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_s_h_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_s_h_d)8 int ostei_k_s_h_d(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__k_s_h_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__k_s_h_d);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int ibra;
26
27 // partition workspace
28 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 0);
29 double * const INT__k_s_i_s = work + (SIMINT_NSHELL_SIMD * 756);
30 double * const INT__k_s_k_s = work + (SIMINT_NSHELL_SIMD * 1764);
31 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3060);
32 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
33 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 15;
34 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 57;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 120;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 198;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 324;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 540;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 660;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 870;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 1230;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 1730;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1895;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 2210;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 2750;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 3500;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 4400;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 4610;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 5051;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 5807;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 6857;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 8117;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 9440;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 9692;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 10280;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 11288;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 12688;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 14368;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_i_s = primwork + 16132;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 17700;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 17988;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 18744;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 20040;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 21840;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 24000;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_i_s = primwork + 26268;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_k_s = primwork + 28284;
68 double * const hrrwork = (double *)(primwork + 29580);
69 double * const HRR_INT__k_s_h_p = hrrwork + 0;
70 double * const HRR_INT__k_s_i_p = hrrwork + 2268;
71
72
73 // Create constants
74 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
75 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
76 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
77 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
78 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
79 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
80 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
81 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
82
83
84 ////////////////////////////////////////
85 // Loop over shells and primitives
86 ////////////////////////////////////////
87
88 real_abcd = 0;
89 istart = 0;
90 for(ab = 0; ab < P.nshell12_clip; ++ab)
91 {
92 const int iend = istart + P.nprim12[ab];
93
94 cd = 0;
95 jstart = 0;
96
97 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
98 {
99 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
100 int jend = jstart;
101 for(i = 0; i < nshellbatch; i++)
102 jend += Q.nprim12[cd+i];
103
104 // Clear the beginning of the workspace (where we are accumulating integrals)
105 memset(work, 0, SIMINT_NSHELL_SIMD * 3060 * sizeof(double));
106 abcd = 0;
107
108
109 for(i = istart; i < iend; ++i)
110 {
111 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
112
113 if(check_screen)
114 {
115 // Skip this whole thing if always insignificant
116 if((P.screen[i] * Q.screen_max) < screen_tol)
117 continue;
118 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
119 }
120
121 icd = 0;
122 iprimcd = 0;
123 nprim_icd = Q.nprim12[cd];
124 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
125 double * restrict PRIM_PTR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
126 double * restrict PRIM_PTR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
127
128
129
130 // Load these one per loop over i
131 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
132 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
133 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
134
135 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
136
137 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
138 {
139 // calculate the shell offsets
140 // these are the offset from the shell pointed to by cd
141 // for each element
142 int shelloffsets[SIMINT_SIMD_LEN] = {0};
143 int lastoffset = 0;
144 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
145
146 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
147 {
148 // Handle if the first element of the vector is a new shell
149 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
150 {
151 nprim_icd += Q.nprim12[cd + (++icd)];
152 PRIM_PTR_INT__k_s_h_s += 756;
153 PRIM_PTR_INT__k_s_i_s += 1008;
154 PRIM_PTR_INT__k_s_k_s += 1296;
155 }
156 iprimcd++;
157 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
158 {
159 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
160 {
161 shelloffsets[n] = shelloffsets[n-1] + 1;
162 lastoffset++;
163 nprim_icd += Q.nprim12[cd + (++icd)];
164 }
165 else
166 shelloffsets[n] = shelloffsets[n-1];
167 iprimcd++;
168 }
169 }
170 else
171 iprimcd += SIMINT_SIMD_LEN;
172
173 // Do we have to compute this vector (or has it been screened out)?
174 // (not_screened != 0 means we have to do this vector)
175 if(check_screen)
176 {
177 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
178 if(vmax < screen_tol)
179 {
180 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
181 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
182 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
183 continue;
184 }
185 }
186
187 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
188 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
189 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
190 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
191
192
193 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
194 SIMINT_DBLTYPE PQ[3];
195 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
196 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
197 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
198 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
199 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
200 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
201
202 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
203 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
204 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
205 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
206 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
207 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
208 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
209
210 // NOTE: Minus sign!
211 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
212 SIMINT_DBLTYPE aop_PQ[3];
213 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
214 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
215 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
216
217 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
218 SIMINT_DBLTYPE aoq_PQ[3];
219 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
220 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
221 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
222 // Put a minus sign here so we don't have to in RR routines
223 a_over_q = SIMINT_NEG(a_over_q);
224
225
226 //////////////////////////////////////////////
227 // Fjt function section
228 // Maximum v value: 14
229 //////////////////////////////////////////////
230 // The parameter to the Fjt function
231 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
232
233
234 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
235
236
237 boys_F_split(PRIM_INT__s_s_s_s, F_x, 14);
238 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
239 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
240 for(n = 0; n <= 14; n++)
241 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
242
243 //////////////////////////////////////////////
244 // Primitive integrals: Vertical recurrance
245 //////////////////////////////////////////////
246
247 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
248 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
249 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
250 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
251 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
252 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
253 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
254 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
255 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
256 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
257 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
258 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, one_over_2q);
259 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
260 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
261 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
262 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
263 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
264 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
265 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
266
267
268
269 // Forming PRIM_INT__p_s_s_s[14 * 3];
270 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
271 {
272
273 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
274 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]);
275
276 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
277 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]);
278
279 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
280 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]);
281
282 }
283
284
285
286 // Forming PRIM_INT__d_s_s_s[13 * 6];
287 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
288 {
289
290 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
291 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]);
292 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]);
293
294 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
295 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]);
296
297 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
298 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]);
299
300 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
301 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]);
302 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]);
303
304 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
305 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]);
306
307 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
308 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]);
309 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]);
310
311 }
312
313
314
315 // Forming PRIM_INT__f_s_s_s[12 * 10];
316 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
317 {
318
319 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
320 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]);
321 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]);
322
323 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
324 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]);
325
326 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
327 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]);
328
329 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
330 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]);
331
332 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
333 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]);
334
335 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
336 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]);
337
338 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
339 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]);
340 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]);
341
342 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
343 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]);
344
345 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
346 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]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
349 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]);
350 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]);
351
352 }
353
354
355 VRR_I_g_s_s_s(
356 PRIM_INT__g_s_s_s,
357 PRIM_INT__f_s_s_s,
358 PRIM_INT__d_s_s_s,
359 P_PA,
360 a_over_p,
361 aop_PQ,
362 one_over_2p,
363 11);
364
365
366 VRR_I_h_s_s_s(
367 PRIM_INT__h_s_s_s,
368 PRIM_INT__g_s_s_s,
369 PRIM_INT__f_s_s_s,
370 P_PA,
371 a_over_p,
372 aop_PQ,
373 one_over_2p,
374 10);
375
376
377 ostei_general_vrr1_I(6, 9,
378 one_over_2p, a_over_p, aop_PQ, P_PA,
379 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
380
381
382 ostei_general_vrr1_I(7, 8,
383 one_over_2p, a_over_p, aop_PQ, P_PA,
384 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
385
386
387 ostei_general_vrr_K(7, 0, 1, 0, 7,
388 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
389 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
390
391
392 ostei_general_vrr_K(6, 0, 1, 0, 7,
393 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
394 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
395
396
397 ostei_general_vrr_K(7, 0, 2, 0, 6,
398 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
399 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);
400
401
402 ostei_general_vrr_K(5, 0, 1, 0, 7,
403 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
404 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
405
406
407 ostei_general_vrr_K(6, 0, 2, 0, 6,
408 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
409 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);
410
411
412 ostei_general_vrr_K(7, 0, 3, 0, 5,
413 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
414 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);
415
416
417 VRR_K_g_s_p_s(
418 PRIM_INT__g_s_p_s,
419 PRIM_INT__g_s_s_s,
420 PRIM_INT__f_s_s_s,
421 Q_PA,
422 aoq_PQ,
423 one_over_2pq,
424 7);
425
426
427 ostei_general_vrr_K(5, 0, 2, 0, 6,
428 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
429 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);
430
431
432 ostei_general_vrr_K(6, 0, 3, 0, 5,
433 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
434 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);
435
436
437 ostei_general_vrr_K(7, 0, 4, 0, 4,
438 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
439 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);
440
441
442 VRR_K_f_s_p_s(
443 PRIM_INT__f_s_p_s,
444 PRIM_INT__f_s_s_s,
445 PRIM_INT__d_s_s_s,
446 Q_PA,
447 aoq_PQ,
448 one_over_2pq,
449 7);
450
451
452 ostei_general_vrr_K(4, 0, 2, 0, 6,
453 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
454 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);
455
456
457 ostei_general_vrr_K(5, 0, 3, 0, 5,
458 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
459 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);
460
461
462 ostei_general_vrr_K(6, 0, 4, 0, 4,
463 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
464 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);
465
466
467 ostei_general_vrr_K(7, 0, 5, 0, 3,
468 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
469 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);
470
471
472
473 // Forming PRIM_INT__d_s_p_s[7 * 18];
474 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
475 {
476
477 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
478 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]);
479 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]);
480
481 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
482 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]);
483
484 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
485 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]);
486
487 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
488 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]);
489 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]);
490
491 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
492 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]);
493 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]);
494
495 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
496 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]);
497
498 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
499 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]);
500 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]);
501
502 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
503 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]);
504
505 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
506 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]);
507 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]);
508
509 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
510 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]);
511
512 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
513 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]);
514 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]);
515
516 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
517 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]);
518
519 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
520 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]);
521
522 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
523 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]);
524 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]);
525
526 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
527 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]);
528 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]);
529
530 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
531 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]);
532
533 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
534 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]);
535
536 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
537 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]);
538 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]);
539
540 }
541
542
543 VRR_K_f_s_d_s(
544 PRIM_INT__f_s_d_s,
545 PRIM_INT__f_s_p_s,
546 PRIM_INT__f_s_s_s,
547 PRIM_INT__d_s_p_s,
548 Q_PA,
549 a_over_q,
550 aoq_PQ,
551 one_over_2pq,
552 one_over_2q,
553 6);
554
555
556 ostei_general_vrr_K(4, 0, 3, 0, 5,
557 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
558 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);
559
560
561 ostei_general_vrr_K(5, 0, 4, 0, 4,
562 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
563 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);
564
565
566 ostei_general_vrr_K(6, 0, 5, 0, 3,
567 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
568 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);
569
570
571 ostei_general_vrr_K(7, 0, 6, 0, 2,
572 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
573 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);
574
575
576
577 // Forming PRIM_INT__p_s_p_s[7 * 9];
578 for(n = 0; n < 7; ++n) // loop over orders of auxiliary function
579 {
580
581 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
582 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]);
583 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]);
584
585 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
586 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]);
587
588 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
589 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]);
590
591 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 1]);
592 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]);
593
594 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
595 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]);
596 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]);
597
598 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
599 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]);
600
601 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__p_s_s_s[n * 3 + 2]);
602 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]);
603
604 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__p_s_s_s[n * 3 + 2]);
605 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]);
606
607 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
608 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]);
609 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]);
610
611 }
612
613
614 VRR_K_d_s_d_s(
615 PRIM_INT__d_s_d_s,
616 PRIM_INT__d_s_p_s,
617 PRIM_INT__d_s_s_s,
618 PRIM_INT__p_s_p_s,
619 Q_PA,
620 a_over_q,
621 aoq_PQ,
622 one_over_2pq,
623 one_over_2q,
624 6);
625
626
627 ostei_general_vrr_K(3, 0, 3, 0, 5,
628 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
629 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);
630
631
632 ostei_general_vrr_K(4, 0, 4, 0, 4,
633 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
634 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);
635
636
637 ostei_general_vrr_K(5, 0, 5, 0, 3,
638 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
639 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);
640
641
642 ostei_general_vrr_K(6, 0, 6, 0, 2,
643 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
644 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);
645
646
647 ostei_general_vrr_K(7, 0, 7, 0, 1,
648 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
649 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);
650
651
652
653
654 ////////////////////////////////////
655 // Accumulate contracted integrals
656 ////////////////////////////////////
657 if(lastoffset == 0)
658 {
659 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
660 contract_all(1008, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
661 contract_all(1296, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
662 }
663 else
664 {
665 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
666 contract(1008, shelloffsets, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
667 contract(1296, shelloffsets, PRIM_INT__k_s_k_s, PRIM_PTR_INT__k_s_k_s);
668 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
669 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
670 PRIM_PTR_INT__k_s_k_s += lastoffset*1296;
671 }
672
673 } // close loop over j
674 } // close loop over i
675
676 //Advance to the next batch
677 jstart = SIMINT_SIMD_ROUND(jend);
678
679 //////////////////////////////////////////////
680 // Contracted integrals: Horizontal recurrance
681 //////////////////////////////////////////////
682
683
684
685
686 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
687 {
688 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
689
690 // set up HRR pointers
691 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
692 double const * restrict HRR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
693 double const * restrict HRR_INT__k_s_k_s = INT__k_s_k_s + abcd * 1296;
694 double * restrict HRR_INT__k_s_h_d = INT__k_s_h_d + real_abcd * 4536;
695
696 // form INT__k_s_h_p
697 ostei_general_hrr_L(7, 0, 5, 1, hCD, HRR_INT__k_s_i_s, HRR_INT__k_s_h_s, HRR_INT__k_s_h_p);
698
699 // form INT__k_s_i_p
700 ostei_general_hrr_L(7, 0, 6, 1, hCD, HRR_INT__k_s_k_s, HRR_INT__k_s_i_s, HRR_INT__k_s_i_p);
701
702 // form INT__k_s_h_d
703 ostei_general_hrr_L(7, 0, 5, 2, hCD, HRR_INT__k_s_i_p, HRR_INT__k_s_h_p, HRR_INT__k_s_h_d);
704
705
706 } // close HRR loop
707
708
709 } // close loop cdbatch
710
711 istart = iend;
712 } // close loop over ab
713
714 return P.nshell12_clip * Q.nshell12_clip;
715 }
716
ostei_s_k_h_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__s_k_h_d)717 int ostei_s_k_h_d(struct simint_multi_shellpair const P,
718 struct simint_multi_shellpair const Q,
719 double screen_tol,
720 double * const restrict work,
721 double * const restrict INT__s_k_h_d)
722 {
723 double P_AB[3*P.nshell12];
724 struct simint_multi_shellpair P_tmp = P;
725 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
726 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
727 P_tmp.AB_x = P_AB;
728 P_tmp.AB_y = P_AB + P.nshell12;
729 P_tmp.AB_z = P_AB + 2*P.nshell12;
730
731 for(int i = 0; i < P.nshell12; i++)
732 {
733 P_tmp.AB_x[i] = -P.AB_x[i];
734 P_tmp.AB_y[i] = -P.AB_y[i];
735 P_tmp.AB_z[i] = -P.AB_z[i];
736 }
737
738 int ret = ostei_k_s_h_d(P_tmp, Q, screen_tol, work, INT__s_k_h_d);
739 double buffer[4536] SIMINT_ALIGN_ARRAY_DBL;
740
741 for(int q = 0; q < ret; q++)
742 {
743 int idx = 0;
744 for(int a = 0; a < 1; ++a)
745 for(int b = 0; b < 36; ++b)
746 for(int c = 0; c < 21; ++c)
747 for(int d = 0; d < 6; ++d)
748 buffer[idx++] = INT__s_k_h_d[q*4536+b*126+a*126+c*6+d];
749
750 memcpy(INT__s_k_h_d+q*4536, buffer, 4536*sizeof(double));
751 }
752
753 return ret;
754 }
755
ostei_k_s_d_h(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__k_s_d_h)756 int ostei_k_s_d_h(struct simint_multi_shellpair const P,
757 struct simint_multi_shellpair const Q,
758 double screen_tol,
759 double * const restrict work,
760 double * const restrict INT__k_s_d_h)
761 {
762 double Q_AB[3*Q.nshell12];
763 struct simint_multi_shellpair Q_tmp = Q;
764 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
765 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
766 Q_tmp.AB_x = Q_AB;
767 Q_tmp.AB_y = Q_AB + Q.nshell12;
768 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
769
770 for(int i = 0; i < Q.nshell12; i++)
771 {
772 Q_tmp.AB_x[i] = -Q.AB_x[i];
773 Q_tmp.AB_y[i] = -Q.AB_y[i];
774 Q_tmp.AB_z[i] = -Q.AB_z[i];
775 }
776
777 int ret = ostei_k_s_h_d(P, Q_tmp, screen_tol, work, INT__k_s_d_h);
778 double buffer[4536] SIMINT_ALIGN_ARRAY_DBL;
779
780 for(int q = 0; q < ret; q++)
781 {
782 int idx = 0;
783 for(int a = 0; a < 36; ++a)
784 for(int b = 0; b < 1; ++b)
785 for(int c = 0; c < 6; ++c)
786 for(int d = 0; d < 21; ++d)
787 buffer[idx++] = INT__k_s_d_h[q*4536+a*126+b*126+d*6+c];
788
789 memcpy(INT__k_s_d_h+q*4536, buffer, 4536*sizeof(double));
790 }
791
792 return ret;
793 }
794
ostei_s_k_d_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_k_d_h)795 int ostei_s_k_d_h(struct simint_multi_shellpair const P,
796 struct simint_multi_shellpair const Q,
797 double screen_tol,
798 double * const restrict work,
799 double * const restrict INT__s_k_d_h)
800 {
801 double P_AB[3*P.nshell12];
802 struct simint_multi_shellpair P_tmp = P;
803 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
804 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
805 P_tmp.AB_x = P_AB;
806 P_tmp.AB_y = P_AB + P.nshell12;
807 P_tmp.AB_z = P_AB + 2*P.nshell12;
808
809 for(int i = 0; i < P.nshell12; i++)
810 {
811 P_tmp.AB_x[i] = -P.AB_x[i];
812 P_tmp.AB_y[i] = -P.AB_y[i];
813 P_tmp.AB_z[i] = -P.AB_z[i];
814 }
815
816 double Q_AB[3*Q.nshell12];
817 struct simint_multi_shellpair Q_tmp = Q;
818 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
819 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
820 Q_tmp.AB_x = Q_AB;
821 Q_tmp.AB_y = Q_AB + Q.nshell12;
822 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
823
824 for(int i = 0; i < Q.nshell12; i++)
825 {
826 Q_tmp.AB_x[i] = -Q.AB_x[i];
827 Q_tmp.AB_y[i] = -Q.AB_y[i];
828 Q_tmp.AB_z[i] = -Q.AB_z[i];
829 }
830
831 int ret = ostei_k_s_h_d(P_tmp, Q_tmp, screen_tol, work, INT__s_k_d_h);
832 double buffer[4536] SIMINT_ALIGN_ARRAY_DBL;
833
834 for(int q = 0; q < ret; q++)
835 {
836 int idx = 0;
837 for(int a = 0; a < 1; ++a)
838 for(int b = 0; b < 36; ++b)
839 for(int c = 0; c < 6; ++c)
840 for(int d = 0; d < 21; ++d)
841 buffer[idx++] = INT__s_k_d_h[q*4536+b*126+a*126+d*6+c];
842
843 memcpy(INT__s_k_d_h+q*4536, buffer, 4536*sizeof(double));
844 }
845
846 return ret;
847 }
848
849