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_p_d_p_d(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__p_d_p_d)8 int ostei_p_d_p_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__p_d_p_d)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__p_d_p_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__s_d_s_d = work + (SIMINT_NSHELL_SIMD * 0);
30 double * const INT__s_d_s_f = work + (SIMINT_NSHELL_SIMD * 36);
31 double * const INT__s_f_s_d = work + (SIMINT_NSHELL_SIMD * 96);
32 double * const INT__s_f_s_f = work + (SIMINT_NSHELL_SIMD * 156);
33 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*256);
34 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_s = primwork + 0;
35 SIMINT_DBLTYPE * const restrict PRIM_INT__s_s_s_p = primwork + 7;
36 SIMINT_DBLTYPE * const restrict PRIM_INT__s_p_s_s = primwork + 16;
37 SIMINT_DBLTYPE * const restrict PRIM_INT__s_p_s_p = primwork + 34;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__s_p_s_d = primwork + 61;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__s_d_s_s = primwork + 97;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__s_d_s_p = primwork + 127;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__s_d_s_d = primwork + 181;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__s_d_s_f = primwork + 253;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__s_f_s_s = primwork + 313;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__s_f_s_p = primwork + 353;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__s_f_s_d = primwork + 443;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__s_f_s_f = primwork + 563;
47 double * const hrrwork = (double *)(primwork + 663);
48 double * const HRR_INT__p_d_s_d = hrrwork + 0;
49 double * const HRR_INT__p_d_s_f = hrrwork + 108;
50
51
52 // Create constants
53 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
54 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
55 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
56 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
57
58
59 ////////////////////////////////////////
60 // Loop over shells and primitives
61 ////////////////////////////////////////
62
63 real_abcd = 0;
64 istart = 0;
65 for(ab = 0; ab < P.nshell12_clip; ++ab)
66 {
67 const int iend = istart + P.nprim12[ab];
68
69 cd = 0;
70 jstart = 0;
71
72 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
73 {
74 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
75 int jend = jstart;
76 for(i = 0; i < nshellbatch; i++)
77 jend += Q.nprim12[cd+i];
78
79 // Clear the beginning of the workspace (where we are accumulating integrals)
80 memset(work, 0, SIMINT_NSHELL_SIMD * 256 * sizeof(double));
81 abcd = 0;
82
83
84 for(i = istart; i < iend; ++i)
85 {
86 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
87
88 if(check_screen)
89 {
90 // Skip this whole thing if always insignificant
91 if((P.screen[i] * Q.screen_max) < screen_tol)
92 continue;
93 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
94 }
95
96 icd = 0;
97 iprimcd = 0;
98 nprim_icd = Q.nprim12[cd];
99 double * restrict PRIM_PTR_INT__s_d_s_d = INT__s_d_s_d + abcd * 36;
100 double * restrict PRIM_PTR_INT__s_d_s_f = INT__s_d_s_f + abcd * 60;
101 double * restrict PRIM_PTR_INT__s_f_s_d = INT__s_f_s_d + abcd * 60;
102 double * restrict PRIM_PTR_INT__s_f_s_f = INT__s_f_s_f + abcd * 100;
103
104
105
106 // Load these one per loop over i
107 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
108 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
109 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
110
111 const SIMINT_DBLTYPE P_PB[3] = { SIMINT_DBLSET1(P.PB_x[i]), SIMINT_DBLSET1(P.PB_y[i]), SIMINT_DBLSET1(P.PB_z[i]) };
112
113 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
114 {
115 // calculate the shell offsets
116 // these are the offset from the shell pointed to by cd
117 // for each element
118 int shelloffsets[SIMINT_SIMD_LEN] = {0};
119 int lastoffset = 0;
120 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
121
122 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
123 {
124 // Handle if the first element of the vector is a new shell
125 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
126 {
127 nprim_icd += Q.nprim12[cd + (++icd)];
128 PRIM_PTR_INT__s_d_s_d += 36;
129 PRIM_PTR_INT__s_d_s_f += 60;
130 PRIM_PTR_INT__s_f_s_d += 60;
131 PRIM_PTR_INT__s_f_s_f += 100;
132 }
133 iprimcd++;
134 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
135 {
136 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
137 {
138 shelloffsets[n] = shelloffsets[n-1] + 1;
139 lastoffset++;
140 nprim_icd += Q.nprim12[cd + (++icd)];
141 }
142 else
143 shelloffsets[n] = shelloffsets[n-1];
144 iprimcd++;
145 }
146 }
147 else
148 iprimcd += SIMINT_SIMD_LEN;
149
150 // Do we have to compute this vector (or has it been screened out)?
151 // (not_screened != 0 means we have to do this vector)
152 if(check_screen)
153 {
154 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
155 if(vmax < screen_tol)
156 {
157 PRIM_PTR_INT__s_d_s_d += lastoffset*36;
158 PRIM_PTR_INT__s_d_s_f += lastoffset*60;
159 PRIM_PTR_INT__s_f_s_d += lastoffset*60;
160 PRIM_PTR_INT__s_f_s_f += lastoffset*100;
161 continue;
162 }
163 }
164
165 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
166 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
167 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
168 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
169
170
171 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
172 SIMINT_DBLTYPE PQ[3];
173 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
174 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
175 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
176 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
177 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
178 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
179
180 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
181 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
182 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
183 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
184 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
185 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
186 const SIMINT_DBLTYPE Q_PB[3] = { SIMINT_DBLLOAD(Q.PB_x, j), SIMINT_DBLLOAD(Q.PB_y, j), SIMINT_DBLLOAD(Q.PB_z, j) };
187
188 // NOTE: Minus sign!
189 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
190 SIMINT_DBLTYPE aop_PQ[3];
191 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
192 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
193 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
194
195 SIMINT_DBLTYPE a_over_q = SIMINT_MUL(alpha, one_over_q);
196 SIMINT_DBLTYPE aoq_PQ[3];
197 aoq_PQ[0] = SIMINT_MUL(a_over_q, PQ[0]);
198 aoq_PQ[1] = SIMINT_MUL(a_over_q, PQ[1]);
199 aoq_PQ[2] = SIMINT_MUL(a_over_q, PQ[2]);
200 // Put a minus sign here so we don't have to in RR routines
201 a_over_q = SIMINT_NEG(a_over_q);
202
203
204 //////////////////////////////////////////////
205 // Fjt function section
206 // Maximum v value: 6
207 //////////////////////////////////////////////
208 // The parameter to the Fjt function
209 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
210
211
212 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
213
214
215 boys_F_split(PRIM_INT__s_s_s_s, F_x, 6);
216 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
217 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
218 for(n = 0; n <= 6; n++)
219 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
220
221 //////////////////////////////////////////////
222 // Primitive integrals: Vertical recurrance
223 //////////////////////////////////////////////
224
225 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
226 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
227 const SIMINT_DBLTYPE vrr_const_1_over_2q = one_over_2q;
228 const SIMINT_DBLTYPE vrr_const_2_over_2q = SIMINT_MUL(const_2, one_over_2q);
229 const SIMINT_DBLTYPE vrr_const_1_over_2pq = one_over_2pq;
230 const SIMINT_DBLTYPE vrr_const_2_over_2pq = SIMINT_MUL(const_2, one_over_2pq);
231 const SIMINT_DBLTYPE vrr_const_3_over_2pq = SIMINT_MUL(const_3, one_over_2pq);
232
233
234
235 // Forming PRIM_INT__s_p_s_s[6 * 3];
236 for(n = 0; n < 6; ++n) // loop over orders of auxiliary function
237 {
238
239 PRIM_INT__s_p_s_s[n * 3 + 0] = SIMINT_MUL(P_PB[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
240 PRIM_INT__s_p_s_s[n * 3 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_s[n * 3 + 0]);
241
242 PRIM_INT__s_p_s_s[n * 3 + 1] = SIMINT_MUL(P_PB[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
243 PRIM_INT__s_p_s_s[n * 3 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_s[n * 3 + 1]);
244
245 PRIM_INT__s_p_s_s[n * 3 + 2] = SIMINT_MUL(P_PB[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
246 PRIM_INT__s_p_s_s[n * 3 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_s[n * 3 + 2]);
247
248 }
249
250
251
252 // Forming PRIM_INT__s_d_s_s[5 * 6];
253 for(n = 0; n < 5; ++n) // loop over orders of auxiliary function
254 {
255
256 PRIM_INT__s_d_s_s[n * 6 + 0] = SIMINT_MUL(P_PB[0], PRIM_INT__s_p_s_s[n * 3 + 0]);
257 PRIM_INT__s_d_s_s[n * 6 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_s[n * 6 + 0]);
258 PRIM_INT__s_d_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__s_d_s_s[n * 6 + 0]);
259
260 PRIM_INT__s_d_s_s[n * 6 + 1] = SIMINT_MUL(P_PB[1], PRIM_INT__s_p_s_s[n * 3 + 0]);
261 PRIM_INT__s_d_s_s[n * 6 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_s[n * 6 + 1]);
262
263 PRIM_INT__s_d_s_s[n * 6 + 2] = SIMINT_MUL(P_PB[2], PRIM_INT__s_p_s_s[n * 3 + 0]);
264 PRIM_INT__s_d_s_s[n * 6 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_s[n * 6 + 2]);
265
266 PRIM_INT__s_d_s_s[n * 6 + 3] = SIMINT_MUL(P_PB[1], PRIM_INT__s_p_s_s[n * 3 + 1]);
267 PRIM_INT__s_d_s_s[n * 6 + 3] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_d_s_s[n * 6 + 3]);
268 PRIM_INT__s_d_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__s_d_s_s[n * 6 + 3]);
269
270 PRIM_INT__s_d_s_s[n * 6 + 4] = SIMINT_MUL(P_PB[2], PRIM_INT__s_p_s_s[n * 3 + 1]);
271 PRIM_INT__s_d_s_s[n * 6 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_d_s_s[n * 6 + 4]);
272
273 PRIM_INT__s_d_s_s[n * 6 + 5] = SIMINT_MUL(P_PB[2], PRIM_INT__s_p_s_s[n * 3 + 2]);
274 PRIM_INT__s_d_s_s[n * 6 + 5] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_d_s_s[n * 6 + 5]);
275 PRIM_INT__s_d_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__s_d_s_s[n * 6 + 5]);
276
277 }
278
279
280
281 // Forming PRIM_INT__s_d_s_p[3 * 18];
282 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
283 {
284
285 PRIM_INT__s_d_s_p[n * 18 + 0] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 0]);
286 PRIM_INT__s_d_s_p[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_d_s_p[n * 18 + 0]);
287 PRIM_INT__s_d_s_p[n * 18 + 0] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_p[n * 18 + 0]);
288
289 PRIM_INT__s_d_s_p[n * 18 + 1] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 0]);
290 PRIM_INT__s_d_s_p[n * 18 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_d_s_p[n * 18 + 1]);
291
292 PRIM_INT__s_d_s_p[n * 18 + 2] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 0]);
293 PRIM_INT__s_d_s_p[n * 18 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_d_s_p[n * 18 + 2]);
294
295 PRIM_INT__s_d_s_p[n * 18 + 3] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 1]);
296 PRIM_INT__s_d_s_p[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 1], PRIM_INT__s_d_s_p[n * 18 + 3]);
297 PRIM_INT__s_d_s_p[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_d_s_p[n * 18 + 3]);
298
299 PRIM_INT__s_d_s_p[n * 18 + 4] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 1]);
300 PRIM_INT__s_d_s_p[n * 18 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 1], PRIM_INT__s_d_s_p[n * 18 + 4]);
301 PRIM_INT__s_d_s_p[n * 18 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_p[n * 18 + 4]);
302
303 PRIM_INT__s_d_s_p[n * 18 + 5] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 1]);
304 PRIM_INT__s_d_s_p[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 1], PRIM_INT__s_d_s_p[n * 18 + 5]);
305
306 PRIM_INT__s_d_s_p[n * 18 + 6] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 2]);
307 PRIM_INT__s_d_s_p[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 2], PRIM_INT__s_d_s_p[n * 18 + 6]);
308 PRIM_INT__s_d_s_p[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_d_s_p[n * 18 + 6]);
309
310 PRIM_INT__s_d_s_p[n * 18 + 7] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 2]);
311 PRIM_INT__s_d_s_p[n * 18 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 2], PRIM_INT__s_d_s_p[n * 18 + 7]);
312
313 PRIM_INT__s_d_s_p[n * 18 + 8] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 2]);
314 PRIM_INT__s_d_s_p[n * 18 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 2], PRIM_INT__s_d_s_p[n * 18 + 8]);
315 PRIM_INT__s_d_s_p[n * 18 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_d_s_p[n * 18 + 8]);
316
317 PRIM_INT__s_d_s_p[n * 18 + 9] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 3]);
318 PRIM_INT__s_d_s_p[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_d_s_p[n * 18 + 9]);
319
320 PRIM_INT__s_d_s_p[n * 18 + 10] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 3]);
321 PRIM_INT__s_d_s_p[n * 18 + 10] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_d_s_p[n * 18 + 10]);
322 PRIM_INT__s_d_s_p[n * 18 + 10] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_d_s_p[n * 18 + 10]);
323
324 PRIM_INT__s_d_s_p[n * 18 + 11] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 3]);
325 PRIM_INT__s_d_s_p[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_d_s_p[n * 18 + 11]);
326
327 PRIM_INT__s_d_s_p[n * 18 + 12] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 4]);
328 PRIM_INT__s_d_s_p[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 4], PRIM_INT__s_d_s_p[n * 18 + 12]);
329
330 PRIM_INT__s_d_s_p[n * 18 + 13] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 4]);
331 PRIM_INT__s_d_s_p[n * 18 + 13] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 4], PRIM_INT__s_d_s_p[n * 18 + 13]);
332 PRIM_INT__s_d_s_p[n * 18 + 13] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_d_s_p[n * 18 + 13]);
333
334 PRIM_INT__s_d_s_p[n * 18 + 14] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 4]);
335 PRIM_INT__s_d_s_p[n * 18 + 14] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 4], PRIM_INT__s_d_s_p[n * 18 + 14]);
336 PRIM_INT__s_d_s_p[n * 18 + 14] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_d_s_p[n * 18 + 14]);
337
338 PRIM_INT__s_d_s_p[n * 18 + 15] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_d_s_s[n * 6 + 5]);
339 PRIM_INT__s_d_s_p[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_d_s_p[n * 18 + 15]);
340
341 PRIM_INT__s_d_s_p[n * 18 + 16] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_d_s_s[n * 6 + 5]);
342 PRIM_INT__s_d_s_p[n * 18 + 16] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_d_s_p[n * 18 + 16]);
343
344 PRIM_INT__s_d_s_p[n * 18 + 17] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_d_s_s[n * 6 + 5]);
345 PRIM_INT__s_d_s_p[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_d_s_p[n * 18 + 17]);
346 PRIM_INT__s_d_s_p[n * 18 + 17] = SIMINT_FMADD( vrr_const_2_over_2pq, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_d_s_p[n * 18 + 17]);
347
348 }
349
350
351
352 // Forming PRIM_INT__s_p_s_p[3 * 9];
353 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
354 {
355
356 PRIM_INT__s_p_s_p[n * 9 + 0] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_s[n * 3 + 0]);
357 PRIM_INT__s_p_s_p[n * 9 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_p[n * 9 + 0]);
358 PRIM_INT__s_p_s_p[n * 9 + 0] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_p[n * 9 + 0]);
359
360 PRIM_INT__s_p_s_p[n * 9 + 1] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_s[n * 3 + 0]);
361 PRIM_INT__s_p_s_p[n * 9 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_p[n * 9 + 1]);
362
363 PRIM_INT__s_p_s_p[n * 9 + 2] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_s[n * 3 + 0]);
364 PRIM_INT__s_p_s_p[n * 9 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_p[n * 9 + 2]);
365
366 PRIM_INT__s_p_s_p[n * 9 + 3] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_s[n * 3 + 1]);
367 PRIM_INT__s_p_s_p[n * 9 + 3] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_p[n * 9 + 3]);
368
369 PRIM_INT__s_p_s_p[n * 9 + 4] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_s[n * 3 + 1]);
370 PRIM_INT__s_p_s_p[n * 9 + 4] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_p[n * 9 + 4]);
371 PRIM_INT__s_p_s_p[n * 9 + 4] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_p[n * 9 + 4]);
372
373 PRIM_INT__s_p_s_p[n * 9 + 5] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_s[n * 3 + 1]);
374 PRIM_INT__s_p_s_p[n * 9 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_p[n * 9 + 5]);
375
376 PRIM_INT__s_p_s_p[n * 9 + 6] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_s[n * 3 + 2]);
377 PRIM_INT__s_p_s_p[n * 9 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_p[n * 9 + 6]);
378
379 PRIM_INT__s_p_s_p[n * 9 + 7] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_s[n * 3 + 2]);
380 PRIM_INT__s_p_s_p[n * 9 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_p[n * 9 + 7]);
381
382 PRIM_INT__s_p_s_p[n * 9 + 8] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_s[n * 3 + 2]);
383 PRIM_INT__s_p_s_p[n * 9 + 8] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_p[n * 9 + 8]);
384 PRIM_INT__s_p_s_p[n * 9 + 8] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_p_s_p[n * 9 + 8]);
385
386 }
387
388
389 VRR_L_s_d_s_d(
390 PRIM_INT__s_d_s_d,
391 PRIM_INT__s_d_s_p,
392 PRIM_INT__s_d_s_s,
393 PRIM_INT__s_p_s_p,
394 Q_PB,
395 a_over_q,
396 aoq_PQ,
397 one_over_2pq,
398 one_over_2q,
399 2);
400
401
402
403 // Forming PRIM_INT__s_s_s_p[3 * 3];
404 for(n = 0; n < 3; ++n) // loop over orders of auxiliary function
405 {
406
407 PRIM_INT__s_s_s_p[n * 3 + 0] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
408 PRIM_INT__s_s_s_p[n * 3 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_p[n * 3 + 0]);
409
410 PRIM_INT__s_s_s_p[n * 3 + 1] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
411 PRIM_INT__s_s_s_p[n * 3 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_p[n * 3 + 1]);
412
413 PRIM_INT__s_s_s_p[n * 3 + 2] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
414 PRIM_INT__s_s_s_p[n * 3 + 2] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_s_s_s[(n+1) * 1 + 0], PRIM_INT__s_s_s_p[n * 3 + 2]);
415
416 }
417
418
419
420 // Forming PRIM_INT__s_p_s_d[2 * 18];
421 for(n = 0; n < 2; ++n) // loop over orders of auxiliary function
422 {
423
424 PRIM_INT__s_p_s_d[n * 18 + 0] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_p[n * 9 + 0]);
425 PRIM_INT__s_p_s_d[n * 18 + 0] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_p[(n+1) * 9 + 0], PRIM_INT__s_p_s_d[n * 18 + 0]);
426 PRIM_INT__s_p_s_d[n * 18 + 0] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_s[n * 3 + 0]), PRIM_INT__s_p_s_d[n * 18 + 0]);
427 PRIM_INT__s_p_s_d[n * 18 + 0] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_p[(n+1) * 3 + 0], PRIM_INT__s_p_s_d[n * 18 + 0]);
428
429 PRIM_INT__s_p_s_d[n * 18 + 1] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 0]);
430 PRIM_INT__s_p_s_d[n * 18 + 1] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 0], PRIM_INT__s_p_s_d[n * 18 + 1]);
431
432 PRIM_INT__s_p_s_d[n * 18 + 3] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 1]);
433 PRIM_INT__s_p_s_d[n * 18 + 3] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 1], PRIM_INT__s_p_s_d[n * 18 + 3]);
434 PRIM_INT__s_p_s_d[n * 18 + 3] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_s[n * 3 + 0]), PRIM_INT__s_p_s_d[n * 18 + 3]);
435
436 PRIM_INT__s_p_s_d[n * 18 + 5] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_p[n * 9 + 2]);
437 PRIM_INT__s_p_s_d[n * 18 + 5] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_p[(n+1) * 9 + 2], PRIM_INT__s_p_s_d[n * 18 + 5]);
438 PRIM_INT__s_p_s_d[n * 18 + 5] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_s[n * 3 + 0]), PRIM_INT__s_p_s_d[n * 18 + 5]);
439
440 PRIM_INT__s_p_s_d[n * 18 + 6] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_p[n * 9 + 3]);
441 PRIM_INT__s_p_s_d[n * 18 + 6] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_p[(n+1) * 9 + 3], PRIM_INT__s_p_s_d[n * 18 + 6]);
442 PRIM_INT__s_p_s_d[n * 18 + 6] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_s[n * 3 + 1]), PRIM_INT__s_p_s_d[n * 18 + 6]);
443
444 PRIM_INT__s_p_s_d[n * 18 + 7] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 3]);
445 PRIM_INT__s_p_s_d[n * 18 + 7] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 3], PRIM_INT__s_p_s_d[n * 18 + 7]);
446 PRIM_INT__s_p_s_d[n * 18 + 7] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_p[(n+1) * 3 + 0], PRIM_INT__s_p_s_d[n * 18 + 7]);
447
448 PRIM_INT__s_p_s_d[n * 18 + 9] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 4]);
449 PRIM_INT__s_p_s_d[n * 18 + 9] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 4], PRIM_INT__s_p_s_d[n * 18 + 9]);
450 PRIM_INT__s_p_s_d[n * 18 + 9] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_s[n * 3 + 1]), PRIM_INT__s_p_s_d[n * 18 + 9]);
451 PRIM_INT__s_p_s_d[n * 18 + 9] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_p[(n+1) * 3 + 1], PRIM_INT__s_p_s_d[n * 18 + 9]);
452
453 PRIM_INT__s_p_s_d[n * 18 + 11] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_p[n * 9 + 5]);
454 PRIM_INT__s_p_s_d[n * 18 + 11] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_p[(n+1) * 9 + 5], PRIM_INT__s_p_s_d[n * 18 + 11]);
455 PRIM_INT__s_p_s_d[n * 18 + 11] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_s[n * 3 + 1]), PRIM_INT__s_p_s_d[n * 18 + 11]);
456
457 PRIM_INT__s_p_s_d[n * 18 + 12] = SIMINT_MUL(Q_PB[0], PRIM_INT__s_p_s_p[n * 9 + 6]);
458 PRIM_INT__s_p_s_d[n * 18 + 12] = SIMINT_FMADD( aoq_PQ[0], PRIM_INT__s_p_s_p[(n+1) * 9 + 6], PRIM_INT__s_p_s_d[n * 18 + 12]);
459 PRIM_INT__s_p_s_d[n * 18 + 12] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_s[n * 3 + 2]), PRIM_INT__s_p_s_d[n * 18 + 12]);
460
461 PRIM_INT__s_p_s_d[n * 18 + 13] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 6]);
462 PRIM_INT__s_p_s_d[n * 18 + 13] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 6], PRIM_INT__s_p_s_d[n * 18 + 13]);
463
464 PRIM_INT__s_p_s_d[n * 18 + 15] = SIMINT_MUL(Q_PB[1], PRIM_INT__s_p_s_p[n * 9 + 7]);
465 PRIM_INT__s_p_s_d[n * 18 + 15] = SIMINT_FMADD( aoq_PQ[1], PRIM_INT__s_p_s_p[(n+1) * 9 + 7], PRIM_INT__s_p_s_d[n * 18 + 15]);
466 PRIM_INT__s_p_s_d[n * 18 + 15] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_s[n * 3 + 2]), PRIM_INT__s_p_s_d[n * 18 + 15]);
467
468 PRIM_INT__s_p_s_d[n * 18 + 17] = SIMINT_MUL(Q_PB[2], PRIM_INT__s_p_s_p[n * 9 + 8]);
469 PRIM_INT__s_p_s_d[n * 18 + 17] = SIMINT_FMADD( aoq_PQ[2], PRIM_INT__s_p_s_p[(n+1) * 9 + 8], PRIM_INT__s_p_s_d[n * 18 + 17]);
470 PRIM_INT__s_p_s_d[n * 18 + 17] = SIMINT_FMADD( vrr_const_1_over_2q, SIMINT_FMADD(a_over_q, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_s[n * 3 + 2]), PRIM_INT__s_p_s_d[n * 18 + 17]);
471 PRIM_INT__s_p_s_d[n * 18 + 17] = SIMINT_FMADD( vrr_const_1_over_2pq, PRIM_INT__s_s_s_p[(n+1) * 3 + 2], PRIM_INT__s_p_s_d[n * 18 + 17]);
472
473 }
474
475
476 VRR_L_s_d_s_f(
477 PRIM_INT__s_d_s_f,
478 PRIM_INT__s_d_s_d,
479 PRIM_INT__s_d_s_p,
480 PRIM_INT__s_p_s_d,
481 Q_PB,
482 a_over_q,
483 aoq_PQ,
484 one_over_2pq,
485 one_over_2q,
486 1);
487
488
489
490 // Forming PRIM_INT__s_f_s_s[4 * 10];
491 for(n = 0; n < 4; ++n) // loop over orders of auxiliary function
492 {
493
494 PRIM_INT__s_f_s_s[n * 10 + 0] = SIMINT_MUL(P_PB[0], PRIM_INT__s_d_s_s[n * 6 + 0]);
495 PRIM_INT__s_f_s_s[n * 10 + 0] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_f_s_s[n * 10 + 0]);
496 PRIM_INT__s_f_s_s[n * 10 + 0] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_p_s_s[(n+1) * 3 + 0], PRIM_INT__s_p_s_s[n * 3 + 0]), PRIM_INT__s_f_s_s[n * 10 + 0]);
497
498 PRIM_INT__s_f_s_s[n * 10 + 1] = SIMINT_MUL(P_PB[1], PRIM_INT__s_d_s_s[n * 6 + 0]);
499 PRIM_INT__s_f_s_s[n * 10 + 1] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_f_s_s[n * 10 + 1]);
500
501 PRIM_INT__s_f_s_s[n * 10 + 2] = SIMINT_MUL(P_PB[2], PRIM_INT__s_d_s_s[n * 6 + 0]);
502 PRIM_INT__s_f_s_s[n * 10 + 2] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 0], PRIM_INT__s_f_s_s[n * 10 + 2]);
503
504 PRIM_INT__s_f_s_s[n * 10 + 3] = SIMINT_MUL(P_PB[0], PRIM_INT__s_d_s_s[n * 6 + 3]);
505 PRIM_INT__s_f_s_s[n * 10 + 3] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_f_s_s[n * 10 + 3]);
506
507 PRIM_INT__s_f_s_s[n * 10 + 4] = SIMINT_MUL(P_PB[2], PRIM_INT__s_d_s_s[n * 6 + 1]);
508 PRIM_INT__s_f_s_s[n * 10 + 4] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 1], PRIM_INT__s_f_s_s[n * 10 + 4]);
509
510 PRIM_INT__s_f_s_s[n * 10 + 5] = SIMINT_MUL(P_PB[0], PRIM_INT__s_d_s_s[n * 6 + 5]);
511 PRIM_INT__s_f_s_s[n * 10 + 5] = SIMINT_FMADD( aop_PQ[0], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_f_s_s[n * 10 + 5]);
512
513 PRIM_INT__s_f_s_s[n * 10 + 6] = SIMINT_MUL(P_PB[1], PRIM_INT__s_d_s_s[n * 6 + 3]);
514 PRIM_INT__s_f_s_s[n * 10 + 6] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_f_s_s[n * 10 + 6]);
515 PRIM_INT__s_f_s_s[n * 10 + 6] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_p_s_s[(n+1) * 3 + 1], PRIM_INT__s_p_s_s[n * 3 + 1]), PRIM_INT__s_f_s_s[n * 10 + 6]);
516
517 PRIM_INT__s_f_s_s[n * 10 + 7] = SIMINT_MUL(P_PB[2], PRIM_INT__s_d_s_s[n * 6 + 3]);
518 PRIM_INT__s_f_s_s[n * 10 + 7] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 3], PRIM_INT__s_f_s_s[n * 10 + 7]);
519
520 PRIM_INT__s_f_s_s[n * 10 + 8] = SIMINT_MUL(P_PB[1], PRIM_INT__s_d_s_s[n * 6 + 5]);
521 PRIM_INT__s_f_s_s[n * 10 + 8] = SIMINT_FMADD( aop_PQ[1], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_f_s_s[n * 10 + 8]);
522
523 PRIM_INT__s_f_s_s[n * 10 + 9] = SIMINT_MUL(P_PB[2], PRIM_INT__s_d_s_s[n * 6 + 5]);
524 PRIM_INT__s_f_s_s[n * 10 + 9] = SIMINT_FMADD( aop_PQ[2], PRIM_INT__s_d_s_s[(n+1) * 6 + 5], PRIM_INT__s_f_s_s[n * 10 + 9]);
525 PRIM_INT__s_f_s_s[n * 10 + 9] = SIMINT_FMADD( vrr_const_2_over_2p, SIMINT_FMADD(a_over_p, PRIM_INT__s_p_s_s[(n+1) * 3 + 2], PRIM_INT__s_p_s_s[n * 3 + 2]), PRIM_INT__s_f_s_s[n * 10 + 9]);
526
527 }
528
529
530 VRR_L_s_f_s_p(
531 PRIM_INT__s_f_s_p,
532 PRIM_INT__s_f_s_s,
533 PRIM_INT__s_d_s_s,
534 Q_PB,
535 aoq_PQ,
536 one_over_2pq,
537 3);
538
539
540 VRR_L_s_f_s_d(
541 PRIM_INT__s_f_s_d,
542 PRIM_INT__s_f_s_p,
543 PRIM_INT__s_f_s_s,
544 PRIM_INT__s_d_s_p,
545 Q_PB,
546 a_over_q,
547 aoq_PQ,
548 one_over_2pq,
549 one_over_2q,
550 2);
551
552
553 ostei_general_vrr_L(0, 3, 0, 3, 1,
554 one_over_2q, a_over_q, one_over_2pq, aoq_PQ, Q_PB,
555 PRIM_INT__s_f_s_d, NULL, PRIM_INT__s_f_s_p, NULL, PRIM_INT__s_d_s_d, PRIM_INT__s_f_s_f);
556
557
558
559
560 ////////////////////////////////////
561 // Accumulate contracted integrals
562 ////////////////////////////////////
563 if(lastoffset == 0)
564 {
565 contract_all(36, PRIM_INT__s_d_s_d, PRIM_PTR_INT__s_d_s_d);
566 contract_all(60, PRIM_INT__s_d_s_f, PRIM_PTR_INT__s_d_s_f);
567 contract_all(60, PRIM_INT__s_f_s_d, PRIM_PTR_INT__s_f_s_d);
568 contract_all(100, PRIM_INT__s_f_s_f, PRIM_PTR_INT__s_f_s_f);
569 }
570 else
571 {
572 contract(36, shelloffsets, PRIM_INT__s_d_s_d, PRIM_PTR_INT__s_d_s_d);
573 contract(60, shelloffsets, PRIM_INT__s_d_s_f, PRIM_PTR_INT__s_d_s_f);
574 contract(60, shelloffsets, PRIM_INT__s_f_s_d, PRIM_PTR_INT__s_f_s_d);
575 contract(100, shelloffsets, PRIM_INT__s_f_s_f, PRIM_PTR_INT__s_f_s_f);
576 PRIM_PTR_INT__s_d_s_d += lastoffset*36;
577 PRIM_PTR_INT__s_d_s_f += lastoffset*60;
578 PRIM_PTR_INT__s_f_s_d += lastoffset*60;
579 PRIM_PTR_INT__s_f_s_f += lastoffset*100;
580 }
581
582 } // close loop over j
583 } // close loop over i
584
585 //Advance to the next batch
586 jstart = SIMINT_SIMD_ROUND(jend);
587
588 //////////////////////////////////////////////
589 // Contracted integrals: Horizontal recurrance
590 //////////////////////////////////////////////
591
592
593 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
594
595
596 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
597 {
598 const double hCD[3] = { Q.AB_x[cd+abcd], Q.AB_y[cd+abcd], Q.AB_z[cd+abcd] };
599
600 // set up HRR pointers
601 double const * restrict HRR_INT__s_d_s_d = INT__s_d_s_d + abcd * 36;
602 double const * restrict HRR_INT__s_d_s_f = INT__s_d_s_f + abcd * 60;
603 double const * restrict HRR_INT__s_f_s_d = INT__s_f_s_d + abcd * 60;
604 double const * restrict HRR_INT__s_f_s_f = INT__s_f_s_f + abcd * 100;
605 double * restrict HRR_INT__p_d_p_d = INT__p_d_p_d + real_abcd * 324;
606
607 // form INT__p_d_s_d
608 for(iket = 0; iket < 6; ++iket)
609 {
610 HRR_INT__p_d_s_d[0 * 6 + iket] = HRR_INT__s_f_s_d[0 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[0 * 6 + iket] );
611
612 HRR_INT__p_d_s_d[1 * 6 + iket] = HRR_INT__s_f_s_d[1 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[1 * 6 + iket] );
613
614 HRR_INT__p_d_s_d[2 * 6 + iket] = HRR_INT__s_f_s_d[2 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[2 * 6 + iket] );
615
616 HRR_INT__p_d_s_d[3 * 6 + iket] = HRR_INT__s_f_s_d[3 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[3 * 6 + iket] );
617
618 HRR_INT__p_d_s_d[4 * 6 + iket] = HRR_INT__s_f_s_d[4 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[4 * 6 + iket] );
619
620 HRR_INT__p_d_s_d[5 * 6 + iket] = HRR_INT__s_f_s_d[5 * 6 + iket] - ( hAB[0] * HRR_INT__s_d_s_d[5 * 6 + iket] );
621
622 HRR_INT__p_d_s_d[6 * 6 + iket] = HRR_INT__s_f_s_d[1 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[0 * 6 + iket] );
623
624 HRR_INT__p_d_s_d[7 * 6 + iket] = HRR_INT__s_f_s_d[3 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[1 * 6 + iket] );
625
626 HRR_INT__p_d_s_d[8 * 6 + iket] = HRR_INT__s_f_s_d[4 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[2 * 6 + iket] );
627
628 HRR_INT__p_d_s_d[9 * 6 + iket] = HRR_INT__s_f_s_d[6 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[3 * 6 + iket] );
629
630 HRR_INT__p_d_s_d[10 * 6 + iket] = HRR_INT__s_f_s_d[7 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[4 * 6 + iket] );
631
632 HRR_INT__p_d_s_d[11 * 6 + iket] = HRR_INT__s_f_s_d[8 * 6 + iket] - ( hAB[1] * HRR_INT__s_d_s_d[5 * 6 + iket] );
633
634 HRR_INT__p_d_s_d[12 * 6 + iket] = HRR_INT__s_f_s_d[2 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[0 * 6 + iket] );
635
636 HRR_INT__p_d_s_d[13 * 6 + iket] = HRR_INT__s_f_s_d[4 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[1 * 6 + iket] );
637
638 HRR_INT__p_d_s_d[14 * 6 + iket] = HRR_INT__s_f_s_d[5 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[2 * 6 + iket] );
639
640 HRR_INT__p_d_s_d[15 * 6 + iket] = HRR_INT__s_f_s_d[7 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[3 * 6 + iket] );
641
642 HRR_INT__p_d_s_d[16 * 6 + iket] = HRR_INT__s_f_s_d[8 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[4 * 6 + iket] );
643
644 HRR_INT__p_d_s_d[17 * 6 + iket] = HRR_INT__s_f_s_d[9 * 6 + iket] - ( hAB[2] * HRR_INT__s_d_s_d[5 * 6 + iket] );
645
646 }
647
648
649 // form INT__p_d_s_f
650 for(iket = 0; iket < 10; ++iket)
651 {
652 HRR_INT__p_d_s_f[0 * 10 + iket] = HRR_INT__s_f_s_f[0 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[0 * 10 + iket] );
653
654 HRR_INT__p_d_s_f[1 * 10 + iket] = HRR_INT__s_f_s_f[1 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[1 * 10 + iket] );
655
656 HRR_INT__p_d_s_f[2 * 10 + iket] = HRR_INT__s_f_s_f[2 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[2 * 10 + iket] );
657
658 HRR_INT__p_d_s_f[3 * 10 + iket] = HRR_INT__s_f_s_f[3 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[3 * 10 + iket] );
659
660 HRR_INT__p_d_s_f[4 * 10 + iket] = HRR_INT__s_f_s_f[4 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[4 * 10 + iket] );
661
662 HRR_INT__p_d_s_f[5 * 10 + iket] = HRR_INT__s_f_s_f[5 * 10 + iket] - ( hAB[0] * HRR_INT__s_d_s_f[5 * 10 + iket] );
663
664 HRR_INT__p_d_s_f[6 * 10 + iket] = HRR_INT__s_f_s_f[1 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[0 * 10 + iket] );
665
666 HRR_INT__p_d_s_f[7 * 10 + iket] = HRR_INT__s_f_s_f[3 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[1 * 10 + iket] );
667
668 HRR_INT__p_d_s_f[8 * 10 + iket] = HRR_INT__s_f_s_f[4 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[2 * 10 + iket] );
669
670 HRR_INT__p_d_s_f[9 * 10 + iket] = HRR_INT__s_f_s_f[6 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[3 * 10 + iket] );
671
672 HRR_INT__p_d_s_f[10 * 10 + iket] = HRR_INT__s_f_s_f[7 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[4 * 10 + iket] );
673
674 HRR_INT__p_d_s_f[11 * 10 + iket] = HRR_INT__s_f_s_f[8 * 10 + iket] - ( hAB[1] * HRR_INT__s_d_s_f[5 * 10 + iket] );
675
676 HRR_INT__p_d_s_f[12 * 10 + iket] = HRR_INT__s_f_s_f[2 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[0 * 10 + iket] );
677
678 HRR_INT__p_d_s_f[13 * 10 + iket] = HRR_INT__s_f_s_f[4 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[1 * 10 + iket] );
679
680 HRR_INT__p_d_s_f[14 * 10 + iket] = HRR_INT__s_f_s_f[5 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[2 * 10 + iket] );
681
682 HRR_INT__p_d_s_f[15 * 10 + iket] = HRR_INT__s_f_s_f[7 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[3 * 10 + iket] );
683
684 HRR_INT__p_d_s_f[16 * 10 + iket] = HRR_INT__s_f_s_f[8 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[4 * 10 + iket] );
685
686 HRR_INT__p_d_s_f[17 * 10 + iket] = HRR_INT__s_f_s_f[9 * 10 + iket] - ( hAB[2] * HRR_INT__s_d_s_f[5 * 10 + iket] );
687
688 }
689
690
691 // form INT__p_d_p_d
692 for(ibra = 0; ibra < 18; ++ibra)
693 {
694 HRR_INT__p_d_p_d[ibra * 18 + 0] = HRR_INT__p_d_s_f[ibra * 10 + 0] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 0] );
695
696 HRR_INT__p_d_p_d[ibra * 18 + 1] = HRR_INT__p_d_s_f[ibra * 10 + 1] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 1] );
697
698 HRR_INT__p_d_p_d[ibra * 18 + 2] = HRR_INT__p_d_s_f[ibra * 10 + 2] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 2] );
699
700 HRR_INT__p_d_p_d[ibra * 18 + 3] = HRR_INT__p_d_s_f[ibra * 10 + 3] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 3] );
701
702 HRR_INT__p_d_p_d[ibra * 18 + 4] = HRR_INT__p_d_s_f[ibra * 10 + 4] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 4] );
703
704 HRR_INT__p_d_p_d[ibra * 18 + 5] = HRR_INT__p_d_s_f[ibra * 10 + 5] - ( hCD[0] * HRR_INT__p_d_s_d[ibra * 6 + 5] );
705
706 HRR_INT__p_d_p_d[ibra * 18 + 6] = HRR_INT__p_d_s_f[ibra * 10 + 1] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 0] );
707
708 HRR_INT__p_d_p_d[ibra * 18 + 7] = HRR_INT__p_d_s_f[ibra * 10 + 3] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 1] );
709
710 HRR_INT__p_d_p_d[ibra * 18 + 8] = HRR_INT__p_d_s_f[ibra * 10 + 4] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 2] );
711
712 HRR_INT__p_d_p_d[ibra * 18 + 9] = HRR_INT__p_d_s_f[ibra * 10 + 6] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 3] );
713
714 HRR_INT__p_d_p_d[ibra * 18 + 10] = HRR_INT__p_d_s_f[ibra * 10 + 7] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 4] );
715
716 HRR_INT__p_d_p_d[ibra * 18 + 11] = HRR_INT__p_d_s_f[ibra * 10 + 8] - ( hCD[1] * HRR_INT__p_d_s_d[ibra * 6 + 5] );
717
718 HRR_INT__p_d_p_d[ibra * 18 + 12] = HRR_INT__p_d_s_f[ibra * 10 + 2] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 0] );
719
720 HRR_INT__p_d_p_d[ibra * 18 + 13] = HRR_INT__p_d_s_f[ibra * 10 + 4] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 1] );
721
722 HRR_INT__p_d_p_d[ibra * 18 + 14] = HRR_INT__p_d_s_f[ibra * 10 + 5] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 2] );
723
724 HRR_INT__p_d_p_d[ibra * 18 + 15] = HRR_INT__p_d_s_f[ibra * 10 + 7] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 3] );
725
726 HRR_INT__p_d_p_d[ibra * 18 + 16] = HRR_INT__p_d_s_f[ibra * 10 + 8] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 4] );
727
728 HRR_INT__p_d_p_d[ibra * 18 + 17] = HRR_INT__p_d_s_f[ibra * 10 + 9] - ( hCD[2] * HRR_INT__p_d_s_d[ibra * 6 + 5] );
729
730 }
731
732
733 } // close HRR loop
734
735
736 } // close loop cdbatch
737
738 istart = iend;
739 } // close loop over ab
740
741 return P.nshell12_clip * Q.nshell12_clip;
742 }
743
744