File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_GeomW3W3_c.c |
Location: | line 746, 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_VdwCSTab_GeomW3W3_VF_c |
51 | * Electrostatics interaction: Ewald |
52 | * VdW interaction: CubicSplineTable |
53 | * Geometry: Water3-Water3 |
54 | * Calculate force/pot: PotentialAndForce |
55 | */ |
56 | void |
57 | nb_kernel_ElecEw_VdwCSTab_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 vfitab; |
100 | real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF; |
101 | real *vftab; |
102 | int ewitab; |
103 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
104 | real *ewtab; |
105 | |
106 | x = xx[0]; |
107 | f = ff[0]; |
108 | |
109 | nri = nlist->nri; |
110 | iinr = nlist->iinr; |
111 | jindex = nlist->jindex; |
112 | jjnr = nlist->jjnr; |
113 | shiftidx = nlist->shift; |
114 | gid = nlist->gid; |
115 | shiftvec = fr->shift_vec[0]; |
116 | fshift = fr->fshift[0]; |
117 | facel = fr->epsfac; |
118 | charge = mdatoms->chargeA; |
119 | nvdwtype = fr->ntype; |
120 | vdwparam = fr->nbfp; |
121 | vdwtype = mdatoms->typeA; |
122 | |
123 | vftab = kernel_data->table_vdw->data; |
124 | vftabscale = kernel_data->table_vdw->scale; |
125 | |
126 | sh_ewald = fr->ic->sh_ewald; |
127 | ewtab = fr->ic->tabq_coul_FDV0; |
128 | ewtabscale = fr->ic->tabq_scale; |
129 | ewtabhalfspace = 0.5/ewtabscale; |
130 | |
131 | /* Setup water-specific parameters */ |
132 | inr = nlist->iinr[0]; |
133 | iq0 = facel*charge[inr+0]; |
134 | iq1 = facel*charge[inr+1]; |
135 | iq2 = facel*charge[inr+2]; |
136 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
137 | |
138 | jq0 = charge[inr+0]; |
139 | jq1 = charge[inr+1]; |
140 | jq2 = charge[inr+2]; |
141 | vdwjidx0 = 2*vdwtype[inr+0]; |
142 | qq00 = iq0*jq0; |
143 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
144 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
145 | qq01 = iq0*jq1; |
146 | qq02 = iq0*jq2; |
147 | qq10 = iq1*jq0; |
148 | qq11 = iq1*jq1; |
149 | qq12 = iq1*jq2; |
150 | qq20 = iq2*jq0; |
151 | qq21 = iq2*jq1; |
152 | qq22 = iq2*jq2; |
153 | |
154 | outeriter = 0; |
155 | inneriter = 0; |
156 | |
157 | /* Start outer loop over neighborlists */ |
158 | for(iidx=0; iidx<nri; iidx++) |
159 | { |
160 | /* Load shift vector for this list */ |
161 | i_shift_offset = DIM3*shiftidx[iidx]; |
162 | shX = shiftvec[i_shift_offset+XX0]; |
163 | shY = shiftvec[i_shift_offset+YY1]; |
164 | shZ = shiftvec[i_shift_offset+ZZ2]; |
165 | |
166 | /* Load limits for loop over neighbors */ |
167 | j_index_start = jindex[iidx]; |
168 | j_index_end = jindex[iidx+1]; |
169 | |
170 | /* Get outer coordinate index */ |
171 | inr = iinr[iidx]; |
172 | i_coord_offset = DIM3*inr; |
173 | |
174 | /* Load i particle coords and add shift vector */ |
175 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
176 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
177 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
178 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
179 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
180 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
181 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
182 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
183 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
184 | |
185 | fix0 = 0.0; |
186 | fiy0 = 0.0; |
187 | fiz0 = 0.0; |
188 | fix1 = 0.0; |
189 | fiy1 = 0.0; |
190 | fiz1 = 0.0; |
191 | fix2 = 0.0; |
192 | fiy2 = 0.0; |
193 | fiz2 = 0.0; |
194 | |
195 | /* Reset potential sums */ |
196 | velecsum = 0.0; |
197 | vvdwsum = 0.0; |
198 | |
199 | /* Start inner kernel loop */ |
200 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
201 | { |
202 | /* Get j neighbor index, and coordinate index */ |
203 | jnr = jjnr[jidx]; |
204 | j_coord_offset = DIM3*jnr; |
205 | |
206 | /* load j atom coordinates */ |
207 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
208 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
209 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
210 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
211 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
212 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
213 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
214 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
215 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
216 | |
217 | /* Calculate displacement vector */ |
218 | dx00 = ix0 - jx0; |
219 | dy00 = iy0 - jy0; |
220 | dz00 = iz0 - jz0; |
221 | dx01 = ix0 - jx1; |
222 | dy01 = iy0 - jy1; |
223 | dz01 = iz0 - jz1; |
224 | dx02 = ix0 - jx2; |
225 | dy02 = iy0 - jy2; |
226 | dz02 = iz0 - jz2; |
227 | dx10 = ix1 - jx0; |
228 | dy10 = iy1 - jy0; |
229 | dz10 = iz1 - jz0; |
230 | dx11 = ix1 - jx1; |
231 | dy11 = iy1 - jy1; |
232 | dz11 = iz1 - jz1; |
233 | dx12 = ix1 - jx2; |
234 | dy12 = iy1 - jy2; |
235 | dz12 = iz1 - jz2; |
236 | dx20 = ix2 - jx0; |
237 | dy20 = iy2 - jy0; |
238 | dz20 = iz2 - jz0; |
239 | dx21 = ix2 - jx1; |
240 | dy21 = iy2 - jy1; |
241 | dz21 = iz2 - jz1; |
242 | dx22 = ix2 - jx2; |
243 | dy22 = iy2 - jy2; |
244 | dz22 = iz2 - jz2; |
245 | |
246 | /* Calculate squared distance and things based on it */ |
247 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
248 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
249 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
250 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
251 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
252 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
253 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
254 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
255 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
256 | |
257 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
258 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
259 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
260 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
261 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
262 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
263 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
264 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
265 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
266 | |
267 | rinvsq00 = rinv00*rinv00; |
268 | rinvsq01 = rinv01*rinv01; |
269 | rinvsq02 = rinv02*rinv02; |
270 | rinvsq10 = rinv10*rinv10; |
271 | rinvsq11 = rinv11*rinv11; |
272 | rinvsq12 = rinv12*rinv12; |
273 | rinvsq20 = rinv20*rinv20; |
274 | rinvsq21 = rinv21*rinv21; |
275 | rinvsq22 = rinv22*rinv22; |
276 | |
277 | /************************** |
278 | * CALCULATE INTERACTIONS * |
279 | **************************/ |
280 | |
281 | r00 = rsq00*rinv00; |
282 | |
283 | /* Calculate table index by multiplying r with table scale and truncate to integer */ |
284 | rt = r00*vftabscale; |
285 | vfitab = rt; |
286 | vfeps = rt-vfitab; |
287 | vfitab = 2*4*vfitab; |
288 | |
289 | /* EWALD ELECTROSTATICS */ |
290 | |
291 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
292 | ewrt = r00*ewtabscale; |
293 | ewitab = ewrt; |
294 | eweps = ewrt-ewitab; |
295 | ewitab = 4*ewitab; |
296 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
297 | velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
298 | felec = qq00*rinv00*(rinvsq00-felec); |
299 | |
300 | /* CUBIC SPLINE TABLE DISPERSION */ |
301 | vfitab += 0; |
302 | Y = vftab[vfitab]; |
303 | F = vftab[vfitab+1]; |
304 | Geps = vfeps*vftab[vfitab+2]; |
305 | Heps2 = vfeps*vfeps*vftab[vfitab+3]; |
306 | Fp = F+Geps+Heps2; |
307 | VV = Y+vfeps*Fp; |
308 | vvdw6 = c6_00*VV; |
309 | FF = Fp+Geps+2.0*Heps2; |
310 | fvdw6 = c6_00*FF; |
311 | |
312 | /* CUBIC SPLINE TABLE REPULSION */ |
313 | Y = vftab[vfitab+4]; |
314 | F = vftab[vfitab+5]; |
315 | Geps = vfeps*vftab[vfitab+6]; |
316 | Heps2 = vfeps*vfeps*vftab[vfitab+7]; |
317 | Fp = F+Geps+Heps2; |
318 | VV = Y+vfeps*Fp; |
319 | vvdw12 = c12_00*VV; |
320 | FF = Fp+Geps+2.0*Heps2; |
321 | fvdw12 = c12_00*FF; |
322 | vvdw = vvdw12+vvdw6; |
323 | fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00; |
324 | |
325 | /* Update potential sums from outer loop */ |
326 | velecsum += velec; |
327 | vvdwsum += vvdw; |
328 | |
329 | fscal = felec+fvdw; |
330 | |
331 | /* Calculate temporary vectorial force */ |
332 | tx = fscal*dx00; |
333 | ty = fscal*dy00; |
334 | tz = fscal*dz00; |
335 | |
336 | /* Update vectorial force */ |
337 | fix0 += tx; |
338 | fiy0 += ty; |
339 | fiz0 += tz; |
340 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
341 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
342 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
343 | |
344 | /************************** |
345 | * CALCULATE INTERACTIONS * |
346 | **************************/ |
347 | |
348 | r01 = rsq01*rinv01; |
349 | |
350 | /* EWALD ELECTROSTATICS */ |
351 | |
352 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
353 | ewrt = r01*ewtabscale; |
354 | ewitab = ewrt; |
355 | eweps = ewrt-ewitab; |
356 | ewitab = 4*ewitab; |
357 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
358 | velec = qq01*(rinv01-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
359 | felec = qq01*rinv01*(rinvsq01-felec); |
360 | |
361 | /* Update potential sums from outer loop */ |
362 | velecsum += velec; |
363 | |
364 | fscal = felec; |
365 | |
366 | /* Calculate temporary vectorial force */ |
367 | tx = fscal*dx01; |
368 | ty = fscal*dy01; |
369 | tz = fscal*dz01; |
370 | |
371 | /* Update vectorial force */ |
372 | fix0 += tx; |
373 | fiy0 += ty; |
374 | fiz0 += tz; |
375 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
376 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
377 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
378 | |
379 | /************************** |
380 | * CALCULATE INTERACTIONS * |
381 | **************************/ |
382 | |
383 | r02 = rsq02*rinv02; |
384 | |
385 | /* EWALD ELECTROSTATICS */ |
386 | |
387 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
388 | ewrt = r02*ewtabscale; |
389 | ewitab = ewrt; |
390 | eweps = ewrt-ewitab; |
391 | ewitab = 4*ewitab; |
392 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
393 | velec = qq02*(rinv02-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
394 | felec = qq02*rinv02*(rinvsq02-felec); |
395 | |
396 | /* Update potential sums from outer loop */ |
397 | velecsum += velec; |
398 | |
399 | fscal = felec; |
400 | |
401 | /* Calculate temporary vectorial force */ |
402 | tx = fscal*dx02; |
403 | ty = fscal*dy02; |
404 | tz = fscal*dz02; |
405 | |
406 | /* Update vectorial force */ |
407 | fix0 += tx; |
408 | fiy0 += ty; |
409 | fiz0 += tz; |
410 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
411 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
412 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
413 | |
414 | /************************** |
415 | * CALCULATE INTERACTIONS * |
416 | **************************/ |
417 | |
418 | r10 = rsq10*rinv10; |
419 | |
420 | /* EWALD ELECTROSTATICS */ |
421 | |
422 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
423 | ewrt = r10*ewtabscale; |
424 | ewitab = ewrt; |
425 | eweps = ewrt-ewitab; |
426 | ewitab = 4*ewitab; |
427 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
428 | velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
429 | felec = qq10*rinv10*(rinvsq10-felec); |
430 | |
431 | /* Update potential sums from outer loop */ |
432 | velecsum += velec; |
433 | |
434 | fscal = felec; |
435 | |
436 | /* Calculate temporary vectorial force */ |
437 | tx = fscal*dx10; |
438 | ty = fscal*dy10; |
439 | tz = fscal*dz10; |
440 | |
441 | /* Update vectorial force */ |
442 | fix1 += tx; |
443 | fiy1 += ty; |
444 | fiz1 += tz; |
445 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
446 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
447 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
448 | |
449 | /************************** |
450 | * CALCULATE INTERACTIONS * |
451 | **************************/ |
452 | |
453 | r11 = rsq11*rinv11; |
454 | |
455 | /* EWALD ELECTROSTATICS */ |
456 | |
457 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
458 | ewrt = r11*ewtabscale; |
459 | ewitab = ewrt; |
460 | eweps = ewrt-ewitab; |
461 | ewitab = 4*ewitab; |
462 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
463 | velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
464 | felec = qq11*rinv11*(rinvsq11-felec); |
465 | |
466 | /* Update potential sums from outer loop */ |
467 | velecsum += velec; |
468 | |
469 | fscal = felec; |
470 | |
471 | /* Calculate temporary vectorial force */ |
472 | tx = fscal*dx11; |
473 | ty = fscal*dy11; |
474 | tz = fscal*dz11; |
475 | |
476 | /* Update vectorial force */ |
477 | fix1 += tx; |
478 | fiy1 += ty; |
479 | fiz1 += tz; |
480 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
481 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
482 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
483 | |
484 | /************************** |
485 | * CALCULATE INTERACTIONS * |
486 | **************************/ |
487 | |
488 | r12 = rsq12*rinv12; |
489 | |
490 | /* EWALD ELECTROSTATICS */ |
491 | |
492 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
493 | ewrt = r12*ewtabscale; |
494 | ewitab = ewrt; |
495 | eweps = ewrt-ewitab; |
496 | ewitab = 4*ewitab; |
497 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
498 | velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
499 | felec = qq12*rinv12*(rinvsq12-felec); |
500 | |
501 | /* Update potential sums from outer loop */ |
502 | velecsum += velec; |
503 | |
504 | fscal = felec; |
505 | |
506 | /* Calculate temporary vectorial force */ |
507 | tx = fscal*dx12; |
508 | ty = fscal*dy12; |
509 | tz = fscal*dz12; |
510 | |
511 | /* Update vectorial force */ |
512 | fix1 += tx; |
513 | fiy1 += ty; |
514 | fiz1 += tz; |
515 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
516 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
517 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
518 | |
519 | /************************** |
520 | * CALCULATE INTERACTIONS * |
521 | **************************/ |
522 | |
523 | r20 = rsq20*rinv20; |
524 | |
525 | /* EWALD ELECTROSTATICS */ |
526 | |
527 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
528 | ewrt = r20*ewtabscale; |
529 | ewitab = ewrt; |
530 | eweps = ewrt-ewitab; |
531 | ewitab = 4*ewitab; |
532 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
533 | velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
534 | felec = qq20*rinv20*(rinvsq20-felec); |
535 | |
536 | /* Update potential sums from outer loop */ |
537 | velecsum += velec; |
538 | |
539 | fscal = felec; |
540 | |
541 | /* Calculate temporary vectorial force */ |
542 | tx = fscal*dx20; |
543 | ty = fscal*dy20; |
544 | tz = fscal*dz20; |
545 | |
546 | /* Update vectorial force */ |
547 | fix2 += tx; |
548 | fiy2 += ty; |
549 | fiz2 += tz; |
550 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
551 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
552 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
553 | |
554 | /************************** |
555 | * CALCULATE INTERACTIONS * |
556 | **************************/ |
557 | |
558 | r21 = rsq21*rinv21; |
559 | |
560 | /* EWALD ELECTROSTATICS */ |
561 | |
562 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
563 | ewrt = r21*ewtabscale; |
564 | ewitab = ewrt; |
565 | eweps = ewrt-ewitab; |
566 | ewitab = 4*ewitab; |
567 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
568 | velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
569 | felec = qq21*rinv21*(rinvsq21-felec); |
570 | |
571 | /* Update potential sums from outer loop */ |
572 | velecsum += velec; |
573 | |
574 | fscal = felec; |
575 | |
576 | /* Calculate temporary vectorial force */ |
577 | tx = fscal*dx21; |
578 | ty = fscal*dy21; |
579 | tz = fscal*dz21; |
580 | |
581 | /* Update vectorial force */ |
582 | fix2 += tx; |
583 | fiy2 += ty; |
584 | fiz2 += tz; |
585 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
586 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
587 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
588 | |
589 | /************************** |
590 | * CALCULATE INTERACTIONS * |
591 | **************************/ |
592 | |
593 | r22 = rsq22*rinv22; |
594 | |
595 | /* EWALD ELECTROSTATICS */ |
596 | |
597 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
598 | ewrt = r22*ewtabscale; |
599 | ewitab = ewrt; |
600 | eweps = ewrt-ewitab; |
601 | ewitab = 4*ewitab; |
602 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
603 | velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
604 | felec = qq22*rinv22*(rinvsq22-felec); |
605 | |
606 | /* Update potential sums from outer loop */ |
607 | velecsum += velec; |
608 | |
609 | fscal = felec; |
610 | |
611 | /* Calculate temporary vectorial force */ |
612 | tx = fscal*dx22; |
613 | ty = fscal*dy22; |
614 | tz = fscal*dz22; |
615 | |
616 | /* Update vectorial force */ |
617 | fix2 += tx; |
618 | fiy2 += ty; |
619 | fiz2 += tz; |
620 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
621 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
622 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
623 | |
624 | /* Inner loop uses 393 flops */ |
625 | } |
626 | /* End of innermost loop */ |
627 | |
628 | tx = ty = tz = 0; |
629 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
630 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
631 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
632 | tx += fix0; |
633 | ty += fiy0; |
634 | tz += fiz0; |
635 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
636 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
637 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
638 | tx += fix1; |
639 | ty += fiy1; |
640 | tz += fiz1; |
641 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
642 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
643 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
644 | tx += fix2; |
645 | ty += fiy2; |
646 | tz += fiz2; |
647 | fshift[i_shift_offset+XX0] += tx; |
648 | fshift[i_shift_offset+YY1] += ty; |
649 | fshift[i_shift_offset+ZZ2] += tz; |
650 | |
651 | ggid = gid[iidx]; |
652 | /* Update potential energies */ |
653 | kernel_data->energygrp_elec[ggid] += velecsum; |
654 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
655 | |
656 | /* Increment number of inner iterations */ |
657 | inneriter += j_index_end - j_index_start; |
658 | |
659 | /* Outer loop uses 32 flops */ |
660 | } |
661 | |
662 | /* Increment number of outer iterations */ |
663 | outeriter += nri; |
664 | |
665 | /* Update outer/inner flops */ |
666 | |
667 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*393)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*32 + inneriter*393; |
668 | } |
669 | /* |
670 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwCSTab_GeomW3W3_F_c |
671 | * Electrostatics interaction: Ewald |
672 | * VdW interaction: CubicSplineTable |
673 | * Geometry: Water3-Water3 |
674 | * Calculate force/pot: Force |
675 | */ |
676 | void |
677 | nb_kernel_ElecEw_VdwCSTab_GeomW3W3_F_c |
678 | (t_nblist * gmx_restrict__restrict nlist, |
679 | rvec * gmx_restrict__restrict xx, |
680 | rvec * gmx_restrict__restrict ff, |
681 | t_forcerec * gmx_restrict__restrict fr, |
682 | t_mdatoms * gmx_restrict__restrict mdatoms, |
683 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
684 | t_nrnb * gmx_restrict__restrict nrnb) |
685 | { |
686 | int i_shift_offset,i_coord_offset,j_coord_offset; |
687 | int j_index_start,j_index_end; |
688 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
689 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
690 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
691 | real *shiftvec,*fshift,*x,*f; |
692 | int vdwioffset0; |
693 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
694 | int vdwioffset1; |
695 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
696 | int vdwioffset2; |
697 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
698 | int vdwjidx0; |
699 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
700 | int vdwjidx1; |
701 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
702 | int vdwjidx2; |
703 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
704 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
705 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
706 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
707 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
708 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
709 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
710 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
711 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
712 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
713 | real velec,felec,velecsum,facel,crf,krf,krf2; |
714 | real *charge; |
715 | int nvdwtype; |
716 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
717 | int *vdwtype; |
718 | real *vdwparam; |
719 | int vfitab; |
720 | real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF; |
721 | real *vftab; |
722 | int ewitab; |
723 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
724 | real *ewtab; |
725 | |
726 | x = xx[0]; |
727 | f = ff[0]; |
728 | |
729 | nri = nlist->nri; |
730 | iinr = nlist->iinr; |
731 | jindex = nlist->jindex; |
732 | jjnr = nlist->jjnr; |
733 | shiftidx = nlist->shift; |
734 | gid = nlist->gid; |
735 | shiftvec = fr->shift_vec[0]; |
736 | fshift = fr->fshift[0]; |
737 | facel = fr->epsfac; |
738 | charge = mdatoms->chargeA; |
739 | nvdwtype = fr->ntype; |
740 | vdwparam = fr->nbfp; |
741 | vdwtype = mdatoms->typeA; |
742 | |
743 | vftab = kernel_data->table_vdw->data; |
744 | vftabscale = kernel_data->table_vdw->scale; |
745 | |
746 | sh_ewald = fr->ic->sh_ewald; |
Value stored to 'sh_ewald' is never read | |
747 | ewtab = fr->ic->tabq_coul_F; |
748 | ewtabscale = fr->ic->tabq_scale; |
749 | ewtabhalfspace = 0.5/ewtabscale; |
750 | |
751 | /* Setup water-specific parameters */ |
752 | inr = nlist->iinr[0]; |
753 | iq0 = facel*charge[inr+0]; |
754 | iq1 = facel*charge[inr+1]; |
755 | iq2 = facel*charge[inr+2]; |
756 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
757 | |
758 | jq0 = charge[inr+0]; |
759 | jq1 = charge[inr+1]; |
760 | jq2 = charge[inr+2]; |
761 | vdwjidx0 = 2*vdwtype[inr+0]; |
762 | qq00 = iq0*jq0; |
763 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
764 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
765 | qq01 = iq0*jq1; |
766 | qq02 = iq0*jq2; |
767 | qq10 = iq1*jq0; |
768 | qq11 = iq1*jq1; |
769 | qq12 = iq1*jq2; |
770 | qq20 = iq2*jq0; |
771 | qq21 = iq2*jq1; |
772 | qq22 = iq2*jq2; |
773 | |
774 | outeriter = 0; |
775 | inneriter = 0; |
776 | |
777 | /* Start outer loop over neighborlists */ |
778 | for(iidx=0; iidx<nri; iidx++) |
779 | { |
780 | /* Load shift vector for this list */ |
781 | i_shift_offset = DIM3*shiftidx[iidx]; |
782 | shX = shiftvec[i_shift_offset+XX0]; |
783 | shY = shiftvec[i_shift_offset+YY1]; |
784 | shZ = shiftvec[i_shift_offset+ZZ2]; |
785 | |
786 | /* Load limits for loop over neighbors */ |
787 | j_index_start = jindex[iidx]; |
788 | j_index_end = jindex[iidx+1]; |
789 | |
790 | /* Get outer coordinate index */ |
791 | inr = iinr[iidx]; |
792 | i_coord_offset = DIM3*inr; |
793 | |
794 | /* Load i particle coords and add shift vector */ |
795 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
796 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
797 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
798 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
799 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
800 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
801 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
802 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
803 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
804 | |
805 | fix0 = 0.0; |
806 | fiy0 = 0.0; |
807 | fiz0 = 0.0; |
808 | fix1 = 0.0; |
809 | fiy1 = 0.0; |
810 | fiz1 = 0.0; |
811 | fix2 = 0.0; |
812 | fiy2 = 0.0; |
813 | fiz2 = 0.0; |
814 | |
815 | /* Start inner kernel loop */ |
816 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
817 | { |
818 | /* Get j neighbor index, and coordinate index */ |
819 | jnr = jjnr[jidx]; |
820 | j_coord_offset = DIM3*jnr; |
821 | |
822 | /* load j atom coordinates */ |
823 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
824 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
825 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
826 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
827 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
828 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
829 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
830 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
831 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
832 | |
833 | /* Calculate displacement vector */ |
834 | dx00 = ix0 - jx0; |
835 | dy00 = iy0 - jy0; |
836 | dz00 = iz0 - jz0; |
837 | dx01 = ix0 - jx1; |
838 | dy01 = iy0 - jy1; |
839 | dz01 = iz0 - jz1; |
840 | dx02 = ix0 - jx2; |
841 | dy02 = iy0 - jy2; |
842 | dz02 = iz0 - jz2; |
843 | dx10 = ix1 - jx0; |
844 | dy10 = iy1 - jy0; |
845 | dz10 = iz1 - jz0; |
846 | dx11 = ix1 - jx1; |
847 | dy11 = iy1 - jy1; |
848 | dz11 = iz1 - jz1; |
849 | dx12 = ix1 - jx2; |
850 | dy12 = iy1 - jy2; |
851 | dz12 = iz1 - jz2; |
852 | dx20 = ix2 - jx0; |
853 | dy20 = iy2 - jy0; |
854 | dz20 = iz2 - jz0; |
855 | dx21 = ix2 - jx1; |
856 | dy21 = iy2 - jy1; |
857 | dz21 = iz2 - jz1; |
858 | dx22 = ix2 - jx2; |
859 | dy22 = iy2 - jy2; |
860 | dz22 = iz2 - jz2; |
861 | |
862 | /* Calculate squared distance and things based on it */ |
863 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
864 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
865 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
866 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
867 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
868 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
869 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
870 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
871 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
872 | |
873 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
874 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
875 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
876 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
877 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
878 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
879 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
880 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
881 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
882 | |
883 | rinvsq00 = rinv00*rinv00; |
884 | rinvsq01 = rinv01*rinv01; |
885 | rinvsq02 = rinv02*rinv02; |
886 | rinvsq10 = rinv10*rinv10; |
887 | rinvsq11 = rinv11*rinv11; |
888 | rinvsq12 = rinv12*rinv12; |
889 | rinvsq20 = rinv20*rinv20; |
890 | rinvsq21 = rinv21*rinv21; |
891 | rinvsq22 = rinv22*rinv22; |
892 | |
893 | /************************** |
894 | * CALCULATE INTERACTIONS * |
895 | **************************/ |
896 | |
897 | r00 = rsq00*rinv00; |
898 | |
899 | /* Calculate table index by multiplying r with table scale and truncate to integer */ |
900 | rt = r00*vftabscale; |
901 | vfitab = rt; |
902 | vfeps = rt-vfitab; |
903 | vfitab = 2*4*vfitab; |
904 | |
905 | /* EWALD ELECTROSTATICS */ |
906 | |
907 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
908 | ewrt = r00*ewtabscale; |
909 | ewitab = ewrt; |
910 | eweps = ewrt-ewitab; |
911 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
912 | felec = qq00*rinv00*(rinvsq00-felec); |
913 | |
914 | /* CUBIC SPLINE TABLE DISPERSION */ |
915 | vfitab += 0; |
916 | F = vftab[vfitab+1]; |
917 | Geps = vfeps*vftab[vfitab+2]; |
918 | Heps2 = vfeps*vfeps*vftab[vfitab+3]; |
919 | Fp = F+Geps+Heps2; |
920 | FF = Fp+Geps+2.0*Heps2; |
921 | fvdw6 = c6_00*FF; |
922 | |
923 | /* CUBIC SPLINE TABLE REPULSION */ |
924 | F = vftab[vfitab+5]; |
925 | Geps = vfeps*vftab[vfitab+6]; |
926 | Heps2 = vfeps*vfeps*vftab[vfitab+7]; |
927 | Fp = F+Geps+Heps2; |
928 | FF = Fp+Geps+2.0*Heps2; |
929 | fvdw12 = c12_00*FF; |
930 | fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00; |
931 | |
932 | fscal = felec+fvdw; |
933 | |
934 | /* Calculate temporary vectorial force */ |
935 | tx = fscal*dx00; |
936 | ty = fscal*dy00; |
937 | tz = fscal*dz00; |
938 | |
939 | /* Update vectorial force */ |
940 | fix0 += tx; |
941 | fiy0 += ty; |
942 | fiz0 += tz; |
943 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
944 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
945 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
946 | |
947 | /************************** |
948 | * CALCULATE INTERACTIONS * |
949 | **************************/ |
950 | |
951 | r01 = rsq01*rinv01; |
952 | |
953 | /* EWALD ELECTROSTATICS */ |
954 | |
955 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
956 | ewrt = r01*ewtabscale; |
957 | ewitab = ewrt; |
958 | eweps = ewrt-ewitab; |
959 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
960 | felec = qq01*rinv01*(rinvsq01-felec); |
961 | |
962 | fscal = felec; |
963 | |
964 | /* Calculate temporary vectorial force */ |
965 | tx = fscal*dx01; |
966 | ty = fscal*dy01; |
967 | tz = fscal*dz01; |
968 | |
969 | /* Update vectorial force */ |
970 | fix0 += tx; |
971 | fiy0 += ty; |
972 | fiz0 += tz; |
973 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
974 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
975 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
976 | |
977 | /************************** |
978 | * CALCULATE INTERACTIONS * |
979 | **************************/ |
980 | |
981 | r02 = rsq02*rinv02; |
982 | |
983 | /* EWALD ELECTROSTATICS */ |
984 | |
985 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
986 | ewrt = r02*ewtabscale; |
987 | ewitab = ewrt; |
988 | eweps = ewrt-ewitab; |
989 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
990 | felec = qq02*rinv02*(rinvsq02-felec); |
991 | |
992 | fscal = felec; |
993 | |
994 | /* Calculate temporary vectorial force */ |
995 | tx = fscal*dx02; |
996 | ty = fscal*dy02; |
997 | tz = fscal*dz02; |
998 | |
999 | /* Update vectorial force */ |
1000 | fix0 += tx; |
1001 | fiy0 += ty; |
1002 | fiz0 += tz; |
1003 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1004 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1005 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1006 | |
1007 | /************************** |
1008 | * CALCULATE INTERACTIONS * |
1009 | **************************/ |
1010 | |
1011 | r10 = rsq10*rinv10; |
1012 | |
1013 | /* EWALD ELECTROSTATICS */ |
1014 | |
1015 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1016 | ewrt = r10*ewtabscale; |
1017 | ewitab = ewrt; |
1018 | eweps = ewrt-ewitab; |
1019 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1020 | felec = qq10*rinv10*(rinvsq10-felec); |
1021 | |
1022 | fscal = felec; |
1023 | |
1024 | /* Calculate temporary vectorial force */ |
1025 | tx = fscal*dx10; |
1026 | ty = fscal*dy10; |
1027 | tz = fscal*dz10; |
1028 | |
1029 | /* Update vectorial force */ |
1030 | fix1 += tx; |
1031 | fiy1 += ty; |
1032 | fiz1 += tz; |
1033 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1034 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1035 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1036 | |
1037 | /************************** |
1038 | * CALCULATE INTERACTIONS * |
1039 | **************************/ |
1040 | |
1041 | r11 = rsq11*rinv11; |
1042 | |
1043 | /* EWALD ELECTROSTATICS */ |
1044 | |
1045 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1046 | ewrt = r11*ewtabscale; |
1047 | ewitab = ewrt; |
1048 | eweps = ewrt-ewitab; |
1049 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1050 | felec = qq11*rinv11*(rinvsq11-felec); |
1051 | |
1052 | fscal = felec; |
1053 | |
1054 | /* Calculate temporary vectorial force */ |
1055 | tx = fscal*dx11; |
1056 | ty = fscal*dy11; |
1057 | tz = fscal*dz11; |
1058 | |
1059 | /* Update vectorial force */ |
1060 | fix1 += tx; |
1061 | fiy1 += ty; |
1062 | fiz1 += tz; |
1063 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1064 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1065 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1066 | |
1067 | /************************** |
1068 | * CALCULATE INTERACTIONS * |
1069 | **************************/ |
1070 | |
1071 | r12 = rsq12*rinv12; |
1072 | |
1073 | /* EWALD ELECTROSTATICS */ |
1074 | |
1075 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1076 | ewrt = r12*ewtabscale; |
1077 | ewitab = ewrt; |
1078 | eweps = ewrt-ewitab; |
1079 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1080 | felec = qq12*rinv12*(rinvsq12-felec); |
1081 | |
1082 | fscal = felec; |
1083 | |
1084 | /* Calculate temporary vectorial force */ |
1085 | tx = fscal*dx12; |
1086 | ty = fscal*dy12; |
1087 | tz = fscal*dz12; |
1088 | |
1089 | /* Update vectorial force */ |
1090 | fix1 += tx; |
1091 | fiy1 += ty; |
1092 | fiz1 += tz; |
1093 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1094 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1095 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1096 | |
1097 | /************************** |
1098 | * CALCULATE INTERACTIONS * |
1099 | **************************/ |
1100 | |
1101 | r20 = rsq20*rinv20; |
1102 | |
1103 | /* EWALD ELECTROSTATICS */ |
1104 | |
1105 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1106 | ewrt = r20*ewtabscale; |
1107 | ewitab = ewrt; |
1108 | eweps = ewrt-ewitab; |
1109 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1110 | felec = qq20*rinv20*(rinvsq20-felec); |
1111 | |
1112 | fscal = felec; |
1113 | |
1114 | /* Calculate temporary vectorial force */ |
1115 | tx = fscal*dx20; |
1116 | ty = fscal*dy20; |
1117 | tz = fscal*dz20; |
1118 | |
1119 | /* Update vectorial force */ |
1120 | fix2 += tx; |
1121 | fiy2 += ty; |
1122 | fiz2 += tz; |
1123 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
1124 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
1125 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
1126 | |
1127 | /************************** |
1128 | * CALCULATE INTERACTIONS * |
1129 | **************************/ |
1130 | |
1131 | r21 = rsq21*rinv21; |
1132 | |
1133 | /* EWALD ELECTROSTATICS */ |
1134 | |
1135 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1136 | ewrt = r21*ewtabscale; |
1137 | ewitab = ewrt; |
1138 | eweps = ewrt-ewitab; |
1139 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1140 | felec = qq21*rinv21*(rinvsq21-felec); |
1141 | |
1142 | fscal = felec; |
1143 | |
1144 | /* Calculate temporary vectorial force */ |
1145 | tx = fscal*dx21; |
1146 | ty = fscal*dy21; |
1147 | tz = fscal*dz21; |
1148 | |
1149 | /* Update vectorial force */ |
1150 | fix2 += tx; |
1151 | fiy2 += ty; |
1152 | fiz2 += tz; |
1153 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
1154 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
1155 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
1156 | |
1157 | /************************** |
1158 | * CALCULATE INTERACTIONS * |
1159 | **************************/ |
1160 | |
1161 | r22 = rsq22*rinv22; |
1162 | |
1163 | /* EWALD ELECTROSTATICS */ |
1164 | |
1165 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1166 | ewrt = r22*ewtabscale; |
1167 | ewitab = ewrt; |
1168 | eweps = ewrt-ewitab; |
1169 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
1170 | felec = qq22*rinv22*(rinvsq22-felec); |
1171 | |
1172 | fscal = felec; |
1173 | |
1174 | /* Calculate temporary vectorial force */ |
1175 | tx = fscal*dx22; |
1176 | ty = fscal*dy22; |
1177 | tz = fscal*dz22; |
1178 | |
1179 | /* Update vectorial force */ |
1180 | fix2 += tx; |
1181 | fiy2 += ty; |
1182 | fiz2 += tz; |
1183 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
1184 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
1185 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
1186 | |
1187 | /* Inner loop uses 322 flops */ |
1188 | } |
1189 | /* End of innermost loop */ |
1190 | |
1191 | tx = ty = tz = 0; |
1192 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
1193 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
1194 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
1195 | tx += fix0; |
1196 | ty += fiy0; |
1197 | tz += fiz0; |
1198 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
1199 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
1200 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
1201 | tx += fix1; |
1202 | ty += fiy1; |
1203 | tz += fiz1; |
1204 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
1205 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
1206 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
1207 | tx += fix2; |
1208 | ty += fiy2; |
1209 | tz += fiz2; |
1210 | fshift[i_shift_offset+XX0] += tx; |
1211 | fshift[i_shift_offset+YY1] += ty; |
1212 | fshift[i_shift_offset+ZZ2] += tz; |
1213 | |
1214 | /* Increment number of inner iterations */ |
1215 | inneriter += j_index_end - j_index_start; |
1216 | |
1217 | /* Outer loop uses 30 flops */ |
1218 | } |
1219 | |
1220 | /* Increment number of outer iterations */ |
1221 | outeriter += nri; |
1222 | |
1223 | /* Update outer/inner flops */ |
1224 | |
1225 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*322)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*30 + inneriter *322; |
1226 | } |