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_i_s_s(struct simint_multi_shellpair const P,struct simint_multi_shellpair const Q,double screen_tol,double * const restrict work,double * const restrict INT__i_i_s_s)8 int ostei_i_i_s_s(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_i_s_s)
13 {
14
15 SIMINT_ASSUME_ALIGN_DBL(work);
16 SIMINT_ASSUME_ALIGN_DBL(INT__i_i_s_s);
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
27 // partition workspace
28 double * const INT__i_s_s_s = work + (SIMINT_NSHELL_SIMD * 0);
29 double * const INT__k_s_s_s = work + (SIMINT_NSHELL_SIMD * 28);
30 double * const INT__l_s_s_s = work + (SIMINT_NSHELL_SIMD * 64);
31 double * const INT__m_s_s_s = work + (SIMINT_NSHELL_SIMD * 109);
32 double * const INT__n_s_s_s = work + (SIMINT_NSHELL_SIMD * 164);
33 double * const INT__o_s_s_s = work + (SIMINT_NSHELL_SIMD * 230);
34 double * const INT__q_s_s_s = work + (SIMINT_NSHELL_SIMD * 308);
35 SIMINT_DBLTYPE * const primwork = (SIMINT_DBLTYPE *)(work + SIMINT_NSHELL_SIMD*399);
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 + 13;
38 SIMINT_DBLTYPE * const restrict PRIM_INT__d_s_s_s = primwork + 49;
39 SIMINT_DBLTYPE * const restrict PRIM_INT__f_s_s_s = primwork + 115;
40 SIMINT_DBLTYPE * const restrict PRIM_INT__g_s_s_s = primwork + 215;
41 SIMINT_DBLTYPE * const restrict PRIM_INT__h_s_s_s = primwork + 350;
42 SIMINT_DBLTYPE * const restrict PRIM_INT__i_s_s_s = primwork + 518;
43 SIMINT_DBLTYPE * const restrict PRIM_INT__k_s_s_s = primwork + 714;
44 SIMINT_DBLTYPE * const restrict PRIM_INT__l_s_s_s = primwork + 930;
45 SIMINT_DBLTYPE * const restrict PRIM_INT__m_s_s_s = primwork + 1155;
46 SIMINT_DBLTYPE * const restrict PRIM_INT__n_s_s_s = primwork + 1375;
47 SIMINT_DBLTYPE * const restrict PRIM_INT__o_s_s_s = primwork + 1573;
48 SIMINT_DBLTYPE * const restrict PRIM_INT__q_s_s_s = primwork + 1729;
49 double * const hrrwork = (double *)(primwork + 1820);
50 double * const HRR_INT__i_p_s_s = hrrwork + 0;
51 double * const HRR_INT__i_d_s_s = hrrwork + 84;
52 double * const HRR_INT__i_f_s_s = hrrwork + 252;
53 double * const HRR_INT__i_g_s_s = hrrwork + 532;
54 double * const HRR_INT__i_h_s_s = hrrwork + 952;
55 double * const HRR_INT__k_p_s_s = hrrwork + 1540;
56 double * const HRR_INT__k_d_s_s = hrrwork + 1648;
57 double * const HRR_INT__k_f_s_s = hrrwork + 1864;
58 double * const HRR_INT__k_g_s_s = hrrwork + 2224;
59 double * const HRR_INT__k_h_s_s = hrrwork + 2764;
60 double * const HRR_INT__l_p_s_s = hrrwork + 3520;
61 double * const HRR_INT__l_d_s_s = hrrwork + 3655;
62 double * const HRR_INT__l_f_s_s = hrrwork + 3925;
63 double * const HRR_INT__l_g_s_s = hrrwork + 4375;
64 double * const HRR_INT__m_p_s_s = hrrwork + 5050;
65 double * const HRR_INT__m_d_s_s = hrrwork + 5215;
66 double * const HRR_INT__m_f_s_s = hrrwork + 5545;
67 double * const HRR_INT__n_p_s_s = hrrwork + 6095;
68 double * const HRR_INT__n_d_s_s = hrrwork + 6293;
69 double * const HRR_INT__o_p_s_s = hrrwork + 6689;
70
71
72 // Create constants
73 const SIMINT_DBLTYPE const_1 = SIMINT_DBLSET1(1);
74 const SIMINT_DBLTYPE const_10 = SIMINT_DBLSET1(10);
75 const SIMINT_DBLTYPE const_11 = SIMINT_DBLSET1(11);
76 const SIMINT_DBLTYPE const_2 = SIMINT_DBLSET1(2);
77 const SIMINT_DBLTYPE const_3 = SIMINT_DBLSET1(3);
78 const SIMINT_DBLTYPE const_4 = SIMINT_DBLSET1(4);
79 const SIMINT_DBLTYPE const_5 = SIMINT_DBLSET1(5);
80 const SIMINT_DBLTYPE const_6 = SIMINT_DBLSET1(6);
81 const SIMINT_DBLTYPE const_7 = SIMINT_DBLSET1(7);
82 const SIMINT_DBLTYPE const_8 = SIMINT_DBLSET1(8);
83 const SIMINT_DBLTYPE const_9 = SIMINT_DBLSET1(9);
84 const SIMINT_DBLTYPE one_half = SIMINT_DBLSET1(0.5);
85
86
87 ////////////////////////////////////////
88 // Loop over shells and primitives
89 ////////////////////////////////////////
90
91 real_abcd = 0;
92 istart = 0;
93 for(ab = 0; ab < P.nshell12_clip; ++ab)
94 {
95 const int iend = istart + P.nprim12[ab];
96
97 cd = 0;
98 jstart = 0;
99
100 for(cd = 0; cd < Q.nshell12_clip; cd += SIMINT_NSHELL_SIMD)
101 {
102 const int nshellbatch = ((cd + SIMINT_NSHELL_SIMD) > Q.nshell12_clip) ? Q.nshell12_clip - cd : SIMINT_NSHELL_SIMD;
103 int jend = jstart;
104 for(i = 0; i < nshellbatch; i++)
105 jend += Q.nprim12[cd+i];
106
107 // Clear the beginning of the workspace (where we are accumulating integrals)
108 memset(work, 0, SIMINT_NSHELL_SIMD * 399 * sizeof(double));
109 abcd = 0;
110
111
112 for(i = istart; i < iend; ++i)
113 {
114 SIMINT_DBLTYPE bra_screen_max; // only used if check_screen
115
116 if(check_screen)
117 {
118 // Skip this whole thing if always insignificant
119 if((P.screen[i] * Q.screen_max) < screen_tol)
120 continue;
121 bra_screen_max = SIMINT_DBLSET1(P.screen[i]);
122 }
123
124 icd = 0;
125 iprimcd = 0;
126 nprim_icd = Q.nprim12[cd];
127 double * restrict PRIM_PTR_INT__i_s_s_s = INT__i_s_s_s + abcd * 28;
128 double * restrict PRIM_PTR_INT__k_s_s_s = INT__k_s_s_s + abcd * 36;
129 double * restrict PRIM_PTR_INT__l_s_s_s = INT__l_s_s_s + abcd * 45;
130 double * restrict PRIM_PTR_INT__m_s_s_s = INT__m_s_s_s + abcd * 55;
131 double * restrict PRIM_PTR_INT__n_s_s_s = INT__n_s_s_s + abcd * 66;
132 double * restrict PRIM_PTR_INT__o_s_s_s = INT__o_s_s_s + abcd * 78;
133 double * restrict PRIM_PTR_INT__q_s_s_s = INT__q_s_s_s + abcd * 91;
134
135
136
137 // Load these one per loop over i
138 const SIMINT_DBLTYPE P_alpha = SIMINT_DBLSET1(P.alpha[i]);
139 const SIMINT_DBLTYPE P_prefac = SIMINT_DBLSET1(P.prefac[i]);
140 const SIMINT_DBLTYPE Pxyz[3] = { SIMINT_DBLSET1(P.x[i]), SIMINT_DBLSET1(P.y[i]), SIMINT_DBLSET1(P.z[i]) };
141
142 const SIMINT_DBLTYPE P_PA[3] = { SIMINT_DBLSET1(P.PA_x[i]), SIMINT_DBLSET1(P.PA_y[i]), SIMINT_DBLSET1(P.PA_z[i]) };
143
144 for(j = jstart; j < jend; j += SIMINT_SIMD_LEN)
145 {
146 // calculate the shell offsets
147 // these are the offset from the shell pointed to by cd
148 // for each element
149 int shelloffsets[SIMINT_SIMD_LEN] = {0};
150 int lastoffset = 0;
151 const int nlane = ( ((j + SIMINT_SIMD_LEN) < jend) ? SIMINT_SIMD_LEN : (jend - j));
152
153 if((iprimcd + SIMINT_SIMD_LEN) >= nprim_icd)
154 {
155 // Handle if the first element of the vector is a new shell
156 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
157 {
158 nprim_icd += Q.nprim12[cd + (++icd)];
159 PRIM_PTR_INT__i_s_s_s += 28;
160 PRIM_PTR_INT__k_s_s_s += 36;
161 PRIM_PTR_INT__l_s_s_s += 45;
162 PRIM_PTR_INT__m_s_s_s += 55;
163 PRIM_PTR_INT__n_s_s_s += 66;
164 PRIM_PTR_INT__o_s_s_s += 78;
165 PRIM_PTR_INT__q_s_s_s += 91;
166 }
167 iprimcd++;
168 for(n = 1; n < SIMINT_SIMD_LEN; ++n)
169 {
170 if(iprimcd >= nprim_icd && ((icd+1) < nshellbatch))
171 {
172 shelloffsets[n] = shelloffsets[n-1] + 1;
173 lastoffset++;
174 nprim_icd += Q.nprim12[cd + (++icd)];
175 }
176 else
177 shelloffsets[n] = shelloffsets[n-1];
178 iprimcd++;
179 }
180 }
181 else
182 iprimcd += SIMINT_SIMD_LEN;
183
184 // Do we have to compute this vector (or has it been screened out)?
185 // (not_screened != 0 means we have to do this vector)
186 if(check_screen)
187 {
188 const double vmax = vector_max(SIMINT_MUL(bra_screen_max, SIMINT_DBLLOAD(Q.screen, j)));
189 if(vmax < screen_tol)
190 {
191 PRIM_PTR_INT__i_s_s_s += lastoffset*28;
192 PRIM_PTR_INT__k_s_s_s += lastoffset*36;
193 PRIM_PTR_INT__l_s_s_s += lastoffset*45;
194 PRIM_PTR_INT__m_s_s_s += lastoffset*55;
195 PRIM_PTR_INT__n_s_s_s += lastoffset*66;
196 PRIM_PTR_INT__o_s_s_s += lastoffset*78;
197 PRIM_PTR_INT__q_s_s_s += lastoffset*91;
198 continue;
199 }
200 }
201
202 const SIMINT_DBLTYPE Q_alpha = SIMINT_DBLLOAD(Q.alpha, j);
203 const SIMINT_DBLTYPE PQalpha_mul = SIMINT_MUL(P_alpha, Q_alpha);
204 const SIMINT_DBLTYPE PQalpha_sum = SIMINT_ADD(P_alpha, Q_alpha);
205 const SIMINT_DBLTYPE one_over_PQalpha_sum = SIMINT_DIV(const_1, PQalpha_sum);
206
207
208 /* construct R2 = (Px - Qx)**2 + (Py - Qy)**2 + (Pz -Qz)**2 */
209 SIMINT_DBLTYPE PQ[3];
210 PQ[0] = SIMINT_SUB(Pxyz[0], SIMINT_DBLLOAD(Q.x, j));
211 PQ[1] = SIMINT_SUB(Pxyz[1], SIMINT_DBLLOAD(Q.y, j));
212 PQ[2] = SIMINT_SUB(Pxyz[2], SIMINT_DBLLOAD(Q.z, j));
213 SIMINT_DBLTYPE R2 = SIMINT_MUL(PQ[0], PQ[0]);
214 R2 = SIMINT_FMADD(PQ[1], PQ[1], R2);
215 R2 = SIMINT_FMADD(PQ[2], PQ[2], R2);
216
217 const SIMINT_DBLTYPE alpha = SIMINT_MUL(PQalpha_mul, one_over_PQalpha_sum); // alpha from MEST
218 const SIMINT_DBLTYPE one_over_p = SIMINT_DIV(const_1, P_alpha);
219 const SIMINT_DBLTYPE one_over_q = SIMINT_DIV(const_1, Q_alpha);
220 const SIMINT_DBLTYPE one_over_2p = SIMINT_MUL(one_half, one_over_p);
221 const SIMINT_DBLTYPE one_over_2q = SIMINT_MUL(one_half, one_over_q);
222 const SIMINT_DBLTYPE one_over_2pq = SIMINT_MUL(one_half, one_over_PQalpha_sum);
223
224 // NOTE: Minus sign!
225 const SIMINT_DBLTYPE a_over_p = SIMINT_MUL(SIMINT_NEG(alpha), one_over_p);
226 SIMINT_DBLTYPE aop_PQ[3];
227 aop_PQ[0] = SIMINT_MUL(a_over_p, PQ[0]);
228 aop_PQ[1] = SIMINT_MUL(a_over_p, PQ[1]);
229 aop_PQ[2] = SIMINT_MUL(a_over_p, PQ[2]);
230
231
232 //////////////////////////////////////////////
233 // Fjt function section
234 // Maximum v value: 12
235 //////////////////////////////////////////////
236 // The parameter to the Fjt function
237 const SIMINT_DBLTYPE F_x = SIMINT_MUL(R2, alpha);
238
239
240 const SIMINT_DBLTYPE Q_prefac = mask_load(nlane, Q.prefac + j);
241
242
243 boys_F_split(PRIM_INT__s_s_s_s, F_x, 12);
244 SIMINT_DBLTYPE prefac = SIMINT_SQRT(one_over_PQalpha_sum);
245 prefac = SIMINT_MUL(SIMINT_MUL(P_prefac, Q_prefac), prefac);
246 for(n = 0; n <= 12; n++)
247 PRIM_INT__s_s_s_s[n] = SIMINT_MUL(PRIM_INT__s_s_s_s[n], prefac);
248
249 //////////////////////////////////////////////
250 // Primitive integrals: Vertical recurrance
251 //////////////////////////////////////////////
252
253 const SIMINT_DBLTYPE vrr_const_1_over_2p = one_over_2p;
254 const SIMINT_DBLTYPE vrr_const_2_over_2p = SIMINT_MUL(const_2, one_over_2p);
255 const SIMINT_DBLTYPE vrr_const_3_over_2p = SIMINT_MUL(const_3, one_over_2p);
256 const SIMINT_DBLTYPE vrr_const_4_over_2p = SIMINT_MUL(const_4, one_over_2p);
257 const SIMINT_DBLTYPE vrr_const_5_over_2p = SIMINT_MUL(const_5, one_over_2p);
258 const SIMINT_DBLTYPE vrr_const_6_over_2p = SIMINT_MUL(const_6, one_over_2p);
259 const SIMINT_DBLTYPE vrr_const_7_over_2p = SIMINT_MUL(const_7, one_over_2p);
260 const SIMINT_DBLTYPE vrr_const_8_over_2p = SIMINT_MUL(const_8, one_over_2p);
261 const SIMINT_DBLTYPE vrr_const_9_over_2p = SIMINT_MUL(const_9, one_over_2p);
262 const SIMINT_DBLTYPE vrr_const_10_over_2p = SIMINT_MUL(const_10, one_over_2p);
263 const SIMINT_DBLTYPE vrr_const_11_over_2p = SIMINT_MUL(const_11, one_over_2p);
264
265
266
267 // Forming PRIM_INT__p_s_s_s[12 * 3];
268 for(n = 0; n < 12; ++n) // loop over orders of auxiliary function
269 {
270
271 PRIM_INT__p_s_s_s[n * 3 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__s_s_s_s[n * 1 + 0]);
272 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]);
273
274 PRIM_INT__p_s_s_s[n * 3 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__s_s_s_s[n * 1 + 0]);
275 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]);
276
277 PRIM_INT__p_s_s_s[n * 3 + 2] = SIMINT_MUL(P_PA[2], PRIM_INT__s_s_s_s[n * 1 + 0]);
278 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]);
279
280 }
281
282
283
284 // Forming PRIM_INT__d_s_s_s[11 * 6];
285 for(n = 0; n < 11; ++n) // loop over orders of auxiliary function
286 {
287
288 PRIM_INT__d_s_s_s[n * 6 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__p_s_s_s[n * 3 + 0]);
289 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]);
290 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]);
291
292 PRIM_INT__d_s_s_s[n * 6 + 3] = SIMINT_MUL(P_PA[1], PRIM_INT__p_s_s_s[n * 3 + 1]);
293 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]);
294 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]);
295
296 PRIM_INT__d_s_s_s[n * 6 + 5] = SIMINT_MUL(P_PA[2], PRIM_INT__p_s_s_s[n * 3 + 2]);
297 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]);
298 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]);
299
300 }
301
302
303
304 // Forming PRIM_INT__f_s_s_s[10 * 10];
305 for(n = 0; n < 10; ++n) // loop over orders of auxiliary function
306 {
307
308 PRIM_INT__f_s_s_s[n * 10 + 0] = SIMINT_MUL(P_PA[0], PRIM_INT__d_s_s_s[n * 6 + 0]);
309 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]);
310 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]);
311
312 PRIM_INT__f_s_s_s[n * 10 + 1] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 0]);
313 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]);
314
315 PRIM_INT__f_s_s_s[n * 10 + 6] = SIMINT_MUL(P_PA[1], PRIM_INT__d_s_s_s[n * 6 + 3]);
316 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]);
317 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]);
318
319 PRIM_INT__f_s_s_s[n * 10 + 9] = SIMINT_MUL(P_PA[2], PRIM_INT__d_s_s_s[n * 6 + 5]);
320 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]);
321 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]);
322
323 }
324
325
326 VRR_I_g_s_s_s(
327 PRIM_INT__g_s_s_s,
328 PRIM_INT__f_s_s_s,
329 PRIM_INT__d_s_s_s,
330 P_PA,
331 a_over_p,
332 aop_PQ,
333 one_over_2p,
334 9);
335
336
337 VRR_I_h_s_s_s(
338 PRIM_INT__h_s_s_s,
339 PRIM_INT__g_s_s_s,
340 PRIM_INT__f_s_s_s,
341 P_PA,
342 a_over_p,
343 aop_PQ,
344 one_over_2p,
345 8);
346
347
348 ostei_general_vrr1_I(6, 7,
349 one_over_2p, a_over_p, aop_PQ, P_PA,
350 PRIM_INT__h_s_s_s, PRIM_INT__g_s_s_s, PRIM_INT__i_s_s_s);
351
352
353 ostei_general_vrr1_I(7, 6,
354 one_over_2p, a_over_p, aop_PQ, P_PA,
355 PRIM_INT__i_s_s_s, PRIM_INT__h_s_s_s, PRIM_INT__k_s_s_s);
356
357
358 ostei_general_vrr1_I(8, 5,
359 one_over_2p, a_over_p, aop_PQ, P_PA,
360 PRIM_INT__k_s_s_s, PRIM_INT__i_s_s_s, PRIM_INT__l_s_s_s);
361
362
363 ostei_general_vrr1_I(9, 4,
364 one_over_2p, a_over_p, aop_PQ, P_PA,
365 PRIM_INT__l_s_s_s, PRIM_INT__k_s_s_s, PRIM_INT__m_s_s_s);
366
367
368 ostei_general_vrr1_I(10, 3,
369 one_over_2p, a_over_p, aop_PQ, P_PA,
370 PRIM_INT__m_s_s_s, PRIM_INT__l_s_s_s, PRIM_INT__n_s_s_s);
371
372
373 ostei_general_vrr1_I(11, 2,
374 one_over_2p, a_over_p, aop_PQ, P_PA,
375 PRIM_INT__n_s_s_s, PRIM_INT__m_s_s_s, PRIM_INT__o_s_s_s);
376
377
378 ostei_general_vrr1_I(12, 1,
379 one_over_2p, a_over_p, aop_PQ, P_PA,
380 PRIM_INT__o_s_s_s, PRIM_INT__n_s_s_s, PRIM_INT__q_s_s_s);
381
382
383
384
385 ////////////////////////////////////
386 // Accumulate contracted integrals
387 ////////////////////////////////////
388 if(lastoffset == 0)
389 {
390 contract_all(28, PRIM_INT__i_s_s_s, PRIM_PTR_INT__i_s_s_s);
391 contract_all(36, PRIM_INT__k_s_s_s, PRIM_PTR_INT__k_s_s_s);
392 contract_all(45, PRIM_INT__l_s_s_s, PRIM_PTR_INT__l_s_s_s);
393 contract_all(55, PRIM_INT__m_s_s_s, PRIM_PTR_INT__m_s_s_s);
394 contract_all(66, PRIM_INT__n_s_s_s, PRIM_PTR_INT__n_s_s_s);
395 contract_all(78, PRIM_INT__o_s_s_s, PRIM_PTR_INT__o_s_s_s);
396 contract_all(91, PRIM_INT__q_s_s_s, PRIM_PTR_INT__q_s_s_s);
397 }
398 else
399 {
400 contract(28, shelloffsets, PRIM_INT__i_s_s_s, PRIM_PTR_INT__i_s_s_s);
401 contract(36, shelloffsets, PRIM_INT__k_s_s_s, PRIM_PTR_INT__k_s_s_s);
402 contract(45, shelloffsets, PRIM_INT__l_s_s_s, PRIM_PTR_INT__l_s_s_s);
403 contract(55, shelloffsets, PRIM_INT__m_s_s_s, PRIM_PTR_INT__m_s_s_s);
404 contract(66, shelloffsets, PRIM_INT__n_s_s_s, PRIM_PTR_INT__n_s_s_s);
405 contract(78, shelloffsets, PRIM_INT__o_s_s_s, PRIM_PTR_INT__o_s_s_s);
406 contract(91, shelloffsets, PRIM_INT__q_s_s_s, PRIM_PTR_INT__q_s_s_s);
407 PRIM_PTR_INT__i_s_s_s += lastoffset*28;
408 PRIM_PTR_INT__k_s_s_s += lastoffset*36;
409 PRIM_PTR_INT__l_s_s_s += lastoffset*45;
410 PRIM_PTR_INT__m_s_s_s += lastoffset*55;
411 PRIM_PTR_INT__n_s_s_s += lastoffset*66;
412 PRIM_PTR_INT__o_s_s_s += lastoffset*78;
413 PRIM_PTR_INT__q_s_s_s += lastoffset*91;
414 }
415
416 } // close loop over j
417 } // close loop over i
418
419 //Advance to the next batch
420 jstart = SIMINT_SIMD_ROUND(jend);
421
422 //////////////////////////////////////////////
423 // Contracted integrals: Horizontal recurrance
424 //////////////////////////////////////////////
425
426
427 const double hAB[3] = { P.AB_x[ab], P.AB_y[ab], P.AB_z[ab] };
428
429
430 for(abcd = 0; abcd < nshellbatch; ++abcd, ++real_abcd)
431 {
432
433 // set up HRR pointers
434 double const * restrict HRR_INT__i_s_s_s = INT__i_s_s_s + abcd * 28;
435 double const * restrict HRR_INT__k_s_s_s = INT__k_s_s_s + abcd * 36;
436 double const * restrict HRR_INT__l_s_s_s = INT__l_s_s_s + abcd * 45;
437 double const * restrict HRR_INT__m_s_s_s = INT__m_s_s_s + abcd * 55;
438 double const * restrict HRR_INT__n_s_s_s = INT__n_s_s_s + abcd * 66;
439 double const * restrict HRR_INT__o_s_s_s = INT__o_s_s_s + abcd * 78;
440 double const * restrict HRR_INT__q_s_s_s = INT__q_s_s_s + abcd * 91;
441 double * restrict HRR_INT__i_i_s_s = INT__i_i_s_s + real_abcd * 784;
442
443 // form INT__i_p_s_s
444 ostei_general_hrr_J(6, 1, 0, 0, hAB, HRR_INT__k_s_s_s, HRR_INT__i_s_s_s, HRR_INT__i_p_s_s);
445
446 // form INT__k_p_s_s
447 ostei_general_hrr_J(7, 1, 0, 0, hAB, HRR_INT__l_s_s_s, HRR_INT__k_s_s_s, HRR_INT__k_p_s_s);
448
449 // form INT__l_p_s_s
450 ostei_general_hrr_J(8, 1, 0, 0, hAB, HRR_INT__m_s_s_s, HRR_INT__l_s_s_s, HRR_INT__l_p_s_s);
451
452 // form INT__m_p_s_s
453 ostei_general_hrr_J(9, 1, 0, 0, hAB, HRR_INT__n_s_s_s, HRR_INT__m_s_s_s, HRR_INT__m_p_s_s);
454
455 // form INT__n_p_s_s
456 ostei_general_hrr_J(10, 1, 0, 0, hAB, HRR_INT__o_s_s_s, HRR_INT__n_s_s_s, HRR_INT__n_p_s_s);
457
458 // form INT__o_p_s_s
459 ostei_general_hrr_J(11, 1, 0, 0, hAB, HRR_INT__q_s_s_s, HRR_INT__o_s_s_s, HRR_INT__o_p_s_s);
460
461 // form INT__i_d_s_s
462 ostei_general_hrr_J(6, 2, 0, 0, hAB, HRR_INT__k_p_s_s, HRR_INT__i_p_s_s, HRR_INT__i_d_s_s);
463
464 // form INT__k_d_s_s
465 ostei_general_hrr_J(7, 2, 0, 0, hAB, HRR_INT__l_p_s_s, HRR_INT__k_p_s_s, HRR_INT__k_d_s_s);
466
467 // form INT__l_d_s_s
468 ostei_general_hrr_J(8, 2, 0, 0, hAB, HRR_INT__m_p_s_s, HRR_INT__l_p_s_s, HRR_INT__l_d_s_s);
469
470 // form INT__m_d_s_s
471 ostei_general_hrr_J(9, 2, 0, 0, hAB, HRR_INT__n_p_s_s, HRR_INT__m_p_s_s, HRR_INT__m_d_s_s);
472
473 // form INT__n_d_s_s
474 ostei_general_hrr_J(10, 2, 0, 0, hAB, HRR_INT__o_p_s_s, HRR_INT__n_p_s_s, HRR_INT__n_d_s_s);
475
476 // form INT__i_f_s_s
477 ostei_general_hrr_J(6, 3, 0, 0, hAB, HRR_INT__k_d_s_s, HRR_INT__i_d_s_s, HRR_INT__i_f_s_s);
478
479 // form INT__k_f_s_s
480 ostei_general_hrr_J(7, 3, 0, 0, hAB, HRR_INT__l_d_s_s, HRR_INT__k_d_s_s, HRR_INT__k_f_s_s);
481
482 // form INT__l_f_s_s
483 ostei_general_hrr_J(8, 3, 0, 0, hAB, HRR_INT__m_d_s_s, HRR_INT__l_d_s_s, HRR_INT__l_f_s_s);
484
485 // form INT__m_f_s_s
486 ostei_general_hrr_J(9, 3, 0, 0, hAB, HRR_INT__n_d_s_s, HRR_INT__m_d_s_s, HRR_INT__m_f_s_s);
487
488 // form INT__i_g_s_s
489 ostei_general_hrr_J(6, 4, 0, 0, hAB, HRR_INT__k_f_s_s, HRR_INT__i_f_s_s, HRR_INT__i_g_s_s);
490
491 // form INT__k_g_s_s
492 ostei_general_hrr_J(7, 4, 0, 0, hAB, HRR_INT__l_f_s_s, HRR_INT__k_f_s_s, HRR_INT__k_g_s_s);
493
494 // form INT__l_g_s_s
495 ostei_general_hrr_J(8, 4, 0, 0, hAB, HRR_INT__m_f_s_s, HRR_INT__l_f_s_s, HRR_INT__l_g_s_s);
496
497 // form INT__i_h_s_s
498 ostei_general_hrr_J(6, 5, 0, 0, hAB, HRR_INT__k_g_s_s, HRR_INT__i_g_s_s, HRR_INT__i_h_s_s);
499
500 // form INT__k_h_s_s
501 ostei_general_hrr_J(7, 5, 0, 0, hAB, HRR_INT__l_g_s_s, HRR_INT__k_g_s_s, HRR_INT__k_h_s_s);
502
503 // form INT__i_i_s_s
504 ostei_general_hrr_J(6, 6, 0, 0, hAB, HRR_INT__k_h_s_s, HRR_INT__i_h_s_s, HRR_INT__i_i_s_s);
505
506
507 } // close HRR loop
508
509
510 } // close loop cdbatch
511
512 istart = iend;
513 } // close loop over ab
514
515 return P.nshell12_clip * Q.nshell12_clip;
516 }
517
518