File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwCSTab_GeomW3P1_c.c |
Location: | line 444, column 5 |
Description: | Value stored to 'gid' 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_ElecRFCut_VdwCSTab_GeomW3P1_VF_c |
51 | * Electrostatics interaction: ReactionField |
52 | * VdW interaction: CubicSplineTable |
53 | * Geometry: Water3-Particle |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecRFCut_VdwCSTab_GeomW3P1_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 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
81 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
82 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
83 | real velec,felec,velecsum,facel,crf,krf,krf2; |
84 | real *charge; |
85 | int nvdwtype; |
86 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
87 | int *vdwtype; |
88 | real *vdwparam; |
89 | int vfitab; |
90 | real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF; |
91 | real *vftab; |
92 | |
93 | x = xx[0]; |
94 | f = ff[0]; |
95 | |
96 | nri = nlist->nri; |
97 | iinr = nlist->iinr; |
98 | jindex = nlist->jindex; |
99 | jjnr = nlist->jjnr; |
100 | shiftidx = nlist->shift; |
101 | gid = nlist->gid; |
102 | shiftvec = fr->shift_vec[0]; |
103 | fshift = fr->fshift[0]; |
104 | facel = fr->epsfac; |
105 | charge = mdatoms->chargeA; |
106 | krf = fr->ic->k_rf; |
107 | krf2 = krf*2.0; |
108 | crf = fr->ic->c_rf; |
109 | nvdwtype = fr->ntype; |
110 | vdwparam = fr->nbfp; |
111 | vdwtype = mdatoms->typeA; |
112 | |
113 | vftab = kernel_data->table_vdw->data; |
114 | vftabscale = kernel_data->table_vdw->scale; |
115 | |
116 | /* Setup water-specific parameters */ |
117 | inr = nlist->iinr[0]; |
118 | iq0 = facel*charge[inr+0]; |
119 | iq1 = facel*charge[inr+1]; |
120 | iq2 = facel*charge[inr+2]; |
121 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
122 | |
123 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
124 | rcutoff = fr->rcoulomb; |
125 | rcutoff2 = rcutoff*rcutoff; |
126 | |
127 | outeriter = 0; |
128 | inneriter = 0; |
129 | |
130 | /* Start outer loop over neighborlists */ |
131 | for(iidx=0; iidx<nri; iidx++) |
132 | { |
133 | /* Load shift vector for this list */ |
134 | i_shift_offset = DIM3*shiftidx[iidx]; |
135 | shX = shiftvec[i_shift_offset+XX0]; |
136 | shY = shiftvec[i_shift_offset+YY1]; |
137 | shZ = shiftvec[i_shift_offset+ZZ2]; |
138 | |
139 | /* Load limits for loop over neighbors */ |
140 | j_index_start = jindex[iidx]; |
141 | j_index_end = jindex[iidx+1]; |
142 | |
143 | /* Get outer coordinate index */ |
144 | inr = iinr[iidx]; |
145 | i_coord_offset = DIM3*inr; |
146 | |
147 | /* Load i particle coords and add shift vector */ |
148 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
149 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
150 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
151 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
152 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
153 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
154 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
155 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
156 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
157 | |
158 | fix0 = 0.0; |
159 | fiy0 = 0.0; |
160 | fiz0 = 0.0; |
161 | fix1 = 0.0; |
162 | fiy1 = 0.0; |
163 | fiz1 = 0.0; |
164 | fix2 = 0.0; |
165 | fiy2 = 0.0; |
166 | fiz2 = 0.0; |
167 | |
168 | /* Reset potential sums */ |
169 | velecsum = 0.0; |
170 | vvdwsum = 0.0; |
171 | |
172 | /* Start inner kernel loop */ |
173 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
174 | { |
175 | /* Get j neighbor index, and coordinate index */ |
176 | jnr = jjnr[jidx]; |
177 | j_coord_offset = DIM3*jnr; |
178 | |
179 | /* load j atom coordinates */ |
180 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
181 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
182 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
183 | |
184 | /* Calculate displacement vector */ |
185 | dx00 = ix0 - jx0; |
186 | dy00 = iy0 - jy0; |
187 | dz00 = iz0 - jz0; |
188 | dx10 = ix1 - jx0; |
189 | dy10 = iy1 - jy0; |
190 | dz10 = iz1 - jz0; |
191 | dx20 = ix2 - jx0; |
192 | dy20 = iy2 - jy0; |
193 | dz20 = iz2 - jz0; |
194 | |
195 | /* Calculate squared distance and things based on it */ |
196 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
197 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
198 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
199 | |
200 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
201 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
202 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
203 | |
204 | rinvsq00 = rinv00*rinv00; |
205 | rinvsq10 = rinv10*rinv10; |
206 | rinvsq20 = rinv20*rinv20; |
207 | |
208 | /* Load parameters for j particles */ |
209 | jq0 = charge[jnr+0]; |
210 | vdwjidx0 = 2*vdwtype[jnr+0]; |
211 | |
212 | /************************** |
213 | * CALCULATE INTERACTIONS * |
214 | **************************/ |
215 | |
216 | if (rsq00<rcutoff2) |
217 | { |
218 | |
219 | r00 = rsq00*rinv00; |
220 | |
221 | qq00 = iq0*jq0; |
222 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
223 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
224 | |
225 | /* Calculate table index by multiplying r with table scale and truncate to integer */ |
226 | rt = r00*vftabscale; |
227 | vfitab = rt; |
228 | vfeps = rt-vfitab; |
229 | vfitab = 2*4*vfitab; |
230 | |
231 | /* REACTION-FIELD ELECTROSTATICS */ |
232 | velec = qq00*(rinv00+krf*rsq00-crf); |
233 | felec = qq00*(rinv00*rinvsq00-krf2); |
234 | |
235 | /* CUBIC SPLINE TABLE DISPERSION */ |
236 | vfitab += 0; |
237 | Y = vftab[vfitab]; |
238 | F = vftab[vfitab+1]; |
239 | Geps = vfeps*vftab[vfitab+2]; |
240 | Heps2 = vfeps*vfeps*vftab[vfitab+3]; |
241 | Fp = F+Geps+Heps2; |
242 | VV = Y+vfeps*Fp; |
243 | vvdw6 = c6_00*VV; |
244 | FF = Fp+Geps+2.0*Heps2; |
245 | fvdw6 = c6_00*FF; |
246 | |
247 | /* CUBIC SPLINE TABLE REPULSION */ |
248 | Y = vftab[vfitab+4]; |
249 | F = vftab[vfitab+5]; |
250 | Geps = vfeps*vftab[vfitab+6]; |
251 | Heps2 = vfeps*vfeps*vftab[vfitab+7]; |
252 | Fp = F+Geps+Heps2; |
253 | VV = Y+vfeps*Fp; |
254 | vvdw12 = c12_00*VV; |
255 | FF = Fp+Geps+2.0*Heps2; |
256 | fvdw12 = c12_00*FF; |
257 | vvdw = vvdw12+vvdw6; |
258 | fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00; |
259 | |
260 | /* Update potential sums from outer loop */ |
261 | velecsum += velec; |
262 | vvdwsum += vvdw; |
263 | |
264 | fscal = felec+fvdw; |
265 | |
266 | /* Calculate temporary vectorial force */ |
267 | tx = fscal*dx00; |
268 | ty = fscal*dy00; |
269 | tz = fscal*dz00; |
270 | |
271 | /* Update vectorial force */ |
272 | fix0 += tx; |
273 | fiy0 += ty; |
274 | fiz0 += tz; |
275 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
276 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
277 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
278 | |
279 | } |
280 | |
281 | /************************** |
282 | * CALCULATE INTERACTIONS * |
283 | **************************/ |
284 | |
285 | if (rsq10<rcutoff2) |
286 | { |
287 | |
288 | qq10 = iq1*jq0; |
289 | |
290 | /* REACTION-FIELD ELECTROSTATICS */ |
291 | velec = qq10*(rinv10+krf*rsq10-crf); |
292 | felec = qq10*(rinv10*rinvsq10-krf2); |
293 | |
294 | /* Update potential sums from outer loop */ |
295 | velecsum += velec; |
296 | |
297 | fscal = felec; |
298 | |
299 | /* Calculate temporary vectorial force */ |
300 | tx = fscal*dx10; |
301 | ty = fscal*dy10; |
302 | tz = fscal*dz10; |
303 | |
304 | /* Update vectorial force */ |
305 | fix1 += tx; |
306 | fiy1 += ty; |
307 | fiz1 += tz; |
308 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
309 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
310 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
311 | |
312 | } |
313 | |
314 | /************************** |
315 | * CALCULATE INTERACTIONS * |
316 | **************************/ |
317 | |
318 | if (rsq20<rcutoff2) |
319 | { |
320 | |
321 | qq20 = iq2*jq0; |
322 | |
323 | /* REACTION-FIELD ELECTROSTATICS */ |
324 | velec = qq20*(rinv20+krf*rsq20-crf); |
325 | felec = qq20*(rinv20*rinvsq20-krf2); |
326 | |
327 | /* Update potential sums from outer loop */ |
328 | velecsum += velec; |
329 | |
330 | fscal = felec; |
331 | |
332 | /* Calculate temporary vectorial force */ |
333 | tx = fscal*dx20; |
334 | ty = fscal*dy20; |
335 | tz = fscal*dz20; |
336 | |
337 | /* Update vectorial force */ |
338 | fix2 += tx; |
339 | fiy2 += ty; |
340 | fiz2 += tz; |
341 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
342 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
343 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
344 | |
345 | } |
346 | |
347 | /* Inner loop uses 130 flops */ |
348 | } |
349 | /* End of innermost loop */ |
350 | |
351 | tx = ty = tz = 0; |
352 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
353 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
354 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
355 | tx += fix0; |
356 | ty += fiy0; |
357 | tz += fiz0; |
358 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
359 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
360 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
361 | tx += fix1; |
362 | ty += fiy1; |
363 | tz += fiz1; |
364 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
365 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
366 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
367 | tx += fix2; |
368 | ty += fiy2; |
369 | tz += fiz2; |
370 | fshift[i_shift_offset+XX0] += tx; |
371 | fshift[i_shift_offset+YY1] += ty; |
372 | fshift[i_shift_offset+ZZ2] += tz; |
373 | |
374 | ggid = gid[iidx]; |
375 | /* Update potential energies */ |
376 | kernel_data->energygrp_elec[ggid] += velecsum; |
377 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
378 | |
379 | /* Increment number of inner iterations */ |
380 | inneriter += j_index_end - j_index_start; |
381 | |
382 | /* Outer loop uses 32 flops */ |
383 | } |
384 | |
385 | /* Increment number of outer iterations */ |
386 | outeriter += nri; |
387 | |
388 | /* Update outer/inner flops */ |
389 | |
390 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*130)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_VF] += outeriter*32 + inneriter *130; |
391 | } |
392 | /* |
393 | * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwCSTab_GeomW3P1_F_c |
394 | * Electrostatics interaction: ReactionField |
395 | * VdW interaction: CubicSplineTable |
396 | * Geometry: Water3-Particle |
397 | * Calculate force/pot: Force |
398 | */ |
399 | void |
400 | nb_kernel_ElecRFCut_VdwCSTab_GeomW3P1_F_c |
401 | (t_nblist * gmx_restrict__restrict nlist, |
402 | rvec * gmx_restrict__restrict xx, |
403 | rvec * gmx_restrict__restrict ff, |
404 | t_forcerec * gmx_restrict__restrict fr, |
405 | t_mdatoms * gmx_restrict__restrict mdatoms, |
406 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
407 | t_nrnb * gmx_restrict__restrict nrnb) |
408 | { |
409 | int i_shift_offset,i_coord_offset,j_coord_offset; |
410 | int j_index_start,j_index_end; |
411 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
412 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
413 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
414 | real *shiftvec,*fshift,*x,*f; |
415 | int vdwioffset0; |
416 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
417 | int vdwioffset1; |
418 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
419 | int vdwioffset2; |
420 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
421 | int vdwjidx0; |
422 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
423 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
424 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
425 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
426 | real velec,felec,velecsum,facel,crf,krf,krf2; |
427 | real *charge; |
428 | int nvdwtype; |
429 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
430 | int *vdwtype; |
431 | real *vdwparam; |
432 | int vfitab; |
433 | real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF; |
434 | real *vftab; |
435 | |
436 | x = xx[0]; |
437 | f = ff[0]; |
438 | |
439 | nri = nlist->nri; |
440 | iinr = nlist->iinr; |
441 | jindex = nlist->jindex; |
442 | jjnr = nlist->jjnr; |
443 | shiftidx = nlist->shift; |
444 | gid = nlist->gid; |
Value stored to 'gid' is never read | |
445 | shiftvec = fr->shift_vec[0]; |
446 | fshift = fr->fshift[0]; |
447 | facel = fr->epsfac; |
448 | charge = mdatoms->chargeA; |
449 | krf = fr->ic->k_rf; |
450 | krf2 = krf*2.0; |
451 | crf = fr->ic->c_rf; |
452 | nvdwtype = fr->ntype; |
453 | vdwparam = fr->nbfp; |
454 | vdwtype = mdatoms->typeA; |
455 | |
456 | vftab = kernel_data->table_vdw->data; |
457 | vftabscale = kernel_data->table_vdw->scale; |
458 | |
459 | /* Setup water-specific parameters */ |
460 | inr = nlist->iinr[0]; |
461 | iq0 = facel*charge[inr+0]; |
462 | iq1 = facel*charge[inr+1]; |
463 | iq2 = facel*charge[inr+2]; |
464 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
465 | |
466 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
467 | rcutoff = fr->rcoulomb; |
468 | rcutoff2 = rcutoff*rcutoff; |
469 | |
470 | outeriter = 0; |
471 | inneriter = 0; |
472 | |
473 | /* Start outer loop over neighborlists */ |
474 | for(iidx=0; iidx<nri; iidx++) |
475 | { |
476 | /* Load shift vector for this list */ |
477 | i_shift_offset = DIM3*shiftidx[iidx]; |
478 | shX = shiftvec[i_shift_offset+XX0]; |
479 | shY = shiftvec[i_shift_offset+YY1]; |
480 | shZ = shiftvec[i_shift_offset+ZZ2]; |
481 | |
482 | /* Load limits for loop over neighbors */ |
483 | j_index_start = jindex[iidx]; |
484 | j_index_end = jindex[iidx+1]; |
485 | |
486 | /* Get outer coordinate index */ |
487 | inr = iinr[iidx]; |
488 | i_coord_offset = DIM3*inr; |
489 | |
490 | /* Load i particle coords and add shift vector */ |
491 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
492 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
493 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
494 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
495 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
496 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
497 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
498 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
499 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
500 | |
501 | fix0 = 0.0; |
502 | fiy0 = 0.0; |
503 | fiz0 = 0.0; |
504 | fix1 = 0.0; |
505 | fiy1 = 0.0; |
506 | fiz1 = 0.0; |
507 | fix2 = 0.0; |
508 | fiy2 = 0.0; |
509 | fiz2 = 0.0; |
510 | |
511 | /* Start inner kernel loop */ |
512 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
513 | { |
514 | /* Get j neighbor index, and coordinate index */ |
515 | jnr = jjnr[jidx]; |
516 | j_coord_offset = DIM3*jnr; |
517 | |
518 | /* load j atom coordinates */ |
519 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
520 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
521 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
522 | |
523 | /* Calculate displacement vector */ |
524 | dx00 = ix0 - jx0; |
525 | dy00 = iy0 - jy0; |
526 | dz00 = iz0 - jz0; |
527 | dx10 = ix1 - jx0; |
528 | dy10 = iy1 - jy0; |
529 | dz10 = iz1 - jz0; |
530 | dx20 = ix2 - jx0; |
531 | dy20 = iy2 - jy0; |
532 | dz20 = iz2 - jz0; |
533 | |
534 | /* Calculate squared distance and things based on it */ |
535 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
536 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
537 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
538 | |
539 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
540 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
541 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
542 | |
543 | rinvsq00 = rinv00*rinv00; |
544 | rinvsq10 = rinv10*rinv10; |
545 | rinvsq20 = rinv20*rinv20; |
546 | |
547 | /* Load parameters for j particles */ |
548 | jq0 = charge[jnr+0]; |
549 | vdwjidx0 = 2*vdwtype[jnr+0]; |
550 | |
551 | /************************** |
552 | * CALCULATE INTERACTIONS * |
553 | **************************/ |
554 | |
555 | if (rsq00<rcutoff2) |
556 | { |
557 | |
558 | r00 = rsq00*rinv00; |
559 | |
560 | qq00 = iq0*jq0; |
561 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
562 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
563 | |
564 | /* Calculate table index by multiplying r with table scale and truncate to integer */ |
565 | rt = r00*vftabscale; |
566 | vfitab = rt; |
567 | vfeps = rt-vfitab; |
568 | vfitab = 2*4*vfitab; |
569 | |
570 | /* REACTION-FIELD ELECTROSTATICS */ |
571 | felec = qq00*(rinv00*rinvsq00-krf2); |
572 | |
573 | /* CUBIC SPLINE TABLE DISPERSION */ |
574 | vfitab += 0; |
575 | F = vftab[vfitab+1]; |
576 | Geps = vfeps*vftab[vfitab+2]; |
577 | Heps2 = vfeps*vfeps*vftab[vfitab+3]; |
578 | Fp = F+Geps+Heps2; |
579 | FF = Fp+Geps+2.0*Heps2; |
580 | fvdw6 = c6_00*FF; |
581 | |
582 | /* CUBIC SPLINE TABLE REPULSION */ |
583 | F = vftab[vfitab+5]; |
584 | Geps = vfeps*vftab[vfitab+6]; |
585 | Heps2 = vfeps*vfeps*vftab[vfitab+7]; |
586 | Fp = F+Geps+Heps2; |
587 | FF = Fp+Geps+2.0*Heps2; |
588 | fvdw12 = c12_00*FF; |
589 | fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00; |
590 | |
591 | fscal = felec+fvdw; |
592 | |
593 | /* Calculate temporary vectorial force */ |
594 | tx = fscal*dx00; |
595 | ty = fscal*dy00; |
596 | tz = fscal*dz00; |
597 | |
598 | /* Update vectorial force */ |
599 | fix0 += tx; |
600 | fiy0 += ty; |
601 | fiz0 += tz; |
602 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
603 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
604 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
605 | |
606 | } |
607 | |
608 | /************************** |
609 | * CALCULATE INTERACTIONS * |
610 | **************************/ |
611 | |
612 | if (rsq10<rcutoff2) |
613 | { |
614 | |
615 | qq10 = iq1*jq0; |
616 | |
617 | /* REACTION-FIELD ELECTROSTATICS */ |
618 | felec = qq10*(rinv10*rinvsq10-krf2); |
619 | |
620 | fscal = felec; |
621 | |
622 | /* Calculate temporary vectorial force */ |
623 | tx = fscal*dx10; |
624 | ty = fscal*dy10; |
625 | tz = fscal*dz10; |
626 | |
627 | /* Update vectorial force */ |
628 | fix1 += tx; |
629 | fiy1 += ty; |
630 | fiz1 += tz; |
631 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
632 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
633 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
634 | |
635 | } |
636 | |
637 | /************************** |
638 | * CALCULATE INTERACTIONS * |
639 | **************************/ |
640 | |
641 | if (rsq20<rcutoff2) |
642 | { |
643 | |
644 | qq20 = iq2*jq0; |
645 | |
646 | /* REACTION-FIELD ELECTROSTATICS */ |
647 | felec = qq20*(rinv20*rinvsq20-krf2); |
648 | |
649 | fscal = felec; |
650 | |
651 | /* Calculate temporary vectorial force */ |
652 | tx = fscal*dx20; |
653 | ty = fscal*dy20; |
654 | tz = fscal*dz20; |
655 | |
656 | /* Update vectorial force */ |
657 | fix2 += tx; |
658 | fiy2 += ty; |
659 | fiz2 += tz; |
660 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
661 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
662 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
663 | |
664 | } |
665 | |
666 | /* Inner loop uses 107 flops */ |
667 | } |
668 | /* End of innermost loop */ |
669 | |
670 | tx = ty = tz = 0; |
671 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
672 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
673 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
674 | tx += fix0; |
675 | ty += fiy0; |
676 | tz += fiz0; |
677 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
678 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
679 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
680 | tx += fix1; |
681 | ty += fiy1; |
682 | tz += fiz1; |
683 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
684 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
685 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
686 | tx += fix2; |
687 | ty += fiy2; |
688 | tz += fiz2; |
689 | fshift[i_shift_offset+XX0] += tx; |
690 | fshift[i_shift_offset+YY1] += ty; |
691 | fshift[i_shift_offset+ZZ2] += tz; |
692 | |
693 | /* Increment number of inner iterations */ |
694 | inneriter += j_index_end - j_index_start; |
695 | |
696 | /* Outer loop uses 30 flops */ |
697 | } |
698 | |
699 | /* Increment number of outer iterations */ |
700 | outeriter += nri; |
701 | |
702 | /* Update outer/inner flops */ |
703 | |
704 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*107)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_F] += outeriter*30 + inneriter *107; |
705 | } |