGN_ARRAY_DBL;
935
936 for(int q = 0; q < ret; q++)
937 {
938 int idx = 0;
939 for(int a = 0; a < 1; ++a)
940 for(int b = 0; b < 21; ++b)
941 for(int c = 0; c < 15; ++c)
942 for(int d = 0; d < 21; ++d)
943 buffer[idx++] = INT__s_h_g_h[q*6615+b*315+a*315+d*15+c];
944
945 memcpy(INT__s_h_g_h+q*6615, buffer, 6615*sizeof(double));
946 }
947
948 return ret;
949 }
950
951