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_k_p_g_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_p_g_f)8 int ostei_k_p_g_f(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__k_p_g_f)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__k_p_g_f);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 540);
31 double * const INT__k_s_i_s = work + (SIMINT_NSHELL_SIMD * 1296);
32 double * const INT__k_s_k_s = work + (SIMINT_NSHELL_SIMD * 2304);
33 double * const INT__l_s_g_s = work + (SIMINT_NSHELL_SIMD * 3600);
34 double * const INT__l_s_h_s = work + (SIMINT_NSHELL_SIMD * 4275);
35 double * const INT__l_s_i_s = work + (SIMINT_NSHELL_SIMD * 5220);
36 double * const INT__l_s_k_s = work + (SIMINT_NSHELL_SIMD * 6480);
37 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*8100);
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 16;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 61;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 124;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 208;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 334;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 550;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 680;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 890;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 1250;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 1750;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1930;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 2245;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 2785;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 3535;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 4435;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 4666;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 5107;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 5863;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 6913;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 8173;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 9496;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 9776;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 10364;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 11372;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 12772;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 14452;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_i_s = primwork + 16216;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 17784;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 18108;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 18864;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 20160;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 21960;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 24120;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_i_s = primwork + 26388;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_k_s = primwork + 28404;
74 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 29700;
75 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 30060;
76 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 31005;
77 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 32625;
78 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_g_s = primwork + 34875;
79 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_h_s = primwork + 37575;
80 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_i_s = primwork + 40410;
81 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_k_s = primwork + 42930;
82 double * const hrrwork = (double *)(primwork + 44550);
83 double * const HRR_INT__k_p_g_s = hrrwork + 0;
84 double * const HRR_INT__k_p_g_p = hrrwork + 1620;
85 double * const HRR_INT__k_p_g_d = hrrwork + 6480;
86 double * const HRR_INT__k_p_h_s = hrrwork + 16200;
87 double * const HRR_INT__k_p_h_p = hrrwork + 18468;
88 double * const HRR_INT__k_p_h_d = hrrwork + 25272;
89 double * const HRR_INT__k_p_i_s = hrrwork + 38880;
90 double * const HRR_INT__k_p_i_p = hrrwork + 41904;
91 double * const HRR_INT__k_p_k_s = hrrwork + 50976;
92
93
94 // Create constants
95 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
96 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
97 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
98 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
99 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
100 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
101 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
102 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
103 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
104
105
106 ////////////////////////////////////////
107 // Loop over shells and primitives
108 ////////////////////////////////////////
109
110 real_abcd = 0;
111 istart = 0;
112 for(ab = 0; ab < P.nshell12_clip; ++ab)
113 {
114 const int iend = istart + P.nprim12[ab];
115
116 cd = 0;
117 jstart = 0;
118
119 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
120 {
121 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
122 int jend = jstart;
123 for(i = 0; i < nshellbatch; i++)
124 jend += Q.nprim12[cd+i];
125
126 // Clear the beginning of the workspace (where we are accumulating integrals)
127 memset(work, 0, SIMINT_NSHELL_SIMD * 8100 * sizeof(double));
128 abcd = 0;
129
130
131 for(i = istart; i < iend; ++i)
132 {
133 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
134
135 if(check_screen)
136 {
137 // Skip this whole thing if always insignificant
138 if((P.screen[i] * Q.screen_max) < screen_tol)
139 continue;
140 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
141 }
142
143 icd = 0;
144 iprimcd = 0;
145 nprim_icd = Q.nprim12[cd];
146 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
147 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
148 double * restrict PRIM_PTR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
149 double * restrict PRIM_PTR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
150 double * restrict PRIM_PTR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
151 double * restrict PRIM_PTR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
152 double * restrict PRIM_PTR_INT__l_s_i_s = INT__l_s_i_s + abcd * 1260;
153 double * restrict PRIM_PTR_INT__l_s_k_s = INT__l_s_k_s + abcd * 1620;
154
155
156
157 // Load these one per loop over i
158 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
159 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
160 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
161
162 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
163
164 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
165 {
166 // calculate the shell offsets
167 // these are the offset from the shell pointed to by cd
168 // for each element
169 int shelloffsets[SIMINT_SIMD_LEN] = {0};
170 int lastoffset = 0;
171 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
172
173 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
174 {
175 // Handle if the first element of the vector is a new shell
176 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
177 {
178 nprim_icd += Q.nprim12[cd + (++icd)];
179 PRIM_PTR_INT__k_s_g_s += 540;
180 PRIM_PTR_INT__k_s_h_s += 756;
181 PRIM_PTR_INT__k_s_i_s += 1008;
182 PRIM_PTR_INT__k_s_k_s += 1296;
183 PRIM_PTR_INT__l_s_g_s += 675;
184 PRIM_PTR_INT__l_s_h_s += 945;
185 PRIM_PTR_INT__l_s_i_s += 1260;
186 PRIM_PTR_INT__l_s_k_s += 1620;
187 }
188 iprimcd++;
189 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
190 {
191 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
192 {
193 shelloffsets[n] = shelloffsets[n-1] + 1;
194 lastoffset++;
195 nprim_icd += Q.nprim12[cd + (++icd)];
196 }
197 else
198 shelloffsets[n] = shelloffsets[n-1];
199 iprimcd++;
200 }
201 }
202 else
203 iprimcd += SIMINT_SIMD_LEN;
204
205 // Do we have to compute this vector (or has it been screened out)?
206 // (not_screened != 0 means we have to do this vector)
207 if(check_screen)
208 {
209 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
210 if(vmax < screen_tol)
211 {
212 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
213 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
214 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
215 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
216 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
217 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
218 PRIM_PTR_INT__l_s_i_s += lastoffset*1260;
219 PRIM_PTR_INT__l_s_k_s += lastoffset*1620;
220 continue;
221 }
222 }
223
224 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
225 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
226 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
227 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
228
229
230 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
231 SIMINT_DBLTYPE PQ[3];
232 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
233 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
234 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
235 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
236 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
237 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
238
239 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
240 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
241 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
242 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
243 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
244 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
245 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
246
247 // NOTE: Minus sign!
248 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
249 SIMINT_DBLTYPE aop_PQ[3];
250 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
251 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
252 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
253
254 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
255 SIMINT_DBLTYPE aoq_PQ[3];
256 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
257 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
258 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
259 // Put a minus sign here so we don't have to in RR routines
260 a_over_q = SIMINT_NEG(a_over_q);
261
262
263 //////////////////////////////////////////////
264 // Fjt function section
265 // Maximum v value: 15
266 //////////////////////////////////////////////
267 // The parameter to the Fjt function
268 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
269
270
271 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
272
273
274 boys_F_split(PRIM_INT__s_s_s_s, F_x, 15);
275 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
276 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
277 for(n = 0; n <= 15; n++)
278 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
279
280 //////////////////////////////////////////////
281 // Primitive integrals: Vertical recurrance
282 //////////////////////////////////////////////
283
284 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
285 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
286 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
287 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
288 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
289 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
290 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
291 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
292 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
293 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
294 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
295 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
296 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
297 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
298 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
299 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
300 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
301 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
302 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
303 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
304 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
305
306
307
308 // Forming PRIM_INT__p_s_s_s[15 * 3];
309 for(n = 0; n < 15; ++n) // loop over orders of auxiliary function
310 {
311
312 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
313 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]);
314
315 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
316 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]);
317
318 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
319 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]);
320
321 }
322
323
324
325 // Forming PRIM_INT__d_s_s_s[14 * 6];
326 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
327 {
328
329 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
330 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]);
331 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]);
332
333 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
334 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]);
335
336 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
337 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]);
338
339 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
340 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]);
341 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]);
342
343 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
344 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]);
345
346 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
347 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]);
348 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]);
349
350 }
351
352
353
354 // Forming PRIM_INT__f_s_s_s[13 * 10];
355 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
356 {
357
358 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
359 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]);
360 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]);
361
362 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
363 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]);
364
365 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
366 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]);
367
368 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
369 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]);
370
371 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
372 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]);
373
374 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
375 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]);
376
377 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
378 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]);
379 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]);
380
381 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
382 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]);
383
384 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
385 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]);
386
387 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
388 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]);
389 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]);
390
391 }
392
393
394 VRR_I_g_s_s_s(
395 PRIM_INT__g_s_s_s,
396 PRIM_INT__f_s_s_s,
397 PRIM_INT__d_s_s_s,
398 P_PA,
399 a_over_p,
400 aop_PQ,
401 one_over_2p,
402 12);
403
404
405 VRR_I_h_s_s_s(
406 PRIM_INT__h_s_s_s,
407 PRIM_INT__g_s_s_s,
408 PRIM_INT__f_s_s_s,
409 P_PA,
410 a_over_p,
411 aop_PQ,
412 one_over_2p,
413 11);
414
415
416 ostei_general_vrr1_I(6, 10,
417 one_over_2p, a_over_p, aop_PQ, P_PA,
418 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
419
420
421 ostei_general_vrr1_I(7, 9,
422 one_over_2p, a_over_p, aop_PQ, P_PA,
423 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
424
425
426 ostei_general_vrr_K(7, 0, 1, 0, 7,
427 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
428 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
429
430
431 ostei_general_vrr_K(6, 0, 1, 0, 7,
432 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
433 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
434
435
436 ostei_general_vrr_K(7, 0, 2, 0, 6,
437 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
438 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);
439
440
441 ostei_general_vrr_K(5, 0, 1, 0, 7,
442 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
443 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
444
445
446 ostei_general_vrr_K(6, 0, 2, 0, 6,
447 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
448 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);
449
450
451 ostei_general_vrr_K(7, 0, 3, 0, 5,
452 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
453 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);
454
455
456 VRR_K_g_s_p_s(
457 PRIM_INT__g_s_p_s,
458 PRIM_INT__g_s_s_s,
459 PRIM_INT__f_s_s_s,
460 Q_PA,
461 aoq_PQ,
462 one_over_2pq,
463 7);
464
465
466 ostei_general_vrr_K(5, 0, 2, 0, 6,
467 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
468 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);
469
470
471 ostei_general_vrr_K(6, 0, 3, 0, 5,
472 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
473 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);
474
475
476 ostei_general_vrr_K(7, 0, 4, 0, 4,
477 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
478 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);
479
480
481 VRR_K_f_s_p_s(
482 PRIM_INT__f_s_p_s,
483 PRIM_INT__f_s_s_s,
484 PRIM_INT__d_s_s_s,
485 Q_PA,
486 aoq_PQ,
487 one_over_2pq,
488 7);
489
490
491 ostei_general_vrr_K(4, 0, 2, 0, 6,
492 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
493 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);
494
495
496 ostei_general_vrr_K(5, 0, 3, 0, 5,
497 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
498 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);
499
500
501 ostei_general_vrr_K(6, 0, 4, 0, 4,
502 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
503 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);
504
505
506 ostei_general_vrr_K(7, 0, 5, 0, 3,
507 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
508 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);
509
510
511 ostei_general_vrr1_I(8, 8,
512 one_over_2p, a_over_p, aop_PQ, P_PA,
513 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
514
515
516 ostei_general_vrr_K(8, 0, 1, 0, 7,
517 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
518 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
519
520
521 ostei_general_vrr_K(8, 0, 2, 0, 6,
522 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
523 PRIM_INT__l_s_p_s, PRIM_INT__l_s_s_s, NULL, PRIM_INT__k_s_p_s, NULL, PRIM_INT__l_s_d_s);
524
525
526 ostei_general_vrr_K(8, 0, 3, 0, 5,
527 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
528 PRIM_INT__l_s_d_s, PRIM_INT__l_s_p_s, NULL, PRIM_INT__k_s_d_s, NULL, PRIM_INT__l_s_f_s);
529
530
531 ostei_general_vrr_K(8, 0, 4, 0, 4,
532 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
533 PRIM_INT__l_s_f_s, PRIM_INT__l_s_d_s, NULL, PRIM_INT__k_s_f_s, NULL, PRIM_INT__l_s_g_s);
534
535
536
537 // Forming PRIM_INT__d_s_p_s[7 * 18];
538 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
539 {
540
541 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
542 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]);
543 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]);
544
545 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
546 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]);
547
548 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
549 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]);
550
551 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
552 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]);
553 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]);
554
555 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
556 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]);
557 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]);
558
559 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
560 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]);
561
562 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
563 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]);
564 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]);
565
566 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
567 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]);
568
569 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
570 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]);
571 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]);
572
573 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
574 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]);
575
576 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
577 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]);
578 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]);
579
580 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
581 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]);
582
583 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
584 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]);
585
586 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
587 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]);
588 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]);
589
590 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
591 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]);
592 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]);
593
594 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
595 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]);
596
597 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
598 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]);
599
600 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
601 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]);
602 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]);
603
604 }
605
606
607 VRR_K_f_s_d_s(
608 PRIM_INT__f_s_d_s,
609 PRIM_INT__f_s_p_s,
610 PRIM_INT__f_s_s_s,
611 PRIM_INT__d_s_p_s,
612 Q_PA,
613 a_over_q,
614 aoq_PQ,
615 one_over_2pq,
616 one_over_2q,
617 6);
618
619
620 ostei_general_vrr_K(4, 0, 3, 0, 5,
621 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
622 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);
623
624
625 ostei_general_vrr_K(5, 0, 4, 0, 4,
626 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
627 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);
628
629
630 ostei_general_vrr_K(6, 0, 5, 0, 3,
631 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
632 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);
633
634
635 ostei_general_vrr_K(7, 0, 6, 0, 2,
636 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
637 PRIM_INT__k_s_h_s, PRIM_INT__k_s_g_s, NULL, PRIM_INT__i_s_h_s, NULL, PRIM_INT__k_s_i_s);
638
639
640 ostei_general_vrr_K(8, 0, 5, 0, 3,
641 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
642 PRIM_INT__l_s_g_s, PRIM_INT__l_s_f_s, NULL, PRIM_INT__k_s_g_s, NULL, PRIM_INT__l_s_h_s);
643
644
645
646 // Forming PRIM_INT__p_s_p_s[7 * 9];
647 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
648 {
649
650 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
651 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]);
652 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]);
653
654 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
655 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]);
656
657 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
658 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]);
659
660 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
661 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]);
662
663 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
664 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]);
665 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]);
666
667 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
668 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]);
669
670 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
671 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]);
672
673 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
674 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]);
675
676 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
677 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]);
678 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]);
679
680 }
681
682
683 VRR_K_d_s_d_s(
684 PRIM_INT__d_s_d_s,
685 PRIM_INT__d_s_p_s,
686 PRIM_INT__d_s_s_s,
687 PRIM_INT__p_s_p_s,
688 Q_PA,
689 a_over_q,
690 aoq_PQ,
691 one_over_2pq,
692 one_over_2q,
693 6);
694
695
696 ostei_general_vrr_K(3, 0, 3, 0, 5,
697 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
698 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);
699
700
701 ostei_general_vrr_K(4, 0, 4, 0, 4,
702 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
703 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);
704
705
706 ostei_general_vrr_K(5, 0, 5, 0, 3,
707 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
708 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);
709
710
711 ostei_general_vrr_K(6, 0, 6, 0, 2,
712 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
713 PRIM_INT__i_s_h_s, PRIM_INT__i_s_g_s, NULL, PRIM_INT__h_s_h_s, NULL, PRIM_INT__i_s_i_s);
714
715
716 ostei_general_vrr_K(7, 0, 7, 0, 1,
717 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
718 PRIM_INT__k_s_i_s, PRIM_INT__k_s_h_s, NULL, PRIM_INT__i_s_i_s, NULL, PRIM_INT__k_s_k_s);
719
720
721 ostei_general_vrr_K(8, 0, 6, 0, 2,
722 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
723 PRIM_INT__l_s_h_s, PRIM_INT__l_s_g_s, NULL, PRIM_INT__k_s_h_s, NULL, PRIM_INT__l_s_i_s);
724
725
726 ostei_general_vrr_K(8, 0, 7, 0, 1,
727 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
728 PRIM_INT__l_s_i_s, PRIM_INT__l_s_h_s, NULL, PRIM_INT__k_s_i_s, NULL, PRIM_INT__l_s_k_s);
729
730
731
732
733 ////////////////////////////////////
734 // Accumulate contracted integrals
735 ////////////////////////////////////
736 if(lastoffset == 0)
737 {
738 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
739 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
740 contract_all(1008, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
741 contract_all(1296, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
742 contract_all(675, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
743 contract_all(945, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
744 contract_all(1260, PRIM_INT__l_s_i_s, PRIM_PTR_INT__l_s_i_s);
745 contract_all(1620, PRIM_INT__l_s_k_s, PRIM_PTR_INT__l_s_k_s);
746 }
747 else
748 {
749 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
750 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
751 contract(1008, shelloffsets, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
752 contract(1296, shelloffsets, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
753 contract(675, shelloffsets, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
754 contract(945, shelloffsets, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
755 contract(1260, shelloffsets, PRIM_INT__l_s_i_s, PRIM_PTR_INT__l_s_i_s);
756 contract(1620, shelloffsets, PRIM_INT__l_s_k_s, PRIM_PTR_INT__l_s_k_s);
757 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
758 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
759 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
760 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
761 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
762 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
763 PRIM_PTR_INT__l_s_i_s += lastoffset*1260;
764 PRIM_PTR_INT__l_s_k_s += lastoffset*1620;
765 }
766
767 } // close loop over j
768 } // close loop over i
769
770 //Advance to the next batch
771 jstart = SIMINT_SIMD_ROUND(jend);
772
773 //////////////////////////////////////////////
774 // Contracted integrals: Horizontal recurrance
775 //////////////////////////////////////////////
776
777
778 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
779
780
781 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
782 {
783 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
784
785 // set up HRR pointers
786 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
787 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
788 double const * restrict HRR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
789 double const * restrict HRR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
790 double const * restrict HRR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
791 double const * restrict HRR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
792 double const * restrict HRR_INT__l_s_i_s = INT__l_s_i_s + abcd * 1260;
793 double const * restrict HRR_INT__l_s_k_s = INT__l_s_k_s + abcd * 1620;
794 double * restrict HRR_INT__k_p_g_f = INT__k_p_g_f + real_abcd * 16200;
795
796 // form INT__k_p_g_s
797 ostei_general_hrr_J(7, 1, 4, 0, hAB, HRR_INT__l_s_g_s, HRR_INT__k_s_g_s, HRR_INT__k_p_g_s);
798
799 // form INT__k_p_h_s
800 ostei_general_hrr_J(7, 1, 5, 0, hAB, HRR_INT__l_s_h_s, HRR_INT__k_s_h_s, HRR_INT__k_p_h_s);
801
802 // form INT__k_p_i_s
803 ostei_general_hrr_J(7, 1, 6, 0, hAB, HRR_INT__l_s_i_s, HRR_INT__k_s_i_s, HRR_INT__k_p_i_s);
804
805 // form INT__k_p_k_s
806 ostei_general_hrr_J(7, 1, 7, 0, hAB, HRR_INT__l_s_k_s, HRR_INT__k_s_k_s, HRR_INT__k_p_k_s);
807
808 // form INT__k_p_g_p
809 HRR_L_g_p(
810 HRR_INT__k_p_g_p,
811 HRR_INT__k_p_g_s,
812 HRR_INT__k_p_h_s,
813 hCD, 108);
814
815 // form INT__k_p_h_p
816 ostei_general_hrr_L(7, 1, 5, 1, hCD, HRR_INT__k_p_i_s, HRR_INT__k_p_h_s, HRR_INT__k_p_h_p);
817
818 // form INT__k_p_i_p
819 ostei_general_hrr_L(7, 1, 6, 1, hCD, HRR_INT__k_p_k_s, HRR_INT__k_p_i_s, HRR_INT__k_p_i_p);
820
821 // form INT__k_p_g_d
822 ostei_general_hrr_L(7, 1, 4, 2, hCD, HRR_INT__k_p_h_p, HRR_INT__k_p_g_p, HRR_INT__k_p_g_d);
823
824 // form INT__k_p_h_d
825 ostei_general_hrr_L(7, 1, 5, 2, hCD, HRR_INT__k_p_i_p, HRR_INT__k_p_h_p, HRR_INT__k_p_h_d);
826
827 // form INT__k_p_g_f
828 ostei_general_hrr_L(7, 1, 4, 3, hCD, HRR_INT__k_p_h_d, HRR_INT__k_p_g_d, HRR_INT__k_p_g_f);
829
830
831 } // close HRR loop
832
833
834 } // close loop cdbatch
835
836 istart = iend;
837 } // close loop over ab
838
839 return P.nshell12_clip * Q.nshell12_clip;
840 }
841
ostei_p_k_g_f(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_k_g_f)842 int ostei_p_k_g_f(struct simint_multi_shellpair const P,
843 struct simint_multi_shellpair const Q,
844 double screen_tol,
845 double * const restrict work,
846 double * const restrict INT__p_k_g_f)
847 {
848 double P_AB[3*P.nshell12];
849 struct simint_multi_shellpair P_tmp = P;
850 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
851 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
852 P_tmp.AB_x = P_AB;
853 P_tmp.AB_y = P_AB + P.nshell12;
854 P_tmp.AB_z = P_AB + 2*P.nshell12;
855
856 for(int i = 0; i < P.nshell12; i++)
857 {
858 P_tmp.AB_x[i] = -P.AB_x[i];
859 P_tmp.AB_y[i] = -P.AB_y[i];
860 P_tmp.AB_z[i] = -P.AB_z[i];
861 }
862
863 int ret = ostei_k_p_g_f(P_tmp, Q, screen_tol, work, INT__p_k_g_f);
864 double buffer[16200] SIMINT_ALIGN_ARRAY_DBL;
865
866 for(int q = 0; q < ret; q++)
867 {
868 int idx = 0;
869 for(int a = 0; a < 3; ++a)
870 for(int b = 0; b < 36; ++b)
871 for(int c = 0; c < 15; ++c)
872 for(int d = 0; d < 10; ++d)
873 buffer[idx++] = INT__p_k_g_f[q*16200+b*450+a*150+c*10+d];
874
875 memcpy(INT__p_k_g_f+q*16200, buffer, 16200*sizeof(double));
876 }
877
878 return ret;
879 }
880
ostei_k_p_f_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_p_f_g)881 int ostei_k_p_f_g(struct simint_multi_shellpair const P,
882 struct simint_multi_shellpair const Q,
883 double screen_tol,
884 double * const restrict work,
885 double * const restrict INT__k_p_f_g)
886 {
887 double Q_AB[3*Q.nshell12];
888 struct simint_multi_shellpair Q_tmp = Q;
889 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
890 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
891 Q_tmp.AB_x = Q_AB;
892 Q_tmp.AB_y = Q_AB + Q.nshell12;
893 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
894
895 for(int i = 0; i < Q.nshell12; i++)
896 {
897 Q_tmp.AB_x[i] = -Q.AB_x[i];
898 Q_tmp.AB_y[i] = -Q.AB_y[i];
899 Q_tmp.AB_z[i] = -Q.AB_z[i];
900 }
901
902 int ret = ostei_k_p_g_f(P, Q_tmp, screen_tol, work, INT__k_p_f_g);
903 double buffer[16200] SIMINT_ALIGN_ARRAY_DBL;
904
905 for(int q = 0; q < ret; q++)
906 {
907 int idx = 0;
908 for(int a = 0; a < 36; ++a)
909 for(int b = 0; b < 3; ++b)
910 for(int c = 0; c < 10; ++c)
911 for(int d = 0; d < 15; ++d)
912 buffer[idx++] = INT__k_p_f_g[q*16200+a*450+b*150+d*10+c];
913
914 memcpy(INT__k_p_f_g+q*16200, buffer, 16200*sizeof(double));
915 }
916
917 return ret;
918 }
919
ostei_p_k_f_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_k_f_g)920 int ostei_p_k_f_g(struct simint_multi_shellpair const P,
921 struct simint_multi_shellpair const Q,
922 double screen_tol,
923 double * const restrict work,
924 double * const restrict INT__p_k_f_g)
925 {
926 double P_AB[3*P.nshell12];
927 struct simint_multi_shellpair P_tmp = P;
928 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
929 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
930 P_tmp.AB_x = P_AB;
931 P_tmp.AB_y = P_AB + P.nshell12;
932 P_tmp.AB_z = P_AB + 2*P.nshell12;
933
934 for(int i = 0; i < P.nshell12; i++)
935 {
936 P_tmp.AB_x[i] = -P.AB_x[i];
937 P_tmp.AB_y[i] = -P.AB_y[i];
938 P_tmp.AB_z[i] = -P.AB_z[i];
939 }
940
941 double Q_AB[3*Q.nshell12];
942 struct simint_multi_shellpair Q_tmp = Q;
943 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
944 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
945 Q_tmp.AB_x = Q_AB;
946 Q_tmp.AB_y = Q_AB + Q.nshell12;
947 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
948
949 for(int i = 0; i < Q.nshell12; i++)
950 {
951 Q_tmp.AB_x[i] = -Q.AB_x[i];
952 Q_tmp.AB_y[i] = -Q.AB_y[i];
953 Q_tmp.AB_z[i] = -Q.AB_z[i];
954 }
955
956 int ret = ostei_k_p_g_f(P_tmp, Q_tmp, screen_tol, work, INT__p_k_f_g);
957 double buffer[16200] SIMINT_ALIGN_ARRAY_DBL;
958
959 for(int q = 0; q < ret; q++)
960 {
961 int idx = 0;
962 for(int a = 0; a < 3; ++a)
963 for(int b = 0; b < 36; ++b)
964 for(int c = 0; c < 10; ++c)
965 for(int d = 0; d < 15; ++d)
966 buffer[idx++] = INT__p_k_f_g[q*16200+b*450+a*150+d*10+c];
967
968 memcpy(INT__p_k_f_g+q*16200, buffer, 16200*sizeof(double));
969 }
970
971 return ret;
972 }
973
974