2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013, 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.
44 #include "../nb_kernel.h"
45 #include "types/simple.h"
50 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_VF_c
51 * Electrostatics interaction: Ewald
52 * VdW interaction: Buckingham
53 * Geometry: Water4-Water4
54 * Calculate force/pot: PotentialAndForce
57 nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_VF_c
58 (t_nblist * gmx_restrict nlist,
59 rvec * gmx_restrict xx,
60 rvec * gmx_restrict ff,
61 t_forcerec * gmx_restrict fr,
62 t_mdatoms * gmx_restrict mdatoms,
63 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
64 t_nrnb * gmx_restrict nrnb)
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;
73 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
75 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
77 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
79 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
81 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
83 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
85 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
87 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
88 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
89 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
90 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
91 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
92 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
93 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
94 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
95 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
96 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
97 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
98 real velec,felec,velecsum,facel,crf,krf,krf2;
101 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
105 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
107 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
114 jindex = nlist->jindex;
116 shiftidx = nlist->shift;
118 shiftvec = fr->shift_vec[0];
119 fshift = fr->fshift[0];
121 charge = mdatoms->chargeA;
122 nvdwtype = fr->ntype;
124 vdwtype = mdatoms->typeA;
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;
131 /* Setup water-specific parameters */
132 inr = nlist->iinr[0];
133 iq1 = facel*charge[inr+1];
134 iq2 = facel*charge[inr+2];
135 iq3 = facel*charge[inr+3];
136 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
141 vdwjidx0 = 3*vdwtype[inr+0];
142 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
143 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
144 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
155 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
156 rcutoff = fr->rcoulomb;
157 rcutoff2 = rcutoff*rcutoff;
159 rswitch = fr->rcoulomb_switch;
160 /* Setup switch parameters */
162 swV3 = -10.0/(d*d*d);
163 swV4 = 15.0/(d*d*d*d);
164 swV5 = -6.0/(d*d*d*d*d);
165 swF2 = -30.0/(d*d*d);
166 swF3 = 60.0/(d*d*d*d);
167 swF4 = -30.0/(d*d*d*d*d);
172 /* Start outer loop over neighborlists */
173 for(iidx=0; iidx<nri; iidx++)
175 /* Load shift vector for this list */
176 i_shift_offset = DIM*shiftidx[iidx];
177 shX = shiftvec[i_shift_offset+XX];
178 shY = shiftvec[i_shift_offset+YY];
179 shZ = shiftvec[i_shift_offset+ZZ];
181 /* Load limits for loop over neighbors */
182 j_index_start = jindex[iidx];
183 j_index_end = jindex[iidx+1];
185 /* Get outer coordinate index */
187 i_coord_offset = DIM*inr;
189 /* Load i particle coords and add shift vector */
190 ix0 = shX + x[i_coord_offset+DIM*0+XX];
191 iy0 = shY + x[i_coord_offset+DIM*0+YY];
192 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
193 ix1 = shX + x[i_coord_offset+DIM*1+XX];
194 iy1 = shY + x[i_coord_offset+DIM*1+YY];
195 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
196 ix2 = shX + x[i_coord_offset+DIM*2+XX];
197 iy2 = shY + x[i_coord_offset+DIM*2+YY];
198 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
199 ix3 = shX + x[i_coord_offset+DIM*3+XX];
200 iy3 = shY + x[i_coord_offset+DIM*3+YY];
201 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
216 /* Reset potential sums */
220 /* Start inner kernel loop */
221 for(jidx=j_index_start; jidx<j_index_end; jidx++)
223 /* Get j neighbor index, and coordinate index */
225 j_coord_offset = DIM*jnr;
227 /* load j atom coordinates */
228 jx0 = x[j_coord_offset+DIM*0+XX];
229 jy0 = x[j_coord_offset+DIM*0+YY];
230 jz0 = x[j_coord_offset+DIM*0+ZZ];
231 jx1 = x[j_coord_offset+DIM*1+XX];
232 jy1 = x[j_coord_offset+DIM*1+YY];
233 jz1 = x[j_coord_offset+DIM*1+ZZ];
234 jx2 = x[j_coord_offset+DIM*2+XX];
235 jy2 = x[j_coord_offset+DIM*2+YY];
236 jz2 = x[j_coord_offset+DIM*2+ZZ];
237 jx3 = x[j_coord_offset+DIM*3+XX];
238 jy3 = x[j_coord_offset+DIM*3+YY];
239 jz3 = x[j_coord_offset+DIM*3+ZZ];
241 /* Calculate displacement vector */
273 /* Calculate squared distance and things based on it */
274 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
275 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
276 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
277 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
278 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
279 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
280 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
281 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
282 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
283 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
285 rinv00 = gmx_invsqrt(rsq00);
286 rinv11 = gmx_invsqrt(rsq11);
287 rinv12 = gmx_invsqrt(rsq12);
288 rinv13 = gmx_invsqrt(rsq13);
289 rinv21 = gmx_invsqrt(rsq21);
290 rinv22 = gmx_invsqrt(rsq22);
291 rinv23 = gmx_invsqrt(rsq23);
292 rinv31 = gmx_invsqrt(rsq31);
293 rinv32 = gmx_invsqrt(rsq32);
294 rinv33 = gmx_invsqrt(rsq33);
296 rinvsq00 = rinv00*rinv00;
297 rinvsq11 = rinv11*rinv11;
298 rinvsq12 = rinv12*rinv12;
299 rinvsq13 = rinv13*rinv13;
300 rinvsq21 = rinv21*rinv21;
301 rinvsq22 = rinv22*rinv22;
302 rinvsq23 = rinv23*rinv23;
303 rinvsq31 = rinv31*rinv31;
304 rinvsq32 = rinv32*rinv32;
305 rinvsq33 = rinv33*rinv33;
307 /**************************
308 * CALCULATE INTERACTIONS *
309 **************************/
316 /* BUCKINGHAM DISPERSION/REPULSION */
317 rinvsix = rinvsq00*rinvsq00*rinvsq00;
318 vvdw6 = c6_00*rinvsix;
320 vvdwexp = cexp1_00*exp(-br);
321 vvdw = vvdwexp - vvdw6*(1.0/6.0);
322 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
325 d = (d>0.0) ? d : 0.0;
327 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
329 dsw = d2*(swF2+d*(swF3+d*swF4));
331 /* Evaluate switch function */
332 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
333 fvdw = fvdw*sw - rinv00*vvdw*dsw;
336 /* Update potential sums from outer loop */
341 /* Calculate temporary vectorial force */
346 /* Update vectorial force */
350 f[j_coord_offset+DIM*0+XX] -= tx;
351 f[j_coord_offset+DIM*0+YY] -= ty;
352 f[j_coord_offset+DIM*0+ZZ] -= tz;
356 /**************************
357 * CALCULATE INTERACTIONS *
358 **************************/
365 /* EWALD ELECTROSTATICS */
367 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
368 ewrt = r11*ewtabscale;
372 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
373 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
374 felec = qq11*rinv11*(rinvsq11-felec);
377 d = (d>0.0) ? d : 0.0;
379 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
381 dsw = d2*(swF2+d*(swF3+d*swF4));
383 /* Evaluate switch function */
384 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
385 felec = felec*sw - rinv11*velec*dsw;
388 /* Update potential sums from outer loop */
393 /* Calculate temporary vectorial force */
398 /* Update vectorial force */
402 f[j_coord_offset+DIM*1+XX] -= tx;
403 f[j_coord_offset+DIM*1+YY] -= ty;
404 f[j_coord_offset+DIM*1+ZZ] -= tz;
408 /**************************
409 * CALCULATE INTERACTIONS *
410 **************************/
417 /* EWALD ELECTROSTATICS */
419 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
420 ewrt = r12*ewtabscale;
424 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
425 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
426 felec = qq12*rinv12*(rinvsq12-felec);
429 d = (d>0.0) ? d : 0.0;
431 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
433 dsw = d2*(swF2+d*(swF3+d*swF4));
435 /* Evaluate switch function */
436 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
437 felec = felec*sw - rinv12*velec*dsw;
440 /* Update potential sums from outer loop */
445 /* Calculate temporary vectorial force */
450 /* Update vectorial force */
454 f[j_coord_offset+DIM*2+XX] -= tx;
455 f[j_coord_offset+DIM*2+YY] -= ty;
456 f[j_coord_offset+DIM*2+ZZ] -= tz;
460 /**************************
461 * CALCULATE INTERACTIONS *
462 **************************/
469 /* EWALD ELECTROSTATICS */
471 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
472 ewrt = r13*ewtabscale;
476 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
477 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
478 felec = qq13*rinv13*(rinvsq13-felec);
481 d = (d>0.0) ? d : 0.0;
483 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
485 dsw = d2*(swF2+d*(swF3+d*swF4));
487 /* Evaluate switch function */
488 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
489 felec = felec*sw - rinv13*velec*dsw;
492 /* Update potential sums from outer loop */
497 /* Calculate temporary vectorial force */
502 /* Update vectorial force */
506 f[j_coord_offset+DIM*3+XX] -= tx;
507 f[j_coord_offset+DIM*3+YY] -= ty;
508 f[j_coord_offset+DIM*3+ZZ] -= tz;
512 /**************************
513 * CALCULATE INTERACTIONS *
514 **************************/
521 /* EWALD ELECTROSTATICS */
523 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
524 ewrt = r21*ewtabscale;
528 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
529 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
530 felec = qq21*rinv21*(rinvsq21-felec);
533 d = (d>0.0) ? d : 0.0;
535 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
537 dsw = d2*(swF2+d*(swF3+d*swF4));
539 /* Evaluate switch function */
540 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
541 felec = felec*sw - rinv21*velec*dsw;
544 /* Update potential sums from outer loop */
549 /* Calculate temporary vectorial force */
554 /* Update vectorial force */
558 f[j_coord_offset+DIM*1+XX] -= tx;
559 f[j_coord_offset+DIM*1+YY] -= ty;
560 f[j_coord_offset+DIM*1+ZZ] -= tz;
564 /**************************
565 * CALCULATE INTERACTIONS *
566 **************************/
573 /* EWALD ELECTROSTATICS */
575 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
576 ewrt = r22*ewtabscale;
580 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
581 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
582 felec = qq22*rinv22*(rinvsq22-felec);
585 d = (d>0.0) ? d : 0.0;
587 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
589 dsw = d2*(swF2+d*(swF3+d*swF4));
591 /* Evaluate switch function */
592 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
593 felec = felec*sw - rinv22*velec*dsw;
596 /* Update potential sums from outer loop */
601 /* Calculate temporary vectorial force */
606 /* Update vectorial force */
610 f[j_coord_offset+DIM*2+XX] -= tx;
611 f[j_coord_offset+DIM*2+YY] -= ty;
612 f[j_coord_offset+DIM*2+ZZ] -= tz;
616 /**************************
617 * CALCULATE INTERACTIONS *
618 **************************/
625 /* EWALD ELECTROSTATICS */
627 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
628 ewrt = r23*ewtabscale;
632 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
633 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
634 felec = qq23*rinv23*(rinvsq23-felec);
637 d = (d>0.0) ? d : 0.0;
639 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
641 dsw = d2*(swF2+d*(swF3+d*swF4));
643 /* Evaluate switch function */
644 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
645 felec = felec*sw - rinv23*velec*dsw;
648 /* Update potential sums from outer loop */
653 /* Calculate temporary vectorial force */
658 /* Update vectorial force */
662 f[j_coord_offset+DIM*3+XX] -= tx;
663 f[j_coord_offset+DIM*3+YY] -= ty;
664 f[j_coord_offset+DIM*3+ZZ] -= tz;
668 /**************************
669 * CALCULATE INTERACTIONS *
670 **************************/
677 /* EWALD ELECTROSTATICS */
679 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
680 ewrt = r31*ewtabscale;
684 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
685 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
686 felec = qq31*rinv31*(rinvsq31-felec);
689 d = (d>0.0) ? d : 0.0;
691 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
693 dsw = d2*(swF2+d*(swF3+d*swF4));
695 /* Evaluate switch function */
696 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
697 felec = felec*sw - rinv31*velec*dsw;
700 /* Update potential sums from outer loop */
705 /* Calculate temporary vectorial force */
710 /* Update vectorial force */
714 f[j_coord_offset+DIM*1+XX] -= tx;
715 f[j_coord_offset+DIM*1+YY] -= ty;
716 f[j_coord_offset+DIM*1+ZZ] -= tz;
720 /**************************
721 * CALCULATE INTERACTIONS *
722 **************************/
729 /* EWALD ELECTROSTATICS */
731 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
732 ewrt = r32*ewtabscale;
736 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
737 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
738 felec = qq32*rinv32*(rinvsq32-felec);
741 d = (d>0.0) ? d : 0.0;
743 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
745 dsw = d2*(swF2+d*(swF3+d*swF4));
747 /* Evaluate switch function */
748 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
749 felec = felec*sw - rinv32*velec*dsw;
752 /* Update potential sums from outer loop */
757 /* Calculate temporary vectorial force */
762 /* Update vectorial force */
766 f[j_coord_offset+DIM*2+XX] -= tx;
767 f[j_coord_offset+DIM*2+YY] -= ty;
768 f[j_coord_offset+DIM*2+ZZ] -= tz;
772 /**************************
773 * CALCULATE INTERACTIONS *
774 **************************/
781 /* EWALD ELECTROSTATICS */
783 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
784 ewrt = r33*ewtabscale;
788 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
789 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
790 felec = qq33*rinv33*(rinvsq33-felec);
793 d = (d>0.0) ? d : 0.0;
795 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
797 dsw = d2*(swF2+d*(swF3+d*swF4));
799 /* Evaluate switch function */
800 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
801 felec = felec*sw - rinv33*velec*dsw;
804 /* Update potential sums from outer loop */
809 /* Calculate temporary vectorial force */
814 /* Update vectorial force */
818 f[j_coord_offset+DIM*3+XX] -= tx;
819 f[j_coord_offset+DIM*3+YY] -= ty;
820 f[j_coord_offset+DIM*3+ZZ] -= tz;
824 /* Inner loop uses 601 flops */
826 /* End of innermost loop */
829 f[i_coord_offset+DIM*0+XX] += fix0;
830 f[i_coord_offset+DIM*0+YY] += fiy0;
831 f[i_coord_offset+DIM*0+ZZ] += fiz0;
835 f[i_coord_offset+DIM*1+XX] += fix1;
836 f[i_coord_offset+DIM*1+YY] += fiy1;
837 f[i_coord_offset+DIM*1+ZZ] += fiz1;
841 f[i_coord_offset+DIM*2+XX] += fix2;
842 f[i_coord_offset+DIM*2+YY] += fiy2;
843 f[i_coord_offset+DIM*2+ZZ] += fiz2;
847 f[i_coord_offset+DIM*3+XX] += fix3;
848 f[i_coord_offset+DIM*3+YY] += fiy3;
849 f[i_coord_offset+DIM*3+ZZ] += fiz3;
853 fshift[i_shift_offset+XX] += tx;
854 fshift[i_shift_offset+YY] += ty;
855 fshift[i_shift_offset+ZZ] += tz;
858 /* Update potential energies */
859 kernel_data->energygrp_elec[ggid] += velecsum;
860 kernel_data->energygrp_vdw[ggid] += vvdwsum;
862 /* Increment number of inner iterations */
863 inneriter += j_index_end - j_index_start;
865 /* Outer loop uses 41 flops */
868 /* Increment number of outer iterations */
871 /* Update outer/inner flops */
873 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*41 + inneriter*601);
876 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_F_c
877 * Electrostatics interaction: Ewald
878 * VdW interaction: Buckingham
879 * Geometry: Water4-Water4
880 * Calculate force/pot: Force
883 nb_kernel_ElecEwSw_VdwBhamSw_GeomW4W4_F_c
884 (t_nblist * gmx_restrict nlist,
885 rvec * gmx_restrict xx,
886 rvec * gmx_restrict ff,
887 t_forcerec * gmx_restrict fr,
888 t_mdatoms * gmx_restrict mdatoms,
889 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
890 t_nrnb * gmx_restrict nrnb)
892 int i_shift_offset,i_coord_offset,j_coord_offset;
893 int j_index_start,j_index_end;
894 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
895 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
896 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
897 real *shiftvec,*fshift,*x,*f;
899 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
901 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
903 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
905 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
907 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
909 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
911 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
913 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
914 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
915 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
916 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
917 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
918 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
919 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
920 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
921 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
922 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
923 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
924 real velec,felec,velecsum,facel,crf,krf,krf2;
927 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
931 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
933 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
940 jindex = nlist->jindex;
942 shiftidx = nlist->shift;
944 shiftvec = fr->shift_vec[0];
945 fshift = fr->fshift[0];
947 charge = mdatoms->chargeA;
948 nvdwtype = fr->ntype;
950 vdwtype = mdatoms->typeA;
952 sh_ewald = fr->ic->sh_ewald;
953 ewtab = fr->ic->tabq_coul_FDV0;
954 ewtabscale = fr->ic->tabq_scale;
955 ewtabhalfspace = 0.5/ewtabscale;
957 /* Setup water-specific parameters */
958 inr = nlist->iinr[0];
959 iq1 = facel*charge[inr+1];
960 iq2 = facel*charge[inr+2];
961 iq3 = facel*charge[inr+3];
962 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
967 vdwjidx0 = 3*vdwtype[inr+0];
968 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
969 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
970 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
981 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
982 rcutoff = fr->rcoulomb;
983 rcutoff2 = rcutoff*rcutoff;
985 rswitch = fr->rcoulomb_switch;
986 /* Setup switch parameters */
988 swV3 = -10.0/(d*d*d);
989 swV4 = 15.0/(d*d*d*d);
990 swV5 = -6.0/(d*d*d*d*d);
991 swF2 = -30.0/(d*d*d);
992 swF3 = 60.0/(d*d*d*d);
993 swF4 = -30.0/(d*d*d*d*d);
998 /* Start outer loop over neighborlists */
999 for(iidx=0; iidx<nri; iidx++)
1001 /* Load shift vector for this list */
1002 i_shift_offset = DIM*shiftidx[iidx];
1003 shX = shiftvec[i_shift_offset+XX];
1004 shY = shiftvec[i_shift_offset+YY];
1005 shZ = shiftvec[i_shift_offset+ZZ];
1007 /* Load limits for loop over neighbors */
1008 j_index_start = jindex[iidx];
1009 j_index_end = jindex[iidx+1];
1011 /* Get outer coordinate index */
1013 i_coord_offset = DIM*inr;
1015 /* Load i particle coords and add shift vector */
1016 ix0 = shX + x[i_coord_offset+DIM*0+XX];
1017 iy0 = shY + x[i_coord_offset+DIM*0+YY];
1018 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
1019 ix1 = shX + x[i_coord_offset+DIM*1+XX];
1020 iy1 = shY + x[i_coord_offset+DIM*1+YY];
1021 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
1022 ix2 = shX + x[i_coord_offset+DIM*2+XX];
1023 iy2 = shY + x[i_coord_offset+DIM*2+YY];
1024 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
1025 ix3 = shX + x[i_coord_offset+DIM*3+XX];
1026 iy3 = shY + x[i_coord_offset+DIM*3+YY];
1027 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
1042 /* Start inner kernel loop */
1043 for(jidx=j_index_start; jidx<j_index_end; jidx++)
1045 /* Get j neighbor index, and coordinate index */
1047 j_coord_offset = DIM*jnr;
1049 /* load j atom coordinates */
1050 jx0 = x[j_coord_offset+DIM*0+XX];
1051 jy0 = x[j_coord_offset+DIM*0+YY];
1052 jz0 = x[j_coord_offset+DIM*0+ZZ];
1053 jx1 = x[j_coord_offset+DIM*1+XX];
1054 jy1 = x[j_coord_offset+DIM*1+YY];
1055 jz1 = x[j_coord_offset+DIM*1+ZZ];
1056 jx2 = x[j_coord_offset+DIM*2+XX];
1057 jy2 = x[j_coord_offset+DIM*2+YY];
1058 jz2 = x[j_coord_offset+DIM*2+ZZ];
1059 jx3 = x[j_coord_offset+DIM*3+XX];
1060 jy3 = x[j_coord_offset+DIM*3+YY];
1061 jz3 = x[j_coord_offset+DIM*3+ZZ];
1063 /* Calculate displacement vector */
1095 /* Calculate squared distance and things based on it */
1096 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
1097 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
1098 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
1099 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
1100 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
1101 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
1102 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
1103 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
1104 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
1105 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
1107 rinv00 = gmx_invsqrt(rsq00);
1108 rinv11 = gmx_invsqrt(rsq11);
1109 rinv12 = gmx_invsqrt(rsq12);
1110 rinv13 = gmx_invsqrt(rsq13);
1111 rinv21 = gmx_invsqrt(rsq21);
1112 rinv22 = gmx_invsqrt(rsq22);
1113 rinv23 = gmx_invsqrt(rsq23);
1114 rinv31 = gmx_invsqrt(rsq31);
1115 rinv32 = gmx_invsqrt(rsq32);
1116 rinv33 = gmx_invsqrt(rsq33);
1118 rinvsq00 = rinv00*rinv00;
1119 rinvsq11 = rinv11*rinv11;
1120 rinvsq12 = rinv12*rinv12;
1121 rinvsq13 = rinv13*rinv13;
1122 rinvsq21 = rinv21*rinv21;
1123 rinvsq22 = rinv22*rinv22;
1124 rinvsq23 = rinv23*rinv23;
1125 rinvsq31 = rinv31*rinv31;
1126 rinvsq32 = rinv32*rinv32;
1127 rinvsq33 = rinv33*rinv33;
1129 /**************************
1130 * CALCULATE INTERACTIONS *
1131 **************************/
1138 /* BUCKINGHAM DISPERSION/REPULSION */
1139 rinvsix = rinvsq00*rinvsq00*rinvsq00;
1140 vvdw6 = c6_00*rinvsix;
1142 vvdwexp = cexp1_00*exp(-br);
1143 vvdw = vvdwexp - vvdw6*(1.0/6.0);
1144 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
1147 d = (d>0.0) ? d : 0.0;
1149 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1151 dsw = d2*(swF2+d*(swF3+d*swF4));
1153 /* Evaluate switch function */
1154 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1155 fvdw = fvdw*sw - rinv00*vvdw*dsw;
1159 /* Calculate temporary vectorial force */
1164 /* Update vectorial force */
1168 f[j_coord_offset+DIM*0+XX] -= tx;
1169 f[j_coord_offset+DIM*0+YY] -= ty;
1170 f[j_coord_offset+DIM*0+ZZ] -= tz;
1174 /**************************
1175 * CALCULATE INTERACTIONS *
1176 **************************/
1183 /* EWALD ELECTROSTATICS */
1185 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1186 ewrt = r11*ewtabscale;
1188 eweps = ewrt-ewitab;
1190 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1191 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1192 felec = qq11*rinv11*(rinvsq11-felec);
1195 d = (d>0.0) ? d : 0.0;
1197 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1199 dsw = d2*(swF2+d*(swF3+d*swF4));
1201 /* Evaluate switch function */
1202 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1203 felec = felec*sw - rinv11*velec*dsw;
1207 /* Calculate temporary vectorial force */
1212 /* Update vectorial force */
1216 f[j_coord_offset+DIM*1+XX] -= tx;
1217 f[j_coord_offset+DIM*1+YY] -= ty;
1218 f[j_coord_offset+DIM*1+ZZ] -= tz;
1222 /**************************
1223 * CALCULATE INTERACTIONS *
1224 **************************/
1231 /* EWALD ELECTROSTATICS */
1233 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1234 ewrt = r12*ewtabscale;
1236 eweps = ewrt-ewitab;
1238 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1239 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1240 felec = qq12*rinv12*(rinvsq12-felec);
1243 d = (d>0.0) ? d : 0.0;
1245 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1247 dsw = d2*(swF2+d*(swF3+d*swF4));
1249 /* Evaluate switch function */
1250 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1251 felec = felec*sw - rinv12*velec*dsw;
1255 /* Calculate temporary vectorial force */
1260 /* Update vectorial force */
1264 f[j_coord_offset+DIM*2+XX] -= tx;
1265 f[j_coord_offset+DIM*2+YY] -= ty;
1266 f[j_coord_offset+DIM*2+ZZ] -= tz;
1270 /**************************
1271 * CALCULATE INTERACTIONS *
1272 **************************/
1279 /* EWALD ELECTROSTATICS */
1281 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1282 ewrt = r13*ewtabscale;
1284 eweps = ewrt-ewitab;
1286 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1287 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1288 felec = qq13*rinv13*(rinvsq13-felec);
1291 d = (d>0.0) ? d : 0.0;
1293 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1295 dsw = d2*(swF2+d*(swF3+d*swF4));
1297 /* Evaluate switch function */
1298 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1299 felec = felec*sw - rinv13*velec*dsw;
1303 /* Calculate temporary vectorial force */
1308 /* Update vectorial force */
1312 f[j_coord_offset+DIM*3+XX] -= tx;
1313 f[j_coord_offset+DIM*3+YY] -= ty;
1314 f[j_coord_offset+DIM*3+ZZ] -= tz;
1318 /**************************
1319 * CALCULATE INTERACTIONS *
1320 **************************/
1327 /* EWALD ELECTROSTATICS */
1329 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1330 ewrt = r21*ewtabscale;
1332 eweps = ewrt-ewitab;
1334 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1335 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1336 felec = qq21*rinv21*(rinvsq21-felec);
1339 d = (d>0.0) ? d : 0.0;
1341 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1343 dsw = d2*(swF2+d*(swF3+d*swF4));
1345 /* Evaluate switch function */
1346 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1347 felec = felec*sw - rinv21*velec*dsw;
1351 /* Calculate temporary vectorial force */
1356 /* Update vectorial force */
1360 f[j_coord_offset+DIM*1+XX] -= tx;
1361 f[j_coord_offset+DIM*1+YY] -= ty;
1362 f[j_coord_offset+DIM*1+ZZ] -= tz;
1366 /**************************
1367 * CALCULATE INTERACTIONS *
1368 **************************/
1375 /* EWALD ELECTROSTATICS */
1377 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1378 ewrt = r22*ewtabscale;
1380 eweps = ewrt-ewitab;
1382 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1383 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1384 felec = qq22*rinv22*(rinvsq22-felec);
1387 d = (d>0.0) ? d : 0.0;
1389 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1391 dsw = d2*(swF2+d*(swF3+d*swF4));
1393 /* Evaluate switch function */
1394 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1395 felec = felec*sw - rinv22*velec*dsw;
1399 /* Calculate temporary vectorial force */
1404 /* Update vectorial force */
1408 f[j_coord_offset+DIM*2+XX] -= tx;
1409 f[j_coord_offset+DIM*2+YY] -= ty;
1410 f[j_coord_offset+DIM*2+ZZ] -= tz;
1414 /**************************
1415 * CALCULATE INTERACTIONS *
1416 **************************/
1423 /* EWALD ELECTROSTATICS */
1425 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1426 ewrt = r23*ewtabscale;
1428 eweps = ewrt-ewitab;
1430 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1431 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1432 felec = qq23*rinv23*(rinvsq23-felec);
1435 d = (d>0.0) ? d : 0.0;
1437 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1439 dsw = d2*(swF2+d*(swF3+d*swF4));
1441 /* Evaluate switch function */
1442 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1443 felec = felec*sw - rinv23*velec*dsw;
1447 /* Calculate temporary vectorial force */
1452 /* Update vectorial force */
1456 f[j_coord_offset+DIM*3+XX] -= tx;
1457 f[j_coord_offset+DIM*3+YY] -= ty;
1458 f[j_coord_offset+DIM*3+ZZ] -= tz;
1462 /**************************
1463 * CALCULATE INTERACTIONS *
1464 **************************/
1471 /* EWALD ELECTROSTATICS */
1473 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1474 ewrt = r31*ewtabscale;
1476 eweps = ewrt-ewitab;
1478 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1479 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1480 felec = qq31*rinv31*(rinvsq31-felec);
1483 d = (d>0.0) ? d : 0.0;
1485 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1487 dsw = d2*(swF2+d*(swF3+d*swF4));
1489 /* Evaluate switch function */
1490 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1491 felec = felec*sw - rinv31*velec*dsw;
1495 /* Calculate temporary vectorial force */
1500 /* Update vectorial force */
1504 f[j_coord_offset+DIM*1+XX] -= tx;
1505 f[j_coord_offset+DIM*1+YY] -= ty;
1506 f[j_coord_offset+DIM*1+ZZ] -= tz;
1510 /**************************
1511 * CALCULATE INTERACTIONS *
1512 **************************/
1519 /* EWALD ELECTROSTATICS */
1521 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1522 ewrt = r32*ewtabscale;
1524 eweps = ewrt-ewitab;
1526 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1527 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1528 felec = qq32*rinv32*(rinvsq32-felec);
1531 d = (d>0.0) ? d : 0.0;
1533 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1535 dsw = d2*(swF2+d*(swF3+d*swF4));
1537 /* Evaluate switch function */
1538 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1539 felec = felec*sw - rinv32*velec*dsw;
1543 /* Calculate temporary vectorial force */
1548 /* Update vectorial force */
1552 f[j_coord_offset+DIM*2+XX] -= tx;
1553 f[j_coord_offset+DIM*2+YY] -= ty;
1554 f[j_coord_offset+DIM*2+ZZ] -= tz;
1558 /**************************
1559 * CALCULATE INTERACTIONS *
1560 **************************/
1567 /* EWALD ELECTROSTATICS */
1569 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1570 ewrt = r33*ewtabscale;
1572 eweps = ewrt-ewitab;
1574 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1575 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1576 felec = qq33*rinv33*(rinvsq33-felec);
1579 d = (d>0.0) ? d : 0.0;
1581 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1583 dsw = d2*(swF2+d*(swF3+d*swF4));
1585 /* Evaluate switch function */
1586 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1587 felec = felec*sw - rinv33*velec*dsw;
1591 /* Calculate temporary vectorial force */
1596 /* Update vectorial force */
1600 f[j_coord_offset+DIM*3+XX] -= tx;
1601 f[j_coord_offset+DIM*3+YY] -= ty;
1602 f[j_coord_offset+DIM*3+ZZ] -= tz;
1606 /* Inner loop uses 581 flops */
1608 /* End of innermost loop */
1611 f[i_coord_offset+DIM*0+XX] += fix0;
1612 f[i_coord_offset+DIM*0+YY] += fiy0;
1613 f[i_coord_offset+DIM*0+ZZ] += fiz0;
1617 f[i_coord_offset+DIM*1+XX] += fix1;
1618 f[i_coord_offset+DIM*1+YY] += fiy1;
1619 f[i_coord_offset+DIM*1+ZZ] += fiz1;
1623 f[i_coord_offset+DIM*2+XX] += fix2;
1624 f[i_coord_offset+DIM*2+YY] += fiy2;
1625 f[i_coord_offset+DIM*2+ZZ] += fiz2;
1629 f[i_coord_offset+DIM*3+XX] += fix3;
1630 f[i_coord_offset+DIM*3+YY] += fiy3;
1631 f[i_coord_offset+DIM*3+ZZ] += fiz3;
1635 fshift[i_shift_offset+XX] += tx;
1636 fshift[i_shift_offset+YY] += ty;
1637 fshift[i_shift_offset+ZZ] += tz;
1639 /* Increment number of inner iterations */
1640 inneriter += j_index_end - j_index_start;
1642 /* Outer loop uses 39 flops */
1645 /* Increment number of outer iterations */
1648 /* Update outer/inner flops */
1650 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*39 + inneriter*581);