File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwBhamSh_GeomW3P1_c.c |
Location: | line 449, 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_ElecEwSh_VdwBhamSh_GeomW3P1_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: Buckingham |
53 | * Geometry: Water3-Particle |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEwSh_VdwBhamSh_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 ewitab; |
90 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
91 | real *ewtab; |
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 | nvdwtype = fr->ntype; |
107 | vdwparam = fr->nbfp; |
108 | vdwtype = mdatoms->typeA; |
109 | |
110 | sh_ewald = fr->ic->sh_ewald; |
111 | ewtab = fr->ic->tabq_coul_FDV0; |
112 | ewtabscale = fr->ic->tabq_scale; |
113 | ewtabhalfspace = 0.5/ewtabscale; |
114 | |
115 | /* Setup water-specific parameters */ |
116 | inr = nlist->iinr[0]; |
117 | iq0 = facel*charge[inr+0]; |
118 | iq1 = facel*charge[inr+1]; |
119 | iq2 = facel*charge[inr+2]; |
120 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
121 | |
122 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
123 | rcutoff = fr->rcoulomb; |
124 | rcutoff2 = rcutoff*rcutoff; |
125 | |
126 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
127 | rvdw = fr->rvdw; |
128 | |
129 | outeriter = 0; |
130 | inneriter = 0; |
131 | |
132 | /* Start outer loop over neighborlists */ |
133 | for(iidx=0; iidx<nri; iidx++) |
134 | { |
135 | /* Load shift vector for this list */ |
136 | i_shift_offset = DIM3*shiftidx[iidx]; |
137 | shX = shiftvec[i_shift_offset+XX0]; |
138 | shY = shiftvec[i_shift_offset+YY1]; |
139 | shZ = shiftvec[i_shift_offset+ZZ2]; |
140 | |
141 | /* Load limits for loop over neighbors */ |
142 | j_index_start = jindex[iidx]; |
143 | j_index_end = jindex[iidx+1]; |
144 | |
145 | /* Get outer coordinate index */ |
146 | inr = iinr[iidx]; |
147 | i_coord_offset = DIM3*inr; |
148 | |
149 | /* Load i particle coords and add shift vector */ |
150 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
151 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
152 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
153 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
154 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
155 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
156 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
157 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
158 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
159 | |
160 | fix0 = 0.0; |
161 | fiy0 = 0.0; |
162 | fiz0 = 0.0; |
163 | fix1 = 0.0; |
164 | fiy1 = 0.0; |
165 | fiz1 = 0.0; |
166 | fix2 = 0.0; |
167 | fiy2 = 0.0; |
168 | fiz2 = 0.0; |
169 | |
170 | /* Reset potential sums */ |
171 | velecsum = 0.0; |
172 | vvdwsum = 0.0; |
173 | |
174 | /* Start inner kernel loop */ |
175 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
176 | { |
177 | /* Get j neighbor index, and coordinate index */ |
178 | jnr = jjnr[jidx]; |
179 | j_coord_offset = DIM3*jnr; |
180 | |
181 | /* load j atom coordinates */ |
182 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
183 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
184 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
185 | |
186 | /* Calculate displacement vector */ |
187 | dx00 = ix0 - jx0; |
188 | dy00 = iy0 - jy0; |
189 | dz00 = iz0 - jz0; |
190 | dx10 = ix1 - jx0; |
191 | dy10 = iy1 - jy0; |
192 | dz10 = iz1 - jz0; |
193 | dx20 = ix2 - jx0; |
194 | dy20 = iy2 - jy0; |
195 | dz20 = iz2 - jz0; |
196 | |
197 | /* Calculate squared distance and things based on it */ |
198 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
199 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
200 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
201 | |
202 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
203 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
204 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
205 | |
206 | rinvsq00 = rinv00*rinv00; |
207 | rinvsq10 = rinv10*rinv10; |
208 | rinvsq20 = rinv20*rinv20; |
209 | |
210 | /* Load parameters for j particles */ |
211 | jq0 = charge[jnr+0]; |
212 | vdwjidx0 = 3*vdwtype[jnr+0]; |
213 | |
214 | /************************** |
215 | * CALCULATE INTERACTIONS * |
216 | **************************/ |
217 | |
218 | if (rsq00<rcutoff2) |
219 | { |
220 | |
221 | r00 = rsq00*rinv00; |
222 | |
223 | qq00 = iq0*jq0; |
224 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
225 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
226 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
227 | |
228 | /* EWALD ELECTROSTATICS */ |
229 | |
230 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
231 | ewrt = r00*ewtabscale; |
232 | ewitab = ewrt; |
233 | eweps = ewrt-ewitab; |
234 | ewitab = 4*ewitab; |
235 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
236 | velec = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
237 | felec = qq00*rinv00*(rinvsq00-felec); |
238 | |
239 | /* BUCKINGHAM DISPERSION/REPULSION */ |
240 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
241 | vvdw6 = c6_00*rinvsix; |
242 | br = cexp2_00*r00; |
243 | vvdwexp = cexp1_00*exp(-br); |
244 | vvdw = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0); |
245 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
246 | |
247 | /* Update potential sums from outer loop */ |
248 | velecsum += velec; |
249 | vvdwsum += vvdw; |
250 | |
251 | fscal = felec+fvdw; |
252 | |
253 | /* Calculate temporary vectorial force */ |
254 | tx = fscal*dx00; |
255 | ty = fscal*dy00; |
256 | tz = fscal*dz00; |
257 | |
258 | /* Update vectorial force */ |
259 | fix0 += tx; |
260 | fiy0 += ty; |
261 | fiz0 += tz; |
262 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
263 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
264 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
265 | |
266 | } |
267 | |
268 | /************************** |
269 | * CALCULATE INTERACTIONS * |
270 | **************************/ |
271 | |
272 | if (rsq10<rcutoff2) |
273 | { |
274 | |
275 | r10 = rsq10*rinv10; |
276 | |
277 | qq10 = iq1*jq0; |
278 | |
279 | /* EWALD ELECTROSTATICS */ |
280 | |
281 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
282 | ewrt = r10*ewtabscale; |
283 | ewitab = ewrt; |
284 | eweps = ewrt-ewitab; |
285 | ewitab = 4*ewitab; |
286 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
287 | velec = qq10*((rinv10-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
288 | felec = qq10*rinv10*(rinvsq10-felec); |
289 | |
290 | /* Update potential sums from outer loop */ |
291 | velecsum += velec; |
292 | |
293 | fscal = felec; |
294 | |
295 | /* Calculate temporary vectorial force */ |
296 | tx = fscal*dx10; |
297 | ty = fscal*dy10; |
298 | tz = fscal*dz10; |
299 | |
300 | /* Update vectorial force */ |
301 | fix1 += tx; |
302 | fiy1 += ty; |
303 | fiz1 += tz; |
304 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
305 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
306 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
307 | |
308 | } |
309 | |
310 | /************************** |
311 | * CALCULATE INTERACTIONS * |
312 | **************************/ |
313 | |
314 | if (rsq20<rcutoff2) |
315 | { |
316 | |
317 | r20 = rsq20*rinv20; |
318 | |
319 | qq20 = iq2*jq0; |
320 | |
321 | /* EWALD ELECTROSTATICS */ |
322 | |
323 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
324 | ewrt = r20*ewtabscale; |
325 | ewitab = ewrt; |
326 | eweps = ewrt-ewitab; |
327 | ewitab = 4*ewitab; |
328 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
329 | velec = qq20*((rinv20-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
330 | felec = qq20*rinv20*(rinvsq20-felec); |
331 | |
332 | /* Update potential sums from outer loop */ |
333 | velecsum += velec; |
334 | |
335 | fscal = felec; |
336 | |
337 | /* Calculate temporary vectorial force */ |
338 | tx = fscal*dx20; |
339 | ty = fscal*dy20; |
340 | tz = fscal*dz20; |
341 | |
342 | /* Update vectorial force */ |
343 | fix2 += tx; |
344 | fiy2 += ty; |
345 | fiz2 += tz; |
346 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
347 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
348 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
349 | |
350 | } |
351 | |
352 | /* Inner loop uses 195 flops */ |
353 | } |
354 | /* End of innermost loop */ |
355 | |
356 | tx = ty = tz = 0; |
357 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
358 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
359 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
360 | tx += fix0; |
361 | ty += fiy0; |
362 | tz += fiz0; |
363 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
364 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
365 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
366 | tx += fix1; |
367 | ty += fiy1; |
368 | tz += fiz1; |
369 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
370 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
371 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
372 | tx += fix2; |
373 | ty += fiy2; |
374 | tz += fiz2; |
375 | fshift[i_shift_offset+XX0] += tx; |
376 | fshift[i_shift_offset+YY1] += ty; |
377 | fshift[i_shift_offset+ZZ2] += tz; |
378 | |
379 | ggid = gid[iidx]; |
380 | /* Update potential energies */ |
381 | kernel_data->energygrp_elec[ggid] += velecsum; |
382 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
383 | |
384 | /* Increment number of inner iterations */ |
385 | inneriter += j_index_end - j_index_start; |
386 | |
387 | /* Outer loop uses 32 flops */ |
388 | } |
389 | |
390 | /* Increment number of outer iterations */ |
391 | outeriter += nri; |
392 | |
393 | /* Update outer/inner flops */ |
394 | |
395 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*195)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_VF] += outeriter*32 + inneriter *195; |
396 | } |
397 | /* |
398 | * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwBhamSh_GeomW3P1_F_c |
399 | * Electrostatics interaction: Ewald |
400 | * VdW interaction: Buckingham |
401 | * Geometry: Water3-Particle |
402 | * Calculate force/pot: Force |
403 | */ |
404 | void |
405 | nb_kernel_ElecEwSh_VdwBhamSh_GeomW3P1_F_c |
406 | (t_nblist * gmx_restrict__restrict nlist, |
407 | rvec * gmx_restrict__restrict xx, |
408 | rvec * gmx_restrict__restrict ff, |
409 | t_forcerec * gmx_restrict__restrict fr, |
410 | t_mdatoms * gmx_restrict__restrict mdatoms, |
411 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
412 | t_nrnb * gmx_restrict__restrict nrnb) |
413 | { |
414 | int i_shift_offset,i_coord_offset,j_coord_offset; |
415 | int j_index_start,j_index_end; |
416 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
417 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
418 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
419 | real *shiftvec,*fshift,*x,*f; |
420 | int vdwioffset0; |
421 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
422 | int vdwioffset1; |
423 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
424 | int vdwioffset2; |
425 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
426 | int vdwjidx0; |
427 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
428 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
429 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
430 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
431 | real velec,felec,velecsum,facel,crf,krf,krf2; |
432 | real *charge; |
433 | int nvdwtype; |
434 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
435 | int *vdwtype; |
436 | real *vdwparam; |
437 | int ewitab; |
438 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
439 | real *ewtab; |
440 | |
441 | x = xx[0]; |
442 | f = ff[0]; |
443 | |
444 | nri = nlist->nri; |
445 | iinr = nlist->iinr; |
446 | jindex = nlist->jindex; |
447 | jjnr = nlist->jjnr; |
448 | shiftidx = nlist->shift; |
449 | gid = nlist->gid; |
Value stored to 'gid' is never read | |
450 | shiftvec = fr->shift_vec[0]; |
451 | fshift = fr->fshift[0]; |
452 | facel = fr->epsfac; |
453 | charge = mdatoms->chargeA; |
454 | nvdwtype = fr->ntype; |
455 | vdwparam = fr->nbfp; |
456 | vdwtype = mdatoms->typeA; |
457 | |
458 | sh_ewald = fr->ic->sh_ewald; |
459 | ewtab = fr->ic->tabq_coul_F; |
460 | ewtabscale = fr->ic->tabq_scale; |
461 | ewtabhalfspace = 0.5/ewtabscale; |
462 | |
463 | /* Setup water-specific parameters */ |
464 | inr = nlist->iinr[0]; |
465 | iq0 = facel*charge[inr+0]; |
466 | iq1 = facel*charge[inr+1]; |
467 | iq2 = facel*charge[inr+2]; |
468 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
469 | |
470 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
471 | rcutoff = fr->rcoulomb; |
472 | rcutoff2 = rcutoff*rcutoff; |
473 | |
474 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
475 | rvdw = fr->rvdw; |
476 | |
477 | outeriter = 0; |
478 | inneriter = 0; |
479 | |
480 | /* Start outer loop over neighborlists */ |
481 | for(iidx=0; iidx<nri; iidx++) |
482 | { |
483 | /* Load shift vector for this list */ |
484 | i_shift_offset = DIM3*shiftidx[iidx]; |
485 | shX = shiftvec[i_shift_offset+XX0]; |
486 | shY = shiftvec[i_shift_offset+YY1]; |
487 | shZ = shiftvec[i_shift_offset+ZZ2]; |
488 | |
489 | /* Load limits for loop over neighbors */ |
490 | j_index_start = jindex[iidx]; |
491 | j_index_end = jindex[iidx+1]; |
492 | |
493 | /* Get outer coordinate index */ |
494 | inr = iinr[iidx]; |
495 | i_coord_offset = DIM3*inr; |
496 | |
497 | /* Load i particle coords and add shift vector */ |
498 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
499 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
500 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
501 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
502 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
503 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
504 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
505 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
506 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
507 | |
508 | fix0 = 0.0; |
509 | fiy0 = 0.0; |
510 | fiz0 = 0.0; |
511 | fix1 = 0.0; |
512 | fiy1 = 0.0; |
513 | fiz1 = 0.0; |
514 | fix2 = 0.0; |
515 | fiy2 = 0.0; |
516 | fiz2 = 0.0; |
517 | |
518 | /* Start inner kernel loop */ |
519 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
520 | { |
521 | /* Get j neighbor index, and coordinate index */ |
522 | jnr = jjnr[jidx]; |
523 | j_coord_offset = DIM3*jnr; |
524 | |
525 | /* load j atom coordinates */ |
526 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
527 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
528 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
529 | |
530 | /* Calculate displacement vector */ |
531 | dx00 = ix0 - jx0; |
532 | dy00 = iy0 - jy0; |
533 | dz00 = iz0 - jz0; |
534 | dx10 = ix1 - jx0; |
535 | dy10 = iy1 - jy0; |
536 | dz10 = iz1 - jz0; |
537 | dx20 = ix2 - jx0; |
538 | dy20 = iy2 - jy0; |
539 | dz20 = iz2 - jz0; |
540 | |
541 | /* Calculate squared distance and things based on it */ |
542 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
543 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
544 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
545 | |
546 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
547 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
548 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
549 | |
550 | rinvsq00 = rinv00*rinv00; |
551 | rinvsq10 = rinv10*rinv10; |
552 | rinvsq20 = rinv20*rinv20; |
553 | |
554 | /* Load parameters for j particles */ |
555 | jq0 = charge[jnr+0]; |
556 | vdwjidx0 = 3*vdwtype[jnr+0]; |
557 | |
558 | /************************** |
559 | * CALCULATE INTERACTIONS * |
560 | **************************/ |
561 | |
562 | if (rsq00<rcutoff2) |
563 | { |
564 | |
565 | r00 = rsq00*rinv00; |
566 | |
567 | qq00 = iq0*jq0; |
568 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
569 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
570 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
571 | |
572 | /* EWALD ELECTROSTATICS */ |
573 | |
574 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
575 | ewrt = r00*ewtabscale; |
576 | ewitab = ewrt; |
577 | eweps = ewrt-ewitab; |
578 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
579 | felec = qq00*rinv00*(rinvsq00-felec); |
580 | |
581 | /* BUCKINGHAM DISPERSION/REPULSION */ |
582 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
583 | vvdw6 = c6_00*rinvsix; |
584 | br = cexp2_00*r00; |
585 | vvdwexp = cexp1_00*exp(-br); |
586 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
587 | |
588 | fscal = felec+fvdw; |
589 | |
590 | /* Calculate temporary vectorial force */ |
591 | tx = fscal*dx00; |
592 | ty = fscal*dy00; |
593 | tz = fscal*dz00; |
594 | |
595 | /* Update vectorial force */ |
596 | fix0 += tx; |
597 | fiy0 += ty; |
598 | fiz0 += tz; |
599 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
600 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
601 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
602 | |
603 | } |
604 | |
605 | /************************** |
606 | * CALCULATE INTERACTIONS * |
607 | **************************/ |
608 | |
609 | if (rsq10<rcutoff2) |
610 | { |
611 | |
612 | r10 = rsq10*rinv10; |
613 | |
614 | qq10 = iq1*jq0; |
615 | |
616 | /* EWALD ELECTROSTATICS */ |
617 | |
618 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
619 | ewrt = r10*ewtabscale; |
620 | ewitab = ewrt; |
621 | eweps = ewrt-ewitab; |
622 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
623 | felec = qq10*rinv10*(rinvsq10-felec); |
624 | |
625 | fscal = felec; |
626 | |
627 | /* Calculate temporary vectorial force */ |
628 | tx = fscal*dx10; |
629 | ty = fscal*dy10; |
630 | tz = fscal*dz10; |
631 | |
632 | /* Update vectorial force */ |
633 | fix1 += tx; |
634 | fiy1 += ty; |
635 | fiz1 += tz; |
636 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
637 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
638 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
639 | |
640 | } |
641 | |
642 | /************************** |
643 | * CALCULATE INTERACTIONS * |
644 | **************************/ |
645 | |
646 | if (rsq20<rcutoff2) |
647 | { |
648 | |
649 | r20 = rsq20*rinv20; |
650 | |
651 | qq20 = iq2*jq0; |
652 | |
653 | /* EWALD ELECTROSTATICS */ |
654 | |
655 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
656 | ewrt = r20*ewtabscale; |
657 | ewitab = ewrt; |
658 | eweps = ewrt-ewitab; |
659 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
660 | felec = qq20*rinv20*(rinvsq20-felec); |
661 | |
662 | fscal = felec; |
663 | |
664 | /* Calculate temporary vectorial force */ |
665 | tx = fscal*dx20; |
666 | ty = fscal*dy20; |
667 | tz = fscal*dz20; |
668 | |
669 | /* Update vectorial force */ |
670 | fix2 += tx; |
671 | fiy2 += ty; |
672 | fiz2 += tz; |
673 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
674 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
675 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
676 | |
677 | } |
678 | |
679 | /* Inner loop uses 137 flops */ |
680 | } |
681 | /* End of innermost loop */ |
682 | |
683 | tx = ty = tz = 0; |
684 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
685 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
686 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
687 | tx += fix0; |
688 | ty += fiy0; |
689 | tz += fiz0; |
690 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
691 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
692 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
693 | tx += fix1; |
694 | ty += fiy1; |
695 | tz += fiz1; |
696 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
697 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
698 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
699 | tx += fix2; |
700 | ty += fiy2; |
701 | tz += fiz2; |
702 | fshift[i_shift_offset+XX0] += tx; |
703 | fshift[i_shift_offset+YY1] += ty; |
704 | fshift[i_shift_offset+ZZ2] += tz; |
705 | |
706 | /* Increment number of inner iterations */ |
707 | inneriter += j_index_end - j_index_start; |
708 | |
709 | /* Outer loop uses 30 flops */ |
710 | } |
711 | |
712 | /* Increment number of outer iterations */ |
713 | outeriter += nri; |
714 | |
715 | /* Update outer/inner flops */ |
716 | |
717 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*137)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_F] += outeriter*30 + inneriter *137; |
718 | } |