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_d_g_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_d_g_d)8 int ostei_k_d_g_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_d_g_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__k_d_g_d);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__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__l_s_g_s = work + (SIMINT_NSHELL_SIMD * 2304);
33 double * const INT__l_s_h_s = work + (SIMINT_NSHELL_SIMD * 2979);
34 double * const INT__l_s_i_s = work + (SIMINT_NSHELL_SIMD * 3924);
35 double * const INT__m_s_g_s = work + (SIMINT_NSHELL_SIMD * 5184);
36 double * const INT__m_s_h_s = work + (SIMINT_NSHELL_SIMD * 6009);
37 double * const INT__m_s_i_s = work + (SIMINT_NSHELL_SIMD * 7164);
38 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*8704);
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 16;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 61;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 145;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 253;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 383;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 563;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 863;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 1043;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1313;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1763;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 2363;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2594;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 2972;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 3602;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 4442;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 5387;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 5667;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 6171;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 7011;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 8131;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 9391;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 10567;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 10891;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 11539;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 12619;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 14059;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 15679;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_i_s = primwork + 17191;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 18199;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 18559;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 19369;
71 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 20719;
72 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_g_s = primwork + 22519;
73 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_h_s = primwork + 24544;
74 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_i_s = primwork + 26434;
75 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_s_s = primwork + 27694;
76 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_p_s = primwork + 28079;
77 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_d_s = primwork + 29069;
78 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_f_s = primwork + 30719;
79 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_g_s = primwork + 32919;
80 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_h_s = primwork + 35394;
81 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_i_s = primwork + 37704;
82 double * const hrrwork = (double *)(primwork + 39244);
83 double * const HRR_INT__k_p_g_s = hrrwork + 0;
84 double * const HRR_INT__k_p_h_s = hrrwork + 1620;
85 double * const HRR_INT__k_p_i_s = hrrwork + 3888;
86 double * const HRR_INT__k_d_g_s = hrrwork + 6912;
87 double * const HRR_INT__k_d_g_p = hrrwork + 10152;
88 double * const HRR_INT__k_d_h_s = hrrwork + 19872;
89 double * const HRR_INT__k_d_h_p = hrrwork + 24408;
90 double * const HRR_INT__k_d_i_s = hrrwork + 38016;
91 double * const HRR_INT__l_p_g_s = hrrwork + 44064;
92 double * const HRR_INT__l_p_h_s = hrrwork + 46089;
93 double * const HRR_INT__l_p_i_s = hrrwork + 48924;
94
95
96 // Create constants
97 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
98 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
99 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
100 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
101 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
102 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
103 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
104 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
105 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
106 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
107
108
109 ////////////////////////////////////////
110 // Loop over shells and primitives
111 ////////////////////////////////////////
112
113 real_abcd = 0;
114 istart = 0;
115 for(ab = 0; ab < P.nshell12_clip; ++ab)
116 {
117 const int iend = istart + P.nprim12[ab];
118
119 cd = 0;
120 jstart = 0;
121
122 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
123 {
124 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
125 int jend = jstart;
126 for(i = 0; i < nshellbatch; i++)
127 jend += Q.nprim12[cd+i];
128
129 // Clear the beginning of the workspace (where we are accumulating integrals)
130 memset(work, 0, SIMINT_NSHELL_SIMD * 8704 * sizeof(double));
131 abcd = 0;
132
133
134 for(i = istart; i < iend; ++i)
135 {
136 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
137
138 if(check_screen)
139 {
140 // Skip this whole thing if always insignificant
141 if((P.screen[i] * Q.screen_max) < screen_tol)
142 continue;
143 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
144 }
145
146 icd = 0;
147 iprimcd = 0;
148 nprim_icd = Q.nprim12[cd];
149 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
150 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
151 double * restrict PRIM_PTR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
152 double * restrict PRIM_PTR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
153 double * restrict PRIM_PTR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
154 double * restrict PRIM_PTR_INT__l_s_i_s = INT__l_s_i_s + abcd * 1260;
155 double * restrict PRIM_PTR_INT__m_s_g_s = INT__m_s_g_s + abcd * 825;
156 double * restrict PRIM_PTR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
157 double * restrict PRIM_PTR_INT__m_s_i_s = INT__m_s_i_s + abcd * 1540;
158
159
160
161 // Load these one per loop over i
162 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
163 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
164 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
165
166 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
167
168 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
169 {
170 // calculate the shell offsets
171 // these are the offset from the shell pointed to by cd
172 // for each element
173 int shelloffsets[SIMINT_SIMD_LEN] = {0};
174 int lastoffset = 0;
175 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
176
177 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
178 {
179 // Handle if the first element of the vector is a new shell
180 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
181 {
182 nprim_icd += Q.nprim12[cd + (++icd)];
183 PRIM_PTR_INT__k_s_g_s += 540;
184 PRIM_PTR_INT__k_s_h_s += 756;
185 PRIM_PTR_INT__k_s_i_s += 1008;
186 PRIM_PTR_INT__l_s_g_s += 675;
187 PRIM_PTR_INT__l_s_h_s += 945;
188 PRIM_PTR_INT__l_s_i_s += 1260;
189 PRIM_PTR_INT__m_s_g_s += 825;
190 PRIM_PTR_INT__m_s_h_s += 1155;
191 PRIM_PTR_INT__m_s_i_s += 1540;
192 }
193 iprimcd++;
194 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
195 {
196 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
197 {
198 shelloffsets[n] = shelloffsets[n-1] + 1;
199 lastoffset++;
200 nprim_icd += Q.nprim12[cd + (++icd)];
201 }
202 else
203 shelloffsets[n] = shelloffsets[n-1];
204 iprimcd++;
205 }
206 }
207 else
208 iprimcd += SIMINT_SIMD_LEN;
209
210 // Do we have to compute this vector (or has it been screened out)?
211 // (not_screened != 0 means we have to do this vector)
212 if(check_screen)
213 {
214 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
215 if(vmax < screen_tol)
216 {
217 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
218 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
219 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
220 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
221 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
222 PRIM_PTR_INT__l_s_i_s += lastoffset*1260;
223 PRIM_PTR_INT__m_s_g_s += lastoffset*825;
224 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
225 PRIM_PTR_INT__m_s_i_s += lastoffset*1540;
226 continue;
227 }
228 }
229
230 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
231 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
232 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
233 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
234
235
236 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
237 SIMINT_DBLTYPE PQ[3];
238 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
239 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
240 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
241 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
242 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
243 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
244
245 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
246 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
247 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
248 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
249 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
250 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
251 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
252
253 // NOTE: Minus sign!
254 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
255 SIMINT_DBLTYPE aop_PQ[3];
256 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
257 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
258 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
259
260 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
261 SIMINT_DBLTYPE aoq_PQ[3];
262 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
263 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
264 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
265 // Put a minus sign here so we don't have to in RR routines
266 a_over_q = SIMINT_NEG(a_over_q);
267
268
269 //////////////////////////////////////////////
270 // Fjt function section
271 // Maximum v value: 15
272 //////////////////////////////////////////////
273 // The parameter to the Fjt function
274 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
275
276
277 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
278
279
280 boys_F_split(PRIM_INT__s_s_s_s, F_x, 15);
281 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
282 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
283 for(n = 0; n <= 15; n++)
284 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
285
286 //////////////////////////////////////////////
287 // Primitive integrals: Vertical recurrance
288 //////////////////////////////////////////////
289
290 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
291 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
292 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
293 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
294 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
295 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
296 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
297 const SIMINT_DBLTYPE vrr_const_8_over_2p = SIMINT_MUL(const_8, one_over_2p);
298 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
299 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
300 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
301 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
302 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
303 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
304 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
305 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
306 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
307 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
308 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
309 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
310 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
311 const SIMINT_DBLTYPE vrr_const_9_over_2pq = SIMINT_MUL(const_9, one_over_2pq);
312
313
314
315 // Forming PRIM_INT__p_s_s_s[15 * 3];
316 for(n = 0; n < 15; ++n) // loop over orders of auxiliary function
317 {
318
319 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
320 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]);
321
322 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
323 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]);
324
325 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
326 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]);
327
328 }
329
330
331
332 // Forming PRIM_INT__d_s_s_s[14 * 6];
333 for(n = 0; n < 14; ++n) // loop over orders of auxiliary function
334 {
335
336 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
337 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]);
338 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]);
339
340 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
341 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]);
342
343 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
344 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]);
345
346 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
347 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]);
348 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]);
349
350 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
351 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]);
352
353 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
354 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]);
355 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]);
356
357 }
358
359
360
361 // Forming PRIM_INT__f_s_s_s[13 * 10];
362 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
363 {
364
365 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
366 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]);
367 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]);
368
369 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
370 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]);
371
372 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
373 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]);
374
375 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
376 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]);
377
378 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
379 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]);
380
381 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
382 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]);
383
384 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
385 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]);
386 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]);
387
388 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
389 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]);
390
391 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
392 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]);
393
394 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
395 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]);
396 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]);
397
398 }
399
400
401 VRR_I_g_s_s_s(
402 PRIM_INT__g_s_s_s,
403 PRIM_INT__f_s_s_s,
404 PRIM_INT__d_s_s_s,
405 P_PA,
406 a_over_p,
407 aop_PQ,
408 one_over_2p,
409 12);
410
411
412 VRR_I_h_s_s_s(
413 PRIM_INT__h_s_s_s,
414 PRIM_INT__g_s_s_s,
415 PRIM_INT__f_s_s_s,
416 P_PA,
417 a_over_p,
418 aop_PQ,
419 one_over_2p,
420 11);
421
422
423 ostei_general_vrr1_I(6, 10,
424 one_over_2p, a_over_p, aop_PQ, P_PA,
425 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
426
427
428 ostei_general_vrr1_I(7, 9,
429 one_over_2p, a_over_p, aop_PQ, P_PA,
430 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
431
432
433 ostei_general_vrr_K(7, 0, 1, 0, 6,
434 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
435 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
436
437
438 ostei_general_vrr_K(6, 0, 1, 0, 6,
439 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
440 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
441
442
443 ostei_general_vrr_K(7, 0, 2, 0, 5,
444 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
445 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);
446
447
448 ostei_general_vrr_K(5, 0, 1, 0, 6,
449 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
450 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
451
452
453 ostei_general_vrr_K(6, 0, 2, 0, 5,
454 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
455 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);
456
457
458 ostei_general_vrr_K(7, 0, 3, 0, 4,
459 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
460 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);
461
462
463 VRR_K_g_s_p_s(
464 PRIM_INT__g_s_p_s,
465 PRIM_INT__g_s_s_s,
466 PRIM_INT__f_s_s_s,
467 Q_PA,
468 aoq_PQ,
469 one_over_2pq,
470 6);
471
472
473 ostei_general_vrr_K(5, 0, 2, 0, 5,
474 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
475 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);
476
477
478 ostei_general_vrr_K(6, 0, 3, 0, 4,
479 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
480 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);
481
482
483 ostei_general_vrr_K(7, 0, 4, 0, 3,
484 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
485 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);
486
487
488 VRR_K_f_s_p_s(
489 PRIM_INT__f_s_p_s,
490 PRIM_INT__f_s_s_s,
491 PRIM_INT__d_s_s_s,
492 Q_PA,
493 aoq_PQ,
494 one_over_2pq,
495 6);
496
497
498 ostei_general_vrr_K(4, 0, 2, 0, 5,
499 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
500 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);
501
502
503 ostei_general_vrr_K(5, 0, 3, 0, 4,
504 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
505 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);
506
507
508 ostei_general_vrr_K(6, 0, 4, 0, 3,
509 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
510 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);
511
512
513 ostei_general_vrr_K(7, 0, 5, 0, 2,
514 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
515 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);
516
517
518 ostei_general_vrr1_I(8, 8,
519 one_over_2p, a_over_p, aop_PQ, P_PA,
520 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
521
522
523 ostei_general_vrr_K(8, 0, 1, 0, 6,
524 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
525 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
526
527
528 ostei_general_vrr_K(8, 0, 2, 0, 5,
529 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
530 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);
531
532
533 ostei_general_vrr_K(8, 0, 3, 0, 4,
534 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
535 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);
536
537
538 ostei_general_vrr_K(8, 0, 4, 0, 3,
539 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
540 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);
541
542
543
544 // Forming PRIM_INT__d_s_p_s[6 * 18];
545 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
546 {
547
548 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
549 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]);
550 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]);
551
552 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
553 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]);
554
555 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
556 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]);
557
558 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
559 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]);
560 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]);
561
562 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
563 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]);
564 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]);
565
566 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
567 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]);
568
569 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
570 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]);
571 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]);
572
573 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
574 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]);
575
576 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
577 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]);
578 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]);
579
580 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
581 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]);
582
583 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
584 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]);
585 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]);
586
587 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
588 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]);
589
590 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
591 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]);
592
593 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
594 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]);
595 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]);
596
597 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
598 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]);
599 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]);
600
601 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
602 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]);
603
604 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
605 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]);
606
607 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
608 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]);
609 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]);
610
611 }
612
613
614 VRR_K_f_s_d_s(
615 PRIM_INT__f_s_d_s,
616 PRIM_INT__f_s_p_s,
617 PRIM_INT__f_s_s_s,
618 PRIM_INT__d_s_p_s,
619 Q_PA,
620 a_over_q,
621 aoq_PQ,
622 one_over_2pq,
623 one_over_2q,
624 5);
625
626
627 ostei_general_vrr_K(4, 0, 3, 0, 4,
628 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
629 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);
630
631
632 ostei_general_vrr_K(5, 0, 4, 0, 3,
633 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
634 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);
635
636
637 ostei_general_vrr_K(6, 0, 5, 0, 2,
638 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
639 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);
640
641
642 ostei_general_vrr_K(7, 0, 6, 0, 1,
643 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
644 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);
645
646
647 ostei_general_vrr_K(8, 0, 5, 0, 2,
648 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
649 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);
650
651
652 ostei_general_vrr1_I(9, 7,
653 one_over_2p, a_over_p, aop_PQ, P_PA,
654 PRIM_INT__l_s_s_s, PRIM_INT__k_s_s_s, PRIM_INT__m_s_s_s);
655
656
657 ostei_general_vrr_K(9, 0, 1, 0, 6,
658 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
659 PRIM_INT__m_s_s_s, NULL, NULL, PRIM_INT__l_s_s_s, NULL, PRIM_INT__m_s_p_s);
660
661
662 ostei_general_vrr_K(9, 0, 2, 0, 5,
663 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
664 PRIM_INT__m_s_p_s, PRIM_INT__m_s_s_s, NULL, PRIM_INT__l_s_p_s, NULL, PRIM_INT__m_s_d_s);
665
666
667 ostei_general_vrr_K(9, 0, 3, 0, 4,
668 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
669 PRIM_INT__m_s_d_s, PRIM_INT__m_s_p_s, NULL, PRIM_INT__l_s_d_s, NULL, PRIM_INT__m_s_f_s);
670
671
672 ostei_general_vrr_K(9, 0, 4, 0, 3,
673 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
674 PRIM_INT__m_s_f_s, PRIM_INT__m_s_d_s, NULL, PRIM_INT__l_s_f_s, NULL, PRIM_INT__m_s_g_s);
675
676
677 ostei_general_vrr_K(8, 0, 6, 0, 1,
678 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
679 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);
680
681
682 ostei_general_vrr_K(9, 0, 5, 0, 2,
683 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
684 PRIM_INT__m_s_g_s, PRIM_INT__m_s_f_s, NULL, PRIM_INT__l_s_g_s, NULL, PRIM_INT__m_s_h_s);
685
686
687 ostei_general_vrr_K(9, 0, 6, 0, 1,
688 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
689 PRIM_INT__m_s_h_s, PRIM_INT__m_s_g_s, NULL, PRIM_INT__l_s_h_s, NULL, PRIM_INT__m_s_i_s);
690
691
692
693
694 ////////////////////////////////////
695 // Accumulate contracted integrals
696 ////////////////////////////////////
697 if(lastoffset == 0)
698 {
699 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
700 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
701 contract_all(1008, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
702 contract_all(675, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
703 contract_all(945, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
704 contract_all(1260, PRIM_INT__l_s_i_s, PRIM_PTR_INT__l_s_i_s);
705 contract_all(825, PRIM_INT__m_s_g_s, PRIM_PTR_INT__m_s_g_s);
706 contract_all(1155, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
707 contract_all(1540, PRIM_INT__m_s_i_s, PRIM_PTR_INT__m_s_i_s);
708 }
709 else
710 {
711 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
712 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
713 contract(1008, shelloffsets, PRIM_INT__k_s_i_s, PRIM_PTR_INT__k_s_i_s);
714 contract(675, shelloffsets, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
715 contract(945, shelloffsets, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
716 contract(1260, shelloffsets, PRIM_INT__l_s_i_s, PRIM_PTR_INT__l_s_i_s);
717 contract(825, shelloffsets, PRIM_INT__m_s_g_s, PRIM_PTR_INT__m_s_g_s);
718 contract(1155, shelloffsets, PRIM_INT__m_s_h_s, PRIM_PTR_INT__m_s_h_s);
719 contract(1540, shelloffsets, PRIM_INT__m_s_i_s, PRIM_PTR_INT__m_s_i_s);
720 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
721 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
722 PRIM_PTR_INT__k_s_i_s += lastoffset*1008;
723 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
724 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
725 PRIM_PTR_INT__l_s_i_s += lastoffset*1260;
726 PRIM_PTR_INT__m_s_g_s += lastoffset*825;
727 PRIM_PTR_INT__m_s_h_s += lastoffset*1155;
728 PRIM_PTR_INT__m_s_i_s += lastoffset*1540;
729 }
730
731 } // close loop over j
732 } // close loop over i
733
734 //Advance to the next batch
735 jstart = SIMINT_SIMD_ROUND(jend);
736
737 //////////////////////////////////////////////
738 // Contracted integrals: Horizontal recurrance
739 //////////////////////////////////////////////
740
741
742 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
743
744
745 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
746 {
747 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
748
749 // set up HRR pointers
750 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
751 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
752 double const * restrict HRR_INT__k_s_i_s = INT__k_s_i_s + abcd * 1008;
753 double const * restrict HRR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
754 double const * restrict HRR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
755 double const * restrict HRR_INT__l_s_i_s = INT__l_s_i_s + abcd * 1260;
756 double const * restrict HRR_INT__m_s_g_s = INT__m_s_g_s + abcd * 825;
757 double const * restrict HRR_INT__m_s_h_s = INT__m_s_h_s + abcd * 1155;
758 double const * restrict HRR_INT__m_s_i_s = INT__m_s_i_s + abcd * 1540;
759 double * restrict HRR_INT__k_d_g_d = INT__k_d_g_d + real_abcd * 19440;
760
761 // form INT__k_p_g_s
762 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);
763
764 // form INT__k_p_h_s
765 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);
766
767 // form INT__k_p_i_s
768 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);
769
770 // form INT__l_p_g_s
771 ostei_general_hrr_J(8, 1, 4, 0, hAB, HRR_INT__m_s_g_s, HRR_INT__l_s_g_s, HRR_INT__l_p_g_s);
772
773 // form INT__l_p_h_s
774 ostei_general_hrr_J(8, 1, 5, 0, hAB, HRR_INT__m_s_h_s, HRR_INT__l_s_h_s, HRR_INT__l_p_h_s);
775
776 // form INT__l_p_i_s
777 ostei_general_hrr_J(8, 1, 6, 0, hAB, HRR_INT__m_s_i_s, HRR_INT__l_s_i_s, HRR_INT__l_p_i_s);
778
779 // form INT__k_d_g_s
780 ostei_general_hrr_J(7, 2, 4, 0, hAB, HRR_INT__l_p_g_s, HRR_INT__k_p_g_s, HRR_INT__k_d_g_s);
781
782 // form INT__k_d_h_s
783 ostei_general_hrr_J(7, 2, 5, 0, hAB, HRR_INT__l_p_h_s, HRR_INT__k_p_h_s, HRR_INT__k_d_h_s);
784
785 // form INT__k_d_i_s
786 ostei_general_hrr_J(7, 2, 6, 0, hAB, HRR_INT__l_p_i_s, HRR_INT__k_p_i_s, HRR_INT__k_d_i_s);
787
788 // form INT__k_d_g_p
789 HRR_L_g_p(
790 HRR_INT__k_d_g_p,
791 HRR_INT__k_d_g_s,
792 HRR_INT__k_d_h_s,
793 hCD, 216);
794
795 // form INT__k_d_h_p
796 ostei_general_hrr_L(7, 2, 5, 1, hCD, HRR_INT__k_d_i_s, HRR_INT__k_d_h_s, HRR_INT__k_d_h_p);
797
798 // form INT__k_d_g_d
799 ostei_general_hrr_L(7, 2, 4, 2, hCD, HRR_INT__k_d_h_p, HRR_INT__k_d_g_p, HRR_INT__k_d_g_d);
800
801
802 } // close HRR loop
803
804
805 } // close loop cdbatch
806
807 istart = iend;
808 } // close loop over ab
809
810 return P.nshell12_clip * Q.nshell12_clip;
811 }
812
ostei_d_k_g_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_k_g_d)813 int ostei_d_k_g_d(struct simint_multi_shellpair const P,
814 struct simint_multi_shellpair const Q,
815 double screen_tol,
816 double * const restrict work,
817 double * const restrict INT__d_k_g_d)
818 {
819 double P_AB[3*P.nshell12];
820 struct simint_multi_shellpair P_tmp = P;
821 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
822 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
823 P_tmp.AB_x = P_AB;
824 P_tmp.AB_y = P_AB + P.nshell12;
825 P_tmp.AB_z = P_AB + 2*P.nshell12;
826
827 for(int i = 0; i < P.nshell12; i++)
828 {
829 P_tmp.AB_x[i] = -P.AB_x[i];
830 P_tmp.AB_y[i] = -P.AB_y[i];
831 P_tmp.AB_z[i] = -P.AB_z[i];
832 }
833
834 int ret = ostei_k_d_g_d(P_tmp, Q, screen_tol, work, INT__d_k_g_d);
835 double buffer[19440] SIMINT_ALIGN_ARRAY_DBL;
836
837 for(int q = 0; q < ret; q++)
838 {
839 int idx = 0;
840 for(int a = 0; a < 6; ++a)
841 for(int b = 0; b < 36; ++b)
842 for(int c = 0; c < 15; ++c)
843 for(int d = 0; d < 6; ++d)
844 buffer[idx++] = INT__d_k_g_d[q*19440+b*540+a*90+c*6+d];
845
846 memcpy(INT__d_k_g_d+q*19440, buffer, 19440*sizeof(double));
847 }
848
849 return ret;
850 }
851
ostei_k_d_d_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_d_d_g)852 int ostei_k_d_d_g(struct simint_multi_shellpair const P,
853 struct simint_multi_shellpair const Q,
854 double screen_tol,
855 double * const restrict work,
856 double * const restrict INT__k_d_d_g)
857 {
858 double Q_AB[3*Q.nshell12];
859 struct simint_multi_shellpair Q_tmp = Q;
860 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
861 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
862 Q_tmp.AB_x = Q_AB;
863 Q_tmp.AB_y = Q_AB + Q.nshell12;
864 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
865
866 for(int i = 0; i < Q.nshell12; i++)
867 {
868 Q_tmp.AB_x[i] = -Q.AB_x[i];
869 Q_tmp.AB_y[i] = -Q.AB_y[i];
870 Q_tmp.AB_z[i] = -Q.AB_z[i];
871 }
872
873 int ret = ostei_k_d_g_d(P, Q_tmp, screen_tol, work, INT__k_d_d_g);
874 double buffer[19440] SIMINT_ALIGN_ARRAY_DBL;
875
876 for(int q = 0; q < ret; q++)
877 {
878 int idx = 0;
879 for(int a = 0; a < 36; ++a)
880 for(int b = 0; b < 6; ++b)
881 for(int c = 0; c < 6; ++c)
882 for(int d = 0; d < 15; ++d)
883 buffer[idx++] = INT__k_d_d_g[q*19440+a*540+b*90+d*6+c];
884
885 memcpy(INT__k_d_d_g+q*19440, buffer, 19440*sizeof(double));
886 }
887
888 return ret;
889 }
890
ostei_d_k_d_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_k_d_g)891 int ostei_d_k_d_g(struct simint_multi_shellpair const P,
892 struct simint_multi_shellpair const Q,
893 double screen_tol,
894 double * const restrict work,
895 double * const restrict INT__d_k_d_g)
896 {
897 double P_AB[3*P.nshell12];
898 struct simint_multi_shellpair P_tmp = P;
899 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
900 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
901 P_tmp.AB_x = P_AB;
902 P_tmp.AB_y = P_AB + P.nshell12;
903 P_tmp.AB_z = P_AB + 2*P.nshell12;
904
905 for(int i = 0; i < P.nshell12; i++)
906 {
907 P_tmp.AB_x[i] = -P.AB_x[i];
908 P_tmp.AB_y[i] = -P.AB_y[i];
909 P_tmp.AB_z[i] = -P.AB_z[i];
910 }
911
912 double Q_AB[3*Q.nshell12];
913 struct simint_multi_shellpair Q_tmp = Q;
914 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
915 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
916 Q_tmp.AB_x = Q_AB;
917 Q_tmp.AB_y = Q_AB + Q.nshell12;
918 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
919
920 for(int i = 0; i < Q.nshell12; i++)
921 {
922 Q_tmp.AB_x[i] = -Q.AB_x[i];
923 Q_tmp.AB_y[i] = -Q.AB_y[i];
924 Q_tmp.AB_z[i] = -Q.AB_z[i];
925 }
926
927 int ret = ostei_k_d_g_d(P_tmp, Q_tmp, screen_tol, work, INT__d_k_d_g);
928 double buffer[19440] SIMINT_ALIGN_ARRAY_DBL;
929
930 for(int q = 0; q < ret; q++)
931 {
932 int idx = 0;
933 for(int a = 0; a < 6; ++a)
934 for(int b = 0; b < 36; ++b)
935 for(int c = 0; c < 6; ++c)
936 for(int d = 0; d < 15; ++d)
937 buffer[idx++] = INT__d_k_d_g[q*19440+b*540+a*90+d*6+c];
938
939 memcpy(INT__d_k_d_g+q*19440, buffer, 19440*sizeof(double));
940 }
941
942 return ret;
943 }
944
945