File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwLJEwSh_GeomW3W3_c.c |
Location: | line 793, column 5 |
Description: | Value stored to 'sh_lj_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_ElecEwSh_VdwLJEwSh_GeomW3W3_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: LJEwald |
53 | * Geometry: Water3-Water3 |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEwSh_VdwLJEwSh_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 | real c6grid_00; |
100 | real c6grid_01; |
101 | real c6grid_02; |
102 | real c6grid_10; |
103 | real c6grid_11; |
104 | real c6grid_12; |
105 | real c6grid_20; |
106 | real c6grid_21; |
107 | real c6grid_22; |
108 | real ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald; |
109 | real *vdwgridparam; |
110 | int ewitab; |
111 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
112 | real *ewtab; |
113 | |
114 | x = xx[0]; |
115 | f = ff[0]; |
116 | |
117 | nri = nlist->nri; |
118 | iinr = nlist->iinr; |
119 | jindex = nlist->jindex; |
120 | jjnr = nlist->jjnr; |
121 | shiftidx = nlist->shift; |
122 | gid = nlist->gid; |
123 | shiftvec = fr->shift_vec[0]; |
124 | fshift = fr->fshift[0]; |
125 | facel = fr->epsfac; |
126 | charge = mdatoms->chargeA; |
127 | nvdwtype = fr->ntype; |
128 | vdwparam = fr->nbfp; |
129 | vdwtype = mdatoms->typeA; |
130 | vdwgridparam = fr->ljpme_c6grid; |
131 | ewclj = fr->ewaldcoeff_lj; |
132 | sh_lj_ewald = fr->ic->sh_lj_ewald; |
133 | ewclj2 = ewclj*ewclj; |
134 | ewclj6 = ewclj2*ewclj2*ewclj2; |
135 | |
136 | sh_ewald = fr->ic->sh_ewald; |
137 | ewtab = fr->ic->tabq_coul_FDV0; |
138 | ewtabscale = fr->ic->tabq_scale; |
139 | ewtabhalfspace = 0.5/ewtabscale; |
140 | |
141 | /* Setup water-specific parameters */ |
142 | inr = nlist->iinr[0]; |
143 | iq0 = facel*charge[inr+0]; |
144 | iq1 = facel*charge[inr+1]; |
145 | iq2 = facel*charge[inr+2]; |
146 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
147 | |
148 | jq0 = charge[inr+0]; |
149 | jq1 = charge[inr+1]; |
150 | jq2 = charge[inr+2]; |
151 | vdwjidx0 = 2*vdwtype[inr+0]; |
152 | qq00 = iq0*jq0; |
153 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
154 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
155 | c6grid_00 = vdwgridparam[vdwioffset0+vdwjidx0]; |
156 | qq01 = iq0*jq1; |
157 | qq02 = iq0*jq2; |
158 | qq10 = iq1*jq0; |
159 | qq11 = iq1*jq1; |
160 | qq12 = iq1*jq2; |
161 | qq20 = iq2*jq0; |
162 | qq21 = iq2*jq1; |
163 | qq22 = iq2*jq2; |
164 | |
165 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
166 | rcutoff = fr->rcoulomb; |
167 | rcutoff2 = rcutoff*rcutoff; |
168 | |
169 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
170 | rvdw = fr->rvdw; |
171 | |
172 | outeriter = 0; |
173 | inneriter = 0; |
174 | |
175 | /* Start outer loop over neighborlists */ |
176 | for(iidx=0; iidx<nri; iidx++) |
177 | { |
178 | /* Load shift vector for this list */ |
179 | i_shift_offset = DIM3*shiftidx[iidx]; |
180 | shX = shiftvec[i_shift_offset+XX0]; |
181 | shY = shiftvec[i_shift_offset+YY1]; |
182 | shZ = shiftvec[i_shift_offset+ZZ2]; |
183 | |
184 | /* Load limits for loop over neighbors */ |
185 | j_index_start = jindex[iidx]; |
186 | j_index_end = jindex[iidx+1]; |
187 | |
188 | /* Get outer coordinate index */ |
189 | inr = iinr[iidx]; |
190 | i_coord_offset = DIM3*inr; |
191 | |
192 | /* Load i particle coords and add shift vector */ |
193 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
194 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
195 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
196 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
197 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
198 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
199 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
200 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
201 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
202 | |
203 | fix0 = 0.0; |
204 | fiy0 = 0.0; |
205 | fiz0 = 0.0; |
206 | fix1 = 0.0; |
207 | fiy1 = 0.0; |
208 | fiz1 = 0.0; |
209 | fix2 = 0.0; |
210 | fiy2 = 0.0; |
211 | fiz2 = 0.0; |
212 | |
213 | /* Reset potential sums */ |
214 | velecsum = 0.0; |
215 | vvdwsum = 0.0; |
216 | |
217 | /* Start inner kernel loop */ |
218 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
219 | { |
220 | /* Get j neighbor index, and coordinate index */ |
221 | jnr = jjnr[jidx]; |
222 | j_coord_offset = DIM3*jnr; |
223 | |
224 | /* load j atom coordinates */ |
225 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
226 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
227 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
228 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
229 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
230 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
231 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
232 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
233 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
234 | |
235 | /* Calculate displacement vector */ |
236 | dx00 = ix0 - jx0; |
237 | dy00 = iy0 - jy0; |
238 | dz00 = iz0 - jz0; |
239 | dx01 = ix0 - jx1; |
240 | dy01 = iy0 - jy1; |
241 | dz01 = iz0 - jz1; |
242 | dx02 = ix0 - jx2; |
243 | dy02 = iy0 - jy2; |
244 | dz02 = iz0 - jz2; |
245 | dx10 = ix1 - jx0; |
246 | dy10 = iy1 - jy0; |
247 | dz10 = iz1 - jz0; |
248 | dx11 = ix1 - jx1; |
249 | dy11 = iy1 - jy1; |
250 | dz11 = iz1 - jz1; |
251 | dx12 = ix1 - jx2; |
252 | dy12 = iy1 - jy2; |
253 | dz12 = iz1 - jz2; |
254 | dx20 = ix2 - jx0; |
255 | dy20 = iy2 - jy0; |
256 | dz20 = iz2 - jz0; |
257 | dx21 = ix2 - jx1; |
258 | dy21 = iy2 - jy1; |
259 | dz21 = iz2 - jz1; |
260 | dx22 = ix2 - jx2; |
261 | dy22 = iy2 - jy2; |
262 | dz22 = iz2 - jz2; |
263 | |
264 | /* Calculate squared distance and things based on it */ |
265 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
266 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
267 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
268 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
269 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
270 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
271 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
272 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
273 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
274 | |
275 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
276 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
277 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
278 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
279 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
280 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
281 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
282 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
283 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
284 | |
285 | rinvsq00 = rinv00*rinv00; |
286 | rinvsq01 = rinv01*rinv01; |
287 | rinvsq02 = rinv02*rinv02; |
288 | rinvsq10 = rinv10*rinv10; |
289 | rinvsq11 = rinv11*rinv11; |
290 | rinvsq12 = rinv12*rinv12; |
291 | rinvsq20 = rinv20*rinv20; |
292 | rinvsq21 = rinv21*rinv21; |
293 | rinvsq22 = rinv22*rinv22; |
294 | |
295 | /************************** |
296 | * CALCULATE INTERACTIONS * |
297 | **************************/ |
298 | |
299 | if (rsq00<rcutoff2) |
300 | { |
301 | |
302 | r00 = rsq00*rinv00; |
303 | |
304 | /* EWALD ELECTROSTATICS */ |
305 | |
306 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
307 | ewrt = r00*ewtabscale; |
308 | ewitab = ewrt; |
309 | eweps = ewrt-ewitab; |
310 | ewitab = 4*ewitab; |
311 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
312 | velec = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
313 | felec = qq00*rinv00*(rinvsq00-felec); |
314 | |
315 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
316 | ewcljrsq = ewclj2*rsq00; |
317 | exponent = exp(-ewcljrsq); |
318 | poly = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5); |
319 | vvdw6 = (c6_00-c6grid_00*(1.0-poly))*rinvsix; |
320 | vvdw12 = c12_00*rinvsix*rinvsix; |
321 | vvdw = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6 - c6grid_00*sh_lj_ewald)*(1.0/6.0); |
322 | fvdw = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00; |
323 | |
324 | /* Update potential sums from outer loop */ |
325 | velecsum += velec; |
326 | vvdwsum += vvdw; |
327 | |
328 | fscal = felec+fvdw; |
329 | |
330 | /* Calculate temporary vectorial force */ |
331 | tx = fscal*dx00; |
332 | ty = fscal*dy00; |
333 | tz = fscal*dz00; |
334 | |
335 | /* Update vectorial force */ |
336 | fix0 += tx; |
337 | fiy0 += ty; |
338 | fiz0 += tz; |
339 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
340 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
341 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
342 | |
343 | } |
344 | |
345 | /************************** |
346 | * CALCULATE INTERACTIONS * |
347 | **************************/ |
348 | |
349 | if (rsq01<rcutoff2) |
350 | { |
351 | |
352 | r01 = rsq01*rinv01; |
353 | |
354 | /* EWALD ELECTROSTATICS */ |
355 | |
356 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
357 | ewrt = r01*ewtabscale; |
358 | ewitab = ewrt; |
359 | eweps = ewrt-ewitab; |
360 | ewitab = 4*ewitab; |
361 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
362 | velec = qq01*((rinv01-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
363 | felec = qq01*rinv01*(rinvsq01-felec); |
364 | |
365 | /* Update potential sums from outer loop */ |
366 | velecsum += velec; |
367 | |
368 | fscal = felec; |
369 | |
370 | /* Calculate temporary vectorial force */ |
371 | tx = fscal*dx01; |
372 | ty = fscal*dy01; |
373 | tz = fscal*dz01; |
374 | |
375 | /* Update vectorial force */ |
376 | fix0 += tx; |
377 | fiy0 += ty; |
378 | fiz0 += tz; |
379 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
380 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
381 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
382 | |
383 | } |
384 | |
385 | /************************** |
386 | * CALCULATE INTERACTIONS * |
387 | **************************/ |
388 | |
389 | if (rsq02<rcutoff2) |
390 | { |
391 | |
392 | r02 = rsq02*rinv02; |
393 | |
394 | /* EWALD ELECTROSTATICS */ |
395 | |
396 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
397 | ewrt = r02*ewtabscale; |
398 | ewitab = ewrt; |
399 | eweps = ewrt-ewitab; |
400 | ewitab = 4*ewitab; |
401 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
402 | velec = qq02*((rinv02-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
403 | felec = qq02*rinv02*(rinvsq02-felec); |
404 | |
405 | /* Update potential sums from outer loop */ |
406 | velecsum += velec; |
407 | |
408 | fscal = felec; |
409 | |
410 | /* Calculate temporary vectorial force */ |
411 | tx = fscal*dx02; |
412 | ty = fscal*dy02; |
413 | tz = fscal*dz02; |
414 | |
415 | /* Update vectorial force */ |
416 | fix0 += tx; |
417 | fiy0 += ty; |
418 | fiz0 += tz; |
419 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
420 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
421 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
422 | |
423 | } |
424 | |
425 | /************************** |
426 | * CALCULATE INTERACTIONS * |
427 | **************************/ |
428 | |
429 | if (rsq10<rcutoff2) |
430 | { |
431 | |
432 | r10 = rsq10*rinv10; |
433 | |
434 | /* EWALD ELECTROSTATICS */ |
435 | |
436 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
437 | ewrt = r10*ewtabscale; |
438 | ewitab = ewrt; |
439 | eweps = ewrt-ewitab; |
440 | ewitab = 4*ewitab; |
441 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
442 | velec = qq10*((rinv10-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
443 | felec = qq10*rinv10*(rinvsq10-felec); |
444 | |
445 | /* Update potential sums from outer loop */ |
446 | velecsum += velec; |
447 | |
448 | fscal = felec; |
449 | |
450 | /* Calculate temporary vectorial force */ |
451 | tx = fscal*dx10; |
452 | ty = fscal*dy10; |
453 | tz = fscal*dz10; |
454 | |
455 | /* Update vectorial force */ |
456 | fix1 += tx; |
457 | fiy1 += ty; |
458 | fiz1 += tz; |
459 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
460 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
461 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
462 | |
463 | } |
464 | |
465 | /************************** |
466 | * CALCULATE INTERACTIONS * |
467 | **************************/ |
468 | |
469 | if (rsq11<rcutoff2) |
470 | { |
471 | |
472 | r11 = rsq11*rinv11; |
473 | |
474 | /* EWALD ELECTROSTATICS */ |
475 | |
476 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
477 | ewrt = r11*ewtabscale; |
478 | ewitab = ewrt; |
479 | eweps = ewrt-ewitab; |
480 | ewitab = 4*ewitab; |
481 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
482 | velec = qq11*((rinv11-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
483 | felec = qq11*rinv11*(rinvsq11-felec); |
484 | |
485 | /* Update potential sums from outer loop */ |
486 | velecsum += velec; |
487 | |
488 | fscal = felec; |
489 | |
490 | /* Calculate temporary vectorial force */ |
491 | tx = fscal*dx11; |
492 | ty = fscal*dy11; |
493 | tz = fscal*dz11; |
494 | |
495 | /* Update vectorial force */ |
496 | fix1 += tx; |
497 | fiy1 += ty; |
498 | fiz1 += tz; |
499 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
500 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
501 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
502 | |
503 | } |
504 | |
505 | /************************** |
506 | * CALCULATE INTERACTIONS * |
507 | **************************/ |
508 | |
509 | if (rsq12<rcutoff2) |
510 | { |
511 | |
512 | r12 = rsq12*rinv12; |
513 | |
514 | /* EWALD ELECTROSTATICS */ |
515 | |
516 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
517 | ewrt = r12*ewtabscale; |
518 | ewitab = ewrt; |
519 | eweps = ewrt-ewitab; |
520 | ewitab = 4*ewitab; |
521 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
522 | velec = qq12*((rinv12-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
523 | felec = qq12*rinv12*(rinvsq12-felec); |
524 | |
525 | /* Update potential sums from outer loop */ |
526 | velecsum += velec; |
527 | |
528 | fscal = felec; |
529 | |
530 | /* Calculate temporary vectorial force */ |
531 | tx = fscal*dx12; |
532 | ty = fscal*dy12; |
533 | tz = fscal*dz12; |
534 | |
535 | /* Update vectorial force */ |
536 | fix1 += tx; |
537 | fiy1 += ty; |
538 | fiz1 += tz; |
539 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
540 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
541 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
542 | |
543 | } |
544 | |
545 | /************************** |
546 | * CALCULATE INTERACTIONS * |
547 | **************************/ |
548 | |
549 | if (rsq20<rcutoff2) |
550 | { |
551 | |
552 | r20 = rsq20*rinv20; |
553 | |
554 | /* EWALD ELECTROSTATICS */ |
555 | |
556 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
557 | ewrt = r20*ewtabscale; |
558 | ewitab = ewrt; |
559 | eweps = ewrt-ewitab; |
560 | ewitab = 4*ewitab; |
561 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
562 | velec = qq20*((rinv20-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
563 | felec = qq20*rinv20*(rinvsq20-felec); |
564 | |
565 | /* Update potential sums from outer loop */ |
566 | velecsum += velec; |
567 | |
568 | fscal = felec; |
569 | |
570 | /* Calculate temporary vectorial force */ |
571 | tx = fscal*dx20; |
572 | ty = fscal*dy20; |
573 | tz = fscal*dz20; |
574 | |
575 | /* Update vectorial force */ |
576 | fix2 += tx; |
577 | fiy2 += ty; |
578 | fiz2 += tz; |
579 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
580 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
581 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
582 | |
583 | } |
584 | |
585 | /************************** |
586 | * CALCULATE INTERACTIONS * |
587 | **************************/ |
588 | |
589 | if (rsq21<rcutoff2) |
590 | { |
591 | |
592 | r21 = rsq21*rinv21; |
593 | |
594 | /* EWALD ELECTROSTATICS */ |
595 | |
596 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
597 | ewrt = r21*ewtabscale; |
598 | ewitab = ewrt; |
599 | eweps = ewrt-ewitab; |
600 | ewitab = 4*ewitab; |
601 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
602 | velec = qq21*((rinv21-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
603 | felec = qq21*rinv21*(rinvsq21-felec); |
604 | |
605 | /* Update potential sums from outer loop */ |
606 | velecsum += velec; |
607 | |
608 | fscal = felec; |
609 | |
610 | /* Calculate temporary vectorial force */ |
611 | tx = fscal*dx21; |
612 | ty = fscal*dy21; |
613 | tz = fscal*dz21; |
614 | |
615 | /* Update vectorial force */ |
616 | fix2 += tx; |
617 | fiy2 += ty; |
618 | fiz2 += tz; |
619 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
620 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
621 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
622 | |
623 | } |
624 | |
625 | /************************** |
626 | * CALCULATE INTERACTIONS * |
627 | **************************/ |
628 | |
629 | if (rsq22<rcutoff2) |
630 | { |
631 | |
632 | r22 = rsq22*rinv22; |
633 | |
634 | /* EWALD ELECTROSTATICS */ |
635 | |
636 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
637 | ewrt = r22*ewtabscale; |
638 | ewitab = ewrt; |
639 | eweps = ewrt-ewitab; |
640 | ewitab = 4*ewitab; |
641 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
642 | velec = qq22*((rinv22-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
643 | felec = qq22*rinv22*(rinvsq22-felec); |
644 | |
645 | /* Update potential sums from outer loop */ |
646 | velecsum += velec; |
647 | |
648 | fscal = felec; |
649 | |
650 | /* Calculate temporary vectorial force */ |
651 | tx = fscal*dx22; |
652 | ty = fscal*dy22; |
653 | tz = fscal*dz22; |
654 | |
655 | /* Update vectorial force */ |
656 | fix2 += tx; |
657 | fiy2 += ty; |
658 | fiz2 += tz; |
659 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
660 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
661 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
662 | |
663 | } |
664 | |
665 | /* Inner loop uses 401 flops */ |
666 | } |
667 | /* End of innermost loop */ |
668 | |
669 | tx = ty = tz = 0; |
670 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
671 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
672 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
673 | tx += fix0; |
674 | ty += fiy0; |
675 | tz += fiz0; |
676 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
677 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
678 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
679 | tx += fix1; |
680 | ty += fiy1; |
681 | tz += fiz1; |
682 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
683 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
684 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
685 | tx += fix2; |
686 | ty += fiy2; |
687 | tz += fiz2; |
688 | fshift[i_shift_offset+XX0] += tx; |
689 | fshift[i_shift_offset+YY1] += ty; |
690 | fshift[i_shift_offset+ZZ2] += tz; |
691 | |
692 | ggid = gid[iidx]; |
693 | /* Update potential energies */ |
694 | kernel_data->energygrp_elec[ggid] += velecsum; |
695 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
696 | |
697 | /* Increment number of inner iterations */ |
698 | inneriter += j_index_end - j_index_start; |
699 | |
700 | /* Outer loop uses 32 flops */ |
701 | } |
702 | |
703 | /* Increment number of outer iterations */ |
704 | outeriter += nri; |
705 | |
706 | /* Update outer/inner flops */ |
707 | |
708 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*401)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*32 + inneriter*401; |
709 | } |
710 | /* |
711 | * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwLJEwSh_GeomW3W3_F_c |
712 | * Electrostatics interaction: Ewald |
713 | * VdW interaction: LJEwald |
714 | * Geometry: Water3-Water3 |
715 | * Calculate force/pot: Force |
716 | */ |
717 | void |
718 | nb_kernel_ElecEwSh_VdwLJEwSh_GeomW3W3_F_c |
719 | (t_nblist * gmx_restrict__restrict nlist, |
720 | rvec * gmx_restrict__restrict xx, |
721 | rvec * gmx_restrict__restrict ff, |
722 | t_forcerec * gmx_restrict__restrict fr, |
723 | t_mdatoms * gmx_restrict__restrict mdatoms, |
724 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
725 | t_nrnb * gmx_restrict__restrict nrnb) |
726 | { |
727 | int i_shift_offset,i_coord_offset,j_coord_offset; |
728 | int j_index_start,j_index_end; |
729 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
730 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
731 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
732 | real *shiftvec,*fshift,*x,*f; |
733 | int vdwioffset0; |
734 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
735 | int vdwioffset1; |
736 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
737 | int vdwioffset2; |
738 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
739 | int vdwjidx0; |
740 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
741 | int vdwjidx1; |
742 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
743 | int vdwjidx2; |
744 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
745 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
746 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
747 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
748 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
749 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
750 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
751 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
752 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
753 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
754 | real velec,felec,velecsum,facel,crf,krf,krf2; |
755 | real *charge; |
756 | int nvdwtype; |
757 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
758 | int *vdwtype; |
759 | real *vdwparam; |
760 | real c6grid_00; |
761 | real c6grid_01; |
762 | real c6grid_02; |
763 | real c6grid_10; |
764 | real c6grid_11; |
765 | real c6grid_12; |
766 | real c6grid_20; |
767 | real c6grid_21; |
768 | real c6grid_22; |
769 | real ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald; |
770 | real *vdwgridparam; |
771 | int ewitab; |
772 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
773 | real *ewtab; |
774 | |
775 | x = xx[0]; |
776 | f = ff[0]; |
777 | |
778 | nri = nlist->nri; |
779 | iinr = nlist->iinr; |
780 | jindex = nlist->jindex; |
781 | jjnr = nlist->jjnr; |
782 | shiftidx = nlist->shift; |
783 | gid = nlist->gid; |
784 | shiftvec = fr->shift_vec[0]; |
785 | fshift = fr->fshift[0]; |
786 | facel = fr->epsfac; |
787 | charge = mdatoms->chargeA; |
788 | nvdwtype = fr->ntype; |
789 | vdwparam = fr->nbfp; |
790 | vdwtype = mdatoms->typeA; |
791 | vdwgridparam = fr->ljpme_c6grid; |
792 | ewclj = fr->ewaldcoeff_lj; |
793 | sh_lj_ewald = fr->ic->sh_lj_ewald; |
Value stored to 'sh_lj_ewald' is never read | |
794 | ewclj2 = ewclj*ewclj; |
795 | ewclj6 = ewclj2*ewclj2*ewclj2; |
796 | |
797 | sh_ewald = fr->ic->sh_ewald; |
798 | ewtab = fr->ic->tabq_coul_F; |
799 | ewtabscale = fr->ic->tabq_scale; |
800 | ewtabhalfspace = 0.5/ewtabscale; |
801 | |
802 | /* Setup water-specific parameters */ |
803 | inr = nlist->iinr[0]; |
804 | iq0 = facel*charge[inr+0]; |
805 | iq1 = facel*charge[inr+1]; |
806 | iq2 = facel*charge[inr+2]; |
807 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
808 | |
809 | jq0 = charge[inr+0]; |
810 | jq1 = charge[inr+1]; |
811 | jq2 = charge[inr+2]; |
812 | vdwjidx0 = 2*vdwtype[inr+0]; |
813 | qq00 = iq0*jq0; |
814 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
815 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
816 | c6grid_00 = vdwgridparam[vdwioffset0+vdwjidx0]; |
817 | qq01 = iq0*jq1; |
818 | qq02 = iq0*jq2; |
819 | qq10 = iq1*jq0; |
820 | qq11 = iq1*jq1; |
821 | qq12 = iq1*jq2; |
822 | qq20 = iq2*jq0; |
823 | qq21 = iq2*jq1; |
824 | qq22 = iq2*jq2; |
825 | |
826 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
827 | rcutoff = fr->rcoulomb; |
828 | rcutoff2 = rcutoff*rcutoff; |
829 | |
830 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
831 | rvdw = fr->rvdw; |
832 | |
833 | outeriter = 0; |
834 | inneriter = 0; |
835 | |
836 | /* Start outer loop over neighborlists */ |
837 | for(iidx=0; iidx<nri; iidx++) |
838 | { |
839 | /* Load shift vector for this list */ |
840 | i_shift_offset = DIM3*shiftidx[iidx]; |
841 | shX = shiftvec[i_shift_offset+XX0]; |
842 | shY = shiftvec[i_shift_offset+YY1]; |
843 | shZ = shiftvec[i_shift_offset+ZZ2]; |
844 | |
845 | /* Load limits for loop over neighbors */ |
846 | j_index_start = jindex[iidx]; |
847 | j_index_end = jindex[iidx+1]; |
848 | |
849 | /* Get outer coordinate index */ |
850 | inr = iinr[iidx]; |
851 | i_coord_offset = DIM3*inr; |
852 | |
853 | /* Load i particle coords and add shift vector */ |
854 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
855 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
856 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
857 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
858 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
859 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
860 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
861 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
862 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
863 | |
864 | fix0 = 0.0; |
865 | fiy0 = 0.0; |
866 | fiz0 = 0.0; |
867 | fix1 = 0.0; |
868 | fiy1 = 0.0; |
869 | fiz1 = 0.0; |
870 | fix2 = 0.0; |
871 | fiy2 = 0.0; |
872 | fiz2 = 0.0; |
873 | |
874 | /* Start inner kernel loop */ |
875 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
876 | { |
877 | /* Get j neighbor index, and coordinate index */ |
878 | jnr = jjnr[jidx]; |
879 | j_coord_offset = DIM3*jnr; |
880 | |
881 | /* load j atom coordinates */ |
882 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
883 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
884 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
885 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
886 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
887 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
888 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
889 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
890 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
891 | |
892 | /* Calculate displacement vector */ |
893 | dx00 = ix0 - jx0; |
894 | dy00 = iy0 - jy0; |
895 | dz00 = iz0 - jz0; |
896 | dx01 = ix0 - jx1; |
897 | dy01 = iy0 - jy1; |
898 | dz01 = iz0 - jz1; |
899 | dx02 = ix0 - jx2; |
900 | dy02 = iy0 - jy2; |
901 | dz02 = iz0 - jz2; |
902 | dx10 = ix1 - jx0; |
903 | dy10 = iy1 - jy0; |
904 | dz10 = iz1 - jz0; |
905 | dx11 = ix1 - jx1; |
906 | dy11 = iy1 - jy1; |
907 | dz11 = iz1 - jz1; |
908 | dx12 = ix1 - jx2; |
909 | dy12 = iy1 - jy2; |
910 | dz12 = iz1 - jz2; |
911 | dx20 = ix2 - jx0; |
912 | dy20 = iy2 - jy0; |
913 | dz20 = iz2 - jz0; |
914 | dx21 = ix2 - jx1; |
915 | dy21 = iy2 - jy1; |
916 | dz21 = iz2 - jz1; |
917 | dx22 = ix2 - jx2; |
918 | dy22 = iy2 - jy2; |
919 | dz22 = iz2 - jz2; |
920 | |
921 | /* Calculate squared distance and things based on it */ |
922 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
923 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
924 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
925 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
926 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
927 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
928 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
929 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
930 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
931 | |
932 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
933 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
934 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
935 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
936 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
937 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
938 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
939 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
940 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
941 | |
942 | rinvsq00 = rinv00*rinv00; |
943 | rinvsq01 = rinv01*rinv01; |
944 | rinvsq02 = rinv02*rinv02; |
945 | rinvsq10 = rinv10*rinv10; |
946 | rinvsq11 = rinv11*rinv11; |
947 | rinvsq12 = rinv12*rinv12; |
948 | rinvsq20 = rinv20*rinv20; |
949 | rinvsq21 = rinv21*rinv21; |
950 | rinvsq22 = rinv22*rinv22; |
951 | |
952 | /************************** |
953 | * CALCULATE INTERACTIONS * |
954 | **************************/ |
955 | |
956 | if (rsq00<rcutoff2) |
957 | { |
958 | |
959 | r00 = rsq00*rinv00; |
960 | |
961 | /* EWALD ELECTROSTATICS */ |
962 | |
963 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
964 | ewrt = r00*ewtabscale; |
965 | ewitab = ewrt; |
966 | eweps = ewrt-ewitab; |
967 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
968 | felec = qq00*rinv00*(rinvsq00-felec); |
969 | |
970 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
971 | ewcljrsq = ewclj2*rsq00; |
972 | exponent = exp(-ewcljrsq); |
973 | poly = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5); |
974 | fvdw = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00; |
975 | |
976 | fscal = felec+fvdw; |
977 | |
978 | /* Calculate temporary vectorial force */ |
979 | tx = fscal*dx00; |
980 | ty = fscal*dy00; |
981 | tz = fscal*dz00; |
982 | |
983 | /* Update vectorial force */ |
984 | fix0 += tx; |
985 | fiy0 += ty; |
986 | fiz0 += tz; |
987 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
988 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
989 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
990 | |
991 | } |
992 | |
993 | /************************** |
994 | * CALCULATE INTERACTIONS * |
995 | **************************/ |
996 | |
997 | if (rsq01<rcutoff2) |
998 | { |
999 | |
1000 | r01 = rsq01*rinv01; |
1001 | |
1002 | /* EWALD ELECTROSTATICS */ |
1003 | |
1004 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1005 | ewrt = r01*ewtabscale; |
1006 | ewitab = ewrt; |
1007 | eweps = ewrt-ewitab; |
1008 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1009 | felec = qq01*rinv01*(rinvsq01-felec); |
1010 | |
1011 | fscal = felec; |
1012 | |
1013 | /* Calculate temporary vectorial force */ |
1014 | tx = fscal*dx01; |
1015 | ty = fscal*dy01; |
1016 | tz = fscal*dz01; |
1017 | |
1018 | /* Update vectorial force */ |
1019 | fix0 += tx; |
1020 | fiy0 += ty; |
1021 | fiz0 += tz; |
1022 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1023 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1024 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1025 | |
1026 | } |
1027 | |
1028 | /************************** |
1029 | * CALCULATE INTERACTIONS * |
1030 | **************************/ |
1031 | |
1032 | if (rsq02<rcutoff2) |
1033 | { |
1034 | |
1035 | r02 = rsq02*rinv02; |
1036 | |
1037 | /* EWALD ELECTROSTATICS */ |
1038 | |
1039 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1040 | ewrt = r02*ewtabscale; |
1041 | ewitab = ewrt; |
1042 | eweps = ewrt-ewitab; |
1043 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1044 | felec = qq02*rinv02*(rinvsq02-felec); |
1045 | |
1046 | fscal = felec; |
1047 | |
1048 | /* Calculate temporary vectorial force */ |
1049 | tx = fscal*dx02; |
1050 | ty = fscal*dy02; |
1051 | tz = fscal*dz02; |
1052 | |
1053 | /* Update vectorial force */ |
1054 | fix0 += tx; |
1055 | fiy0 += ty; |
1056 | fiz0 += tz; |
1057 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1058 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1059 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1060 | |
1061 | } |
1062 | |
1063 | /************************** |
1064 | * CALCULATE INTERACTIONS * |
1065 | **************************/ |
1066 | |
1067 | if (rsq10<rcutoff2) |
1068 | { |
1069 | |
1070 | r10 = rsq10*rinv10; |
1071 | |
1072 | /* EWALD ELECTROSTATICS */ |
1073 | |
1074 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1075 | ewrt = r10*ewtabscale; |
1076 | ewitab = ewrt; |
1077 | eweps = ewrt-ewitab; |
1078 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1079 | felec = qq10*rinv10*(rinvsq10-felec); |
1080 | |
1081 | fscal = felec; |
1082 | |
1083 | /* Calculate temporary vectorial force */ |
1084 | tx = fscal*dx10; |
1085 | ty = fscal*dy10; |
1086 | tz = fscal*dz10; |
1087 | |
1088 | /* Update vectorial force */ |
1089 | fix1 += tx; |
1090 | fiy1 += ty; |
1091 | fiz1 += tz; |
1092 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1093 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1094 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1095 | |
1096 | } |
1097 | |
1098 | /************************** |
1099 | * CALCULATE INTERACTIONS * |
1100 | **************************/ |
1101 | |
1102 | if (rsq11<rcutoff2) |
1103 | { |
1104 | |
1105 | r11 = rsq11*rinv11; |
1106 | |
1107 | /* EWALD ELECTROSTATICS */ |
1108 | |
1109 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1110 | ewrt = r11*ewtabscale; |
1111 | ewitab = ewrt; |
1112 | eweps = ewrt-ewitab; |
1113 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1114 | felec = qq11*rinv11*(rinvsq11-felec); |
1115 | |
1116 | fscal = felec; |
1117 | |
1118 | /* Calculate temporary vectorial force */ |
1119 | tx = fscal*dx11; |
1120 | ty = fscal*dy11; |
1121 | tz = fscal*dz11; |
1122 | |
1123 | /* Update vectorial force */ |
1124 | fix1 += tx; |
1125 | fiy1 += ty; |
1126 | fiz1 += tz; |
1127 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1128 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1129 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1130 | |
1131 | } |
1132 | |
1133 | /************************** |
1134 | * CALCULATE INTERACTIONS * |
1135 | **************************/ |
1136 | |
1137 | if (rsq12<rcutoff2) |
1138 | { |
1139 | |
1140 | r12 = rsq12*rinv12; |
1141 | |
1142 | /* EWALD ELECTROSTATICS */ |
1143 | |
1144 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1145 | ewrt = r12*ewtabscale; |
1146 | ewitab = ewrt; |
1147 | eweps = ewrt-ewitab; |
1148 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1149 | felec = qq12*rinv12*(rinvsq12-felec); |
1150 | |
1151 | fscal = felec; |
1152 | |
1153 | /* Calculate temporary vectorial force */ |
1154 | tx = fscal*dx12; |
1155 | ty = fscal*dy12; |
1156 | tz = fscal*dz12; |
1157 | |
1158 | /* Update vectorial force */ |
1159 | fix1 += tx; |
1160 | fiy1 += ty; |
1161 | fiz1 += tz; |
1162 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1163 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1164 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1165 | |
1166 | } |
1167 | |
1168 | /************************** |
1169 | * CALCULATE INTERACTIONS * |
1170 | **************************/ |
1171 | |
1172 | if (rsq20<rcutoff2) |
1173 | { |
1174 | |
1175 | r20 = rsq20*rinv20; |
1176 | |
1177 | /* EWALD ELECTROSTATICS */ |
1178 | |
1179 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1180 | ewrt = r20*ewtabscale; |
1181 | ewitab = ewrt; |
1182 | eweps = ewrt-ewitab; |
1183 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1184 | felec = qq20*rinv20*(rinvsq20-felec); |
1185 | |
1186 | fscal = felec; |
1187 | |
1188 | /* Calculate temporary vectorial force */ |
1189 | tx = fscal*dx20; |
1190 | ty = fscal*dy20; |
1191 | tz = fscal*dz20; |
1192 | |
1193 | /* Update vectorial force */ |
1194 | fix2 += tx; |
1195 | fiy2 += ty; |
1196 | fiz2 += tz; |
1197 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1198 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1199 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1200 | |
1201 | } |
1202 | |
1203 | /************************** |
1204 | * CALCULATE INTERACTIONS * |
1205 | **************************/ |
1206 | |
1207 | if (rsq21<rcutoff2) |
1208 | { |
1209 | |
1210 | r21 = rsq21*rinv21; |
1211 | |
1212 | /* EWALD ELECTROSTATICS */ |
1213 | |
1214 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1215 | ewrt = r21*ewtabscale; |
1216 | ewitab = ewrt; |
1217 | eweps = ewrt-ewitab; |
1218 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1219 | felec = qq21*rinv21*(rinvsq21-felec); |
1220 | |
1221 | fscal = felec; |
1222 | |
1223 | /* Calculate temporary vectorial force */ |
1224 | tx = fscal*dx21; |
1225 | ty = fscal*dy21; |
1226 | tz = fscal*dz21; |
1227 | |
1228 | /* Update vectorial force */ |
1229 | fix2 += tx; |
1230 | fiy2 += ty; |
1231 | fiz2 += tz; |
1232 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1233 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1234 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1235 | |
1236 | } |
1237 | |
1238 | /************************** |
1239 | * CALCULATE INTERACTIONS * |
1240 | **************************/ |
1241 | |
1242 | if (rsq22<rcutoff2) |
1243 | { |
1244 | |
1245 | r22 = rsq22*rinv22; |
1246 | |
1247 | /* EWALD ELECTROSTATICS */ |
1248 | |
1249 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1250 | ewrt = r22*ewtabscale; |
1251 | ewitab = ewrt; |
1252 | eweps = ewrt-ewitab; |
1253 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1254 | felec = qq22*rinv22*(rinvsq22-felec); |
1255 | |
1256 | fscal = felec; |
1257 | |
1258 | /* Calculate temporary vectorial force */ |
1259 | tx = fscal*dx22; |
1260 | ty = fscal*dy22; |
1261 | tz = fscal*dz22; |
1262 | |
1263 | /* Update vectorial force */ |
1264 | fix2 += tx; |
1265 | fiy2 += ty; |
1266 | fiz2 += tz; |
1267 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1268 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1269 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1270 | |
1271 | } |
1272 | |
1273 | /* Inner loop uses 318 flops */ |
1274 | } |
1275 | /* End of innermost loop */ |
1276 | |
1277 | tx = ty = tz = 0; |
1278 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
1279 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
1280 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
1281 | tx += fix0; |
1282 | ty += fiy0; |
1283 | tz += fiz0; |
1284 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
1285 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
1286 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
1287 | tx += fix1; |
1288 | ty += fiy1; |
1289 | tz += fiz1; |
1290 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
1291 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
1292 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
1293 | tx += fix2; |
1294 | ty += fiy2; |
1295 | tz += fiz2; |
1296 | fshift[i_shift_offset+XX0] += tx; |
1297 | fshift[i_shift_offset+YY1] += ty; |
1298 | fshift[i_shift_offset+ZZ2] += tz; |
1299 | |
1300 | /* Increment number of inner iterations */ |
1301 | inneriter += j_index_end - j_index_start; |
1302 | |
1303 | /* Outer loop uses 30 flops */ |
1304 | } |
1305 | |
1306 | /* Increment number of outer iterations */ |
1307 | outeriter += nri; |
1308 | |
1309 | /* Update outer/inner flops */ |
1310 | |
1311 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*318)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*30 + inneriter *318; |
1312 | } |