File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwBhamSh_GeomW3W3_c.c |
Location: | line 797, column 5 |
Description: | Value stored to 'sh_vdw_invrcut6' 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_GeomW3W3_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: Buckingham |
53 | * Geometry: Water3-Water3 |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEwSh_VdwBhamSh_GeomW3W3_VF_c |
58 | (t_nblist * gmx_restrict__restrict nlist, |
59 | rvec * gmx_restrict__restrict xx, |
60 | rvec * gmx_restrict__restrict ff, |
61 | t_forcerec * gmx_restrict__restrict fr, |
62 | t_mdatoms * gmx_restrict__restrict mdatoms, |
63 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
64 | t_nrnb * gmx_restrict__restrict nrnb) |
65 | { |
66 | int i_shift_offset,i_coord_offset,j_coord_offset; |
67 | int j_index_start,j_index_end; |
68 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
69 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
70 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
71 | real *shiftvec,*fshift,*x,*f; |
72 | int vdwioffset0; |
73 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
74 | int vdwioffset1; |
75 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
76 | int vdwioffset2; |
77 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
78 | int vdwjidx0; |
79 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
80 | int vdwjidx1; |
81 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
82 | int vdwjidx2; |
83 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
84 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
85 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
86 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
87 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
88 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
89 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
90 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
91 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
92 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
93 | real velec,felec,velecsum,facel,crf,krf,krf2; |
94 | real *charge; |
95 | int nvdwtype; |
96 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
97 | int *vdwtype; |
98 | real *vdwparam; |
99 | int ewitab; |
100 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
101 | real *ewtab; |
102 | |
103 | x = xx[0]; |
104 | f = ff[0]; |
105 | |
106 | nri = nlist->nri; |
107 | iinr = nlist->iinr; |
108 | jindex = nlist->jindex; |
109 | jjnr = nlist->jjnr; |
110 | shiftidx = nlist->shift; |
111 | gid = nlist->gid; |
112 | shiftvec = fr->shift_vec[0]; |
113 | fshift = fr->fshift[0]; |
114 | facel = fr->epsfac; |
115 | charge = mdatoms->chargeA; |
116 | nvdwtype = fr->ntype; |
117 | vdwparam = fr->nbfp; |
118 | vdwtype = mdatoms->typeA; |
119 | |
120 | sh_ewald = fr->ic->sh_ewald; |
121 | ewtab = fr->ic->tabq_coul_FDV0; |
122 | ewtabscale = fr->ic->tabq_scale; |
123 | ewtabhalfspace = 0.5/ewtabscale; |
124 | |
125 | /* Setup water-specific parameters */ |
126 | inr = nlist->iinr[0]; |
127 | iq0 = facel*charge[inr+0]; |
128 | iq1 = facel*charge[inr+1]; |
129 | iq2 = facel*charge[inr+2]; |
130 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
131 | |
132 | jq0 = charge[inr+0]; |
133 | jq1 = charge[inr+1]; |
134 | jq2 = charge[inr+2]; |
135 | vdwjidx0 = 3*vdwtype[inr+0]; |
136 | qq00 = iq0*jq0; |
137 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
138 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
139 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
140 | qq01 = iq0*jq1; |
141 | qq02 = iq0*jq2; |
142 | qq10 = iq1*jq0; |
143 | qq11 = iq1*jq1; |
144 | qq12 = iq1*jq2; |
145 | qq20 = iq2*jq0; |
146 | qq21 = iq2*jq1; |
147 | qq22 = iq2*jq2; |
148 | |
149 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
150 | rcutoff = fr->rcoulomb; |
151 | rcutoff2 = rcutoff*rcutoff; |
152 | |
153 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
154 | rvdw = fr->rvdw; |
155 | |
156 | outeriter = 0; |
157 | inneriter = 0; |
158 | |
159 | /* Start outer loop over neighborlists */ |
160 | for(iidx=0; iidx<nri; iidx++) |
161 | { |
162 | /* Load shift vector for this list */ |
163 | i_shift_offset = DIM3*shiftidx[iidx]; |
164 | shX = shiftvec[i_shift_offset+XX0]; |
165 | shY = shiftvec[i_shift_offset+YY1]; |
166 | shZ = shiftvec[i_shift_offset+ZZ2]; |
167 | |
168 | /* Load limits for loop over neighbors */ |
169 | j_index_start = jindex[iidx]; |
170 | j_index_end = jindex[iidx+1]; |
171 | |
172 | /* Get outer coordinate index */ |
173 | inr = iinr[iidx]; |
174 | i_coord_offset = DIM3*inr; |
175 | |
176 | /* Load i particle coords and add shift vector */ |
177 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
178 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
179 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
180 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
181 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
182 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
183 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
184 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
185 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
186 | |
187 | fix0 = 0.0; |
188 | fiy0 = 0.0; |
189 | fiz0 = 0.0; |
190 | fix1 = 0.0; |
191 | fiy1 = 0.0; |
192 | fiz1 = 0.0; |
193 | fix2 = 0.0; |
194 | fiy2 = 0.0; |
195 | fiz2 = 0.0; |
196 | |
197 | /* Reset potential sums */ |
198 | velecsum = 0.0; |
199 | vvdwsum = 0.0; |
200 | |
201 | /* Start inner kernel loop */ |
202 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
203 | { |
204 | /* Get j neighbor index, and coordinate index */ |
205 | jnr = jjnr[jidx]; |
206 | j_coord_offset = DIM3*jnr; |
207 | |
208 | /* load j atom coordinates */ |
209 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
210 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
211 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
212 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
213 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
214 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
215 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
216 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
217 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
218 | |
219 | /* Calculate displacement vector */ |
220 | dx00 = ix0 - jx0; |
221 | dy00 = iy0 - jy0; |
222 | dz00 = iz0 - jz0; |
223 | dx01 = ix0 - jx1; |
224 | dy01 = iy0 - jy1; |
225 | dz01 = iz0 - jz1; |
226 | dx02 = ix0 - jx2; |
227 | dy02 = iy0 - jy2; |
228 | dz02 = iz0 - jz2; |
229 | dx10 = ix1 - jx0; |
230 | dy10 = iy1 - jy0; |
231 | dz10 = iz1 - jz0; |
232 | dx11 = ix1 - jx1; |
233 | dy11 = iy1 - jy1; |
234 | dz11 = iz1 - jz1; |
235 | dx12 = ix1 - jx2; |
236 | dy12 = iy1 - jy2; |
237 | dz12 = iz1 - jz2; |
238 | dx20 = ix2 - jx0; |
239 | dy20 = iy2 - jy0; |
240 | dz20 = iz2 - jz0; |
241 | dx21 = ix2 - jx1; |
242 | dy21 = iy2 - jy1; |
243 | dz21 = iz2 - jz1; |
244 | dx22 = ix2 - jx2; |
245 | dy22 = iy2 - jy2; |
246 | dz22 = iz2 - jz2; |
247 | |
248 | /* Calculate squared distance and things based on it */ |
249 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
250 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
251 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
252 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
253 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
254 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
255 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
256 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
257 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
258 | |
259 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
260 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
261 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
262 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
263 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
264 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
265 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
266 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
267 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
268 | |
269 | rinvsq00 = rinv00*rinv00; |
270 | rinvsq01 = rinv01*rinv01; |
271 | rinvsq02 = rinv02*rinv02; |
272 | rinvsq10 = rinv10*rinv10; |
273 | rinvsq11 = rinv11*rinv11; |
274 | rinvsq12 = rinv12*rinv12; |
275 | rinvsq20 = rinv20*rinv20; |
276 | rinvsq21 = rinv21*rinv21; |
277 | rinvsq22 = rinv22*rinv22; |
278 | |
279 | /************************** |
280 | * CALCULATE INTERACTIONS * |
281 | **************************/ |
282 | |
283 | if (rsq00<rcutoff2) |
284 | { |
285 | |
286 | r00 = rsq00*rinv00; |
287 | |
288 | /* EWALD ELECTROSTATICS */ |
289 | |
290 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
291 | ewrt = r00*ewtabscale; |
292 | ewitab = ewrt; |
293 | eweps = ewrt-ewitab; |
294 | ewitab = 4*ewitab; |
295 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
296 | velec = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
297 | felec = qq00*rinv00*(rinvsq00-felec); |
298 | |
299 | /* BUCKINGHAM DISPERSION/REPULSION */ |
300 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
301 | vvdw6 = c6_00*rinvsix; |
302 | br = cexp2_00*r00; |
303 | vvdwexp = cexp1_00*exp(-br); |
304 | vvdw = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0); |
305 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
306 | |
307 | /* Update potential sums from outer loop */ |
308 | velecsum += velec; |
309 | vvdwsum += vvdw; |
310 | |
311 | fscal = felec+fvdw; |
312 | |
313 | /* Calculate temporary vectorial force */ |
314 | tx = fscal*dx00; |
315 | ty = fscal*dy00; |
316 | tz = fscal*dz00; |
317 | |
318 | /* Update vectorial force */ |
319 | fix0 += tx; |
320 | fiy0 += ty; |
321 | fiz0 += tz; |
322 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
323 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
324 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
325 | |
326 | } |
327 | |
328 | /************************** |
329 | * CALCULATE INTERACTIONS * |
330 | **************************/ |
331 | |
332 | if (rsq01<rcutoff2) |
333 | { |
334 | |
335 | r01 = rsq01*rinv01; |
336 | |
337 | /* EWALD ELECTROSTATICS */ |
338 | |
339 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
340 | ewrt = r01*ewtabscale; |
341 | ewitab = ewrt; |
342 | eweps = ewrt-ewitab; |
343 | ewitab = 4*ewitab; |
344 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
345 | velec = qq01*((rinv01-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
346 | felec = qq01*rinv01*(rinvsq01-felec); |
347 | |
348 | /* Update potential sums from outer loop */ |
349 | velecsum += velec; |
350 | |
351 | fscal = felec; |
352 | |
353 | /* Calculate temporary vectorial force */ |
354 | tx = fscal*dx01; |
355 | ty = fscal*dy01; |
356 | tz = fscal*dz01; |
357 | |
358 | /* Update vectorial force */ |
359 | fix0 += tx; |
360 | fiy0 += ty; |
361 | fiz0 += tz; |
362 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
363 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
364 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
365 | |
366 | } |
367 | |
368 | /************************** |
369 | * CALCULATE INTERACTIONS * |
370 | **************************/ |
371 | |
372 | if (rsq02<rcutoff2) |
373 | { |
374 | |
375 | r02 = rsq02*rinv02; |
376 | |
377 | /* EWALD ELECTROSTATICS */ |
378 | |
379 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
380 | ewrt = r02*ewtabscale; |
381 | ewitab = ewrt; |
382 | eweps = ewrt-ewitab; |
383 | ewitab = 4*ewitab; |
384 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
385 | velec = qq02*((rinv02-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
386 | felec = qq02*rinv02*(rinvsq02-felec); |
387 | |
388 | /* Update potential sums from outer loop */ |
389 | velecsum += velec; |
390 | |
391 | fscal = felec; |
392 | |
393 | /* Calculate temporary vectorial force */ |
394 | tx = fscal*dx02; |
395 | ty = fscal*dy02; |
396 | tz = fscal*dz02; |
397 | |
398 | /* Update vectorial force */ |
399 | fix0 += tx; |
400 | fiy0 += ty; |
401 | fiz0 += tz; |
402 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
403 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
404 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
405 | |
406 | } |
407 | |
408 | /************************** |
409 | * CALCULATE INTERACTIONS * |
410 | **************************/ |
411 | |
412 | if (rsq10<rcutoff2) |
413 | { |
414 | |
415 | r10 = rsq10*rinv10; |
416 | |
417 | /* EWALD ELECTROSTATICS */ |
418 | |
419 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
420 | ewrt = r10*ewtabscale; |
421 | ewitab = ewrt; |
422 | eweps = ewrt-ewitab; |
423 | ewitab = 4*ewitab; |
424 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
425 | velec = qq10*((rinv10-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
426 | felec = qq10*rinv10*(rinvsq10-felec); |
427 | |
428 | /* Update potential sums from outer loop */ |
429 | velecsum += velec; |
430 | |
431 | fscal = felec; |
432 | |
433 | /* Calculate temporary vectorial force */ |
434 | tx = fscal*dx10; |
435 | ty = fscal*dy10; |
436 | tz = fscal*dz10; |
437 | |
438 | /* Update vectorial force */ |
439 | fix1 += tx; |
440 | fiy1 += ty; |
441 | fiz1 += tz; |
442 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
443 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
444 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
445 | |
446 | } |
447 | |
448 | /************************** |
449 | * CALCULATE INTERACTIONS * |
450 | **************************/ |
451 | |
452 | if (rsq11<rcutoff2) |
453 | { |
454 | |
455 | r11 = rsq11*rinv11; |
456 | |
457 | /* EWALD ELECTROSTATICS */ |
458 | |
459 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
460 | ewrt = r11*ewtabscale; |
461 | ewitab = ewrt; |
462 | eweps = ewrt-ewitab; |
463 | ewitab = 4*ewitab; |
464 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
465 | velec = qq11*((rinv11-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
466 | felec = qq11*rinv11*(rinvsq11-felec); |
467 | |
468 | /* Update potential sums from outer loop */ |
469 | velecsum += velec; |
470 | |
471 | fscal = felec; |
472 | |
473 | /* Calculate temporary vectorial force */ |
474 | tx = fscal*dx11; |
475 | ty = fscal*dy11; |
476 | tz = fscal*dz11; |
477 | |
478 | /* Update vectorial force */ |
479 | fix1 += tx; |
480 | fiy1 += ty; |
481 | fiz1 += tz; |
482 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
483 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
484 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
485 | |
486 | } |
487 | |
488 | /************************** |
489 | * CALCULATE INTERACTIONS * |
490 | **************************/ |
491 | |
492 | if (rsq12<rcutoff2) |
493 | { |
494 | |
495 | r12 = rsq12*rinv12; |
496 | |
497 | /* EWALD ELECTROSTATICS */ |
498 | |
499 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
500 | ewrt = r12*ewtabscale; |
501 | ewitab = ewrt; |
502 | eweps = ewrt-ewitab; |
503 | ewitab = 4*ewitab; |
504 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
505 | velec = qq12*((rinv12-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
506 | felec = qq12*rinv12*(rinvsq12-felec); |
507 | |
508 | /* Update potential sums from outer loop */ |
509 | velecsum += velec; |
510 | |
511 | fscal = felec; |
512 | |
513 | /* Calculate temporary vectorial force */ |
514 | tx = fscal*dx12; |
515 | ty = fscal*dy12; |
516 | tz = fscal*dz12; |
517 | |
518 | /* Update vectorial force */ |
519 | fix1 += tx; |
520 | fiy1 += ty; |
521 | fiz1 += tz; |
522 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
523 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
524 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
525 | |
526 | } |
527 | |
528 | /************************** |
529 | * CALCULATE INTERACTIONS * |
530 | **************************/ |
531 | |
532 | if (rsq20<rcutoff2) |
533 | { |
534 | |
535 | r20 = rsq20*rinv20; |
536 | |
537 | /* EWALD ELECTROSTATICS */ |
538 | |
539 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
540 | ewrt = r20*ewtabscale; |
541 | ewitab = ewrt; |
542 | eweps = ewrt-ewitab; |
543 | ewitab = 4*ewitab; |
544 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
545 | velec = qq20*((rinv20-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
546 | felec = qq20*rinv20*(rinvsq20-felec); |
547 | |
548 | /* Update potential sums from outer loop */ |
549 | velecsum += velec; |
550 | |
551 | fscal = felec; |
552 | |
553 | /* Calculate temporary vectorial force */ |
554 | tx = fscal*dx20; |
555 | ty = fscal*dy20; |
556 | tz = fscal*dz20; |
557 | |
558 | /* Update vectorial force */ |
559 | fix2 += tx; |
560 | fiy2 += ty; |
561 | fiz2 += tz; |
562 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
563 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
564 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
565 | |
566 | } |
567 | |
568 | /************************** |
569 | * CALCULATE INTERACTIONS * |
570 | **************************/ |
571 | |
572 | if (rsq21<rcutoff2) |
573 | { |
574 | |
575 | r21 = rsq21*rinv21; |
576 | |
577 | /* EWALD ELECTROSTATICS */ |
578 | |
579 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
580 | ewrt = r21*ewtabscale; |
581 | ewitab = ewrt; |
582 | eweps = ewrt-ewitab; |
583 | ewitab = 4*ewitab; |
584 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
585 | velec = qq21*((rinv21-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
586 | felec = qq21*rinv21*(rinvsq21-felec); |
587 | |
588 | /* Update potential sums from outer loop */ |
589 | velecsum += velec; |
590 | |
591 | fscal = felec; |
592 | |
593 | /* Calculate temporary vectorial force */ |
594 | tx = fscal*dx21; |
595 | ty = fscal*dy21; |
596 | tz = fscal*dz21; |
597 | |
598 | /* Update vectorial force */ |
599 | fix2 += tx; |
600 | fiy2 += ty; |
601 | fiz2 += tz; |
602 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
603 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
604 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
605 | |
606 | } |
607 | |
608 | /************************** |
609 | * CALCULATE INTERACTIONS * |
610 | **************************/ |
611 | |
612 | if (rsq22<rcutoff2) |
613 | { |
614 | |
615 | r22 = rsq22*rinv22; |
616 | |
617 | /* EWALD ELECTROSTATICS */ |
618 | |
619 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
620 | ewrt = r22*ewtabscale; |
621 | ewitab = ewrt; |
622 | eweps = ewrt-ewitab; |
623 | ewitab = 4*ewitab; |
624 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
625 | velec = qq22*((rinv22-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
626 | felec = qq22*rinv22*(rinvsq22-felec); |
627 | |
628 | /* Update potential sums from outer loop */ |
629 | velecsum += velec; |
630 | |
631 | fscal = felec; |
632 | |
633 | /* Calculate temporary vectorial force */ |
634 | tx = fscal*dx22; |
635 | ty = fscal*dy22; |
636 | tz = fscal*dz22; |
637 | |
638 | /* Update vectorial force */ |
639 | fix2 += tx; |
640 | fiy2 += ty; |
641 | fiz2 += tz; |
642 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
643 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
644 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
645 | |
646 | } |
647 | |
648 | /* Inner loop uses 438 flops */ |
649 | } |
650 | /* End of innermost loop */ |
651 | |
652 | tx = ty = tz = 0; |
653 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
654 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
655 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
656 | tx += fix0; |
657 | ty += fiy0; |
658 | tz += fiz0; |
659 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
660 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
661 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
662 | tx += fix1; |
663 | ty += fiy1; |
664 | tz += fiz1; |
665 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
666 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
667 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
668 | tx += fix2; |
669 | ty += fiy2; |
670 | tz += fiz2; |
671 | fshift[i_shift_offset+XX0] += tx; |
672 | fshift[i_shift_offset+YY1] += ty; |
673 | fshift[i_shift_offset+ZZ2] += tz; |
674 | |
675 | ggid = gid[iidx]; |
676 | /* Update potential energies */ |
677 | kernel_data->energygrp_elec[ggid] += velecsum; |
678 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
679 | |
680 | /* Increment number of inner iterations */ |
681 | inneriter += j_index_end - j_index_start; |
682 | |
683 | /* Outer loop uses 32 flops */ |
684 | } |
685 | |
686 | /* Increment number of outer iterations */ |
687 | outeriter += nri; |
688 | |
689 | /* Update outer/inner flops */ |
690 | |
691 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*438)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*32 + inneriter*438; |
692 | } |
693 | /* |
694 | * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwBhamSh_GeomW3W3_F_c |
695 | * Electrostatics interaction: Ewald |
696 | * VdW interaction: Buckingham |
697 | * Geometry: Water3-Water3 |
698 | * Calculate force/pot: Force |
699 | */ |
700 | void |
701 | nb_kernel_ElecEwSh_VdwBhamSh_GeomW3W3_F_c |
702 | (t_nblist * gmx_restrict__restrict nlist, |
703 | rvec * gmx_restrict__restrict xx, |
704 | rvec * gmx_restrict__restrict ff, |
705 | t_forcerec * gmx_restrict__restrict fr, |
706 | t_mdatoms * gmx_restrict__restrict mdatoms, |
707 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
708 | t_nrnb * gmx_restrict__restrict nrnb) |
709 | { |
710 | int i_shift_offset,i_coord_offset,j_coord_offset; |
711 | int j_index_start,j_index_end; |
712 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
713 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
714 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
715 | real *shiftvec,*fshift,*x,*f; |
716 | int vdwioffset0; |
717 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
718 | int vdwioffset1; |
719 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
720 | int vdwioffset2; |
721 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
722 | int vdwjidx0; |
723 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
724 | int vdwjidx1; |
725 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
726 | int vdwjidx2; |
727 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
728 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
729 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
730 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
731 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
732 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
733 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
734 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
735 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
736 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
737 | real velec,felec,velecsum,facel,crf,krf,krf2; |
738 | real *charge; |
739 | int nvdwtype; |
740 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
741 | int *vdwtype; |
742 | real *vdwparam; |
743 | int ewitab; |
744 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
745 | real *ewtab; |
746 | |
747 | x = xx[0]; |
748 | f = ff[0]; |
749 | |
750 | nri = nlist->nri; |
751 | iinr = nlist->iinr; |
752 | jindex = nlist->jindex; |
753 | jjnr = nlist->jjnr; |
754 | shiftidx = nlist->shift; |
755 | gid = nlist->gid; |
756 | shiftvec = fr->shift_vec[0]; |
757 | fshift = fr->fshift[0]; |
758 | facel = fr->epsfac; |
759 | charge = mdatoms->chargeA; |
760 | nvdwtype = fr->ntype; |
761 | vdwparam = fr->nbfp; |
762 | vdwtype = mdatoms->typeA; |
763 | |
764 | sh_ewald = fr->ic->sh_ewald; |
765 | ewtab = fr->ic->tabq_coul_F; |
766 | ewtabscale = fr->ic->tabq_scale; |
767 | ewtabhalfspace = 0.5/ewtabscale; |
768 | |
769 | /* Setup water-specific parameters */ |
770 | inr = nlist->iinr[0]; |
771 | iq0 = facel*charge[inr+0]; |
772 | iq1 = facel*charge[inr+1]; |
773 | iq2 = facel*charge[inr+2]; |
774 | vdwioffset0 = 3*nvdwtype*vdwtype[inr+0]; |
775 | |
776 | jq0 = charge[inr+0]; |
777 | jq1 = charge[inr+1]; |
778 | jq2 = charge[inr+2]; |
779 | vdwjidx0 = 3*vdwtype[inr+0]; |
780 | qq00 = iq0*jq0; |
781 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
782 | cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
783 | cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2]; |
784 | qq01 = iq0*jq1; |
785 | qq02 = iq0*jq2; |
786 | qq10 = iq1*jq0; |
787 | qq11 = iq1*jq1; |
788 | qq12 = iq1*jq2; |
789 | qq20 = iq2*jq0; |
790 | qq21 = iq2*jq1; |
791 | qq22 = iq2*jq2; |
792 | |
793 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
794 | rcutoff = fr->rcoulomb; |
795 | rcutoff2 = rcutoff*rcutoff; |
796 | |
797 | sh_vdw_invrcut6 = fr->ic->sh_invrc6; |
Value stored to 'sh_vdw_invrcut6' is never read | |
798 | rvdw = fr->rvdw; |
799 | |
800 | outeriter = 0; |
801 | inneriter = 0; |
802 | |
803 | /* Start outer loop over neighborlists */ |
804 | for(iidx=0; iidx<nri; iidx++) |
805 | { |
806 | /* Load shift vector for this list */ |
807 | i_shift_offset = DIM3*shiftidx[iidx]; |
808 | shX = shiftvec[i_shift_offset+XX0]; |
809 | shY = shiftvec[i_shift_offset+YY1]; |
810 | shZ = shiftvec[i_shift_offset+ZZ2]; |
811 | |
812 | /* Load limits for loop over neighbors */ |
813 | j_index_start = jindex[iidx]; |
814 | j_index_end = jindex[iidx+1]; |
815 | |
816 | /* Get outer coordinate index */ |
817 | inr = iinr[iidx]; |
818 | i_coord_offset = DIM3*inr; |
819 | |
820 | /* Load i particle coords and add shift vector */ |
821 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
822 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
823 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
824 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
825 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
826 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
827 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
828 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
829 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
830 | |
831 | fix0 = 0.0; |
832 | fiy0 = 0.0; |
833 | fiz0 = 0.0; |
834 | fix1 = 0.0; |
835 | fiy1 = 0.0; |
836 | fiz1 = 0.0; |
837 | fix2 = 0.0; |
838 | fiy2 = 0.0; |
839 | fiz2 = 0.0; |
840 | |
841 | /* Start inner kernel loop */ |
842 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
843 | { |
844 | /* Get j neighbor index, and coordinate index */ |
845 | jnr = jjnr[jidx]; |
846 | j_coord_offset = DIM3*jnr; |
847 | |
848 | /* load j atom coordinates */ |
849 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
850 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
851 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
852 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
853 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
854 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
855 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
856 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
857 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
858 | |
859 | /* Calculate displacement vector */ |
860 | dx00 = ix0 - jx0; |
861 | dy00 = iy0 - jy0; |
862 | dz00 = iz0 - jz0; |
863 | dx01 = ix0 - jx1; |
864 | dy01 = iy0 - jy1; |
865 | dz01 = iz0 - jz1; |
866 | dx02 = ix0 - jx2; |
867 | dy02 = iy0 - jy2; |
868 | dz02 = iz0 - jz2; |
869 | dx10 = ix1 - jx0; |
870 | dy10 = iy1 - jy0; |
871 | dz10 = iz1 - jz0; |
872 | dx11 = ix1 - jx1; |
873 | dy11 = iy1 - jy1; |
874 | dz11 = iz1 - jz1; |
875 | dx12 = ix1 - jx2; |
876 | dy12 = iy1 - jy2; |
877 | dz12 = iz1 - jz2; |
878 | dx20 = ix2 - jx0; |
879 | dy20 = iy2 - jy0; |
880 | dz20 = iz2 - jz0; |
881 | dx21 = ix2 - jx1; |
882 | dy21 = iy2 - jy1; |
883 | dz21 = iz2 - jz1; |
884 | dx22 = ix2 - jx2; |
885 | dy22 = iy2 - jy2; |
886 | dz22 = iz2 - jz2; |
887 | |
888 | /* Calculate squared distance and things based on it */ |
889 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
890 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
891 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
892 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
893 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
894 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
895 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
896 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
897 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
898 | |
899 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
900 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
901 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
902 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
903 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
904 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
905 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
906 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
907 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
908 | |
909 | rinvsq00 = rinv00*rinv00; |
910 | rinvsq01 = rinv01*rinv01; |
911 | rinvsq02 = rinv02*rinv02; |
912 | rinvsq10 = rinv10*rinv10; |
913 | rinvsq11 = rinv11*rinv11; |
914 | rinvsq12 = rinv12*rinv12; |
915 | rinvsq20 = rinv20*rinv20; |
916 | rinvsq21 = rinv21*rinv21; |
917 | rinvsq22 = rinv22*rinv22; |
918 | |
919 | /************************** |
920 | * CALCULATE INTERACTIONS * |
921 | **************************/ |
922 | |
923 | if (rsq00<rcutoff2) |
924 | { |
925 | |
926 | r00 = rsq00*rinv00; |
927 | |
928 | /* EWALD ELECTROSTATICS */ |
929 | |
930 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
931 | ewrt = r00*ewtabscale; |
932 | ewitab = ewrt; |
933 | eweps = ewrt-ewitab; |
934 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
935 | felec = qq00*rinv00*(rinvsq00-felec); |
936 | |
937 | /* BUCKINGHAM DISPERSION/REPULSION */ |
938 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
939 | vvdw6 = c6_00*rinvsix; |
940 | br = cexp2_00*r00; |
941 | vvdwexp = cexp1_00*exp(-br); |
942 | fvdw = (br*vvdwexp-vvdw6)*rinvsq00; |
943 | |
944 | fscal = felec+fvdw; |
945 | |
946 | /* Calculate temporary vectorial force */ |
947 | tx = fscal*dx00; |
948 | ty = fscal*dy00; |
949 | tz = fscal*dz00; |
950 | |
951 | /* Update vectorial force */ |
952 | fix0 += tx; |
953 | fiy0 += ty; |
954 | fiz0 += tz; |
955 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
956 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
957 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
958 | |
959 | } |
960 | |
961 | /************************** |
962 | * CALCULATE INTERACTIONS * |
963 | **************************/ |
964 | |
965 | if (rsq01<rcutoff2) |
966 | { |
967 | |
968 | r01 = rsq01*rinv01; |
969 | |
970 | /* EWALD ELECTROSTATICS */ |
971 | |
972 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
973 | ewrt = r01*ewtabscale; |
974 | ewitab = ewrt; |
975 | eweps = ewrt-ewitab; |
976 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
977 | felec = qq01*rinv01*(rinvsq01-felec); |
978 | |
979 | fscal = felec; |
980 | |
981 | /* Calculate temporary vectorial force */ |
982 | tx = fscal*dx01; |
983 | ty = fscal*dy01; |
984 | tz = fscal*dz01; |
985 | |
986 | /* Update vectorial force */ |
987 | fix0 += tx; |
988 | fiy0 += ty; |
989 | fiz0 += tz; |
990 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
991 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
992 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
993 | |
994 | } |
995 | |
996 | /************************** |
997 | * CALCULATE INTERACTIONS * |
998 | **************************/ |
999 | |
1000 | if (rsq02<rcutoff2) |
1001 | { |
1002 | |
1003 | r02 = rsq02*rinv02; |
1004 | |
1005 | /* EWALD ELECTROSTATICS */ |
1006 | |
1007 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1008 | ewrt = r02*ewtabscale; |
1009 | ewitab = ewrt; |
1010 | eweps = ewrt-ewitab; |
1011 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1012 | felec = qq02*rinv02*(rinvsq02-felec); |
1013 | |
1014 | fscal = felec; |
1015 | |
1016 | /* Calculate temporary vectorial force */ |
1017 | tx = fscal*dx02; |
1018 | ty = fscal*dy02; |
1019 | tz = fscal*dz02; |
1020 | |
1021 | /* Update vectorial force */ |
1022 | fix0 += tx; |
1023 | fiy0 += ty; |
1024 | fiz0 += tz; |
1025 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1026 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1027 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1028 | |
1029 | } |
1030 | |
1031 | /************************** |
1032 | * CALCULATE INTERACTIONS * |
1033 | **************************/ |
1034 | |
1035 | if (rsq10<rcutoff2) |
1036 | { |
1037 | |
1038 | r10 = rsq10*rinv10; |
1039 | |
1040 | /* EWALD ELECTROSTATICS */ |
1041 | |
1042 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1043 | ewrt = r10*ewtabscale; |
1044 | ewitab = ewrt; |
1045 | eweps = ewrt-ewitab; |
1046 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1047 | felec = qq10*rinv10*(rinvsq10-felec); |
1048 | |
1049 | fscal = felec; |
1050 | |
1051 | /* Calculate temporary vectorial force */ |
1052 | tx = fscal*dx10; |
1053 | ty = fscal*dy10; |
1054 | tz = fscal*dz10; |
1055 | |
1056 | /* Update vectorial force */ |
1057 | fix1 += tx; |
1058 | fiy1 += ty; |
1059 | fiz1 += tz; |
1060 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1061 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1062 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1063 | |
1064 | } |
1065 | |
1066 | /************************** |
1067 | * CALCULATE INTERACTIONS * |
1068 | **************************/ |
1069 | |
1070 | if (rsq11<rcutoff2) |
1071 | { |
1072 | |
1073 | r11 = rsq11*rinv11; |
1074 | |
1075 | /* EWALD ELECTROSTATICS */ |
1076 | |
1077 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1078 | ewrt = r11*ewtabscale; |
1079 | ewitab = ewrt; |
1080 | eweps = ewrt-ewitab; |
1081 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1082 | felec = qq11*rinv11*(rinvsq11-felec); |
1083 | |
1084 | fscal = felec; |
1085 | |
1086 | /* Calculate temporary vectorial force */ |
1087 | tx = fscal*dx11; |
1088 | ty = fscal*dy11; |
1089 | tz = fscal*dz11; |
1090 | |
1091 | /* Update vectorial force */ |
1092 | fix1 += tx; |
1093 | fiy1 += ty; |
1094 | fiz1 += tz; |
1095 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1096 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1097 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1098 | |
1099 | } |
1100 | |
1101 | /************************** |
1102 | * CALCULATE INTERACTIONS * |
1103 | **************************/ |
1104 | |
1105 | if (rsq12<rcutoff2) |
1106 | { |
1107 | |
1108 | r12 = rsq12*rinv12; |
1109 | |
1110 | /* EWALD ELECTROSTATICS */ |
1111 | |
1112 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1113 | ewrt = r12*ewtabscale; |
1114 | ewitab = ewrt; |
1115 | eweps = ewrt-ewitab; |
1116 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1117 | felec = qq12*rinv12*(rinvsq12-felec); |
1118 | |
1119 | fscal = felec; |
1120 | |
1121 | /* Calculate temporary vectorial force */ |
1122 | tx = fscal*dx12; |
1123 | ty = fscal*dy12; |
1124 | tz = fscal*dz12; |
1125 | |
1126 | /* Update vectorial force */ |
1127 | fix1 += tx; |
1128 | fiy1 += ty; |
1129 | fiz1 += tz; |
1130 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1131 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1132 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1133 | |
1134 | } |
1135 | |
1136 | /************************** |
1137 | * CALCULATE INTERACTIONS * |
1138 | **************************/ |
1139 | |
1140 | if (rsq20<rcutoff2) |
1141 | { |
1142 | |
1143 | r20 = rsq20*rinv20; |
1144 | |
1145 | /* EWALD ELECTROSTATICS */ |
1146 | |
1147 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1148 | ewrt = r20*ewtabscale; |
1149 | ewitab = ewrt; |
1150 | eweps = ewrt-ewitab; |
1151 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1152 | felec = qq20*rinv20*(rinvsq20-felec); |
1153 | |
1154 | fscal = felec; |
1155 | |
1156 | /* Calculate temporary vectorial force */ |
1157 | tx = fscal*dx20; |
1158 | ty = fscal*dy20; |
1159 | tz = fscal*dz20; |
1160 | |
1161 | /* Update vectorial force */ |
1162 | fix2 += tx; |
1163 | fiy2 += ty; |
1164 | fiz2 += tz; |
1165 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1166 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1167 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1168 | |
1169 | } |
1170 | |
1171 | /************************** |
1172 | * CALCULATE INTERACTIONS * |
1173 | **************************/ |
1174 | |
1175 | if (rsq21<rcutoff2) |
1176 | { |
1177 | |
1178 | r21 = rsq21*rinv21; |
1179 | |
1180 | /* EWALD ELECTROSTATICS */ |
1181 | |
1182 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1183 | ewrt = r21*ewtabscale; |
1184 | ewitab = ewrt; |
1185 | eweps = ewrt-ewitab; |
1186 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1187 | felec = qq21*rinv21*(rinvsq21-felec); |
1188 | |
1189 | fscal = felec; |
1190 | |
1191 | /* Calculate temporary vectorial force */ |
1192 | tx = fscal*dx21; |
1193 | ty = fscal*dy21; |
1194 | tz = fscal*dz21; |
1195 | |
1196 | /* Update vectorial force */ |
1197 | fix2 += tx; |
1198 | fiy2 += ty; |
1199 | fiz2 += tz; |
1200 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1201 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1202 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1203 | |
1204 | } |
1205 | |
1206 | /************************** |
1207 | * CALCULATE INTERACTIONS * |
1208 | **************************/ |
1209 | |
1210 | if (rsq22<rcutoff2) |
1211 | { |
1212 | |
1213 | r22 = rsq22*rinv22; |
1214 | |
1215 | /* EWALD ELECTROSTATICS */ |
1216 | |
1217 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1218 | ewrt = r22*ewtabscale; |
1219 | ewitab = ewrt; |
1220 | eweps = ewrt-ewitab; |
1221 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1222 | felec = qq22*rinv22*(rinvsq22-felec); |
1223 | |
1224 | fscal = felec; |
1225 | |
1226 | /* Calculate temporary vectorial force */ |
1227 | tx = fscal*dx22; |
1228 | ty = fscal*dy22; |
1229 | tz = fscal*dz22; |
1230 | |
1231 | /* Update vectorial force */ |
1232 | fix2 += tx; |
1233 | fiy2 += ty; |
1234 | fiz2 += tz; |
1235 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1236 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1237 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1238 | |
1239 | } |
1240 | |
1241 | /* Inner loop uses 332 flops */ |
1242 | } |
1243 | /* End of innermost loop */ |
1244 | |
1245 | tx = ty = tz = 0; |
1246 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
1247 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
1248 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
1249 | tx += fix0; |
1250 | ty += fiy0; |
1251 | tz += fiz0; |
1252 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
1253 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
1254 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
1255 | tx += fix1; |
1256 | ty += fiy1; |
1257 | tz += fiz1; |
1258 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
1259 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
1260 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
1261 | tx += fix2; |
1262 | ty += fiy2; |
1263 | tz += fiz2; |
1264 | fshift[i_shift_offset+XX0] += tx; |
1265 | fshift[i_shift_offset+YY1] += ty; |
1266 | fshift[i_shift_offset+ZZ2] += tz; |
1267 | |
1268 | /* Increment number of inner iterations */ |
1269 | inneriter += j_index_end - j_index_start; |
1270 | |
1271 | /* Outer loop uses 30 flops */ |
1272 | } |
1273 | |
1274 | /* Increment number of outer iterations */ |
1275 | outeriter += nri; |
1276 | |
1277 | /* Update outer/inner flops */ |
1278 | |
1279 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*332)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*30 + inneriter *332; |
1280 | } |