File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJ_GeomW3P1_c.c |
Location: | line 110, 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_VdwLJ_GeomW3P1_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: LennardJones |
53 | * Geometry: Water3-Particle |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEw_VdwLJ_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; |
Value stored to 'sh_ewald' is never read | |
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 = 2*nvdwtype*vdwtype[inr+0]; |
121 | |
122 | outeriter = 0; |
123 | inneriter = 0; |
124 | |
125 | /* Start outer loop over neighborlists */ |
126 | for(iidx=0; iidx<nri; iidx++) |
127 | { |
128 | /* Load shift vector for this list */ |
129 | i_shift_offset = DIM3*shiftidx[iidx]; |
130 | shX = shiftvec[i_shift_offset+XX0]; |
131 | shY = shiftvec[i_shift_offset+YY1]; |
132 | shZ = shiftvec[i_shift_offset+ZZ2]; |
133 | |
134 | /* Load limits for loop over neighbors */ |
135 | j_index_start = jindex[iidx]; |
136 | j_index_end = jindex[iidx+1]; |
137 | |
138 | /* Get outer coordinate index */ |
139 | inr = iinr[iidx]; |
140 | i_coord_offset = DIM3*inr; |
141 | |
142 | /* Load i particle coords and add shift vector */ |
143 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
144 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
145 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
146 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
147 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
148 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
149 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
150 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
151 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
152 | |
153 | fix0 = 0.0; |
154 | fiy0 = 0.0; |
155 | fiz0 = 0.0; |
156 | fix1 = 0.0; |
157 | fiy1 = 0.0; |
158 | fiz1 = 0.0; |
159 | fix2 = 0.0; |
160 | fiy2 = 0.0; |
161 | fiz2 = 0.0; |
162 | |
163 | /* Reset potential sums */ |
164 | velecsum = 0.0; |
165 | vvdwsum = 0.0; |
166 | |
167 | /* Start inner kernel loop */ |
168 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
169 | { |
170 | /* Get j neighbor index, and coordinate index */ |
171 | jnr = jjnr[jidx]; |
172 | j_coord_offset = DIM3*jnr; |
173 | |
174 | /* load j atom coordinates */ |
175 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
176 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
177 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
178 | |
179 | /* Calculate displacement vector */ |
180 | dx00 = ix0 - jx0; |
181 | dy00 = iy0 - jy0; |
182 | dz00 = iz0 - jz0; |
183 | dx10 = ix1 - jx0; |
184 | dy10 = iy1 - jy0; |
185 | dz10 = iz1 - jz0; |
186 | dx20 = ix2 - jx0; |
187 | dy20 = iy2 - jy0; |
188 | dz20 = iz2 - jz0; |
189 | |
190 | /* Calculate squared distance and things based on it */ |
191 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
192 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
193 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
194 | |
195 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
196 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
197 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
198 | |
199 | rinvsq00 = rinv00*rinv00; |
200 | rinvsq10 = rinv10*rinv10; |
201 | rinvsq20 = rinv20*rinv20; |
202 | |
203 | /* Load parameters for j particles */ |
204 | jq0 = charge[jnr+0]; |
205 | vdwjidx0 = 2*vdwtype[jnr+0]; |
206 | |
207 | /************************** |
208 | * CALCULATE INTERACTIONS * |
209 | **************************/ |
210 | |
211 | r00 = rsq00*rinv00; |
212 | |
213 | qq00 = iq0*jq0; |
214 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
215 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
216 | |
217 | /* EWALD ELECTROSTATICS */ |
218 | |
219 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
220 | ewrt = r00*ewtabscale; |
221 | ewitab = ewrt; |
222 | eweps = ewrt-ewitab; |
223 | ewitab = 4*ewitab; |
224 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
225 | velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
226 | felec = qq00*rinv00*(rinvsq00-felec); |
227 | |
228 | /* LENNARD-JONES DISPERSION/REPULSION */ |
229 | |
230 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
231 | vvdw6 = c6_00*rinvsix; |
232 | vvdw12 = c12_00*rinvsix*rinvsix; |
233 | vvdw = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0); |
234 | fvdw = (vvdw12-vvdw6)*rinvsq00; |
235 | |
236 | /* Update potential sums from outer loop */ |
237 | velecsum += velec; |
238 | vvdwsum += vvdw; |
239 | |
240 | fscal = felec+fvdw; |
241 | |
242 | /* Calculate temporary vectorial force */ |
243 | tx = fscal*dx00; |
244 | ty = fscal*dy00; |
245 | tz = fscal*dz00; |
246 | |
247 | /* Update vectorial force */ |
248 | fix0 += tx; |
249 | fiy0 += ty; |
250 | fiz0 += tz; |
251 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
252 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
253 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
254 | |
255 | /************************** |
256 | * CALCULATE INTERACTIONS * |
257 | **************************/ |
258 | |
259 | r10 = rsq10*rinv10; |
260 | |
261 | qq10 = iq1*jq0; |
262 | |
263 | /* EWALD ELECTROSTATICS */ |
264 | |
265 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
266 | ewrt = r10*ewtabscale; |
267 | ewitab = ewrt; |
268 | eweps = ewrt-ewitab; |
269 | ewitab = 4*ewitab; |
270 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
271 | velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
272 | felec = qq10*rinv10*(rinvsq10-felec); |
273 | |
274 | /* Update potential sums from outer loop */ |
275 | velecsum += velec; |
276 | |
277 | fscal = felec; |
278 | |
279 | /* Calculate temporary vectorial force */ |
280 | tx = fscal*dx10; |
281 | ty = fscal*dy10; |
282 | tz = fscal*dz10; |
283 | |
284 | /* Update vectorial force */ |
285 | fix1 += tx; |
286 | fiy1 += ty; |
287 | fiz1 += tz; |
288 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
289 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
290 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
291 | |
292 | /************************** |
293 | * CALCULATE INTERACTIONS * |
294 | **************************/ |
295 | |
296 | r20 = rsq20*rinv20; |
297 | |
298 | qq20 = iq2*jq0; |
299 | |
300 | /* EWALD ELECTROSTATICS */ |
301 | |
302 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
303 | ewrt = r20*ewtabscale; |
304 | ewitab = ewrt; |
305 | eweps = ewrt-ewitab; |
306 | ewitab = 4*ewitab; |
307 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
308 | velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
309 | felec = qq20*rinv20*(rinvsq20-felec); |
310 | |
311 | /* Update potential sums from outer loop */ |
312 | velecsum += velec; |
313 | |
314 | fscal = felec; |
315 | |
316 | /* Calculate temporary vectorial force */ |
317 | tx = fscal*dx20; |
318 | ty = fscal*dy20; |
319 | tz = fscal*dz20; |
320 | |
321 | /* Update vectorial force */ |
322 | fix2 += tx; |
323 | fiy2 += ty; |
324 | fiz2 += tz; |
325 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
326 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
327 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
328 | |
329 | /* Inner loop uses 135 flops */ |
330 | } |
331 | /* End of innermost loop */ |
332 | |
333 | tx = ty = tz = 0; |
334 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
335 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
336 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
337 | tx += fix0; |
338 | ty += fiy0; |
339 | tz += fiz0; |
340 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
341 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
342 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
343 | tx += fix1; |
344 | ty += fiy1; |
345 | tz += fiz1; |
346 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
347 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
348 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
349 | tx += fix2; |
350 | ty += fiy2; |
351 | tz += fiz2; |
352 | fshift[i_shift_offset+XX0] += tx; |
353 | fshift[i_shift_offset+YY1] += ty; |
354 | fshift[i_shift_offset+ZZ2] += tz; |
355 | |
356 | ggid = gid[iidx]; |
357 | /* Update potential energies */ |
358 | kernel_data->energygrp_elec[ggid] += velecsum; |
359 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
360 | |
361 | /* Increment number of inner iterations */ |
362 | inneriter += j_index_end - j_index_start; |
363 | |
364 | /* Outer loop uses 32 flops */ |
365 | } |
366 | |
367 | /* Increment number of outer iterations */ |
368 | outeriter += nri; |
369 | |
370 | /* Update outer/inner flops */ |
371 | |
372 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*135)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_VF] += outeriter*32 + inneriter *135; |
373 | } |
374 | /* |
375 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJ_GeomW3P1_F_c |
376 | * Electrostatics interaction: Ewald |
377 | * VdW interaction: LennardJones |
378 | * Geometry: Water3-Particle |
379 | * Calculate force/pot: Force |
380 | */ |
381 | void |
382 | nb_kernel_ElecEw_VdwLJ_GeomW3P1_F_c |
383 | (t_nblist * gmx_restrict__restrict nlist, |
384 | rvec * gmx_restrict__restrict xx, |
385 | rvec * gmx_restrict__restrict ff, |
386 | t_forcerec * gmx_restrict__restrict fr, |
387 | t_mdatoms * gmx_restrict__restrict mdatoms, |
388 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
389 | t_nrnb * gmx_restrict__restrict nrnb) |
390 | { |
391 | int i_shift_offset,i_coord_offset,j_coord_offset; |
392 | int j_index_start,j_index_end; |
393 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
394 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
395 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
396 | real *shiftvec,*fshift,*x,*f; |
397 | int vdwioffset0; |
398 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
399 | int vdwioffset1; |
400 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
401 | int vdwioffset2; |
402 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
403 | int vdwjidx0; |
404 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
405 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
406 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
407 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
408 | real velec,felec,velecsum,facel,crf,krf,krf2; |
409 | real *charge; |
410 | int nvdwtype; |
411 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
412 | int *vdwtype; |
413 | real *vdwparam; |
414 | int ewitab; |
415 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
416 | real *ewtab; |
417 | |
418 | x = xx[0]; |
419 | f = ff[0]; |
420 | |
421 | nri = nlist->nri; |
422 | iinr = nlist->iinr; |
423 | jindex = nlist->jindex; |
424 | jjnr = nlist->jjnr; |
425 | shiftidx = nlist->shift; |
426 | gid = nlist->gid; |
427 | shiftvec = fr->shift_vec[0]; |
428 | fshift = fr->fshift[0]; |
429 | facel = fr->epsfac; |
430 | charge = mdatoms->chargeA; |
431 | nvdwtype = fr->ntype; |
432 | vdwparam = fr->nbfp; |
433 | vdwtype = mdatoms->typeA; |
434 | |
435 | sh_ewald = fr->ic->sh_ewald; |
436 | ewtab = fr->ic->tabq_coul_F; |
437 | ewtabscale = fr->ic->tabq_scale; |
438 | ewtabhalfspace = 0.5/ewtabscale; |
439 | |
440 | /* Setup water-specific parameters */ |
441 | inr = nlist->iinr[0]; |
442 | iq0 = facel*charge[inr+0]; |
443 | iq1 = facel*charge[inr+1]; |
444 | iq2 = facel*charge[inr+2]; |
445 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
446 | |
447 | outeriter = 0; |
448 | inneriter = 0; |
449 | |
450 | /* Start outer loop over neighborlists */ |
451 | for(iidx=0; iidx<nri; iidx++) |
452 | { |
453 | /* Load shift vector for this list */ |
454 | i_shift_offset = DIM3*shiftidx[iidx]; |
455 | shX = shiftvec[i_shift_offset+XX0]; |
456 | shY = shiftvec[i_shift_offset+YY1]; |
457 | shZ = shiftvec[i_shift_offset+ZZ2]; |
458 | |
459 | /* Load limits for loop over neighbors */ |
460 | j_index_start = jindex[iidx]; |
461 | j_index_end = jindex[iidx+1]; |
462 | |
463 | /* Get outer coordinate index */ |
464 | inr = iinr[iidx]; |
465 | i_coord_offset = DIM3*inr; |
466 | |
467 | /* Load i particle coords and add shift vector */ |
468 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
469 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
470 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
471 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
472 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
473 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
474 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
475 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
476 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
477 | |
478 | fix0 = 0.0; |
479 | fiy0 = 0.0; |
480 | fiz0 = 0.0; |
481 | fix1 = 0.0; |
482 | fiy1 = 0.0; |
483 | fiz1 = 0.0; |
484 | fix2 = 0.0; |
485 | fiy2 = 0.0; |
486 | fiz2 = 0.0; |
487 | |
488 | /* Start inner kernel loop */ |
489 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
490 | { |
491 | /* Get j neighbor index, and coordinate index */ |
492 | jnr = jjnr[jidx]; |
493 | j_coord_offset = DIM3*jnr; |
494 | |
495 | /* load j atom coordinates */ |
496 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
497 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
498 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
499 | |
500 | /* Calculate displacement vector */ |
501 | dx00 = ix0 - jx0; |
502 | dy00 = iy0 - jy0; |
503 | dz00 = iz0 - jz0; |
504 | dx10 = ix1 - jx0; |
505 | dy10 = iy1 - jy0; |
506 | dz10 = iz1 - jz0; |
507 | dx20 = ix2 - jx0; |
508 | dy20 = iy2 - jy0; |
509 | dz20 = iz2 - jz0; |
510 | |
511 | /* Calculate squared distance and things based on it */ |
512 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
513 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
514 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
515 | |
516 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
517 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
518 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
519 | |
520 | rinvsq00 = rinv00*rinv00; |
521 | rinvsq10 = rinv10*rinv10; |
522 | rinvsq20 = rinv20*rinv20; |
523 | |
524 | /* Load parameters for j particles */ |
525 | jq0 = charge[jnr+0]; |
526 | vdwjidx0 = 2*vdwtype[jnr+0]; |
527 | |
528 | /************************** |
529 | * CALCULATE INTERACTIONS * |
530 | **************************/ |
531 | |
532 | r00 = rsq00*rinv00; |
533 | |
534 | qq00 = iq0*jq0; |
535 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
536 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
537 | |
538 | /* EWALD ELECTROSTATICS */ |
539 | |
540 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
541 | ewrt = r00*ewtabscale; |
542 | ewitab = ewrt; |
543 | eweps = ewrt-ewitab; |
544 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
545 | felec = qq00*rinv00*(rinvsq00-felec); |
546 | |
547 | /* LENNARD-JONES DISPERSION/REPULSION */ |
548 | |
549 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
550 | fvdw = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00; |
551 | |
552 | fscal = felec+fvdw; |
553 | |
554 | /* Calculate temporary vectorial force */ |
555 | tx = fscal*dx00; |
556 | ty = fscal*dy00; |
557 | tz = fscal*dz00; |
558 | |
559 | /* Update vectorial force */ |
560 | fix0 += tx; |
561 | fiy0 += ty; |
562 | fiz0 += tz; |
563 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
564 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
565 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
566 | |
567 | /************************** |
568 | * CALCULATE INTERACTIONS * |
569 | **************************/ |
570 | |
571 | r10 = rsq10*rinv10; |
572 | |
573 | qq10 = iq1*jq0; |
574 | |
575 | /* EWALD ELECTROSTATICS */ |
576 | |
577 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
578 | ewrt = r10*ewtabscale; |
579 | ewitab = ewrt; |
580 | eweps = ewrt-ewitab; |
581 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
582 | felec = qq10*rinv10*(rinvsq10-felec); |
583 | |
584 | fscal = felec; |
585 | |
586 | /* Calculate temporary vectorial force */ |
587 | tx = fscal*dx10; |
588 | ty = fscal*dy10; |
589 | tz = fscal*dz10; |
590 | |
591 | /* Update vectorial force */ |
592 | fix1 += tx; |
593 | fiy1 += ty; |
594 | fiz1 += tz; |
595 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
596 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
597 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
598 | |
599 | /************************** |
600 | * CALCULATE INTERACTIONS * |
601 | **************************/ |
602 | |
603 | r20 = rsq20*rinv20; |
604 | |
605 | qq20 = iq2*jq0; |
606 | |
607 | /* EWALD ELECTROSTATICS */ |
608 | |
609 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
610 | ewrt = r20*ewtabscale; |
611 | ewitab = ewrt; |
612 | eweps = ewrt-ewitab; |
613 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
614 | felec = qq20*rinv20*(rinvsq20-felec); |
615 | |
616 | fscal = felec; |
617 | |
618 | /* Calculate temporary vectorial force */ |
619 | tx = fscal*dx20; |
620 | ty = fscal*dy20; |
621 | tz = fscal*dz20; |
622 | |
623 | /* Update vectorial force */ |
624 | fix2 += tx; |
625 | fiy2 += ty; |
626 | fiz2 += tz; |
627 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
628 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
629 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
630 | |
631 | /* Inner loop uses 109 flops */ |
632 | } |
633 | /* End of innermost loop */ |
634 | |
635 | tx = ty = tz = 0; |
636 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
637 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
638 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
639 | tx += fix0; |
640 | ty += fiy0; |
641 | tz += fiz0; |
642 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
643 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
644 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
645 | tx += fix1; |
646 | ty += fiy1; |
647 | tz += fiz1; |
648 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
649 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
650 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
651 | tx += fix2; |
652 | ty += fiy2; |
653 | tz += fiz2; |
654 | fshift[i_shift_offset+XX0] += tx; |
655 | fshift[i_shift_offset+YY1] += ty; |
656 | fshift[i_shift_offset+ZZ2] += tz; |
657 | |
658 | /* Increment number of inner iterations */ |
659 | inneriter += j_index_end - j_index_start; |
660 | |
661 | /* Outer loop uses 30 flops */ |
662 | } |
663 | |
664 | /* Increment number of outer iterations */ |
665 | outeriter += nri; |
666 | |
667 | /* Update outer/inner flops */ |
668 | |
669 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*109)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_F] += outeriter*30 + inneriter *109; |
670 | } |