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 "types/simple.h"
44 #include "gromacs/math/vec.h"
48 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW4W4_VF_c
49 * Electrostatics interaction: Ewald
50 * VdW interaction: None
51 * Geometry: Water4-Water4
52 * Calculate force/pot: PotentialAndForce
55 nb_kernel_ElecEwSw_VdwNone_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 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
73 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
75 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
77 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
79 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
81 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
82 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
83 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
84 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
85 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
86 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
87 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
88 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
89 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
90 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
91 real velec,felec,velecsum,facel,crf,krf,krf2;
94 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
96 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
103 jindex = nlist->jindex;
105 shiftidx = nlist->shift;
107 shiftvec = fr->shift_vec[0];
108 fshift = fr->fshift[0];
110 charge = mdatoms->chargeA;
112 sh_ewald = fr->ic->sh_ewald;
113 ewtab = fr->ic->tabq_coul_FDV0;
114 ewtabscale = fr->ic->tabq_scale;
115 ewtabhalfspace = 0.5/ewtabscale;
117 /* Setup water-specific parameters */
118 inr = nlist->iinr[0];
119 iq1 = facel*charge[inr+1];
120 iq2 = facel*charge[inr+2];
121 iq3 = facel*charge[inr+3];
136 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
137 rcutoff = fr->rcoulomb;
138 rcutoff2 = rcutoff*rcutoff;
140 rswitch = fr->rcoulomb_switch;
141 /* Setup switch parameters */
143 swV3 = -10.0/(d*d*d);
144 swV4 = 15.0/(d*d*d*d);
145 swV5 = -6.0/(d*d*d*d*d);
146 swF2 = -30.0/(d*d*d);
147 swF3 = 60.0/(d*d*d*d);
148 swF4 = -30.0/(d*d*d*d*d);
153 /* Start outer loop over neighborlists */
154 for(iidx=0; iidx<nri; iidx++)
156 /* Load shift vector for this list */
157 i_shift_offset = DIM*shiftidx[iidx];
158 shX = shiftvec[i_shift_offset+XX];
159 shY = shiftvec[i_shift_offset+YY];
160 shZ = shiftvec[i_shift_offset+ZZ];
162 /* Load limits for loop over neighbors */
163 j_index_start = jindex[iidx];
164 j_index_end = jindex[iidx+1];
166 /* Get outer coordinate index */
168 i_coord_offset = DIM*inr;
170 /* Load i particle coords and add shift vector */
171 ix1 = shX + x[i_coord_offset+DIM*1+XX];
172 iy1 = shY + x[i_coord_offset+DIM*1+YY];
173 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
174 ix2 = shX + x[i_coord_offset+DIM*2+XX];
175 iy2 = shY + x[i_coord_offset+DIM*2+YY];
176 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
177 ix3 = shX + x[i_coord_offset+DIM*3+XX];
178 iy3 = shY + x[i_coord_offset+DIM*3+YY];
179 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
191 /* Reset potential sums */
194 /* Start inner kernel loop */
195 for(jidx=j_index_start; jidx<j_index_end; jidx++)
197 /* Get j neighbor index, and coordinate index */
199 j_coord_offset = DIM*jnr;
201 /* load j atom coordinates */
202 jx1 = x[j_coord_offset+DIM*1+XX];
203 jy1 = x[j_coord_offset+DIM*1+YY];
204 jz1 = x[j_coord_offset+DIM*1+ZZ];
205 jx2 = x[j_coord_offset+DIM*2+XX];
206 jy2 = x[j_coord_offset+DIM*2+YY];
207 jz2 = x[j_coord_offset+DIM*2+ZZ];
208 jx3 = x[j_coord_offset+DIM*3+XX];
209 jy3 = x[j_coord_offset+DIM*3+YY];
210 jz3 = x[j_coord_offset+DIM*3+ZZ];
212 /* Calculate displacement vector */
241 /* Calculate squared distance and things based on it */
242 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
243 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
244 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
245 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
246 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
247 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
248 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
249 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
250 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
252 rinv11 = gmx_invsqrt(rsq11);
253 rinv12 = gmx_invsqrt(rsq12);
254 rinv13 = gmx_invsqrt(rsq13);
255 rinv21 = gmx_invsqrt(rsq21);
256 rinv22 = gmx_invsqrt(rsq22);
257 rinv23 = gmx_invsqrt(rsq23);
258 rinv31 = gmx_invsqrt(rsq31);
259 rinv32 = gmx_invsqrt(rsq32);
260 rinv33 = gmx_invsqrt(rsq33);
262 rinvsq11 = rinv11*rinv11;
263 rinvsq12 = rinv12*rinv12;
264 rinvsq13 = rinv13*rinv13;
265 rinvsq21 = rinv21*rinv21;
266 rinvsq22 = rinv22*rinv22;
267 rinvsq23 = rinv23*rinv23;
268 rinvsq31 = rinv31*rinv31;
269 rinvsq32 = rinv32*rinv32;
270 rinvsq33 = rinv33*rinv33;
272 /**************************
273 * CALCULATE INTERACTIONS *
274 **************************/
281 /* EWALD ELECTROSTATICS */
283 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
284 ewrt = r11*ewtabscale;
288 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
289 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
290 felec = qq11*rinv11*(rinvsq11-felec);
293 d = (d>0.0) ? d : 0.0;
295 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
297 dsw = d2*(swF2+d*(swF3+d*swF4));
299 /* Evaluate switch function */
300 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
301 felec = felec*sw - rinv11*velec*dsw;
304 /* Update potential sums from outer loop */
309 /* Calculate temporary vectorial force */
314 /* Update vectorial force */
318 f[j_coord_offset+DIM*1+XX] -= tx;
319 f[j_coord_offset+DIM*1+YY] -= ty;
320 f[j_coord_offset+DIM*1+ZZ] -= tz;
324 /**************************
325 * CALCULATE INTERACTIONS *
326 **************************/
333 /* EWALD ELECTROSTATICS */
335 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
336 ewrt = r12*ewtabscale;
340 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
341 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
342 felec = qq12*rinv12*(rinvsq12-felec);
345 d = (d>0.0) ? d : 0.0;
347 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
349 dsw = d2*(swF2+d*(swF3+d*swF4));
351 /* Evaluate switch function */
352 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
353 felec = felec*sw - rinv12*velec*dsw;
356 /* Update potential sums from outer loop */
361 /* Calculate temporary vectorial force */
366 /* Update vectorial force */
370 f[j_coord_offset+DIM*2+XX] -= tx;
371 f[j_coord_offset+DIM*2+YY] -= ty;
372 f[j_coord_offset+DIM*2+ZZ] -= tz;
376 /**************************
377 * CALCULATE INTERACTIONS *
378 **************************/
385 /* EWALD ELECTROSTATICS */
387 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
388 ewrt = r13*ewtabscale;
392 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
393 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
394 felec = qq13*rinv13*(rinvsq13-felec);
397 d = (d>0.0) ? d : 0.0;
399 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
401 dsw = d2*(swF2+d*(swF3+d*swF4));
403 /* Evaluate switch function */
404 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
405 felec = felec*sw - rinv13*velec*dsw;
408 /* Update potential sums from outer loop */
413 /* Calculate temporary vectorial force */
418 /* Update vectorial force */
422 f[j_coord_offset+DIM*3+XX] -= tx;
423 f[j_coord_offset+DIM*3+YY] -= ty;
424 f[j_coord_offset+DIM*3+ZZ] -= tz;
428 /**************************
429 * CALCULATE INTERACTIONS *
430 **************************/
437 /* EWALD ELECTROSTATICS */
439 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
440 ewrt = r21*ewtabscale;
444 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
445 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
446 felec = qq21*rinv21*(rinvsq21-felec);
449 d = (d>0.0) ? d : 0.0;
451 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
453 dsw = d2*(swF2+d*(swF3+d*swF4));
455 /* Evaluate switch function */
456 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
457 felec = felec*sw - rinv21*velec*dsw;
460 /* Update potential sums from outer loop */
465 /* Calculate temporary vectorial force */
470 /* Update vectorial force */
474 f[j_coord_offset+DIM*1+XX] -= tx;
475 f[j_coord_offset+DIM*1+YY] -= ty;
476 f[j_coord_offset+DIM*1+ZZ] -= tz;
480 /**************************
481 * CALCULATE INTERACTIONS *
482 **************************/
489 /* EWALD ELECTROSTATICS */
491 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
492 ewrt = r22*ewtabscale;
496 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
497 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
498 felec = qq22*rinv22*(rinvsq22-felec);
501 d = (d>0.0) ? d : 0.0;
503 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
505 dsw = d2*(swF2+d*(swF3+d*swF4));
507 /* Evaluate switch function */
508 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
509 felec = felec*sw - rinv22*velec*dsw;
512 /* Update potential sums from outer loop */
517 /* Calculate temporary vectorial force */
522 /* Update vectorial force */
526 f[j_coord_offset+DIM*2+XX] -= tx;
527 f[j_coord_offset+DIM*2+YY] -= ty;
528 f[j_coord_offset+DIM*2+ZZ] -= tz;
532 /**************************
533 * CALCULATE INTERACTIONS *
534 **************************/
541 /* EWALD ELECTROSTATICS */
543 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
544 ewrt = r23*ewtabscale;
548 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
549 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
550 felec = qq23*rinv23*(rinvsq23-felec);
553 d = (d>0.0) ? d : 0.0;
555 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
557 dsw = d2*(swF2+d*(swF3+d*swF4));
559 /* Evaluate switch function */
560 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
561 felec = felec*sw - rinv23*velec*dsw;
564 /* Update potential sums from outer loop */
569 /* Calculate temporary vectorial force */
574 /* Update vectorial force */
578 f[j_coord_offset+DIM*3+XX] -= tx;
579 f[j_coord_offset+DIM*3+YY] -= ty;
580 f[j_coord_offset+DIM*3+ZZ] -= tz;
584 /**************************
585 * CALCULATE INTERACTIONS *
586 **************************/
593 /* EWALD ELECTROSTATICS */
595 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
596 ewrt = r31*ewtabscale;
600 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
601 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
602 felec = qq31*rinv31*(rinvsq31-felec);
605 d = (d>0.0) ? d : 0.0;
607 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
609 dsw = d2*(swF2+d*(swF3+d*swF4));
611 /* Evaluate switch function */
612 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
613 felec = felec*sw - rinv31*velec*dsw;
616 /* Update potential sums from outer loop */
621 /* Calculate temporary vectorial force */
626 /* Update vectorial force */
630 f[j_coord_offset+DIM*1+XX] -= tx;
631 f[j_coord_offset+DIM*1+YY] -= ty;
632 f[j_coord_offset+DIM*1+ZZ] -= tz;
636 /**************************
637 * CALCULATE INTERACTIONS *
638 **************************/
645 /* EWALD ELECTROSTATICS */
647 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
648 ewrt = r32*ewtabscale;
652 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
653 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
654 felec = qq32*rinv32*(rinvsq32-felec);
657 d = (d>0.0) ? d : 0.0;
659 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
661 dsw = d2*(swF2+d*(swF3+d*swF4));
663 /* Evaluate switch function */
664 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
665 felec = felec*sw - rinv32*velec*dsw;
668 /* Update potential sums from outer loop */
673 /* Calculate temporary vectorial force */
678 /* Update vectorial force */
682 f[j_coord_offset+DIM*2+XX] -= tx;
683 f[j_coord_offset+DIM*2+YY] -= ty;
684 f[j_coord_offset+DIM*2+ZZ] -= tz;
688 /**************************
689 * CALCULATE INTERACTIONS *
690 **************************/
697 /* EWALD ELECTROSTATICS */
699 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
700 ewrt = r33*ewtabscale;
704 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
705 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
706 felec = qq33*rinv33*(rinvsq33-felec);
709 d = (d>0.0) ? d : 0.0;
711 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
713 dsw = d2*(swF2+d*(swF3+d*swF4));
715 /* Evaluate switch function */
716 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
717 felec = felec*sw - rinv33*velec*dsw;
720 /* Update potential sums from outer loop */
725 /* Calculate temporary vectorial force */
730 /* Update vectorial force */
734 f[j_coord_offset+DIM*3+XX] -= tx;
735 f[j_coord_offset+DIM*3+YY] -= ty;
736 f[j_coord_offset+DIM*3+ZZ] -= tz;
740 /* Inner loop uses 522 flops */
742 /* End of innermost loop */
745 f[i_coord_offset+DIM*1+XX] += fix1;
746 f[i_coord_offset+DIM*1+YY] += fiy1;
747 f[i_coord_offset+DIM*1+ZZ] += fiz1;
751 f[i_coord_offset+DIM*2+XX] += fix2;
752 f[i_coord_offset+DIM*2+YY] += fiy2;
753 f[i_coord_offset+DIM*2+ZZ] += fiz2;
757 f[i_coord_offset+DIM*3+XX] += fix3;
758 f[i_coord_offset+DIM*3+YY] += fiy3;
759 f[i_coord_offset+DIM*3+ZZ] += fiz3;
763 fshift[i_shift_offset+XX] += tx;
764 fshift[i_shift_offset+YY] += ty;
765 fshift[i_shift_offset+ZZ] += tz;
768 /* Update potential energies */
769 kernel_data->energygrp_elec[ggid] += velecsum;
771 /* Increment number of inner iterations */
772 inneriter += j_index_end - j_index_start;
774 /* Outer loop uses 31 flops */
777 /* Increment number of outer iterations */
780 /* Update outer/inner flops */
782 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_VF,outeriter*31 + inneriter*522);
785 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW4W4_F_c
786 * Electrostatics interaction: Ewald
787 * VdW interaction: None
788 * Geometry: Water4-Water4
789 * Calculate force/pot: Force
792 nb_kernel_ElecEwSw_VdwNone_GeomW4W4_F_c
793 (t_nblist * gmx_restrict nlist,
794 rvec * gmx_restrict xx,
795 rvec * gmx_restrict ff,
796 t_forcerec * gmx_restrict fr,
797 t_mdatoms * gmx_restrict mdatoms,
798 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
799 t_nrnb * gmx_restrict nrnb)
801 int i_shift_offset,i_coord_offset,j_coord_offset;
802 int j_index_start,j_index_end;
803 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
804 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
805 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
806 real *shiftvec,*fshift,*x,*f;
808 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
810 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
812 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
814 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
816 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
818 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
819 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
820 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
821 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
822 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
823 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
824 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
825 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
826 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
827 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
828 real velec,felec,velecsum,facel,crf,krf,krf2;
831 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
833 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
840 jindex = nlist->jindex;
842 shiftidx = nlist->shift;
844 shiftvec = fr->shift_vec[0];
845 fshift = fr->fshift[0];
847 charge = mdatoms->chargeA;
849 sh_ewald = fr->ic->sh_ewald;
850 ewtab = fr->ic->tabq_coul_FDV0;
851 ewtabscale = fr->ic->tabq_scale;
852 ewtabhalfspace = 0.5/ewtabscale;
854 /* Setup water-specific parameters */
855 inr = nlist->iinr[0];
856 iq1 = facel*charge[inr+1];
857 iq2 = facel*charge[inr+2];
858 iq3 = facel*charge[inr+3];
873 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
874 rcutoff = fr->rcoulomb;
875 rcutoff2 = rcutoff*rcutoff;
877 rswitch = fr->rcoulomb_switch;
878 /* Setup switch parameters */
880 swV3 = -10.0/(d*d*d);
881 swV4 = 15.0/(d*d*d*d);
882 swV5 = -6.0/(d*d*d*d*d);
883 swF2 = -30.0/(d*d*d);
884 swF3 = 60.0/(d*d*d*d);
885 swF4 = -30.0/(d*d*d*d*d);
890 /* Start outer loop over neighborlists */
891 for(iidx=0; iidx<nri; iidx++)
893 /* Load shift vector for this list */
894 i_shift_offset = DIM*shiftidx[iidx];
895 shX = shiftvec[i_shift_offset+XX];
896 shY = shiftvec[i_shift_offset+YY];
897 shZ = shiftvec[i_shift_offset+ZZ];
899 /* Load limits for loop over neighbors */
900 j_index_start = jindex[iidx];
901 j_index_end = jindex[iidx+1];
903 /* Get outer coordinate index */
905 i_coord_offset = DIM*inr;
907 /* Load i particle coords and add shift vector */
908 ix1 = shX + x[i_coord_offset+DIM*1+XX];
909 iy1 = shY + x[i_coord_offset+DIM*1+YY];
910 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
911 ix2 = shX + x[i_coord_offset+DIM*2+XX];
912 iy2 = shY + x[i_coord_offset+DIM*2+YY];
913 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
914 ix3 = shX + x[i_coord_offset+DIM*3+XX];
915 iy3 = shY + x[i_coord_offset+DIM*3+YY];
916 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
928 /* Start inner kernel loop */
929 for(jidx=j_index_start; jidx<j_index_end; jidx++)
931 /* Get j neighbor index, and coordinate index */
933 j_coord_offset = DIM*jnr;
935 /* load j atom coordinates */
936 jx1 = x[j_coord_offset+DIM*1+XX];
937 jy1 = x[j_coord_offset+DIM*1+YY];
938 jz1 = x[j_coord_offset+DIM*1+ZZ];
939 jx2 = x[j_coord_offset+DIM*2+XX];
940 jy2 = x[j_coord_offset+DIM*2+YY];
941 jz2 = x[j_coord_offset+DIM*2+ZZ];
942 jx3 = x[j_coord_offset+DIM*3+XX];
943 jy3 = x[j_coord_offset+DIM*3+YY];
944 jz3 = x[j_coord_offset+DIM*3+ZZ];
946 /* Calculate displacement vector */
975 /* Calculate squared distance and things based on it */
976 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
977 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
978 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
979 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
980 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
981 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
982 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
983 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
984 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
986 rinv11 = gmx_invsqrt(rsq11);
987 rinv12 = gmx_invsqrt(rsq12);
988 rinv13 = gmx_invsqrt(rsq13);
989 rinv21 = gmx_invsqrt(rsq21);
990 rinv22 = gmx_invsqrt(rsq22);
991 rinv23 = gmx_invsqrt(rsq23);
992 rinv31 = gmx_invsqrt(rsq31);
993 rinv32 = gmx_invsqrt(rsq32);
994 rinv33 = gmx_invsqrt(rsq33);
996 rinvsq11 = rinv11*rinv11;
997 rinvsq12 = rinv12*rinv12;
998 rinvsq13 = rinv13*rinv13;
999 rinvsq21 = rinv21*rinv21;
1000 rinvsq22 = rinv22*rinv22;
1001 rinvsq23 = rinv23*rinv23;
1002 rinvsq31 = rinv31*rinv31;
1003 rinvsq32 = rinv32*rinv32;
1004 rinvsq33 = rinv33*rinv33;
1006 /**************************
1007 * CALCULATE INTERACTIONS *
1008 **************************/
1015 /* EWALD ELECTROSTATICS */
1017 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1018 ewrt = r11*ewtabscale;
1020 eweps = ewrt-ewitab;
1022 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1023 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1024 felec = qq11*rinv11*(rinvsq11-felec);
1027 d = (d>0.0) ? d : 0.0;
1029 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1031 dsw = d2*(swF2+d*(swF3+d*swF4));
1033 /* Evaluate switch function */
1034 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1035 felec = felec*sw - rinv11*velec*dsw;
1039 /* Calculate temporary vectorial force */
1044 /* Update vectorial force */
1048 f[j_coord_offset+DIM*1+XX] -= tx;
1049 f[j_coord_offset+DIM*1+YY] -= ty;
1050 f[j_coord_offset+DIM*1+ZZ] -= tz;
1054 /**************************
1055 * CALCULATE INTERACTIONS *
1056 **************************/
1063 /* EWALD ELECTROSTATICS */
1065 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1066 ewrt = r12*ewtabscale;
1068 eweps = ewrt-ewitab;
1070 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1071 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1072 felec = qq12*rinv12*(rinvsq12-felec);
1075 d = (d>0.0) ? d : 0.0;
1077 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1079 dsw = d2*(swF2+d*(swF3+d*swF4));
1081 /* Evaluate switch function */
1082 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1083 felec = felec*sw - rinv12*velec*dsw;
1087 /* Calculate temporary vectorial force */
1092 /* Update vectorial force */
1096 f[j_coord_offset+DIM*2+XX] -= tx;
1097 f[j_coord_offset+DIM*2+YY] -= ty;
1098 f[j_coord_offset+DIM*2+ZZ] -= tz;
1102 /**************************
1103 * CALCULATE INTERACTIONS *
1104 **************************/
1111 /* EWALD ELECTROSTATICS */
1113 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1114 ewrt = r13*ewtabscale;
1116 eweps = ewrt-ewitab;
1118 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1119 velec = qq13*(rinv13-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1120 felec = qq13*rinv13*(rinvsq13-felec);
1123 d = (d>0.0) ? d : 0.0;
1125 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1127 dsw = d2*(swF2+d*(swF3+d*swF4));
1129 /* Evaluate switch function */
1130 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1131 felec = felec*sw - rinv13*velec*dsw;
1135 /* Calculate temporary vectorial force */
1140 /* Update vectorial force */
1144 f[j_coord_offset+DIM*3+XX] -= tx;
1145 f[j_coord_offset+DIM*3+YY] -= ty;
1146 f[j_coord_offset+DIM*3+ZZ] -= tz;
1150 /**************************
1151 * CALCULATE INTERACTIONS *
1152 **************************/
1159 /* EWALD ELECTROSTATICS */
1161 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1162 ewrt = r21*ewtabscale;
1164 eweps = ewrt-ewitab;
1166 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1167 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1168 felec = qq21*rinv21*(rinvsq21-felec);
1171 d = (d>0.0) ? d : 0.0;
1173 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1175 dsw = d2*(swF2+d*(swF3+d*swF4));
1177 /* Evaluate switch function */
1178 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1179 felec = felec*sw - rinv21*velec*dsw;
1183 /* Calculate temporary vectorial force */
1188 /* Update vectorial force */
1192 f[j_coord_offset+DIM*1+XX] -= tx;
1193 f[j_coord_offset+DIM*1+YY] -= ty;
1194 f[j_coord_offset+DIM*1+ZZ] -= tz;
1198 /**************************
1199 * CALCULATE INTERACTIONS *
1200 **************************/
1207 /* EWALD ELECTROSTATICS */
1209 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1210 ewrt = r22*ewtabscale;
1212 eweps = ewrt-ewitab;
1214 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1215 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1216 felec = qq22*rinv22*(rinvsq22-felec);
1219 d = (d>0.0) ? d : 0.0;
1221 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1223 dsw = d2*(swF2+d*(swF3+d*swF4));
1225 /* Evaluate switch function */
1226 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1227 felec = felec*sw - rinv22*velec*dsw;
1231 /* Calculate temporary vectorial force */
1236 /* Update vectorial force */
1240 f[j_coord_offset+DIM*2+XX] -= tx;
1241 f[j_coord_offset+DIM*2+YY] -= ty;
1242 f[j_coord_offset+DIM*2+ZZ] -= tz;
1246 /**************************
1247 * CALCULATE INTERACTIONS *
1248 **************************/
1255 /* EWALD ELECTROSTATICS */
1257 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1258 ewrt = r23*ewtabscale;
1260 eweps = ewrt-ewitab;
1262 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1263 velec = qq23*(rinv23-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1264 felec = qq23*rinv23*(rinvsq23-felec);
1267 d = (d>0.0) ? d : 0.0;
1269 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1271 dsw = d2*(swF2+d*(swF3+d*swF4));
1273 /* Evaluate switch function */
1274 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1275 felec = felec*sw - rinv23*velec*dsw;
1279 /* Calculate temporary vectorial force */
1284 /* Update vectorial force */
1288 f[j_coord_offset+DIM*3+XX] -= tx;
1289 f[j_coord_offset+DIM*3+YY] -= ty;
1290 f[j_coord_offset+DIM*3+ZZ] -= tz;
1294 /**************************
1295 * CALCULATE INTERACTIONS *
1296 **************************/
1303 /* EWALD ELECTROSTATICS */
1305 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1306 ewrt = r31*ewtabscale;
1308 eweps = ewrt-ewitab;
1310 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1311 velec = qq31*(rinv31-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1312 felec = qq31*rinv31*(rinvsq31-felec);
1315 d = (d>0.0) ? d : 0.0;
1317 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1319 dsw = d2*(swF2+d*(swF3+d*swF4));
1321 /* Evaluate switch function */
1322 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1323 felec = felec*sw - rinv31*velec*dsw;
1327 /* Calculate temporary vectorial force */
1332 /* Update vectorial force */
1336 f[j_coord_offset+DIM*1+XX] -= tx;
1337 f[j_coord_offset+DIM*1+YY] -= ty;
1338 f[j_coord_offset+DIM*1+ZZ] -= tz;
1342 /**************************
1343 * CALCULATE INTERACTIONS *
1344 **************************/
1351 /* EWALD ELECTROSTATICS */
1353 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1354 ewrt = r32*ewtabscale;
1356 eweps = ewrt-ewitab;
1358 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1359 velec = qq32*(rinv32-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1360 felec = qq32*rinv32*(rinvsq32-felec);
1363 d = (d>0.0) ? d : 0.0;
1365 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1367 dsw = d2*(swF2+d*(swF3+d*swF4));
1369 /* Evaluate switch function */
1370 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1371 felec = felec*sw - rinv32*velec*dsw;
1375 /* Calculate temporary vectorial force */
1380 /* Update vectorial force */
1384 f[j_coord_offset+DIM*2+XX] -= tx;
1385 f[j_coord_offset+DIM*2+YY] -= ty;
1386 f[j_coord_offset+DIM*2+ZZ] -= tz;
1390 /**************************
1391 * CALCULATE INTERACTIONS *
1392 **************************/
1399 /* EWALD ELECTROSTATICS */
1401 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1402 ewrt = r33*ewtabscale;
1404 eweps = ewrt-ewitab;
1406 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1407 velec = qq33*(rinv33-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1408 felec = qq33*rinv33*(rinvsq33-felec);
1411 d = (d>0.0) ? d : 0.0;
1413 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1415 dsw = d2*(swF2+d*(swF3+d*swF4));
1417 /* Evaluate switch function */
1418 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1419 felec = felec*sw - rinv33*velec*dsw;
1423 /* Calculate temporary vectorial force */
1428 /* Update vectorial force */
1432 f[j_coord_offset+DIM*3+XX] -= tx;
1433 f[j_coord_offset+DIM*3+YY] -= ty;
1434 f[j_coord_offset+DIM*3+ZZ] -= tz;
1438 /* Inner loop uses 504 flops */
1440 /* End of innermost loop */
1443 f[i_coord_offset+DIM*1+XX] += fix1;
1444 f[i_coord_offset+DIM*1+YY] += fiy1;
1445 f[i_coord_offset+DIM*1+ZZ] += fiz1;
1449 f[i_coord_offset+DIM*2+XX] += fix2;
1450 f[i_coord_offset+DIM*2+YY] += fiy2;
1451 f[i_coord_offset+DIM*2+ZZ] += fiz2;
1455 f[i_coord_offset+DIM*3+XX] += fix3;
1456 f[i_coord_offset+DIM*3+YY] += fiy3;
1457 f[i_coord_offset+DIM*3+ZZ] += fiz3;
1461 fshift[i_shift_offset+XX] += tx;
1462 fshift[i_shift_offset+YY] += ty;
1463 fshift[i_shift_offset+ZZ] += tz;
1465 /* Increment number of inner iterations */
1466 inneriter += j_index_end - j_index_start;
1468 /* Outer loop uses 30 flops */
1471 /* Increment number of outer iterations */
1474 /* Update outer/inner flops */
1476 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_F,outeriter*30 + inneriter*504);