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_i_d_g_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_d_g_p)8 int ostei_i_d_g_p(struct simint_multi_shellpair const P,
9 struct simint_multi_shellpair const Q,
10 double screen_tol,
11 double * const restrict work,
12 double * const restrict INT__i_d_g_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_d_g_p);
17 int ab, cd, abcd;
18 int istart, jstart;
19 int iprimcd, nprim_icd, icd;
20 const int check_screen = (screen_tol > 0.0);
21 int i, j;
22 int n;
23 int not_screened;
24 int real_abcd;
25 int iket;
26 int ibra;
27
28 // partition workspace
29 double * const INT__i_s_g_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__i_s_h_s = work + (SIMINT_NSHELL_SIMD * 420);
31 double * const INT__k_s_g_s = work + (SIMINT_NSHELL_SIMD * 1008);
32 double * const INT__k_s_h_s = work + (SIMINT_NSHELL_SIMD * 1548);
33 double * const INT__l_s_g_s = work + (SIMINT_NSHELL_SIMD * 2304);
34 double * const INT__l_s_h_s = work + (SIMINT_NSHELL_SIMD * 2979);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*3924);
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_s_s = primwork + 14;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 53;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_p_s = primwork + 125;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 215;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_p_s = primwork + 325;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_d_s = primwork + 475;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 715;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_p_s = primwork + 865;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_d_s = primwork + 1090;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_f_s = primwork + 1450;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 1900;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_p_s = primwork + 2089;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_d_s = primwork + 2404;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_f_s = primwork + 2908;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_g_s = primwork + 3538;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 4168;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_p_s = primwork + 4392;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_d_s = primwork + 4812;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_f_s = primwork + 5484;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_g_s = primwork + 6324;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_h_s = primwork + 7164;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 7752;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_p_s = primwork + 8004;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_d_s = primwork + 8544;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_f_s = primwork + 9408;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_g_s = primwork + 10488;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_h_s = primwork + 11568;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 12324;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_p_s = primwork + 12594;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_d_s = primwork + 13269;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_f_s = primwork + 14349;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_g_s = primwork + 15699;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_h_s = primwork + 17049;
70 double * const hrrwork = (double *)(primwork + 17994);
71 double * const HRR_INT__i_p_g_s = hrrwork + 0;
72 double * const HRR_INT__i_p_h_s = hrrwork + 1260;
73 double * const HRR_INT__i_d_g_s = hrrwork + 3024;
74 double * const HRR_INT__i_d_h_s = hrrwork + 5544;
75 double * const HRR_INT__k_p_g_s = hrrwork + 9072;
76 double * const HRR_INT__k_p_h_s = hrrwork + 10692;
77
78
79 // Create constants
80 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
81 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
82 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
83 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
84 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
85 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
86 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
87 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
88 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
89
90
91 ////////////////////////////////////////
92 // Loop over shells and primitives
93 ////////////////////////////////////////
94
95 real_abcd = 0;
96 istart = 0;
97 for(ab = 0; ab < P.nshell12_clip; ++ab)
98 {
99 const int iend = istart + P.nprim12[ab];
100
101 cd = 0;
102 jstart = 0;
103
104 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
105 {
106 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
107 int jend = jstart;
108 for(i = 0; i < nshellbatch; i++)
109 jend += Q.nprim12[cd+i];
110
111 // Clear the beginning of the workspace (where we are accumulating integrals)
112 memset(work, 0, SIMINT_NSHELL_SIMD * 3924 * sizeof(double));
113 abcd = 0;
114
115
116 for(i = istart; i < iend; ++i)
117 {
118 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
119
120 if(check_screen)
121 {
122 // Skip this whole thing if always insignificant
123 if((P.screen[i] * Q.screen_max) < screen_tol)
124 continue;
125 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
126 }
127
128 icd = 0;
129 iprimcd = 0;
130 nprim_icd = Q.nprim12[cd];
131 double * restrict PRIM_PTR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
132 double * restrict PRIM_PTR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
133 double * restrict PRIM_PTR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
134 double * restrict PRIM_PTR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
135 double * restrict PRIM_PTR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
136 double * restrict PRIM_PTR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
137
138
139
140 // Load these one per loop over i
141 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
142 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
143 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
144
145 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
146
147 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
148 {
149 // calculate the shell offsets
150 // these are the offset from the shell pointed to by cd
151 // for each element
152 int shelloffsets[SIMINT_SIMD_LEN] = {0};
153 int lastoffset = 0;
154 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
155
156 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
157 {
158 // Handle if the first element of the vector is a new shell
159 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
160 {
161 nprim_icd += Q.nprim12[cd + (++icd)];
162 PRIM_PTR_INT__i_s_g_s += 420;
163 PRIM_PTR_INT__i_s_h_s += 588;
164 PRIM_PTR_INT__k_s_g_s += 540;
165 PRIM_PTR_INT__k_s_h_s += 756;
166 PRIM_PTR_INT__l_s_g_s += 675;
167 PRIM_PTR_INT__l_s_h_s += 945;
168 }
169 iprimcd++;
170 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
171 {
172 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
173 {
174 shelloffsets[n] = shelloffsets[n-1] + 1;
175 lastoffset++;
176 nprim_icd += Q.nprim12[cd + (++icd)];
177 }
178 else
179 shelloffsets[n] = shelloffsets[n-1];
180 iprimcd++;
181 }
182 }
183 else
184 iprimcd += SIMINT_SIMD_LEN;
185
186 // Do we have to compute this vector (or has it been screened out)?
187 // (not_screened != 0 means we have to do this vector)
188 if(check_screen)
189 {
190 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
191 if(vmax < screen_tol)
192 {
193 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
194 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
195 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
196 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
197 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
198 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
199 continue;
200 }
201 }
202
203 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
204 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
205 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
206 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
207
208
209 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
210 SIMINT_DBLTYPE PQ[3];
211 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
212 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
213 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
214 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
215 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
216 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
217
218 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
219 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
220 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
221 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
222 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
223 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
224 const SIMINT_DBLTYPE Q_PA[3] = { SIMINT_DBLLOAD(Q.PA_x, j), SIMINT_DBLLOAD(Q.PA_y, j), SIMINT_DBLLOAD(Q.PA_z, j) };
225
226 // NOTE: Minus sign!
227 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
228 SIMINT_DBLTYPE aop_PQ[3];
229 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
230 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
231 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
232
233 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
234 SIMINT_DBLTYPE aoq_PQ[3];
235 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
236 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
237 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
238 // Put a minus sign here so we don't have to in RR routines
239 a_over_q = SIMINT_NEG(a_over_q);
240
241
242 //////////////////////////////////////////////
243 // Fjt function section
244 // Maximum v value: 13
245 //////////////////////////////////////////////
246 // The parameter to the Fjt function
247 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
248
249
250 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
251
252
253 boys_F_split(PRIM_INT__s_s_s_s, F_x, 13);
254 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
255 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
256 for(n = 0; n <= 13; n++)
257 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
258
259 //////////////////////////////////////////////
260 // Primitive integrals: Vertical recurrance
261 //////////////////////////////////////////////
262
263 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
264 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
265 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
266 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
267 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
268 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
269 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
270 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
271 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
272 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
273 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
274 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
275 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
276 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
277 const SIMINT_DBLTYPE vrr_const_4_over_2pq = SIMINT_MUL(const_4, one_over_2pq);
278 const SIMINT_DBLTYPE vrr_const_5_over_2pq = SIMINT_MUL(const_5, one_over_2pq);
279 const SIMINT_DBLTYPE vrr_const_6_over_2pq = SIMINT_MUL(const_6, one_over_2pq);
280 const SIMINT_DBLTYPE vrr_const_7_over_2pq = SIMINT_MUL(const_7, one_over_2pq);
281 const SIMINT_DBLTYPE vrr_const_8_over_2pq = SIMINT_MUL(const_8, one_over_2pq);
282
283
284
285 // Forming PRIM_INT__p_s_s_s[13 * 3];
286 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
287 {
288
289 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
290 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]);
291
292 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
293 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]);
294
295 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
296 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]);
297
298 }
299
300
301
302 // Forming PRIM_INT__d_s_s_s[12 * 6];
303 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
304 {
305
306 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
307 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]);
308 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]);
309
310 PRIM_INT__d_s_s_s[n * 6 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 0]);
311 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]);
312
313 PRIM_INT__d_s_s_s[n * 6 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 0]);
314 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]);
315
316 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
317 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]);
318 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]);
319
320 PRIM_INT__d_s_s_s[n * 6 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 1]);
321 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]);
322
323 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
324 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]);
325 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]);
326
327 }
328
329
330
331 // Forming PRIM_INT__f_s_s_s[11 * 10];
332 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
333 {
334
335 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
336 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]);
337 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]);
338
339 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
340 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]);
341
342 PRIM_INT__f_s_s_s[n * 10 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
343 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]);
344
345 PRIM_INT__f_s_s_s[n * 10 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
346 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]);
347
348 PRIM_INT__f_s_s_s[n * 10 + 4] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
349 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]);
350
351 PRIM_INT__f_s_s_s[n * 10 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
352 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]);
353
354 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
355 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]);
356 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]);
357
358 PRIM_INT__f_s_s_s[n * 10 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
359 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]);
360
361 PRIM_INT__f_s_s_s[n * 10 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
362 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]);
363
364 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
365 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]);
366 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]);
367
368 }
369
370
371 VRR_I_g_s_s_s(
372 PRIM_INT__g_s_s_s,
373 PRIM_INT__f_s_s_s,
374 PRIM_INT__d_s_s_s,
375 P_PA,
376 a_over_p,
377 aop_PQ,
378 one_over_2p,
379 10);
380
381
382 VRR_I_h_s_s_s(
383 PRIM_INT__h_s_s_s,
384 PRIM_INT__g_s_s_s,
385 PRIM_INT__f_s_s_s,
386 P_PA,
387 a_over_p,
388 aop_PQ,
389 one_over_2p,
390 9);
391
392
393 ostei_general_vrr1_I(6, 8,
394 one_over_2p, a_over_p, aop_PQ, P_PA,
395 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
396
397
398 ostei_general_vrr_K(6, 0, 1, 0, 5,
399 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
400 PRIM_INT__i_s_s_s, NULL, NULL, PRIM_INT__h_s_s_s, NULL, PRIM_INT__i_s_p_s);
401
402
403 ostei_general_vrr_K(5, 0, 1, 0, 5,
404 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
405 PRIM_INT__h_s_s_s, NULL, NULL, PRIM_INT__g_s_s_s, NULL, PRIM_INT__h_s_p_s);
406
407
408 ostei_general_vrr_K(6, 0, 2, 0, 4,
409 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
410 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);
411
412
413 VRR_K_g_s_p_s(
414 PRIM_INT__g_s_p_s,
415 PRIM_INT__g_s_s_s,
416 PRIM_INT__f_s_s_s,
417 Q_PA,
418 aoq_PQ,
419 one_over_2pq,
420 5);
421
422
423 ostei_general_vrr_K(5, 0, 2, 0, 4,
424 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
425 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);
426
427
428 ostei_general_vrr_K(6, 0, 3, 0, 3,
429 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
430 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);
431
432
433 VRR_K_f_s_p_s(
434 PRIM_INT__f_s_p_s,
435 PRIM_INT__f_s_s_s,
436 PRIM_INT__d_s_s_s,
437 Q_PA,
438 aoq_PQ,
439 one_over_2pq,
440 5);
441
442
443 ostei_general_vrr_K(4, 0, 2, 0, 4,
444 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
445 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);
446
447
448 ostei_general_vrr_K(5, 0, 3, 0, 3,
449 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
450 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);
451
452
453 ostei_general_vrr_K(6, 0, 4, 0, 2,
454 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
455 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);
456
457
458
459 // Forming PRIM_INT__d_s_p_s[5 * 18];
460 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
461 {
462
463 PRIM_INT__d_s_p_s[n * 18 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
464 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]);
465 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]);
466
467 PRIM_INT__d_s_p_s[n * 18 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
468 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]);
469
470 PRIM_INT__d_s_p_s[n * 18 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 0]);
471 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]);
472
473 PRIM_INT__d_s_p_s[n * 18 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 1]);
474 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]);
475 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]);
476
477 PRIM_INT__d_s_p_s[n * 18 + 4] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 1]);
478 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]);
479 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]);
480
481 PRIM_INT__d_s_p_s[n * 18 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 1]);
482 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]);
483
484 PRIM_INT__d_s_p_s[n * 18 + 6] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 2]);
485 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]);
486 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]);
487
488 PRIM_INT__d_s_p_s[n * 18 + 7] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 2]);
489 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]);
490
491 PRIM_INT__d_s_p_s[n * 18 + 8] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 2]);
492 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]);
493 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]);
494
495 PRIM_INT__d_s_p_s[n * 18 + 9] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 3]);
496 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]);
497
498 PRIM_INT__d_s_p_s[n * 18 + 10] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
499 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]);
500 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]);
501
502 PRIM_INT__d_s_p_s[n * 18 + 11] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 3]);
503 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]);
504
505 PRIM_INT__d_s_p_s[n * 18 + 12] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 4]);
506 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]);
507
508 PRIM_INT__d_s_p_s[n * 18 + 13] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 4]);
509 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]);
510 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]);
511
512 PRIM_INT__d_s_p_s[n * 18 + 14] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 4]);
513 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]);
514 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]);
515
516 PRIM_INT__d_s_p_s[n * 18 + 15] = SIMINT_MUL(Q_PA[0], PRIM_INT__d_s_s_s[n * 6 + 5]);
517 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]);
518
519 PRIM_INT__d_s_p_s[n * 18 + 16] = SIMINT_MUL(Q_PA[1], PRIM_INT__d_s_s_s[n * 6 + 5]);
520 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]);
521
522 PRIM_INT__d_s_p_s[n * 18 + 17] = SIMINT_MUL(Q_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
523 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]);
524 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]);
525
526 }
527
528
529 VRR_K_f_s_d_s(
530 PRIM_INT__f_s_d_s,
531 PRIM_INT__f_s_p_s,
532 PRIM_INT__f_s_s_s,
533 PRIM_INT__d_s_p_s,
534 Q_PA,
535 a_over_q,
536 aoq_PQ,
537 one_over_2pq,
538 one_over_2q,
539 4);
540
541
542 ostei_general_vrr_K(4, 0, 3, 0, 3,
543 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
544 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);
545
546
547 ostei_general_vrr_K(5, 0, 4, 0, 2,
548 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
549 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);
550
551
552 ostei_general_vrr_K(6, 0, 5, 0, 1,
553 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
554 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);
555
556
557 ostei_general_vrr1_I(7, 7,
558 one_over_2p, a_over_p, aop_PQ, P_PA,
559 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
560
561
562 ostei_general_vrr_K(7, 0, 1, 0, 5,
563 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
564 PRIM_INT__k_s_s_s, NULL, NULL, PRIM_INT__i_s_s_s, NULL, PRIM_INT__k_s_p_s);
565
566
567 ostei_general_vrr_K(7, 0, 2, 0, 4,
568 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
569 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);
570
571
572 ostei_general_vrr_K(7, 0, 3, 0, 3,
573 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
574 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);
575
576
577 ostei_general_vrr_K(7, 0, 4, 0, 2,
578 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
579 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);
580
581
582 ostei_general_vrr_K(7, 0, 5, 0, 1,
583 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
584 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);
585
586
587 ostei_general_vrr1_I(8, 6,
588 one_over_2p, a_over_p, aop_PQ, P_PA,
589 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
590
591
592 ostei_general_vrr_K(8, 0, 1, 0, 5,
593 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
594 PRIM_INT__l_s_s_s, NULL, NULL, PRIM_INT__k_s_s_s, NULL, PRIM_INT__l_s_p_s);
595
596
597 ostei_general_vrr_K(8, 0, 2, 0, 4,
598 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
599 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);
600
601
602 ostei_general_vrr_K(8, 0, 3, 0, 3,
603 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
604 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);
605
606
607 ostei_general_vrr_K(8, 0, 4, 0, 2,
608 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
609 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);
610
611
612 ostei_general_vrr_K(8, 0, 5, 0, 1,
613 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PA,
614 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);
615
616
617
618
619 ////////////////////////////////////
620 // Accumulate contracted integrals
621 ////////////////////////////////////
622 if(lastoffset == 0)
623 {
624 contract_all(420, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
625 contract_all(588, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
626 contract_all(540, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
627 contract_all(756, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
628 contract_all(675, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
629 contract_all(945, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
630 }
631 else
632 {
633 contract(420, shelloffsets, PRIM_INT__i_s_g_s, PRIM_PTR_INT__i_s_g_s);
634 contract(588, shelloffsets, PRIM_INT__i_s_h_s, PRIM_PTR_INT__i_s_h_s);
635 contract(540, shelloffsets, PRIM_INT__k_s_g_s, PRIM_PTR_INT__k_s_g_s);
636 contract(756, shelloffsets, PRIM_INT__k_s_h_s, PRIM_PTR_INT__k_s_h_s);
637 contract(675, shelloffsets, PRIM_INT__l_s_g_s, PRIM_PTR_INT__l_s_g_s);
638 contract(945, shelloffsets, PRIM_INT__l_s_h_s, PRIM_PTR_INT__l_s_h_s);
639 PRIM_PTR_INT__i_s_g_s += lastoffset*420;
640 PRIM_PTR_INT__i_s_h_s += lastoffset*588;
641 PRIM_PTR_INT__k_s_g_s += lastoffset*540;
642 PRIM_PTR_INT__k_s_h_s += lastoffset*756;
643 PRIM_PTR_INT__l_s_g_s += lastoffset*675;
644 PRIM_PTR_INT__l_s_h_s += lastoffset*945;
645 }
646
647 } // close loop over j
648 } // close loop over i
649
650 //Advance to the next batch
651 jstart = SIMINT_SIMD_ROUND(jend);
652
653 //////////////////////////////////////////////
654 // Contracted integrals: Horizontal recurrance
655 //////////////////////////////////////////////
656
657
658 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
659
660
661 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
662 {
663 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
664
665 // set up HRR pointers
666 double const * restrict HRR_INT__i_s_g_s = INT__i_s_g_s + abcd * 420;
667 double const * restrict HRR_INT__i_s_h_s = INT__i_s_h_s + abcd * 588;
668 double const * restrict HRR_INT__k_s_g_s = INT__k_s_g_s + abcd * 540;
669 double const * restrict HRR_INT__k_s_h_s = INT__k_s_h_s + abcd * 756;
670 double const * restrict HRR_INT__l_s_g_s = INT__l_s_g_s + abcd * 675;
671 double const * restrict HRR_INT__l_s_h_s = INT__l_s_h_s + abcd * 945;
672 double * restrict HRR_INT__i_d_g_p = INT__i_d_g_p + real_abcd * 7560;
673
674 // form INT__i_p_g_s
675 ostei_general_hrr_J(6, 1, 4, 0, hAB, HRR_INT__k_s_g_s, HRR_INT__i_s_g_s, HRR_INT__i_p_g_s);
676
677 // form INT__i_p_h_s
678 ostei_general_hrr_J(6, 1, 5, 0, hAB, HRR_INT__k_s_h_s, HRR_INT__i_s_h_s, HRR_INT__i_p_h_s);
679
680 // form INT__k_p_g_s
681 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);
682
683 // form INT__k_p_h_s
684 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);
685
686 // form INT__i_d_g_s
687 ostei_general_hrr_J(6, 2, 4, 0, hAB, HRR_INT__k_p_g_s, HRR_INT__i_p_g_s, HRR_INT__i_d_g_s);
688
689 // form INT__i_d_h_s
690 ostei_general_hrr_J(6, 2, 5, 0, hAB, HRR_INT__k_p_h_s, HRR_INT__i_p_h_s, HRR_INT__i_d_h_s);
691
692 // form INT__i_d_g_p
693 HRR_L_g_p(
694 HRR_INT__i_d_g_p,
695 HRR_INT__i_d_g_s,
696 HRR_INT__i_d_h_s,
697 hCD, 168);
698
699
700 } // close HRR loop
701
702
703 } // close loop cdbatch
704
705 istart = iend;
706 } // close loop over ab
707
708 return P.nshell12_clip * Q.nshell12_clip;
709 }
710
ostei_d_i_g_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_i_g_p)711 int ostei_d_i_g_p(struct simint_multi_shellpair const P,
712 struct simint_multi_shellpair const Q,
713 double screen_tol,
714 double * const restrict work,
715 double * const restrict INT__d_i_g_p)
716 {
717 double P_AB[3*P.nshell12];
718 struct simint_multi_shellpair P_tmp = P;
719 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
720 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
721 P_tmp.AB_x = P_AB;
722 P_tmp.AB_y = P_AB + P.nshell12;
723 P_tmp.AB_z = P_AB + 2*P.nshell12;
724
725 for(int i = 0; i < P.nshell12; i++)
726 {
727 P_tmp.AB_x[i] = -P.AB_x[i];
728 P_tmp.AB_y[i] = -P.AB_y[i];
729 P_tmp.AB_z[i] = -P.AB_z[i];
730 }
731
732 int ret = ostei_i_d_g_p(P_tmp, Q, screen_tol, work, INT__d_i_g_p);
733 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
734
735 for(int q = 0; q < ret; q++)
736 {
737 int idx = 0;
738 for(int a = 0; a < 6; ++a)
739 for(int b = 0; b < 28; ++b)
740 for(int c = 0; c < 15; ++c)
741 for(int d = 0; d < 3; ++d)
742 buffer[idx++] = INT__d_i_g_p[q*7560+b*270+a*45+c*3+d];
743
744 memcpy(INT__d_i_g_p+q*7560, buffer, 7560*sizeof(double));
745 }
746
747 return ret;
748 }
749
ostei_i_d_p_g(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_d_p_g)750 int ostei_i_d_p_g(struct simint_multi_shellpair const P,
751 struct simint_multi_shellpair const Q,
752 double screen_tol,
753 double * const restrict work,
754 double * const restrict INT__i_d_p_g)
755 {
756 double Q_AB[3*Q.nshell12];
757 struct simint_multi_shellpair Q_tmp = Q;
758 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
759 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
760 Q_tmp.AB_x = Q_AB;
761 Q_tmp.AB_y = Q_AB + Q.nshell12;
762 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
763
764 for(int i = 0; i < Q.nshell12; i++)
765 {
766 Q_tmp.AB_x[i] = -Q.AB_x[i];
767 Q_tmp.AB_y[i] = -Q.AB_y[i];
768 Q_tmp.AB_z[i] = -Q.AB_z[i];
769 }
770
771 int ret = ostei_i_d_g_p(P, Q_tmp, screen_tol, work, INT__i_d_p_g);
772 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
773
774 for(int q = 0; q < ret; q++)
775 {
776 int idx = 0;
777 for(int a = 0; a < 28; ++a)
778 for(int b = 0; b < 6; ++b)
779 for(int c = 0; c < 3; ++c)
780 for(int d = 0; d < 15; ++d)
781 buffer[idx++] = INT__i_d_p_g[q*7560+a*270+b*45+d*3+c];
782
783 memcpy(INT__i_d_p_g+q*7560, buffer, 7560*sizeof(double));
784 }
785
786 return ret;
787 }
788
ostei_d_i_p_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_i_p_g)789 int ostei_d_i_p_g(struct simint_multi_shellpair const P,
790 struct simint_multi_shellpair const Q,
791 double screen_tol,
792 double * const restrict work,
793 double * const restrict INT__d_i_p_g)
794 {
795 double P_AB[3*P.nshell12];
796 struct simint_multi_shellpair P_tmp = P;
797 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
798 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
799 P_tmp.AB_x = P_AB;
800 P_tmp.AB_y = P_AB + P.nshell12;
801 P_tmp.AB_z = P_AB + 2*P.nshell12;
802
803 for(int i = 0; i < P.nshell12; i++)
804 {
805 P_tmp.AB_x[i] = -P.AB_x[i];
806 P_tmp.AB_y[i] = -P.AB_y[i];
807 P_tmp.AB_z[i] = -P.AB_z[i];
808 }
809
810 double Q_AB[3*Q.nshell12];
811 struct simint_multi_shellpair Q_tmp = Q;
812 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
813 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
814 Q_tmp.AB_x = Q_AB;
815 Q_tmp.AB_y = Q_AB + Q.nshell12;
816 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
817
818 for(int i = 0; i < Q.nshell12; i++)
819 {
820 Q_tmp.AB_x[i] = -Q.AB_x[i];
821 Q_tmp.AB_y[i] = -Q.AB_y[i];
822 Q_tmp.AB_z[i] = -Q.AB_z[i];
823 }
824
825 int ret = ostei_i_d_g_p(P_tmp, Q_tmp, screen_tol, work, INT__d_i_p_g);
826 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
827
828 for(int q = 0; q < ret; q++)
829 {
830 int idx = 0;
831 for(int a = 0; a < 6; ++a)
832 for(int b = 0; b < 28; ++b)
833 for(int c = 0; c < 3; ++c)
834 for(int d = 0; d < 15; ++d)
835 buffer[idx++] = INT__d_i_p_g[q*7560+b*270+a*45+d*3+c];
836
837 memcpy(INT__d_i_p_g+q*7560, buffer, 7560*sizeof(double));
838 }
839
840 return ret;
841 }
842
843