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_g_d_i_p(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_d_i_p)8 int ostei_g_d_i_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__g_d_i_p)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__g_d_i_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__g_s_i_s = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__g_s_k_s = work + (SIMINT_NSHELL_SIMD * 420);
31 double * const INT__h_s_i_s = work + (SIMINT_NSHELL_SIMD * 960);
32 double * const INT__h_s_k_s = work + (SIMINT_NSHELL_SIMD * 1548);
33 double * const INT__i_s_i_s = work + (SIMINT_NSHELL_SIMD * 2304);
34 double * const INT__i_s_k_s = work + (SIMINT_NSHELL_SIMD * 3088);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*4096);
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_p_s = primwork + 14;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_d_s = primwork + 53;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_f_s = primwork + 125;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_g_s = primwork + 235;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_h_s = primwork + 385;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_i_s = primwork + 574;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_k_s = primwork + 798;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_p_s = primwork + 1050;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_d_s = primwork + 1104;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_f_s = primwork + 1212;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_g_s = primwork + 1392;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_h_s = primwork + 1662;
49 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_i_s = primwork + 2040;
50 SIMINT_DBLTYPE * const restrict PRIM_INT__p_s_k_s = primwork + 2544;
51 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_d_s = primwork + 3192;
52 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_f_s = primwork + 3372;
53 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_g_s = primwork + 3672;
54 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_h_s = primwork + 4122;
55 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_i_s = primwork + 4752;
56 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_k_s = primwork + 5592;
57 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_f_s = primwork + 6672;
58 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_g_s = primwork + 7072;
59 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_h_s = primwork + 7672;
60 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_i_s = primwork + 8512;
61 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_k_s = primwork + 9632;
62 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_g_s = primwork + 11072;
63 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_h_s = primwork + 11747;
64 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_i_s = primwork + 12692;
65 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_k_s = primwork + 13952;
66 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_h_s = primwork + 15572;
67 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_i_s = primwork + 16454;
68 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_k_s = primwork + 17630;
69 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_i_s = primwork + 19142;
70 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_k_s = primwork + 19926;
71 double * const hrrwork = (double *)(primwork + 20934);
72 double * const HRR_INT__g_p_i_s = hrrwork + 0;
73 double * const HRR_INT__g_p_k_s = hrrwork + 1260;
74 double * const HRR_INT__g_d_i_s = hrrwork + 2880;
75 double * const HRR_INT__g_d_k_s = hrrwork + 5400;
76 double * const HRR_INT__h_p_i_s = hrrwork + 8640;
77 double * const HRR_INT__h_p_k_s = hrrwork + 10404;
78
79
80 // Create constants
81 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
82 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
83 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
84 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
85 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
86 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
87 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
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 * 4096 * 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__g_s_i_s = INT__g_s_i_s + abcd * 420;
132 double * restrict PRIM_PTR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
133 double * restrict PRIM_PTR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
134 double * restrict PRIM_PTR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
135 double * restrict PRIM_PTR_INT__i_s_i_s = INT__i_s_i_s + abcd * 784;
136 double * restrict PRIM_PTR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
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__g_s_i_s += 420;
163 PRIM_PTR_INT__g_s_k_s += 540;
164 PRIM_PTR_INT__h_s_i_s += 588;
165 PRIM_PTR_INT__h_s_k_s += 756;
166 PRIM_PTR_INT__i_s_i_s += 784;
167 PRIM_PTR_INT__i_s_k_s += 1008;
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__g_s_i_s += lastoffset*420;
194 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
195 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
196 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
197 PRIM_PTR_INT__i_s_i_s += lastoffset*784;
198 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
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_1_over_2q = one_over_2q;
269 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
270 const SIMINT_DBLTYPE vrr_const_3_over_2q = SIMINT_MUL(const_3, one_over_2q);
271 const SIMINT_DBLTYPE vrr_const_4_over_2q = SIMINT_MUL(const_4, one_over_2q);
272 const SIMINT_DBLTYPE vrr_const_5_over_2q = SIMINT_MUL(const_5, one_over_2q);
273 const SIMINT_DBLTYPE vrr_const_6_over_2q = SIMINT_MUL(const_6, 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
282
283
284 // Forming PRIM_INT__s_s_p_s[13 * 3];
285 for(n = 0; n < 13; ++n) // loop over orders of auxiliary function
286 {
287
288 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
289 PRIM_INT__s_s_p_s[n * 3 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]);
290
291 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
292 PRIM_INT__s_s_p_s[n * 3 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 1]);
293
294 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
295 PRIM_INT__s_s_p_s[n * 3 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_p_s[n * 3 + 2]);
296
297 }
298
299
300
301 // Forming PRIM_INT__s_s_d_s[12 * 6];
302 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
303 {
304
305 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
306 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 0]);
307 PRIM_INT__s_s_d_s[n * 6 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 0]);
308
309 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
310 PRIM_INT__s_s_d_s[n * 6 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 1]);
311
312 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
313 PRIM_INT__s_s_d_s[n * 6 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_d_s[n * 6 + 2]);
314
315 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
316 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 3]);
317 PRIM_INT__s_s_d_s[n * 6 + 3] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 3]);
318
319 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
320 PRIM_INT__s_s_d_s[n * 6 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_d_s[n * 6 + 4]);
321
322 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
323 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_d_s[n * 6 + 5]);
324 PRIM_INT__s_s_d_s[n * 6 + 5] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_s[n * 1 + 0]), PRIM_INT__s_s_d_s[n * 6 + 5]);
325
326 }
327
328
329
330 // Forming PRIM_INT__s_s_f_s[11 * 10];
331 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
332 {
333
334 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
335 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 0]);
336 PRIM_INT__s_s_f_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__s_s_p_s[n * 3 + 0]), PRIM_INT__s_s_f_s[n * 10 + 0]);
337
338 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
339 PRIM_INT__s_s_f_s[n * 10 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 1]);
340
341 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
342 PRIM_INT__s_s_f_s[n * 10 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__s_s_f_s[n * 10 + 2]);
343
344 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
345 PRIM_INT__s_s_f_s[n * 10 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 3]);
346
347 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
348 PRIM_INT__s_s_f_s[n * 10 + 4] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__s_s_f_s[n * 10 + 4]);
349
350 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_MUL(Q_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
351 PRIM_INT__s_s_f_s[n * 10 + 5] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 5]);
352
353 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
354 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 6]);
355 PRIM_INT__s_s_f_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__s_s_p_s[n * 3 + 1]), PRIM_INT__s_s_f_s[n * 10 + 6]);
356
357 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
358 PRIM_INT__s_s_f_s[n * 10 + 7] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__s_s_f_s[n * 10 + 7]);
359
360 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_MUL(Q_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
361 PRIM_INT__s_s_f_s[n * 10 + 8] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 8]);
362
363 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_MUL(Q_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
364 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__s_s_f_s[n * 10 + 9]);
365 PRIM_INT__s_s_f_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__s_s_p_s[n * 3 + 2]), PRIM_INT__s_s_f_s[n * 10 + 9]);
366
367 }
368
369
370 VRR_K_s_s_g_s(
371 PRIM_INT__s_s_g_s,
372 PRIM_INT__s_s_f_s,
373 PRIM_INT__s_s_d_s,
374 Q_PA,
375 a_over_q,
376 aoq_PQ,
377 one_over_2q,
378 10);
379
380
381 VRR_K_s_s_h_s(
382 PRIM_INT__s_s_h_s,
383 PRIM_INT__s_s_g_s,
384 PRIM_INT__s_s_f_s,
385 Q_PA,
386 a_over_q,
387 aoq_PQ,
388 one_over_2q,
389 9);
390
391
392 ostei_general_vrr1_K(6, 8,
393 one_over_2q, a_over_q, aoq_PQ, Q_PA,
394 PRIM_INT__s_s_h_s, PRIM_INT__s_s_g_s, PRIM_INT__s_s_i_s);
395
396
397 ostei_general_vrr_I(1, 0, 6, 0, 6,
398 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
399 PRIM_INT__s_s_i_s, NULL, NULL, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_i_s);
400
401
402 ostei_general_vrr_I(1, 0, 5, 0, 6,
403 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
404 PRIM_INT__s_s_h_s, NULL, NULL, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_h_s);
405
406
407 ostei_general_vrr_I(2, 0, 6, 0, 5,
408 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
409 PRIM_INT__p_s_i_s, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_i_s);
410
411
412 VRR_I_p_s_g_s(
413 PRIM_INT__p_s_g_s,
414 PRIM_INT__s_s_g_s,
415 PRIM_INT__s_s_f_s,
416 P_PA,
417 aop_PQ,
418 one_over_2pq,
419 6);
420
421
422 ostei_general_vrr_I(2, 0, 5, 0, 5,
423 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
424 PRIM_INT__p_s_h_s, PRIM_INT__s_s_h_s, NULL, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_h_s);
425
426
427 ostei_general_vrr_I(3, 0, 6, 0, 4,
428 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
429 PRIM_INT__d_s_i_s, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_i_s);
430
431
432 VRR_I_p_s_f_s(
433 PRIM_INT__p_s_f_s,
434 PRIM_INT__s_s_f_s,
435 PRIM_INT__s_s_d_s,
436 P_PA,
437 aop_PQ,
438 one_over_2pq,
439 6);
440
441
442 ostei_general_vrr_I(2, 0, 4, 0, 5,
443 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
444 PRIM_INT__p_s_g_s, PRIM_INT__s_s_g_s, NULL, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_g_s);
445
446
447 ostei_general_vrr_I(3, 0, 5, 0, 4,
448 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
449 PRIM_INT__d_s_h_s, PRIM_INT__p_s_h_s, NULL, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_h_s);
450
451
452 ostei_general_vrr_I(4, 0, 6, 0, 3,
453 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
454 PRIM_INT__f_s_i_s, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_i_s);
455
456
457 ostei_general_vrr1_K(7, 7,
458 one_over_2q, a_over_q, aoq_PQ, Q_PA,
459 PRIM_INT__s_s_i_s, PRIM_INT__s_s_h_s, PRIM_INT__s_s_k_s);
460
461
462 ostei_general_vrr_I(1, 0, 7, 0, 6,
463 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
464 PRIM_INT__s_s_k_s, NULL, NULL, PRIM_INT__s_s_i_s, NULL, PRIM_INT__p_s_k_s);
465
466
467 ostei_general_vrr_I(2, 0, 7, 0, 5,
468 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
469 PRIM_INT__p_s_k_s, PRIM_INT__s_s_k_s, NULL, PRIM_INT__p_s_i_s, NULL, PRIM_INT__d_s_k_s);
470
471
472 ostei_general_vrr_I(3, 0, 7, 0, 4,
473 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
474 PRIM_INT__d_s_k_s, PRIM_INT__p_s_k_s, NULL, PRIM_INT__d_s_i_s, NULL, PRIM_INT__f_s_k_s);
475
476
477 ostei_general_vrr_I(4, 0, 7, 0, 3,
478 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
479 PRIM_INT__f_s_k_s, PRIM_INT__d_s_k_s, NULL, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_k_s);
480
481
482
483 // Forming PRIM_INT__p_s_d_s[6 * 18];
484 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
485 {
486
487 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 0]);
488 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
489 PRIM_INT__p_s_d_s[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 0]);
490
491 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 1]);
492 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 1]);
493 PRIM_INT__p_s_d_s[n * 18 + 1] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 1]);
494
495 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 2]);
496 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 2]);
497 PRIM_INT__p_s_d_s[n * 18 + 2] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 2]);
498
499 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 3]);
500 PRIM_INT__p_s_d_s[n * 18 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 3]);
501
502 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 4]);
503 PRIM_INT__p_s_d_s[n * 18 + 4] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 4]);
504
505 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_d_s[n * 6 + 5]);
506 PRIM_INT__p_s_d_s[n * 18 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 5]);
507
508 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 0]);
509 PRIM_INT__p_s_d_s[n * 18 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 6]);
510
511 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 1]);
512 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 7]);
513 PRIM_INT__p_s_d_s[n * 18 + 7] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 7]);
514
515 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 2]);
516 PRIM_INT__p_s_d_s[n * 18 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 8]);
517
518 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 3]);
519 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 9]);
520 PRIM_INT__p_s_d_s[n * 18 + 9] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 9]);
521
522 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 4]);
523 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 10]);
524 PRIM_INT__p_s_d_s[n * 18 + 10] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 10]);
525
526 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_d_s[n * 6 + 5]);
527 PRIM_INT__p_s_d_s[n * 18 + 11] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 11]);
528
529 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 0]);
530 PRIM_INT__p_s_d_s[n * 18 + 12] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 0], PRIM_INT__p_s_d_s[n * 18 + 12]);
531
532 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 1]);
533 PRIM_INT__p_s_d_s[n * 18 + 13] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 1], PRIM_INT__p_s_d_s[n * 18 + 13]);
534
535 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 2]);
536 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 2], PRIM_INT__p_s_d_s[n * 18 + 14]);
537 PRIM_INT__p_s_d_s[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_d_s[n * 18 + 14]);
538
539 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 3]);
540 PRIM_INT__p_s_d_s[n * 18 + 15] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 3], PRIM_INT__p_s_d_s[n * 18 + 15]);
541
542 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 4]);
543 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 4], PRIM_INT__p_s_d_s[n * 18 + 16]);
544 PRIM_INT__p_s_d_s[n * 18 + 16] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_d_s[n * 18 + 16]);
545
546 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_d_s[n * 6 + 5]);
547 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_d_s[(n+1) * 6 + 5], PRIM_INT__p_s_d_s[n * 18 + 17]);
548 PRIM_INT__p_s_d_s[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_d_s[n * 18 + 17]);
549
550 }
551
552
553 VRR_I_d_s_f_s(
554 PRIM_INT__d_s_f_s,
555 PRIM_INT__p_s_f_s,
556 PRIM_INT__s_s_f_s,
557 PRIM_INT__p_s_d_s,
558 P_PA,
559 a_over_p,
560 aop_PQ,
561 one_over_2p,
562 one_over_2pq,
563 5);
564
565
566 ostei_general_vrr_I(3, 0, 4, 0, 4,
567 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
568 PRIM_INT__d_s_g_s, PRIM_INT__p_s_g_s, NULL, PRIM_INT__d_s_f_s, NULL, PRIM_INT__f_s_g_s);
569
570
571 ostei_general_vrr_I(4, 0, 5, 0, 3,
572 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
573 PRIM_INT__f_s_h_s, PRIM_INT__d_s_h_s, NULL, PRIM_INT__f_s_g_s, NULL, PRIM_INT__g_s_h_s);
574
575
576 ostei_general_vrr_I(5, 0, 6, 0, 2,
577 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
578 PRIM_INT__g_s_i_s, PRIM_INT__f_s_i_s, NULL, PRIM_INT__g_s_h_s, NULL, PRIM_INT__h_s_i_s);
579
580
581 ostei_general_vrr_I(5, 0, 7, 0, 2,
582 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
583 PRIM_INT__g_s_k_s, PRIM_INT__f_s_k_s, NULL, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_k_s);
584
585
586
587 // Forming PRIM_INT__p_s_p_s[6 * 9];
588 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
589 {
590
591 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 0]);
592 PRIM_INT__p_s_p_s[n * 9 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 0]);
593 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]);
594
595 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 1]);
596 PRIM_INT__p_s_p_s[n * 9 + 1] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 1]);
597
598 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_p_s[n * 3 + 2]);
599 PRIM_INT__p_s_p_s[n * 9 + 2] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 2]);
600
601 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 0]);
602 PRIM_INT__p_s_p_s[n * 9 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 3]);
603
604 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 1]);
605 PRIM_INT__p_s_p_s[n * 9 + 4] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 4]);
606 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]);
607
608 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_p_s[n * 3 + 2]);
609 PRIM_INT__p_s_p_s[n * 9 + 5] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 5]);
610
611 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 0]);
612 PRIM_INT__p_s_p_s[n * 9 + 6] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 0], PRIM_INT__p_s_p_s[n * 9 + 6]);
613
614 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 1]);
615 PRIM_INT__p_s_p_s[n * 9 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 1], PRIM_INT__p_s_p_s[n * 9 + 7]);
616
617 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_p_s[n * 3 + 2]);
618 PRIM_INT__p_s_p_s[n * 9 + 8] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_p_s[(n+1) * 3 + 2], PRIM_INT__p_s_p_s[n * 9 + 8]);
619 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]);
620
621 }
622
623
624 VRR_I_d_s_d_s(
625 PRIM_INT__d_s_d_s,
626 PRIM_INT__p_s_d_s,
627 PRIM_INT__s_s_d_s,
628 PRIM_INT__p_s_p_s,
629 P_PA,
630 a_over_p,
631 aop_PQ,
632 one_over_2p,
633 one_over_2pq,
634 5);
635
636
637 ostei_general_vrr_I(3, 0, 3, 0, 4,
638 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
639 PRIM_INT__d_s_f_s, PRIM_INT__p_s_f_s, NULL, PRIM_INT__d_s_d_s, NULL, PRIM_INT__f_s_f_s);
640
641
642 ostei_general_vrr_I(4, 0, 4, 0, 3,
643 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
644 PRIM_INT__f_s_g_s, PRIM_INT__d_s_g_s, NULL, PRIM_INT__f_s_f_s, NULL, PRIM_INT__g_s_g_s);
645
646
647 ostei_general_vrr_I(5, 0, 5, 0, 2,
648 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
649 PRIM_INT__g_s_h_s, PRIM_INT__f_s_h_s, NULL, PRIM_INT__g_s_g_s, NULL, PRIM_INT__h_s_h_s);
650
651
652 ostei_general_vrr_I(6, 0, 6, 0, 1,
653 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
654 PRIM_INT__h_s_i_s, PRIM_INT__g_s_i_s, NULL, PRIM_INT__h_s_h_s, NULL, PRIM_INT__i_s_i_s);
655
656
657 ostei_general_vrr_I(6, 0, 7, 0, 1,
658 one_over_2p, a_over_p, one_over_2pq, aop_PQ, P_PA,
659 PRIM_INT__h_s_k_s, PRIM_INT__g_s_k_s, NULL, PRIM_INT__h_s_i_s, NULL, PRIM_INT__i_s_k_s);
660
661
662
663
664 ////////////////////////////////////
665 // Accumulate contracted integrals
666 ////////////////////////////////////
667 if(lastoffset == 0)
668 {
669 contract_all(420, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
670 contract_all(540, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
671 contract_all(588, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
672 contract_all(756, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
673 contract_all(784, PRIM_INT__i_s_i_s, PRIM_PTR_INT__i_s_i_s);
674 contract_all(1008, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
675 }
676 else
677 {
678 contract(420, shelloffsets, PRIM_INT__g_s_i_s, PRIM_PTR_INT__g_s_i_s);
679 contract(540, shelloffsets, PRIM_INT__g_s_k_s, PRIM_PTR_INT__g_s_k_s);
680 contract(588, shelloffsets, PRIM_INT__h_s_i_s, PRIM_PTR_INT__h_s_i_s);
681 contract(756, shelloffsets, PRIM_INT__h_s_k_s, PRIM_PTR_INT__h_s_k_s);
682 contract(784, shelloffsets, PRIM_INT__i_s_i_s, PRIM_PTR_INT__i_s_i_s);
683 contract(1008, shelloffsets, PRIM_INT__i_s_k_s, PRIM_PTR_INT__i_s_k_s);
684 PRIM_PTR_INT__g_s_i_s += lastoffset*420;
685 PRIM_PTR_INT__g_s_k_s += lastoffset*540;
686 PRIM_PTR_INT__h_s_i_s += lastoffset*588;
687 PRIM_PTR_INT__h_s_k_s += lastoffset*756;
688 PRIM_PTR_INT__i_s_i_s += lastoffset*784;
689 PRIM_PTR_INT__i_s_k_s += lastoffset*1008;
690 }
691
692 } // close loop over j
693 } // close loop over i
694
695 //Advance to the next batch
696 jstart = SIMINT_SIMD_ROUND(jend);
697
698 //////////////////////////////////////////////
699 // Contracted integrals: Horizontal recurrance
700 //////////////////////////////////////////////
701
702
703 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
704
705
706 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
707 {
708 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
709
710 // set up HRR pointers
711 double const * restrict HRR_INT__g_s_i_s = INT__g_s_i_s + abcd * 420;
712 double const * restrict HRR_INT__g_s_k_s = INT__g_s_k_s + abcd * 540;
713 double const * restrict HRR_INT__h_s_i_s = INT__h_s_i_s + abcd * 588;
714 double const * restrict HRR_INT__h_s_k_s = INT__h_s_k_s + abcd * 756;
715 double const * restrict HRR_INT__i_s_i_s = INT__i_s_i_s + abcd * 784;
716 double const * restrict HRR_INT__i_s_k_s = INT__i_s_k_s + abcd * 1008;
717 double * restrict HRR_INT__g_d_i_p = INT__g_d_i_p + real_abcd * 7560;
718
719 // form INT__g_p_i_s
720 HRR_J_g_p(
721 HRR_INT__g_p_i_s,
722 HRR_INT__g_s_i_s,
723 HRR_INT__h_s_i_s,
724 hAB, 28);
725
726 // form INT__g_p_k_s
727 HRR_J_g_p(
728 HRR_INT__g_p_k_s,
729 HRR_INT__g_s_k_s,
730 HRR_INT__h_s_k_s,
731 hAB, 36);
732
733 // form INT__h_p_i_s
734 ostei_general_hrr_J(5, 1, 6, 0, hAB, HRR_INT__i_s_i_s, HRR_INT__h_s_i_s, HRR_INT__h_p_i_s);
735
736 // form INT__h_p_k_s
737 ostei_general_hrr_J(5, 1, 7, 0, hAB, HRR_INT__i_s_k_s, HRR_INT__h_s_k_s, HRR_INT__h_p_k_s);
738
739 // form INT__g_d_i_s
740 ostei_general_hrr_J(4, 2, 6, 0, hAB, HRR_INT__h_p_i_s, HRR_INT__g_p_i_s, HRR_INT__g_d_i_s);
741
742 // form INT__g_d_k_s
743 ostei_general_hrr_J(4, 2, 7, 0, hAB, HRR_INT__h_p_k_s, HRR_INT__g_p_k_s, HRR_INT__g_d_k_s);
744
745 // form INT__g_d_i_p
746 ostei_general_hrr_L(4, 2, 6, 1, hCD, HRR_INT__g_d_k_s, HRR_INT__g_d_i_s, HRR_INT__g_d_i_p);
747
748
749 } // close HRR loop
750
751
752 } // close loop cdbatch
753
754 istart = iend;
755 } // close loop over ab
756
757 return P.nshell12_clip * Q.nshell12_clip;
758 }
759
ostei_d_g_i_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_g_i_p)760 int ostei_d_g_i_p(struct simint_multi_shellpair const P,
761 struct simint_multi_shellpair const Q,
762 double screen_tol,
763 double * const restrict work,
764 double * const restrict INT__d_g_i_p)
765 {
766 double P_AB[3*P.nshell12];
767 struct simint_multi_shellpair P_tmp = P;
768 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
769 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
770 P_tmp.AB_x = P_AB;
771 P_tmp.AB_y = P_AB + P.nshell12;
772 P_tmp.AB_z = P_AB + 2*P.nshell12;
773
774 for(int i = 0; i < P.nshell12; i++)
775 {
776 P_tmp.AB_x[i] = -P.AB_x[i];
777 P_tmp.AB_y[i] = -P.AB_y[i];
778 P_tmp.AB_z[i] = -P.AB_z[i];
779 }
780
781 int ret = ostei_g_d_i_p(P_tmp, Q, screen_tol, work, INT__d_g_i_p);
782 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
783
784 for(int q = 0; q < ret; q++)
785 {
786 int idx = 0;
787 for(int a = 0; a < 6; ++a)
788 for(int b = 0; b < 15; ++b)
789 for(int c = 0; c < 28; ++c)
790 for(int d = 0; d < 3; ++d)
791 buffer[idx++] = INT__d_g_i_p[q*7560+b*504+a*84+c*3+d];
792
793 memcpy(INT__d_g_i_p+q*7560, buffer, 7560*sizeof(double));
794 }
795
796 return ret;
797 }
798
ostei_g_d_p_i(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__g_d_p_i)799 int ostei_g_d_p_i(struct simint_multi_shellpair const P,
800 struct simint_multi_shellpair const Q,
801 double screen_tol,
802 double * const restrict work,
803 double * const restrict INT__g_d_p_i)
804 {
805 double Q_AB[3*Q.nshell12];
806 struct simint_multi_shellpair Q_tmp = Q;
807 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
808 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
809 Q_tmp.AB_x = Q_AB;
810 Q_tmp.AB_y = Q_AB + Q.nshell12;
811 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
812
813 for(int i = 0; i < Q.nshell12; i++)
814 {
815 Q_tmp.AB_x[i] = -Q.AB_x[i];
816 Q_tmp.AB_y[i] = -Q.AB_y[i];
817 Q_tmp.AB_z[i] = -Q.AB_z[i];
818 }
819
820 int ret = ostei_g_d_i_p(P, Q_tmp, screen_tol, work, INT__g_d_p_i);
821 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
822
823 for(int q = 0; q < ret; q++)
824 {
825 int idx = 0;
826 for(int a = 0; a < 15; ++a)
827 for(int b = 0; b < 6; ++b)
828 for(int c = 0; c < 3; ++c)
829 for(int d = 0; d < 28; ++d)
830 buffer[idx++] = INT__g_d_p_i[q*7560+a*504+b*84+d*3+c];
831
832 memcpy(INT__g_d_p_i+q*7560, buffer, 7560*sizeof(double));
833 }
834
835 return ret;
836 }
837
ostei_d_g_p_i(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__d_g_p_i)838 int ostei_d_g_p_i(struct simint_multi_shellpair const P,
839 struct simint_multi_shellpair const Q,
840 double screen_tol,
841 double * const restrict work,
842 double * const restrict INT__d_g_p_i)
843 {
844 double P_AB[3*P.nshell12];
845 struct simint_multi_shellpair P_tmp = P;
846 P_tmp.PA_x = P.PB_x; P_tmp.PA_y = P.PB_y; P_tmp.PA_z = P.PB_z;
847 P_tmp.PB_x = P.PA_x; P_tmp.PB_y = P.PA_y; P_tmp.PB_z = P.PA_z;
848 P_tmp.AB_x = P_AB;
849 P_tmp.AB_y = P_AB + P.nshell12;
850 P_tmp.AB_z = P_AB + 2*P.nshell12;
851
852 for(int i = 0; i < P.nshell12; i++)
853 {
854 P_tmp.AB_x[i] = -P.AB_x[i];
855 P_tmp.AB_y[i] = -P.AB_y[i];
856 P_tmp.AB_z[i] = -P.AB_z[i];
857 }
858
859 double Q_AB[3*Q.nshell12];
860 struct simint_multi_shellpair Q_tmp = Q;
861 Q_tmp.PA_x = Q.PB_x; Q_tmp.PA_y = Q.PB_y; Q_tmp.PA_z = Q.PB_z;
862 Q_tmp.PB_x = Q.PA_x; Q_tmp.PB_y = Q.PA_y; Q_tmp.PB_z = Q.PA_z;
863 Q_tmp.AB_x = Q_AB;
864 Q_tmp.AB_y = Q_AB + Q.nshell12;
865 Q_tmp.AB_z = Q_AB + 2*Q.nshell12;
866
867 for(int i = 0; i < Q.nshell12; i++)
868 {
869 Q_tmp.AB_x[i] = -Q.AB_x[i];
870 Q_tmp.AB_y[i] = -Q.AB_y[i];
871 Q_tmp.AB_z[i] = -Q.AB_z[i];
872 }
873
874 int ret = ostei_g_d_i_p(P_tmp, Q_tmp, screen_tol, work, INT__d_g_p_i);
875 double buffer[7560] SIMINT_ALIGN_ARRAY_DBL;
876
877 for(int q = 0; q < ret; q++)
878 {
879 int idx = 0;
880 for(int a = 0; a < 6; ++a)
881 for(int b = 0; b < 15; ++b)
882 for(int c = 0; c < 3; ++c)
883 for(int d = 0; d < 28; ++d)
884 buffer[idx++] = INT__d_g_p_i[q*7560+b*504+a*84+d*3+c];
885
886 memcpy(INT__d_g_p_i+q*7560, buffer, 7560*sizeof(double));
887 }
888
889 return ret;
890 }
891
892