File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwBham_GeomW3W3_c.c |
Location: | line 120, column 5 |
Description: | Value stored to 'sh_ewald' is never read |
1 | /* |
2 | * This file is part of the GROMACS molecular simulation package. |
3 | * |
4 | * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by |
5 | * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl, |
6 | * and including many others, as listed in the AUTHORS file in the |
7 | * top-level source directory and at http://www.gromacs.org. |
8 | * |
9 | * GROMACS is free software; you can redistribute it and/or |
10 | * modify it under the terms of the GNU Lesser General Public License |
11 | * as published by the Free Software Foundation; either version 2.1 |
12 | * of the License, or (at your option) any later version. |
13 | * |
14 | * GROMACS is distributed in the hope that it will be useful, |
15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
17 | * Lesser General Public License for more details. |
18 | * |
19 | * You should have received a copy of the GNU Lesser General Public |
20 | * License along with GROMACS; if not, see |
21 | * http://www.gnu.org/licenses, or write to the Free Software Foundation, |
22 | * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. |
23 | * |
24 | * If you want to redistribute modifications to GROMACS, please |
25 | * consider that scientific software is very special. Version |
26 | * control is crucial - bugs must be traceable. We will be happy to |
27 | * consider code for inclusion in the official distribution, but |
28 | * derived work must not be called official GROMACS. Details are found |
29 | * in the README & COPYING files - if they are missing, get the |
30 | * official version at http://www.gromacs.org. |
31 | * |
32 | * To help us fund GROMACS development, we humbly ask that you cite |
33 | * the research papers on the package. Check out http://www.gromacs.org. |
34 | */ |
35 | /* |
36 | * Note: this file was generated by the GROMACS c kernel generator. |
37 | */ |
38 | #ifdef HAVE_CONFIG_H1 |
39 | #include <config.h> |
40 | #endif |
41 | |
42 | #include <math.h> |
43 | |
44 | #include "../nb_kernel.h" |
45 | #include "types/simple.h" |
46 | #include "gromacs/math/vec.h" |
47 | #include "nrnb.h" |
48 | |
49 | /* |
50 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwBham_GeomW3W3_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: Buckingham |
53 | * Geometry: Water3-Water3 |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEw_VdwBham_GeomW3W3_VF_c |
58 | (t_nblist * gmx_restrict__restrict nlist, |
59 | rvec * gmx_restrict__restrict xx, |
60 | rvec * gmx_restrict__restrict ff, |
61 | t_forcerec * gmx_restrict__restrict fr, |
62 | t_mdatoms * gmx_restrict__restrict mdatoms, |
63 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
64 | t_nrnb * gmx_restrict__restrict nrnb) |
65 | { |
66 | int i_shift_offset,i_coord_offset,j_coord_offset; |
67 | int j_index_start,j_index_end; |
68 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
69 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
70 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
71 | real *shiftvec,*fshift,*x,*f; |
72 | int vdwioffset0; |
73 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
74 | int vdwioffset1; |
75 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
76 | int vdwioffset2; |
77 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
78 | int vdwjidx0; |
79 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
80 | int vdwjidx1; |
81 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
82 | int vdwjidx2; |
83 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
84 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
85 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
86 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
87 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
88 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
89 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
90 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
91 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
92 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
93 | real velec,felec,velecsum,facel,crf,krf,krf2; |
94 | real *charge; |
95 | int nvdwtype; |
96 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
97 | int *vdwtype; |
98 | real *vdwparam; |
99 | int ewitab; |
100 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
101 | real *ewtab; |
102 | |
103 | x = xx[0]; |
104 | f = ff[0]; |
105 | |
106 | nri = nlist->nri; |
107 | iinr = nlist->iinr; |
108 | jindex = nlist->jindex; |
109 | jjnr = nlist->jjnr; |
110 | shiftidx = nlist->shift; |
111 | gid = nlist->gid; |
112 | shiftvec = fr->shift_vec[0]; |
113 | fshift = fr->fshift[0]; |
114 | facel = fr->epsfac; |
115 | charge = mdatoms->chargeA; |
116 | nvdwtype = fr->ntype; |
117 | vdwparam = fr->nbfp; |
118 | vdwtype = mdatoms->typeA; |
119 | |
120 | sh_ewald = fr->ic->sh_ewald; |
Value stored to 'sh_ewald' is never read | |
121 | ewtab = fr->ic->tabq_coul_FDV0; |
122 | ewtabscale = fr->ic->tabq_scale; |
123 | ewtabhalfspace = 0.5/ewtabscale; |
124 | |
125 | /* Setup water-specific parameters */ |
126 | inr = nlist->iinr[0]; |
127 | iq0 = facel*charge[inr+0]; |
128 | iq1 = facel*charge[inr+1]; |
129 | iq2 = facel*charge[inr+2]; |
130 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
131 | |
132 | jq0 = charge[inr+0]; |
133 | jq1 = charge[inr+1]; |
134 | jq2 = charge[inr+2]; |
135 | vdwjidx0 = 3*vdwtype[inr+0]; |
136 | qq00 = iq0*jq0; |
137 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
138 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
139 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
140 | qq01 = iq0*jq1; |
141 | qq02 = iq0*jq2; |
142 | qq10 = iq1*jq0; |
143 | qq11 = iq1*jq1; |
144 | qq12 = iq1*jq2; |
145 | qq20 = iq2*jq0; |
146 | qq21 = iq2*jq1; |
147 | qq22 = iq2*jq2; |
148 | |
149 | outeriter = 0; |
150 | inneriter = 0; |
151 | |
152 | /* Start outer loop over neighborlists */ |
153 | for(iidx=0; iidx<nri; iidx++) |
154 | { |
155 | /* Load shift vector for this list */ |
156 | i_shift_offset = DIM3*shiftidx[iidx]; |
157 | shX = shiftvec[i_shift_offset+XX0]; |
158 | shY = shiftvec[i_shift_offset+YY1]; |
159 | shZ = shiftvec[i_shift_offset+ZZ2]; |
160 | |
161 | /* Load limits for loop over neighbors */ |
162 | j_index_start = jindex[iidx]; |
163 | j_index_end = jindex[iidx+1]; |
164 | |
165 | /* Get outer coordinate index */ |
166 | inr = iinr[iidx]; |
167 | i_coord_offset = DIM3*inr; |
168 | |
169 | /* Load i particle coords and add shift vector */ |
170 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
171 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
172 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
173 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
174 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
175 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
176 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
177 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
178 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
179 | |
180 | fix0 = 0.0; |
181 | fiy0 = 0.0; |
182 | fiz0 = 0.0; |
183 | fix1 = 0.0; |
184 | fiy1 = 0.0; |
185 | fiz1 = 0.0; |
186 | fix2 = 0.0; |
187 | fiy2 = 0.0; |
188 | fiz2 = 0.0; |
189 | |
190 | /* Reset potential sums */ |
191 | velecsum = 0.0; |
192 | vvdwsum = 0.0; |
193 | |
194 | /* Start inner kernel loop */ |
195 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
196 | { |
197 | /* Get j neighbor index, and coordinate index */ |
198 | jnr = jjnr[jidx]; |
199 | j_coord_offset = DIM3*jnr; |
200 | |
201 | /* load j atom coordinates */ |
202 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
203 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
204 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
205 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
206 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
207 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
208 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
209 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
210 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
211 | |
212 | /* Calculate displacement vector */ |
213 | dx00 = ix0 - jx0; |
214 | dy00 = iy0 - jy0; |
215 | dz00 = iz0 - jz0; |
216 | dx01 = ix0 - jx1; |
217 | dy01 = iy0 - jy1; |
218 | dz01 = iz0 - jz1; |
219 | dx02 = ix0 - jx2; |
220 | dy02 = iy0 - jy2; |
221 | dz02 = iz0 - jz2; |
222 | dx10 = ix1 - jx0; |
223 | dy10 = iy1 - jy0; |
224 | dz10 = iz1 - jz0; |
225 | dx11 = ix1 - jx1; |
226 | dy11 = iy1 - jy1; |
227 | dz11 = iz1 - jz1; |
228 | dx12 = ix1 - jx2; |
229 | dy12 = iy1 - jy2; |
230 | dz12 = iz1 - jz2; |
231 | dx20 = ix2 - jx0; |
232 | dy20 = iy2 - jy0; |
233 | dz20 = iz2 - jz0; |
234 | dx21 = ix2 - jx1; |
235 | dy21 = iy2 - jy1; |
236 | dz21 = iz2 - jz1; |
237 | dx22 = ix2 - jx2; |
238 | dy22 = iy2 - jy2; |
239 | dz22 = iz2 - jz2; |
240 | |
241 | /* Calculate squared distance and things based on it */ |
242 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
243 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
244 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
245 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
246 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
247 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
248 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
249 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
250 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
251 | |
252 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
253 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
254 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
255 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
256 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
257 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
258 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
259 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
260 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
261 | |
262 | rinvsq00 = rinv00*rinv00; |
263 | rinvsq01 = rinv01*rinv01; |
264 | rinvsq02 = rinv02*rinv02; |
265 | rinvsq10 = rinv10*rinv10; |
266 | rinvsq11 = rinv11*rinv11; |
267 | rinvsq12 = rinv12*rinv12; |
268 | rinvsq20 = rinv20*rinv20; |
269 | rinvsq21 = rinv21*rinv21; |
270 | rinvsq22 = rinv22*rinv22; |
271 | |
272 | /************************** |
273 | * CALCULATE INTERACTIONS * |
274 | **************************/ |
275 | |
276 | r00 = rsq00*rinv00; |
277 | |
278 | /* EWALD ELECTROSTATICS */ |
279 | |
280 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
281 | ewrt = r00*ewtabscale; |
282 | ewitab = ewrt; |
283 | eweps = ewrt-ewitab; |
284 | ewitab = 4*ewitab; |
285 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
286 | velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
287 | felec = qq00*rinv00*(rinvsq00-felec); |
288 | |
289 | /* BUCKINGHAM DISPERSION/REPULSION */ |
290 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
291 | vvdw6 = c6_00*rinvsix; |
292 | br = cexp2_00*r00; |
293 | vvdwexp = cexp1_00*exp(-br); |
294 | vvdw = vvdwexp - vvdw6*(1.0/6.0); |
295 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
296 | |
297 | /* Update potential sums from outer loop */ |
298 | velecsum += velec; |
299 | vvdwsum += vvdw; |
300 | |
301 | fscal = felec+fvdw; |
302 | |
303 | /* Calculate temporary vectorial force */ |
304 | tx = fscal*dx00; |
305 | ty = fscal*dy00; |
306 | tz = fscal*dz00; |
307 | |
308 | /* Update vectorial force */ |
309 | fix0 += tx; |
310 | fiy0 += ty; |
311 | fiz0 += tz; |
312 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
313 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
314 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
315 | |
316 | /************************** |
317 | * CALCULATE INTERACTIONS * |
318 | **************************/ |
319 | |
320 | r01 = rsq01*rinv01; |
321 | |
322 | /* EWALD ELECTROSTATICS */ |
323 | |
324 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
325 | ewrt = r01*ewtabscale; |
326 | ewitab = ewrt; |
327 | eweps = ewrt-ewitab; |
328 | ewitab = 4*ewitab; |
329 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
330 | velec = qq01*(rinv01-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
331 | felec = qq01*rinv01*(rinvsq01-felec); |
332 | |
333 | /* Update potential sums from outer loop */ |
334 | velecsum += velec; |
335 | |
336 | fscal = felec; |
337 | |
338 | /* Calculate temporary vectorial force */ |
339 | tx = fscal*dx01; |
340 | ty = fscal*dy01; |
341 | tz = fscal*dz01; |
342 | |
343 | /* Update vectorial force */ |
344 | fix0 += tx; |
345 | fiy0 += ty; |
346 | fiz0 += tz; |
347 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
348 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
349 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
350 | |
351 | /************************** |
352 | * CALCULATE INTERACTIONS * |
353 | **************************/ |
354 | |
355 | r02 = rsq02*rinv02; |
356 | |
357 | /* EWALD ELECTROSTATICS */ |
358 | |
359 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
360 | ewrt = r02*ewtabscale; |
361 | ewitab = ewrt; |
362 | eweps = ewrt-ewitab; |
363 | ewitab = 4*ewitab; |
364 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
365 | velec = qq02*(rinv02-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
366 | felec = qq02*rinv02*(rinvsq02-felec); |
367 | |
368 | /* Update potential sums from outer loop */ |
369 | velecsum += velec; |
370 | |
371 | fscal = felec; |
372 | |
373 | /* Calculate temporary vectorial force */ |
374 | tx = fscal*dx02; |
375 | ty = fscal*dy02; |
376 | tz = fscal*dz02; |
377 | |
378 | /* Update vectorial force */ |
379 | fix0 += tx; |
380 | fiy0 += ty; |
381 | fiz0 += tz; |
382 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
383 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
384 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
385 | |
386 | /************************** |
387 | * CALCULATE INTERACTIONS * |
388 | **************************/ |
389 | |
390 | r10 = rsq10*rinv10; |
391 | |
392 | /* EWALD ELECTROSTATICS */ |
393 | |
394 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
395 | ewrt = r10*ewtabscale; |
396 | ewitab = ewrt; |
397 | eweps = ewrt-ewitab; |
398 | ewitab = 4*ewitab; |
399 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
400 | velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
401 | felec = qq10*rinv10*(rinvsq10-felec); |
402 | |
403 | /* Update potential sums from outer loop */ |
404 | velecsum += velec; |
405 | |
406 | fscal = felec; |
407 | |
408 | /* Calculate temporary vectorial force */ |
409 | tx = fscal*dx10; |
410 | ty = fscal*dy10; |
411 | tz = fscal*dz10; |
412 | |
413 | /* Update vectorial force */ |
414 | fix1 += tx; |
415 | fiy1 += ty; |
416 | fiz1 += tz; |
417 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
418 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
419 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
420 | |
421 | /************************** |
422 | * CALCULATE INTERACTIONS * |
423 | **************************/ |
424 | |
425 | r11 = rsq11*rinv11; |
426 | |
427 | /* EWALD ELECTROSTATICS */ |
428 | |
429 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
430 | ewrt = r11*ewtabscale; |
431 | ewitab = ewrt; |
432 | eweps = ewrt-ewitab; |
433 | ewitab = 4*ewitab; |
434 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
435 | velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
436 | felec = qq11*rinv11*(rinvsq11-felec); |
437 | |
438 | /* Update potential sums from outer loop */ |
439 | velecsum += velec; |
440 | |
441 | fscal = felec; |
442 | |
443 | /* Calculate temporary vectorial force */ |
444 | tx = fscal*dx11; |
445 | ty = fscal*dy11; |
446 | tz = fscal*dz11; |
447 | |
448 | /* Update vectorial force */ |
449 | fix1 += tx; |
450 | fiy1 += ty; |
451 | fiz1 += tz; |
452 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
453 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
454 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
455 | |
456 | /************************** |
457 | * CALCULATE INTERACTIONS * |
458 | **************************/ |
459 | |
460 | r12 = rsq12*rinv12; |
461 | |
462 | /* EWALD ELECTROSTATICS */ |
463 | |
464 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
465 | ewrt = r12*ewtabscale; |
466 | ewitab = ewrt; |
467 | eweps = ewrt-ewitab; |
468 | ewitab = 4*ewitab; |
469 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
470 | velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
471 | felec = qq12*rinv12*(rinvsq12-felec); |
472 | |
473 | /* Update potential sums from outer loop */ |
474 | velecsum += velec; |
475 | |
476 | fscal = felec; |
477 | |
478 | /* Calculate temporary vectorial force */ |
479 | tx = fscal*dx12; |
480 | ty = fscal*dy12; |
481 | tz = fscal*dz12; |
482 | |
483 | /* Update vectorial force */ |
484 | fix1 += tx; |
485 | fiy1 += ty; |
486 | fiz1 += tz; |
487 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
488 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
489 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
490 | |
491 | /************************** |
492 | * CALCULATE INTERACTIONS * |
493 | **************************/ |
494 | |
495 | r20 = rsq20*rinv20; |
496 | |
497 | /* EWALD ELECTROSTATICS */ |
498 | |
499 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
500 | ewrt = r20*ewtabscale; |
501 | ewitab = ewrt; |
502 | eweps = ewrt-ewitab; |
503 | ewitab = 4*ewitab; |
504 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
505 | velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
506 | felec = qq20*rinv20*(rinvsq20-felec); |
507 | |
508 | /* Update potential sums from outer loop */ |
509 | velecsum += velec; |
510 | |
511 | fscal = felec; |
512 | |
513 | /* Calculate temporary vectorial force */ |
514 | tx = fscal*dx20; |
515 | ty = fscal*dy20; |
516 | tz = fscal*dz20; |
517 | |
518 | /* Update vectorial force */ |
519 | fix2 += tx; |
520 | fiy2 += ty; |
521 | fiz2 += tz; |
522 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
523 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
524 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
525 | |
526 | /************************** |
527 | * CALCULATE INTERACTIONS * |
528 | **************************/ |
529 | |
530 | r21 = rsq21*rinv21; |
531 | |
532 | /* EWALD ELECTROSTATICS */ |
533 | |
534 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
535 | ewrt = r21*ewtabscale; |
536 | ewitab = ewrt; |
537 | eweps = ewrt-ewitab; |
538 | ewitab = 4*ewitab; |
539 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
540 | velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
541 | felec = qq21*rinv21*(rinvsq21-felec); |
542 | |
543 | /* Update potential sums from outer loop */ |
544 | velecsum += velec; |
545 | |
546 | fscal = felec; |
547 | |
548 | /* Calculate temporary vectorial force */ |
549 | tx = fscal*dx21; |
550 | ty = fscal*dy21; |
551 | tz = fscal*dz21; |
552 | |
553 | /* Update vectorial force */ |
554 | fix2 += tx; |
555 | fiy2 += ty; |
556 | fiz2 += tz; |
557 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
558 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
559 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
560 | |
561 | /************************** |
562 | * CALCULATE INTERACTIONS * |
563 | **************************/ |
564 | |
565 | r22 = rsq22*rinv22; |
566 | |
567 | /* EWALD ELECTROSTATICS */ |
568 | |
569 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
570 | ewrt = r22*ewtabscale; |
571 | ewitab = ewrt; |
572 | eweps = ewrt-ewitab; |
573 | ewitab = 4*ewitab; |
574 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
575 | velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
576 | felec = qq22*rinv22*(rinvsq22-felec); |
577 | |
578 | /* Update potential sums from outer loop */ |
579 | velecsum += velec; |
580 | |
581 | fscal = felec; |
582 | |
583 | /* Calculate temporary vectorial force */ |
584 | tx = fscal*dx22; |
585 | ty = fscal*dy22; |
586 | tz = fscal*dz22; |
587 | |
588 | /* Update vectorial force */ |
589 | fix2 += tx; |
590 | fiy2 += ty; |
591 | fiz2 += tz; |
592 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
593 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
594 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
595 | |
596 | /* Inner loop uses 398 flops */ |
597 | } |
598 | /* End of innermost loop */ |
599 | |
600 | tx = ty = tz = 0; |
601 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
602 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
603 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
604 | tx += fix0; |
605 | ty += fiy0; |
606 | tz += fiz0; |
607 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
608 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
609 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
610 | tx += fix1; |
611 | ty += fiy1; |
612 | tz += fiz1; |
613 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
614 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
615 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
616 | tx += fix2; |
617 | ty += fiy2; |
618 | tz += fiz2; |
619 | fshift[i_shift_offset+XX0] += tx; |
620 | fshift[i_shift_offset+YY1] += ty; |
621 | fshift[i_shift_offset+ZZ2] += tz; |
622 | |
623 | ggid = gid[iidx]; |
624 | /* Update potential energies */ |
625 | kernel_data->energygrp_elec[ggid] += velecsum; |
626 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
627 | |
628 | /* Increment number of inner iterations */ |
629 | inneriter += j_index_end - j_index_start; |
630 | |
631 | /* Outer loop uses 32 flops */ |
632 | } |
633 | |
634 | /* Increment number of outer iterations */ |
635 | outeriter += nri; |
636 | |
637 | /* Update outer/inner flops */ |
638 | |
639 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*398)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*32 + inneriter*398; |
640 | } |
641 | /* |
642 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwBham_GeomW3W3_F_c |
643 | * Electrostatics interaction: Ewald |
644 | * VdW interaction: Buckingham |
645 | * Geometry: Water3-Water3 |
646 | * Calculate force/pot: Force |
647 | */ |
648 | void |
649 | nb_kernel_ElecEw_VdwBham_GeomW3W3_F_c |
650 | (t_nblist * gmx_restrict__restrict nlist, |
651 | rvec * gmx_restrict__restrict xx, |
652 | rvec * gmx_restrict__restrict ff, |
653 | t_forcerec * gmx_restrict__restrict fr, |
654 | t_mdatoms * gmx_restrict__restrict mdatoms, |
655 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
656 | t_nrnb * gmx_restrict__restrict nrnb) |
657 | { |
658 | int i_shift_offset,i_coord_offset,j_coord_offset; |
659 | int j_index_start,j_index_end; |
660 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
661 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
662 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
663 | real *shiftvec,*fshift,*x,*f; |
664 | int vdwioffset0; |
665 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
666 | int vdwioffset1; |
667 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
668 | int vdwioffset2; |
669 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
670 | int vdwjidx0; |
671 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
672 | int vdwjidx1; |
673 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
674 | int vdwjidx2; |
675 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
676 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
677 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
678 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
679 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
680 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
681 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
682 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
683 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
684 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
685 | real velec,felec,velecsum,facel,crf,krf,krf2; |
686 | real *charge; |
687 | int nvdwtype; |
688 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
689 | int *vdwtype; |
690 | real *vdwparam; |
691 | int ewitab; |
692 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
693 | real *ewtab; |
694 | |
695 | x = xx[0]; |
696 | f = ff[0]; |
697 | |
698 | nri = nlist->nri; |
699 | iinr = nlist->iinr; |
700 | jindex = nlist->jindex; |
701 | jjnr = nlist->jjnr; |
702 | shiftidx = nlist->shift; |
703 | gid = nlist->gid; |
704 | shiftvec = fr->shift_vec[0]; |
705 | fshift = fr->fshift[0]; |
706 | facel = fr->epsfac; |
707 | charge = mdatoms->chargeA; |
708 | nvdwtype = fr->ntype; |
709 | vdwparam = fr->nbfp; |
710 | vdwtype = mdatoms->typeA; |
711 | |
712 | sh_ewald = fr->ic->sh_ewald; |
713 | ewtab = fr->ic->tabq_coul_F; |
714 | ewtabscale = fr->ic->tabq_scale; |
715 | ewtabhalfspace = 0.5/ewtabscale; |
716 | |
717 | /* Setup water-specific parameters */ |
718 | inr = nlist->iinr[0]; |
719 | iq0 = facel*charge[inr+0]; |
720 | iq1 = facel*charge[inr+1]; |
721 | iq2 = facel*charge[inr+2]; |
722 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
723 | |
724 | jq0 = charge[inr+0]; |
725 | jq1 = charge[inr+1]; |
726 | jq2 = charge[inr+2]; |
727 | vdwjidx0 = 3*vdwtype[inr+0]; |
728 | qq00 = iq0*jq0; |
729 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
730 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
731 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
732 | qq01 = iq0*jq1; |
733 | qq02 = iq0*jq2; |
734 | qq10 = iq1*jq0; |
735 | qq11 = iq1*jq1; |
736 | qq12 = iq1*jq2; |
737 | qq20 = iq2*jq0; |
738 | qq21 = iq2*jq1; |
739 | qq22 = iq2*jq2; |
740 | |
741 | outeriter = 0; |
742 | inneriter = 0; |
743 | |
744 | /* Start outer loop over neighborlists */ |
745 | for(iidx=0; iidx<nri; iidx++) |
746 | { |
747 | /* Load shift vector for this list */ |
748 | i_shift_offset = DIM3*shiftidx[iidx]; |
749 | shX = shiftvec[i_shift_offset+XX0]; |
750 | shY = shiftvec[i_shift_offset+YY1]; |
751 | shZ = shiftvec[i_shift_offset+ZZ2]; |
752 | |
753 | /* Load limits for loop over neighbors */ |
754 | j_index_start = jindex[iidx]; |
755 | j_index_end = jindex[iidx+1]; |
756 | |
757 | /* Get outer coordinate index */ |
758 | inr = iinr[iidx]; |
759 | i_coord_offset = DIM3*inr; |
760 | |
761 | /* Load i particle coords and add shift vector */ |
762 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
763 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
764 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
765 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
766 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
767 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
768 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
769 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
770 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
771 | |
772 | fix0 = 0.0; |
773 | fiy0 = 0.0; |
774 | fiz0 = 0.0; |
775 | fix1 = 0.0; |
776 | fiy1 = 0.0; |
777 | fiz1 = 0.0; |
778 | fix2 = 0.0; |
779 | fiy2 = 0.0; |
780 | fiz2 = 0.0; |
781 | |
782 | /* Start inner kernel loop */ |
783 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
784 | { |
785 | /* Get j neighbor index, and coordinate index */ |
786 | jnr = jjnr[jidx]; |
787 | j_coord_offset = DIM3*jnr; |
788 | |
789 | /* load j atom coordinates */ |
790 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
791 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
792 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
793 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
794 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
795 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
796 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
797 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
798 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
799 | |
800 | /* Calculate displacement vector */ |
801 | dx00 = ix0 - jx0; |
802 | dy00 = iy0 - jy0; |
803 | dz00 = iz0 - jz0; |
804 | dx01 = ix0 - jx1; |
805 | dy01 = iy0 - jy1; |
806 | dz01 = iz0 - jz1; |
807 | dx02 = ix0 - jx2; |
808 | dy02 = iy0 - jy2; |
809 | dz02 = iz0 - jz2; |
810 | dx10 = ix1 - jx0; |
811 | dy10 = iy1 - jy0; |
812 | dz10 = iz1 - jz0; |
813 | dx11 = ix1 - jx1; |
814 | dy11 = iy1 - jy1; |
815 | dz11 = iz1 - jz1; |
816 | dx12 = ix1 - jx2; |
817 | dy12 = iy1 - jy2; |
818 | dz12 = iz1 - jz2; |
819 | dx20 = ix2 - jx0; |
820 | dy20 = iy2 - jy0; |
821 | dz20 = iz2 - jz0; |
822 | dx21 = ix2 - jx1; |
823 | dy21 = iy2 - jy1; |
824 | dz21 = iz2 - jz1; |
825 | dx22 = ix2 - jx2; |
826 | dy22 = iy2 - jy2; |
827 | dz22 = iz2 - jz2; |
828 | |
829 | /* Calculate squared distance and things based on it */ |
830 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
831 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
832 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
833 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
834 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
835 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
836 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
837 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
838 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
839 | |
840 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
841 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
842 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
843 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
844 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
845 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
846 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
847 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
848 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
849 | |
850 | rinvsq00 = rinv00*rinv00; |
851 | rinvsq01 = rinv01*rinv01; |
852 | rinvsq02 = rinv02*rinv02; |
853 | rinvsq10 = rinv10*rinv10; |
854 | rinvsq11 = rinv11*rinv11; |
855 | rinvsq12 = rinv12*rinv12; |
856 | rinvsq20 = rinv20*rinv20; |
857 | rinvsq21 = rinv21*rinv21; |
858 | rinvsq22 = rinv22*rinv22; |
859 | |
860 | /************************** |
861 | * CALCULATE INTERACTIONS * |
862 | **************************/ |
863 | |
864 | r00 = rsq00*rinv00; |
865 | |
866 | /* EWALD ELECTROSTATICS */ |
867 | |
868 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
869 | ewrt = r00*ewtabscale; |
870 | ewitab = ewrt; |
871 | eweps = ewrt-ewitab; |
872 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
873 | felec = qq00*rinv00*(rinvsq00-felec); |
874 | |
875 | /* BUCKINGHAM DISPERSION/REPULSION */ |
876 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
877 | vvdw6 = c6_00*rinvsix; |
878 | br = cexp2_00*r00; |
879 | vvdwexp = cexp1_00*exp(-br); |
880 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
881 | |
882 | fscal = felec+fvdw; |
883 | |
884 | /* Calculate temporary vectorial force */ |
885 | tx = fscal*dx00; |
886 | ty = fscal*dy00; |
887 | tz = fscal*dz00; |
888 | |
889 | /* Update vectorial force */ |
890 | fix0 += tx; |
891 | fiy0 += ty; |
892 | fiz0 += tz; |
893 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
894 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
895 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
896 | |
897 | /************************** |
898 | * CALCULATE INTERACTIONS * |
899 | **************************/ |
900 | |
901 | r01 = rsq01*rinv01; |
902 | |
903 | /* EWALD ELECTROSTATICS */ |
904 | |
905 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
906 | ewrt = r01*ewtabscale; |
907 | ewitab = ewrt; |
908 | eweps = ewrt-ewitab; |
909 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
910 | felec = qq01*rinv01*(rinvsq01-felec); |
911 | |
912 | fscal = felec; |
913 | |
914 | /* Calculate temporary vectorial force */ |
915 | tx = fscal*dx01; |
916 | ty = fscal*dy01; |
917 | tz = fscal*dz01; |
918 | |
919 | /* Update vectorial force */ |
920 | fix0 += tx; |
921 | fiy0 += ty; |
922 | fiz0 += tz; |
923 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
924 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
925 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
926 | |
927 | /************************** |
928 | * CALCULATE INTERACTIONS * |
929 | **************************/ |
930 | |
931 | r02 = rsq02*rinv02; |
932 | |
933 | /* EWALD ELECTROSTATICS */ |
934 | |
935 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
936 | ewrt = r02*ewtabscale; |
937 | ewitab = ewrt; |
938 | eweps = ewrt-ewitab; |
939 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
940 | felec = qq02*rinv02*(rinvsq02-felec); |
941 | |
942 | fscal = felec; |
943 | |
944 | /* Calculate temporary vectorial force */ |
945 | tx = fscal*dx02; |
946 | ty = fscal*dy02; |
947 | tz = fscal*dz02; |
948 | |
949 | /* Update vectorial force */ |
950 | fix0 += tx; |
951 | fiy0 += ty; |
952 | fiz0 += tz; |
953 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
954 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
955 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
956 | |
957 | /************************** |
958 | * CALCULATE INTERACTIONS * |
959 | **************************/ |
960 | |
961 | r10 = rsq10*rinv10; |
962 | |
963 | /* EWALD ELECTROSTATICS */ |
964 | |
965 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
966 | ewrt = r10*ewtabscale; |
967 | ewitab = ewrt; |
968 | eweps = ewrt-ewitab; |
969 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
970 | felec = qq10*rinv10*(rinvsq10-felec); |
971 | |
972 | fscal = felec; |
973 | |
974 | /* Calculate temporary vectorial force */ |
975 | tx = fscal*dx10; |
976 | ty = fscal*dy10; |
977 | tz = fscal*dz10; |
978 | |
979 | /* Update vectorial force */ |
980 | fix1 += tx; |
981 | fiy1 += ty; |
982 | fiz1 += tz; |
983 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
984 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
985 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
986 | |
987 | /************************** |
988 | * CALCULATE INTERACTIONS * |
989 | **************************/ |
990 | |
991 | r11 = rsq11*rinv11; |
992 | |
993 | /* EWALD ELECTROSTATICS */ |
994 | |
995 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
996 | ewrt = r11*ewtabscale; |
997 | ewitab = ewrt; |
998 | eweps = ewrt-ewitab; |
999 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1000 | felec = qq11*rinv11*(rinvsq11-felec); |
1001 | |
1002 | fscal = felec; |
1003 | |
1004 | /* Calculate temporary vectorial force */ |
1005 | tx = fscal*dx11; |
1006 | ty = fscal*dy11; |
1007 | tz = fscal*dz11; |
1008 | |
1009 | /* Update vectorial force */ |
1010 | fix1 += tx; |
1011 | fiy1 += ty; |
1012 | fiz1 += tz; |
1013 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1014 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1015 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1016 | |
1017 | /************************** |
1018 | * CALCULATE INTERACTIONS * |
1019 | **************************/ |
1020 | |
1021 | r12 = rsq12*rinv12; |
1022 | |
1023 | /* EWALD ELECTROSTATICS */ |
1024 | |
1025 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1026 | ewrt = r12*ewtabscale; |
1027 | ewitab = ewrt; |
1028 | eweps = ewrt-ewitab; |
1029 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1030 | felec = qq12*rinv12*(rinvsq12-felec); |
1031 | |
1032 | fscal = felec; |
1033 | |
1034 | /* Calculate temporary vectorial force */ |
1035 | tx = fscal*dx12; |
1036 | ty = fscal*dy12; |
1037 | tz = fscal*dz12; |
1038 | |
1039 | /* Update vectorial force */ |
1040 | fix1 += tx; |
1041 | fiy1 += ty; |
1042 | fiz1 += tz; |
1043 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1044 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1045 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1046 | |
1047 | /************************** |
1048 | * CALCULATE INTERACTIONS * |
1049 | **************************/ |
1050 | |
1051 | r20 = rsq20*rinv20; |
1052 | |
1053 | /* EWALD ELECTROSTATICS */ |
1054 | |
1055 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1056 | ewrt = r20*ewtabscale; |
1057 | ewitab = ewrt; |
1058 | eweps = ewrt-ewitab; |
1059 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1060 | felec = qq20*rinv20*(rinvsq20-felec); |
1061 | |
1062 | fscal = felec; |
1063 | |
1064 | /* Calculate temporary vectorial force */ |
1065 | tx = fscal*dx20; |
1066 | ty = fscal*dy20; |
1067 | tz = fscal*dz20; |
1068 | |
1069 | /* Update vectorial force */ |
1070 | fix2 += tx; |
1071 | fiy2 += ty; |
1072 | fiz2 += tz; |
1073 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1074 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1075 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1076 | |
1077 | /************************** |
1078 | * CALCULATE INTERACTIONS * |
1079 | **************************/ |
1080 | |
1081 | r21 = rsq21*rinv21; |
1082 | |
1083 | /* EWALD ELECTROSTATICS */ |
1084 | |
1085 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1086 | ewrt = r21*ewtabscale; |
1087 | ewitab = ewrt; |
1088 | eweps = ewrt-ewitab; |
1089 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1090 | felec = qq21*rinv21*(rinvsq21-felec); |
1091 | |
1092 | fscal = felec; |
1093 | |
1094 | /* Calculate temporary vectorial force */ |
1095 | tx = fscal*dx21; |
1096 | ty = fscal*dy21; |
1097 | tz = fscal*dz21; |
1098 | |
1099 | /* Update vectorial force */ |
1100 | fix2 += tx; |
1101 | fiy2 += ty; |
1102 | fiz2 += tz; |
1103 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1104 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1105 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1106 | |
1107 | /************************** |
1108 | * CALCULATE INTERACTIONS * |
1109 | **************************/ |
1110 | |
1111 | r22 = rsq22*rinv22; |
1112 | |
1113 | /* EWALD ELECTROSTATICS */ |
1114 | |
1115 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1116 | ewrt = r22*ewtabscale; |
1117 | ewitab = ewrt; |
1118 | eweps = ewrt-ewitab; |
1119 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1120 | felec = qq22*rinv22*(rinvsq22-felec); |
1121 | |
1122 | fscal = felec; |
1123 | |
1124 | /* Calculate temporary vectorial force */ |
1125 | tx = fscal*dx22; |
1126 | ty = fscal*dy22; |
1127 | tz = fscal*dz22; |
1128 | |
1129 | /* Update vectorial force */ |
1130 | fix2 += tx; |
1131 | fiy2 += ty; |
1132 | fiz2 += tz; |
1133 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1134 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1135 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1136 | |
1137 | /* Inner loop uses 332 flops */ |
1138 | } |
1139 | /* End of innermost loop */ |
1140 | |
1141 | tx = ty = tz = 0; |
1142 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
1143 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
1144 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
1145 | tx += fix0; |
1146 | ty += fiy0; |
1147 | tz += fiz0; |
1148 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
1149 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
1150 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
1151 | tx += fix1; |
1152 | ty += fiy1; |
1153 | tz += fiz1; |
1154 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
1155 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
1156 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
1157 | tx += fix2; |
1158 | ty += fiy2; |
1159 | tz += fiz2; |
1160 | fshift[i_shift_offset+XX0] += tx; |
1161 | fshift[i_shift_offset+YY1] += ty; |
1162 | fshift[i_shift_offset+ZZ2] += tz; |
1163 | |
1164 | /* Increment number of inner iterations */ |
1165 | inneriter += j_index_end - j_index_start; |
1166 | |
1167 | /* Outer loop uses 30 flops */ |
1168 | } |
1169 | |
1170 | /* Increment number of outer iterations */ |
1171 | outeriter += nri; |
1172 | |
1173 | /* Update outer/inner flops */ |
1174 | |
1175 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*332)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*30 + inneriter *332; |
1176 | } |