2 * This file is part of the GROMACS molecular simulation package.
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.
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.
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.
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.
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.
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.
36 * Note: this file was generated by the GROMACS c kernel generator.
42 #include "../nb_kernel.h"
43 #include "gromacs/legacyheaders/types/simple.h"
44 #include "gromacs/math/vec.h"
45 #include "gromacs/legacyheaders/nrnb.h"
48 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_VF_c
49 * Electrostatics interaction: Ewald
50 * VdW interaction: Buckingham
51 * Geometry: Water4-Water4
52 * Calculate force/pot: PotentialAndForce
55 nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_VF_c
56 (t_nblist * gmx_restrict nlist,
57 rvec * gmx_restrict xx,
58 rvec * gmx_restrict ff,
59 t_forcerec * gmx_restrict fr,
60 t_mdatoms * gmx_restrict mdatoms,
61 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
62 t_nrnb * gmx_restrict nrnb)
64 int i_shift_offset,i_coord_offset,j_coord_offset;
65 int j_index_start,j_index_end;
66 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
67 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
68 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
69 real *shiftvec,*fshift,*x,*f;
71 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
73 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
75 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
77 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
79 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
81 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
83 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
85 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
86 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
87 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
88 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
89 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
90 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
91 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
92 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
93 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
94 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
95 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
96 real velec,felec,velecsum,facel,crf,krf,krf2;
99 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
103 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
105 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
112 jindex = nlist->jindex;
114 shiftidx = nlist->shift;
116 shiftvec = fr->shift_vec[0];
117 fshift = fr->fshift[0];
119 charge = mdatoms->chargeA;
120 nvdwtype = fr->ntype;
122 vdwtype = mdatoms->typeA;
124 sh_ewald = fr->ic->sh_ewald;
125 ewtab = fr->ic->tabq_coul_FDV0;
126 ewtabscale = fr->ic->tabq_scale;
127 ewtabhalfspace = 0.5/ewtabscale;
129 /* Setup water-specific parameters */
130 inr = nlist->iinr[0];
131 iq1 = facel*charge[inr+1];
132 iq2 = facel*charge[inr+2];
133 iq3 = facel*charge[inr+3];
134 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
139 vdwjidx0 = 3*vdwtype[inr+0];
140 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
141 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
142 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
153 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
154 rcutoff = fr->rcoulomb;
155 rcutoff2 = rcutoff*rcutoff;
157 rswitch = fr->rcoulomb_switch;
158 /* Setup switch parameters */
160 swV3 = -10.0/(d*d*d);
161 swV4 = 15.0/(d*d*d*d);
162 swV5 = -6.0/(d*d*d*d*d);
163 swF2 = -30.0/(d*d*d);
164 swF3 = 60.0/(d*d*d*d);
165 swF4 = -30.0/(d*d*d*d*d);
170 /* Start outer loop over neighborlists */
171 for(iidx=0; iidx<nri; iidx++)
173 /* Load shift vector for this list */
174 i_shift_offset = DIM*shiftidx[iidx];
175 shX = shiftvec[i_shift_offset+XX];
176 shY = shiftvec[i_shift_offset+YY];
177 shZ = shiftvec[i_shift_offset+ZZ];
179 /* Load limits for loop over neighbors */
180 j_index_start = jindex[iidx];
181 j_index_end = jindex[iidx+1];
183 /* Get outer coordinate index */
185 i_coord_offset = DIM*inr;
187 /* Load i particle coords and add shift vector */
188 ix0 = shX + x[i_coord_offset+DIM*0+XX];
189 iy0 = shY + x[i_coord_offset+DIM*0+YY];
190 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
191 ix1 = shX + x[i_coord_offset+DIM*1+XX];
192 iy1 = shY + x[i_coord_offset+DIM*1+YY];
193 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
194 ix2 = shX + x[i_coord_offset+DIM*2+XX];
195 iy2 = shY + x[i_coord_offset+DIM*2+YY];
196 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
197 ix3 = shX + x[i_coord_offset+DIM*3+XX];
198 iy3 = shY + x[i_coord_offset+DIM*3+YY];
199 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
214 /* Reset potential sums */
218 /* Start inner kernel loop */
219 for(jidx=j_index_start; jidx<j_index_end; jidx++)
221 /* Get j neighbor index, and coordinate index */
223 j_coord_offset = DIM*jnr;
225 /* load j atom coordinates */
226 jx0 = x[j_coord_offset+DIM*0+XX];
227 jy0 = x[j_coord_offset+DIM*0+YY];
228 jz0 = x[j_coord_offset+DIM*0+ZZ];
229 jx1 = x[j_coord_offset+DIM*1+XX];
230 jy1 = x[j_coord_offset+DIM*1+YY];
231 jz1 = x[j_coord_offset+DIM*1+ZZ];
232 jx2 = x[j_coord_offset+DIM*2+XX];
233 jy2 = x[j_coord_offset+DIM*2+YY];
234 jz2 = x[j_coord_offset+DIM*2+ZZ];
235 jx3 = x[j_coord_offset+DIM*3+XX];
236 jy3 = x[j_coord_offset+DIM*3+YY];
237 jz3 = x[j_coord_offset+DIM*3+ZZ];
239 /* Calculate displacement vector */
271 /* Calculate squared distance and things based on it */
272 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
273 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
274 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
275 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
276 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
277 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
278 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
279 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
280 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
281 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
283 rinv00 = gmx_invsqrt(rsq00);
284 rinv11 = gmx_invsqrt(rsq11);
285 rinv12 = gmx_invsqrt(rsq12);
286 rinv13 = gmx_invsqrt(rsq13);
287 rinv21 = gmx_invsqrt(rsq21);
288 rinv22 = gmx_invsqrt(rsq22);
289 rinv23 = gmx_invsqrt(rsq23);
290 rinv31 = gmx_invsqrt(rsq31);
291 rinv32 = gmx_invsqrt(rsq32);
292 rinv33 = gmx_invsqrt(rsq33);
294 rinvsq00 = rinv00*rinv00;
295 rinvsq11 = rinv11*rinv11;
296 rinvsq12 = rinv12*rinv12;
297 rinvsq13 = rinv13*rinv13;
298 rinvsq21 = rinv21*rinv21;
299 rinvsq22 = rinv22*rinv22;
300 rinvsq23 = rinv23*rinv23;
301 rinvsq31 = rinv31*rinv31;
302 rinvsq32 = rinv32*rinv32;
303 rinvsq33 = rinv33*rinv33;
305 /**************************
306 * CALCULATE INTERACTIONS *
307 **************************/
314 /* BUCKINGHAM DISPERSION/REPULSION */
315 rinvsix = rinvsq00*rinvsq00*rinvsq00;
316 vvdw6 = c6_00*rinvsix;
318 vvdwexp = cexp1_00*exp(-br);
319 vvdw = vvdwexp - vvdw6*(1.0/6.0);
320 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
323 d = (d>0.0) ? d : 0.0;
325 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
327 dsw = d2*(swF2+d*(swF3+d*swF4));
329 /* Evaluate switch function */
330 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
331 fvdw = fvdw*sw - rinv00*vvdw*dsw;
334 /* Update potential sums from outer loop */
339 /* Calculate temporary vectorial force */
344 /* Update vectorial force */
348 f[j_coord_offset+DIM*0+XX] -= tx;
349 f[j_coord_offset+DIM*0+YY] -= ty;
350 f[j_coord_offset+DIM*0+ZZ] -= tz;
354 /**************************
355 * CALCULATE INTERACTIONS *
356 **************************/
363 /* EWALD ELECTROSTATICS */
365 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
366 ewrt = r11*ewtabscale;
370 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
371 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
372 felec = qq11*rinv11*(rinvsq11-felec);
375 d = (d>0.0) ? d : 0.0;
377 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
379 dsw = d2*(swF2+d*(swF3+d*swF4));
381 /* Evaluate switch function */
382 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
383 felec = felec*sw - rinv11*velec*dsw;
386 /* Update potential sums from outer loop */
391 /* Calculate temporary vectorial force */
396 /* Update vectorial force */
400 f[j_coord_offset+DIM*1+XX] -= tx;
401 f[j_coord_offset+DIM*1+YY] -= ty;
402 f[j_coord_offset+DIM*1+ZZ] -= tz;
406 /**************************
407 * CALCULATE INTERACTIONS *
408 **************************/
415 /* EWALD ELECTROSTATICS */
417 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
418 ewrt = r12*ewtabscale;
422 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
423 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
424 felec = qq12*rinv12*(rinvsq12-felec);
427 d = (d>0.0) ? d : 0.0;
429 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
431 dsw = d2*(swF2+d*(swF3+d*swF4));
433 /* Evaluate switch function */
434 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
435 felec = felec*sw - rinv12*velec*dsw;
438 /* Update potential sums from outer loop */
443 /* Calculate temporary vectorial force */
448 /* Update vectorial force */
452 f[j_coord_offset+DIM*2+XX] -= tx;
453 f[j_coord_offset+DIM*2+YY] -= ty;
454 f[j_coord_offset+DIM*2+ZZ] -= tz;
458 /**************************
459 * CALCULATE INTERACTIONS *
460 **************************/
467 /* EWALD ELECTROSTATICS */
469 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
470 ewrt = r13*ewtabscale;
474 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
475 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
476 felec = qq13*rinv13*(rinvsq13-felec);
479 d = (d>0.0) ? d : 0.0;
481 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
483 dsw = d2*(swF2+d*(swF3+d*swF4));
485 /* Evaluate switch function */
486 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
487 felec = felec*sw - rinv13*velec*dsw;
490 /* Update potential sums from outer loop */
495 /* Calculate temporary vectorial force */
500 /* Update vectorial force */
504 f[j_coord_offset+DIM*3+XX] -= tx;
505 f[j_coord_offset+DIM*3+YY] -= ty;
506 f[j_coord_offset+DIM*3+ZZ] -= tz;
510 /**************************
511 * CALCULATE INTERACTIONS *
512 **************************/
519 /* EWALD ELECTROSTATICS */
521 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
522 ewrt = r21*ewtabscale;
526 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
527 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
528 felec = qq21*rinv21*(rinvsq21-felec);
531 d = (d>0.0) ? d : 0.0;
533 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
535 dsw = d2*(swF2+d*(swF3+d*swF4));
537 /* Evaluate switch function */
538 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
539 felec = felec*sw - rinv21*velec*dsw;
542 /* Update potential sums from outer loop */
547 /* Calculate temporary vectorial force */
552 /* Update vectorial force */
556 f[j_coord_offset+DIM*1+XX] -= tx;
557 f[j_coord_offset+DIM*1+YY] -= ty;
558 f[j_coord_offset+DIM*1+ZZ] -= tz;
562 /**************************
563 * CALCULATE INTERACTIONS *
564 **************************/
571 /* EWALD ELECTROSTATICS */
573 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
574 ewrt = r22*ewtabscale;
578 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
579 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
580 felec = qq22*rinv22*(rinvsq22-felec);
583 d = (d>0.0) ? d : 0.0;
585 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
587 dsw = d2*(swF2+d*(swF3+d*swF4));
589 /* Evaluate switch function */
590 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
591 felec = felec*sw - rinv22*velec*dsw;
594 /* Update potential sums from outer loop */
599 /* Calculate temporary vectorial force */
604 /* Update vectorial force */
608 f[j_coord_offset+DIM*2+XX] -= tx;
609 f[j_coord_offset+DIM*2+YY] -= ty;
610 f[j_coord_offset+DIM*2+ZZ] -= tz;
614 /**************************
615 * CALCULATE INTERACTIONS *
616 **************************/
623 /* EWALD ELECTROSTATICS */
625 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
626 ewrt = r23*ewtabscale;
630 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
631 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
632 felec = qq23*rinv23*(rinvsq23-felec);
635 d = (d>0.0) ? d : 0.0;
637 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
639 dsw = d2*(swF2+d*(swF3+d*swF4));
641 /* Evaluate switch function */
642 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
643 felec = felec*sw - rinv23*velec*dsw;
646 /* Update potential sums from outer loop */
651 /* Calculate temporary vectorial force */
656 /* Update vectorial force */
660 f[j_coord_offset+DIM*3+XX] -= tx;
661 f[j_coord_offset+DIM*3+YY] -= ty;
662 f[j_coord_offset+DIM*3+ZZ] -= tz;
666 /**************************
667 * CALCULATE INTERACTIONS *
668 **************************/
675 /* EWALD ELECTROSTATICS */
677 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
678 ewrt = r31*ewtabscale;
682 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
683 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
684 felec = qq31*rinv31*(rinvsq31-felec);
687 d = (d>0.0) ? d : 0.0;
689 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
691 dsw = d2*(swF2+d*(swF3+d*swF4));
693 /* Evaluate switch function */
694 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
695 felec = felec*sw - rinv31*velec*dsw;
698 /* Update potential sums from outer loop */
703 /* Calculate temporary vectorial force */
708 /* Update vectorial force */
712 f[j_coord_offset+DIM*1+XX] -= tx;
713 f[j_coord_offset+DIM*1+YY] -= ty;
714 f[j_coord_offset+DIM*1+ZZ] -= tz;
718 /**************************
719 * CALCULATE INTERACTIONS *
720 **************************/
727 /* EWALD ELECTROSTATICS */
729 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
730 ewrt = r32*ewtabscale;
734 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
735 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
736 felec = qq32*rinv32*(rinvsq32-felec);
739 d = (d>0.0) ? d : 0.0;
741 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
743 dsw = d2*(swF2+d*(swF3+d*swF4));
745 /* Evaluate switch function */
746 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
747 felec = felec*sw - rinv32*velec*dsw;
750 /* Update potential sums from outer loop */
755 /* Calculate temporary vectorial force */
760 /* Update vectorial force */
764 f[j_coord_offset+DIM*2+XX] -= tx;
765 f[j_coord_offset+DIM*2+YY] -= ty;
766 f[j_coord_offset+DIM*2+ZZ] -= tz;
770 /**************************
771 * CALCULATE INTERACTIONS *
772 **************************/
779 /* EWALD ELECTROSTATICS */
781 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
782 ewrt = r33*ewtabscale;
786 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
787 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
788 felec = qq33*rinv33*(rinvsq33-felec);
791 d = (d>0.0) ? d : 0.0;
793 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
795 dsw = d2*(swF2+d*(swF3+d*swF4));
797 /* Evaluate switch function */
798 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
799 felec = felec*sw - rinv33*velec*dsw;
802 /* Update potential sums from outer loop */
807 /* Calculate temporary vectorial force */
812 /* Update vectorial force */
816 f[j_coord_offset+DIM*3+XX] -= tx;
817 f[j_coord_offset+DIM*3+YY] -= ty;
818 f[j_coord_offset+DIM*3+ZZ] -= tz;
822 /* Inner loop uses 601 flops */
824 /* End of innermost loop */
827 f[i_coord_offset+DIM*0+XX] += fix0;
828 f[i_coord_offset+DIM*0+YY] += fiy0;
829 f[i_coord_offset+DIM*0+ZZ] += fiz0;
833 f[i_coord_offset+DIM*1+XX] += fix1;
834 f[i_coord_offset+DIM*1+YY] += fiy1;
835 f[i_coord_offset+DIM*1+ZZ] += fiz1;
839 f[i_coord_offset+DIM*2+XX] += fix2;
840 f[i_coord_offset+DIM*2+YY] += fiy2;
841 f[i_coord_offset+DIM*2+ZZ] += fiz2;
845 f[i_coord_offset+DIM*3+XX] += fix3;
846 f[i_coord_offset+DIM*3+YY] += fiy3;
847 f[i_coord_offset+DIM*3+ZZ] += fiz3;
851 fshift[i_shift_offset+XX] += tx;
852 fshift[i_shift_offset+YY] += ty;
853 fshift[i_shift_offset+ZZ] += tz;
856 /* Update potential energies */
857 kernel_data->energygrp_elec[ggid] += velecsum;
858 kernel_data->energygrp_vdw[ggid] += vvdwsum;
860 /* Increment number of inner iterations */
861 inneriter += j_index_end - j_index_start;
863 /* Outer loop uses 41 flops */
866 /* Increment number of outer iterations */
869 /* Update outer/inner flops */
871 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*41 + inneriter*601);
874 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_F_c
875 * Electrostatics interaction: Ewald
876 * VdW interaction: Buckingham
877 * Geometry: Water4-Water4
878 * Calculate force/pot: Force
881 nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_F_c
882 (t_nblist * gmx_restrict nlist,
883 rvec * gmx_restrict xx,
884 rvec * gmx_restrict ff,
885 t_forcerec * gmx_restrict fr,
886 t_mdatoms * gmx_restrict mdatoms,
887 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
888 t_nrnb * gmx_restrict nrnb)
890 int i_shift_offset,i_coord_offset,j_coord_offset;
891 int j_index_start,j_index_end;
892 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
893 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
894 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
895 real *shiftvec,*fshift,*x,*f;
897 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
899 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
901 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
903 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
905 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
907 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
909 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
911 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
912 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
913 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
914 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
915 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
916 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
917 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
918 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
919 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
920 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
921 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
922 real velec,felec,velecsum,facel,crf,krf,krf2;
925 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
929 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
931 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
938 jindex = nlist->jindex;
940 shiftidx = nlist->shift;
942 shiftvec = fr->shift_vec[0];
943 fshift = fr->fshift[0];
945 charge = mdatoms->chargeA;
946 nvdwtype = fr->ntype;
948 vdwtype = mdatoms->typeA;
950 sh_ewald = fr->ic->sh_ewald;
951 ewtab = fr->ic->tabq_coul_FDV0;
952 ewtabscale = fr->ic->tabq_scale;
953 ewtabhalfspace = 0.5/ewtabscale;
955 /* Setup water-specific parameters */
956 inr = nlist->iinr[0];
957 iq1 = facel*charge[inr+1];
958 iq2 = facel*charge[inr+2];
959 iq3 = facel*charge[inr+3];
960 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
965 vdwjidx0 = 3*vdwtype[inr+0];
966 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
967 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
968 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
979 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
980 rcutoff = fr->rcoulomb;
981 rcutoff2 = rcutoff*rcutoff;
983 rswitch = fr->rcoulomb_switch;
984 /* Setup switch parameters */
986 swV3 = -10.0/(d*d*d);
987 swV4 = 15.0/(d*d*d*d);
988 swV5 = -6.0/(d*d*d*d*d);
989 swF2 = -30.0/(d*d*d);
990 swF3 = 60.0/(d*d*d*d);
991 swF4 = -30.0/(d*d*d*d*d);
996 /* Start outer loop over neighborlists */
997 for(iidx=0; iidx<nri; iidx++)
999 /* Load shift vector for this list */
1000 i_shift_offset = DIM*shiftidx[iidx];
1001 shX = shiftvec[i_shift_offset+XX];
1002 shY = shiftvec[i_shift_offset+YY];
1003 shZ = shiftvec[i_shift_offset+ZZ];
1005 /* Load limits for loop over neighbors */
1006 j_index_start = jindex[iidx];
1007 j_index_end = jindex[iidx+1];
1009 /* Get outer coordinate index */
1011 i_coord_offset = DIM*inr;
1013 /* Load i particle coords and add shift vector */
1014 ix0 = shX + x[i_coord_offset+DIM*0+XX];
1015 iy0 = shY + x[i_coord_offset+DIM*0+YY];
1016 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
1017 ix1 = shX + x[i_coord_offset+DIM*1+XX];
1018 iy1 = shY + x[i_coord_offset+DIM*1+YY];
1019 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
1020 ix2 = shX + x[i_coord_offset+DIM*2+XX];
1021 iy2 = shY + x[i_coord_offset+DIM*2+YY];
1022 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
1023 ix3 = shX + x[i_coord_offset+DIM*3+XX];
1024 iy3 = shY + x[i_coord_offset+DIM*3+YY];
1025 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
1040 /* Start inner kernel loop */
1041 for(jidx=j_index_start; jidx<j_index_end; jidx++)
1043 /* Get j neighbor index, and coordinate index */
1045 j_coord_offset = DIM*jnr;
1047 /* load j atom coordinates */
1048 jx0 = x[j_coord_offset+DIM*0+XX];
1049 jy0 = x[j_coord_offset+DIM*0+YY];
1050 jz0 = x[j_coord_offset+DIM*0+ZZ];
1051 jx1 = x[j_coord_offset+DIM*1+XX];
1052 jy1 = x[j_coord_offset+DIM*1+YY];
1053 jz1 = x[j_coord_offset+DIM*1+ZZ];
1054 jx2 = x[j_coord_offset+DIM*2+XX];
1055 jy2 = x[j_coord_offset+DIM*2+YY];
1056 jz2 = x[j_coord_offset+DIM*2+ZZ];
1057 jx3 = x[j_coord_offset+DIM*3+XX];
1058 jy3 = x[j_coord_offset+DIM*3+YY];
1059 jz3 = x[j_coord_offset+DIM*3+ZZ];
1061 /* Calculate displacement vector */
1093 /* Calculate squared distance and things based on it */
1094 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
1095 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
1096 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
1097 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
1098 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
1099 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
1100 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
1101 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
1102 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
1103 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
1105 rinv00 = gmx_invsqrt(rsq00);
1106 rinv11 = gmx_invsqrt(rsq11);
1107 rinv12 = gmx_invsqrt(rsq12);
1108 rinv13 = gmx_invsqrt(rsq13);
1109 rinv21 = gmx_invsqrt(rsq21);
1110 rinv22 = gmx_invsqrt(rsq22);
1111 rinv23 = gmx_invsqrt(rsq23);
1112 rinv31 = gmx_invsqrt(rsq31);
1113 rinv32 = gmx_invsqrt(rsq32);
1114 rinv33 = gmx_invsqrt(rsq33);
1116 rinvsq00 = rinv00*rinv00;
1117 rinvsq11 = rinv11*rinv11;
1118 rinvsq12 = rinv12*rinv12;
1119 rinvsq13 = rinv13*rinv13;
1120 rinvsq21 = rinv21*rinv21;
1121 rinvsq22 = rinv22*rinv22;
1122 rinvsq23 = rinv23*rinv23;
1123 rinvsq31 = rinv31*rinv31;
1124 rinvsq32 = rinv32*rinv32;
1125 rinvsq33 = rinv33*rinv33;
1127 /**************************
1128 * CALCULATE INTERACTIONS *
1129 **************************/
1136 /* BUCKINGHAM DISPERSION/REPULSION */
1137 rinvsix = rinvsq00*rinvsq00*rinvsq00;
1138 vvdw6 = c6_00*rinvsix;
1140 vvdwexp = cexp1_00*exp(-br);
1141 vvdw = vvdwexp - vvdw6*(1.0/6.0);
1142 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
1145 d = (d>0.0) ? d : 0.0;
1147 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1149 dsw = d2*(swF2+d*(swF3+d*swF4));
1151 /* Evaluate switch function */
1152 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1153 fvdw = fvdw*sw - rinv00*vvdw*dsw;
1157 /* Calculate temporary vectorial force */
1162 /* Update vectorial force */
1166 f[j_coord_offset+DIM*0+XX] -= tx;
1167 f[j_coord_offset+DIM*0+YY] -= ty;
1168 f[j_coord_offset+DIM*0+ZZ] -= tz;
1172 /**************************
1173 * CALCULATE INTERACTIONS *
1174 **************************/
1181 /* EWALD ELECTROSTATICS */
1183 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1184 ewrt = r11*ewtabscale;
1186 eweps = ewrt-ewitab;
1188 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1189 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1190 felec = qq11*rinv11*(rinvsq11-felec);
1193 d = (d>0.0) ? d : 0.0;
1195 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1197 dsw = d2*(swF2+d*(swF3+d*swF4));
1199 /* Evaluate switch function */
1200 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1201 felec = felec*sw - rinv11*velec*dsw;
1205 /* Calculate temporary vectorial force */
1210 /* Update vectorial force */
1214 f[j_coord_offset+DIM*1+XX] -= tx;
1215 f[j_coord_offset+DIM*1+YY] -= ty;
1216 f[j_coord_offset+DIM*1+ZZ] -= tz;
1220 /**************************
1221 * CALCULATE INTERACTIONS *
1222 **************************/
1229 /* EWALD ELECTROSTATICS */
1231 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1232 ewrt = r12*ewtabscale;
1234 eweps = ewrt-ewitab;
1236 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1237 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1238 felec = qq12*rinv12*(rinvsq12-felec);
1241 d = (d>0.0) ? d : 0.0;
1243 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1245 dsw = d2*(swF2+d*(swF3+d*swF4));
1247 /* Evaluate switch function */
1248 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1249 felec = felec*sw - rinv12*velec*dsw;
1253 /* Calculate temporary vectorial force */
1258 /* Update vectorial force */
1262 f[j_coord_offset+DIM*2+XX] -= tx;
1263 f[j_coord_offset+DIM*2+YY] -= ty;
1264 f[j_coord_offset+DIM*2+ZZ] -= tz;
1268 /**************************
1269 * CALCULATE INTERACTIONS *
1270 **************************/
1277 /* EWALD ELECTROSTATICS */
1279 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1280 ewrt = r13*ewtabscale;
1282 eweps = ewrt-ewitab;
1284 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1285 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1286 felec = qq13*rinv13*(rinvsq13-felec);
1289 d = (d>0.0) ? d : 0.0;
1291 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1293 dsw = d2*(swF2+d*(swF3+d*swF4));
1295 /* Evaluate switch function */
1296 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1297 felec = felec*sw - rinv13*velec*dsw;
1301 /* Calculate temporary vectorial force */
1306 /* Update vectorial force */
1310 f[j_coord_offset+DIM*3+XX] -= tx;
1311 f[j_coord_offset+DIM*3+YY] -= ty;
1312 f[j_coord_offset+DIM*3+ZZ] -= tz;
1316 /**************************
1317 * CALCULATE INTERACTIONS *
1318 **************************/
1325 /* EWALD ELECTROSTATICS */
1327 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1328 ewrt = r21*ewtabscale;
1330 eweps = ewrt-ewitab;
1332 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1333 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1334 felec = qq21*rinv21*(rinvsq21-felec);
1337 d = (d>0.0) ? d : 0.0;
1339 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1341 dsw = d2*(swF2+d*(swF3+d*swF4));
1343 /* Evaluate switch function */
1344 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1345 felec = felec*sw - rinv21*velec*dsw;
1349 /* Calculate temporary vectorial force */
1354 /* Update vectorial force */
1358 f[j_coord_offset+DIM*1+XX] -= tx;
1359 f[j_coord_offset+DIM*1+YY] -= ty;
1360 f[j_coord_offset+DIM*1+ZZ] -= tz;
1364 /**************************
1365 * CALCULATE INTERACTIONS *
1366 **************************/
1373 /* EWALD ELECTROSTATICS */
1375 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1376 ewrt = r22*ewtabscale;
1378 eweps = ewrt-ewitab;
1380 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1381 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1382 felec = qq22*rinv22*(rinvsq22-felec);
1385 d = (d>0.0) ? d : 0.0;
1387 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1389 dsw = d2*(swF2+d*(swF3+d*swF4));
1391 /* Evaluate switch function */
1392 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1393 felec = felec*sw - rinv22*velec*dsw;
1397 /* Calculate temporary vectorial force */
1402 /* Update vectorial force */
1406 f[j_coord_offset+DIM*2+XX] -= tx;
1407 f[j_coord_offset+DIM*2+YY] -= ty;
1408 f[j_coord_offset+DIM*2+ZZ] -= tz;
1412 /**************************
1413 * CALCULATE INTERACTIONS *
1414 **************************/
1421 /* EWALD ELECTROSTATICS */
1423 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1424 ewrt = r23*ewtabscale;
1426 eweps = ewrt-ewitab;
1428 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1429 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1430 felec = qq23*rinv23*(rinvsq23-felec);
1433 d = (d>0.0) ? d : 0.0;
1435 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1437 dsw = d2*(swF2+d*(swF3+d*swF4));
1439 /* Evaluate switch function */
1440 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1441 felec = felec*sw - rinv23*velec*dsw;
1445 /* Calculate temporary vectorial force */
1450 /* Update vectorial force */
1454 f[j_coord_offset+DIM*3+XX] -= tx;
1455 f[j_coord_offset+DIM*3+YY] -= ty;
1456 f[j_coord_offset+DIM*3+ZZ] -= tz;
1460 /**************************
1461 * CALCULATE INTERACTIONS *
1462 **************************/
1469 /* EWALD ELECTROSTATICS */
1471 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1472 ewrt = r31*ewtabscale;
1474 eweps = ewrt-ewitab;
1476 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1477 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1478 felec = qq31*rinv31*(rinvsq31-felec);
1481 d = (d>0.0) ? d : 0.0;
1483 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1485 dsw = d2*(swF2+d*(swF3+d*swF4));
1487 /* Evaluate switch function */
1488 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1489 felec = felec*sw - rinv31*velec*dsw;
1493 /* Calculate temporary vectorial force */
1498 /* Update vectorial force */
1502 f[j_coord_offset+DIM*1+XX] -= tx;
1503 f[j_coord_offset+DIM*1+YY] -= ty;
1504 f[j_coord_offset+DIM*1+ZZ] -= tz;
1508 /**************************
1509 * CALCULATE INTERACTIONS *
1510 **************************/
1517 /* EWALD ELECTROSTATICS */
1519 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1520 ewrt = r32*ewtabscale;
1522 eweps = ewrt-ewitab;
1524 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1525 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1526 felec = qq32*rinv32*(rinvsq32-felec);
1529 d = (d>0.0) ? d : 0.0;
1531 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1533 dsw = d2*(swF2+d*(swF3+d*swF4));
1535 /* Evaluate switch function */
1536 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1537 felec = felec*sw - rinv32*velec*dsw;
1541 /* Calculate temporary vectorial force */
1546 /* Update vectorial force */
1550 f[j_coord_offset+DIM*2+XX] -= tx;
1551 f[j_coord_offset+DIM*2+YY] -= ty;
1552 f[j_coord_offset+DIM*2+ZZ] -= tz;
1556 /**************************
1557 * CALCULATE INTERACTIONS *
1558 **************************/
1565 /* EWALD ELECTROSTATICS */
1567 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1568 ewrt = r33*ewtabscale;
1570 eweps = ewrt-ewitab;
1572 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1573 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1574 felec = qq33*rinv33*(rinvsq33-felec);
1577 d = (d>0.0) ? d : 0.0;
1579 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1581 dsw = d2*(swF2+d*(swF3+d*swF4));
1583 /* Evaluate switch function */
1584 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1585 felec = felec*sw - rinv33*velec*dsw;
1589 /* Calculate temporary vectorial force */
1594 /* Update vectorial force */
1598 f[j_coord_offset+DIM*3+XX] -= tx;
1599 f[j_coord_offset+DIM*3+YY] -= ty;
1600 f[j_coord_offset+DIM*3+ZZ] -= tz;
1604 /* Inner loop uses 581 flops */
1606 /* End of innermost loop */
1609 f[i_coord_offset+DIM*0+XX] += fix0;
1610 f[i_coord_offset+DIM*0+YY] += fiy0;
1611 f[i_coord_offset+DIM*0+ZZ] += fiz0;
1615 f[i_coord_offset+DIM*1+XX] += fix1;
1616 f[i_coord_offset+DIM*1+YY] += fiy1;
1617 f[i_coord_offset+DIM*1+ZZ] += fiz1;
1621 f[i_coord_offset+DIM*2+XX] += fix2;
1622 f[i_coord_offset+DIM*2+YY] += fiy2;
1623 f[i_coord_offset+DIM*2+ZZ] += fiz2;
1627 f[i_coord_offset+DIM*3+XX] += fix3;
1628 f[i_coord_offset+DIM*3+YY] += fiy3;
1629 f[i_coord_offset+DIM*3+ZZ] += fiz3;
1633 fshift[i_shift_offset+XX] += tx;
1634 fshift[i_shift_offset+YY] += ty;
1635 fshift[i_shift_offset+ZZ] += tz;
1637 /* Increment number of inner iterations */
1638 inneriter += j_index_end - j_index_start;
1640 /* Outer loop uses 39 flops */
1643 /* Increment number of outer iterations */
1646 /* Update outer/inner flops */
1648 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*39 + inneriter*581);