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.
44 #include "../nb_kernel.h"
45 #include "gromacs/legacyheaders/types/simple.h"
46 #include "gromacs/math/vec.h"
47 #include "gromacs/legacyheaders/nrnb.h"
50 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW3W3_VF_c
51 * Electrostatics interaction: Ewald
52 * VdW interaction: None
53 * Geometry: Water3-Water3
54 * Calculate force/pot: PotentialAndForce
57 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_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 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;
84 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
85 real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01;
86 real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02;
87 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
88 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
89 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
90 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
91 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
92 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
93 real velec,felec,velecsum,facel,crf,krf,krf2;
96 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
98 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
105 jindex = nlist->jindex;
107 shiftidx = nlist->shift;
109 shiftvec = fr->shift_vec[0];
110 fshift = fr->fshift[0];
112 charge = mdatoms->chargeA;
114 sh_ewald = fr->ic->sh_ewald;
115 ewtab = fr->ic->tabq_coul_FDV0;
116 ewtabscale = fr->ic->tabq_scale;
117 ewtabhalfspace = 0.5/ewtabscale;
119 /* Setup water-specific parameters */
120 inr = nlist->iinr[0];
121 iq0 = facel*charge[inr+0];
122 iq1 = facel*charge[inr+1];
123 iq2 = facel*charge[inr+2];
138 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
139 rcutoff = fr->rcoulomb;
140 rcutoff2 = rcutoff*rcutoff;
142 rswitch = fr->rcoulomb_switch;
143 /* Setup switch parameters */
145 swV3 = -10.0/(d*d*d);
146 swV4 = 15.0/(d*d*d*d);
147 swV5 = -6.0/(d*d*d*d*d);
148 swF2 = -30.0/(d*d*d);
149 swF3 = 60.0/(d*d*d*d);
150 swF4 = -30.0/(d*d*d*d*d);
155 /* Start outer loop over neighborlists */
156 for(iidx=0; iidx<nri; iidx++)
158 /* Load shift vector for this list */
159 i_shift_offset = DIM*shiftidx[iidx];
160 shX = shiftvec[i_shift_offset+XX];
161 shY = shiftvec[i_shift_offset+YY];
162 shZ = shiftvec[i_shift_offset+ZZ];
164 /* Load limits for loop over neighbors */
165 j_index_start = jindex[iidx];
166 j_index_end = jindex[iidx+1];
168 /* Get outer coordinate index */
170 i_coord_offset = DIM*inr;
172 /* Load i particle coords and add shift vector */
173 ix0 = shX + x[i_coord_offset+DIM*0+XX];
174 iy0 = shY + x[i_coord_offset+DIM*0+YY];
175 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
176 ix1 = shX + x[i_coord_offset+DIM*1+XX];
177 iy1 = shY + x[i_coord_offset+DIM*1+YY];
178 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
179 ix2 = shX + x[i_coord_offset+DIM*2+XX];
180 iy2 = shY + x[i_coord_offset+DIM*2+YY];
181 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
193 /* Reset potential sums */
196 /* Start inner kernel loop */
197 for(jidx=j_index_start; jidx<j_index_end; jidx++)
199 /* Get j neighbor index, and coordinate index */
201 j_coord_offset = DIM*jnr;
203 /* load j atom coordinates */
204 jx0 = x[j_coord_offset+DIM*0+XX];
205 jy0 = x[j_coord_offset+DIM*0+YY];
206 jz0 = x[j_coord_offset+DIM*0+ZZ];
207 jx1 = x[j_coord_offset+DIM*1+XX];
208 jy1 = x[j_coord_offset+DIM*1+YY];
209 jz1 = x[j_coord_offset+DIM*1+ZZ];
210 jx2 = x[j_coord_offset+DIM*2+XX];
211 jy2 = x[j_coord_offset+DIM*2+YY];
212 jz2 = x[j_coord_offset+DIM*2+ZZ];
214 /* Calculate displacement vector */
243 /* Calculate squared distance and things based on it */
244 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
245 rsq01 = dx01*dx01+dy01*dy01+dz01*dz01;
246 rsq02 = dx02*dx02+dy02*dy02+dz02*dz02;
247 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
248 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
249 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
250 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
251 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
252 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
254 rinv00 = gmx_invsqrt(rsq00);
255 rinv01 = gmx_invsqrt(rsq01);
256 rinv02 = gmx_invsqrt(rsq02);
257 rinv10 = gmx_invsqrt(rsq10);
258 rinv11 = gmx_invsqrt(rsq11);
259 rinv12 = gmx_invsqrt(rsq12);
260 rinv20 = gmx_invsqrt(rsq20);
261 rinv21 = gmx_invsqrt(rsq21);
262 rinv22 = gmx_invsqrt(rsq22);
264 rinvsq00 = rinv00*rinv00;
265 rinvsq01 = rinv01*rinv01;
266 rinvsq02 = rinv02*rinv02;
267 rinvsq10 = rinv10*rinv10;
268 rinvsq11 = rinv11*rinv11;
269 rinvsq12 = rinv12*rinv12;
270 rinvsq20 = rinv20*rinv20;
271 rinvsq21 = rinv21*rinv21;
272 rinvsq22 = rinv22*rinv22;
274 /**************************
275 * CALCULATE INTERACTIONS *
276 **************************/
283 /* EWALD ELECTROSTATICS */
285 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
286 ewrt = r00*ewtabscale;
290 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
291 velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
292 felec = qq00*rinv00*(rinvsq00-felec);
295 d = (d>0.0) ? d : 0.0;
297 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
299 dsw = d2*(swF2+d*(swF3+d*swF4));
301 /* Evaluate switch function */
302 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
303 felec = felec*sw - rinv00*velec*dsw;
306 /* Update potential sums from outer loop */
311 /* Calculate temporary vectorial force */
316 /* Update vectorial force */
320 f[j_coord_offset+DIM*0+XX] -= tx;
321 f[j_coord_offset+DIM*0+YY] -= ty;
322 f[j_coord_offset+DIM*0+ZZ] -= tz;
326 /**************************
327 * CALCULATE INTERACTIONS *
328 **************************/
335 /* EWALD ELECTROSTATICS */
337 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
338 ewrt = r01*ewtabscale;
342 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
343 velec = qq01*(rinv01-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
344 felec = qq01*rinv01*(rinvsq01-felec);
347 d = (d>0.0) ? d : 0.0;
349 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
351 dsw = d2*(swF2+d*(swF3+d*swF4));
353 /* Evaluate switch function */
354 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
355 felec = felec*sw - rinv01*velec*dsw;
358 /* Update potential sums from outer loop */
363 /* Calculate temporary vectorial force */
368 /* Update vectorial force */
372 f[j_coord_offset+DIM*1+XX] -= tx;
373 f[j_coord_offset+DIM*1+YY] -= ty;
374 f[j_coord_offset+DIM*1+ZZ] -= tz;
378 /**************************
379 * CALCULATE INTERACTIONS *
380 **************************/
387 /* EWALD ELECTROSTATICS */
389 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
390 ewrt = r02*ewtabscale;
394 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
395 velec = qq02*(rinv02-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
396 felec = qq02*rinv02*(rinvsq02-felec);
399 d = (d>0.0) ? d : 0.0;
401 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
403 dsw = d2*(swF2+d*(swF3+d*swF4));
405 /* Evaluate switch function */
406 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
407 felec = felec*sw - rinv02*velec*dsw;
410 /* Update potential sums from outer loop */
415 /* Calculate temporary vectorial force */
420 /* Update vectorial force */
424 f[j_coord_offset+DIM*2+XX] -= tx;
425 f[j_coord_offset+DIM*2+YY] -= ty;
426 f[j_coord_offset+DIM*2+ZZ] -= tz;
430 /**************************
431 * CALCULATE INTERACTIONS *
432 **************************/
439 /* EWALD ELECTROSTATICS */
441 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
442 ewrt = r10*ewtabscale;
446 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
447 velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
448 felec = qq10*rinv10*(rinvsq10-felec);
451 d = (d>0.0) ? d : 0.0;
453 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
455 dsw = d2*(swF2+d*(swF3+d*swF4));
457 /* Evaluate switch function */
458 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
459 felec = felec*sw - rinv10*velec*dsw;
462 /* Update potential sums from outer loop */
467 /* Calculate temporary vectorial force */
472 /* Update vectorial force */
476 f[j_coord_offset+DIM*0+XX] -= tx;
477 f[j_coord_offset+DIM*0+YY] -= ty;
478 f[j_coord_offset+DIM*0+ZZ] -= tz;
482 /**************************
483 * CALCULATE INTERACTIONS *
484 **************************/
491 /* EWALD ELECTROSTATICS */
493 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
494 ewrt = r11*ewtabscale;
498 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
499 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
500 felec = qq11*rinv11*(rinvsq11-felec);
503 d = (d>0.0) ? d : 0.0;
505 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
507 dsw = d2*(swF2+d*(swF3+d*swF4));
509 /* Evaluate switch function */
510 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
511 felec = felec*sw - rinv11*velec*dsw;
514 /* Update potential sums from outer loop */
519 /* Calculate temporary vectorial force */
524 /* Update vectorial force */
528 f[j_coord_offset+DIM*1+XX] -= tx;
529 f[j_coord_offset+DIM*1+YY] -= ty;
530 f[j_coord_offset+DIM*1+ZZ] -= tz;
534 /**************************
535 * CALCULATE INTERACTIONS *
536 **************************/
543 /* EWALD ELECTROSTATICS */
545 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
546 ewrt = r12*ewtabscale;
550 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
551 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
552 felec = qq12*rinv12*(rinvsq12-felec);
555 d = (d>0.0) ? d : 0.0;
557 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
559 dsw = d2*(swF2+d*(swF3+d*swF4));
561 /* Evaluate switch function */
562 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
563 felec = felec*sw - rinv12*velec*dsw;
566 /* Update potential sums from outer loop */
571 /* Calculate temporary vectorial force */
576 /* Update vectorial force */
580 f[j_coord_offset+DIM*2+XX] -= tx;
581 f[j_coord_offset+DIM*2+YY] -= ty;
582 f[j_coord_offset+DIM*2+ZZ] -= tz;
586 /**************************
587 * CALCULATE INTERACTIONS *
588 **************************/
595 /* EWALD ELECTROSTATICS */
597 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
598 ewrt = r20*ewtabscale;
602 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
603 velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
604 felec = qq20*rinv20*(rinvsq20-felec);
607 d = (d>0.0) ? d : 0.0;
609 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
611 dsw = d2*(swF2+d*(swF3+d*swF4));
613 /* Evaluate switch function */
614 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
615 felec = felec*sw - rinv20*velec*dsw;
618 /* Update potential sums from outer loop */
623 /* Calculate temporary vectorial force */
628 /* Update vectorial force */
632 f[j_coord_offset+DIM*0+XX] -= tx;
633 f[j_coord_offset+DIM*0+YY] -= ty;
634 f[j_coord_offset+DIM*0+ZZ] -= tz;
638 /**************************
639 * CALCULATE INTERACTIONS *
640 **************************/
647 /* EWALD ELECTROSTATICS */
649 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
650 ewrt = r21*ewtabscale;
654 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
655 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
656 felec = qq21*rinv21*(rinvsq21-felec);
659 d = (d>0.0) ? d : 0.0;
661 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
663 dsw = d2*(swF2+d*(swF3+d*swF4));
665 /* Evaluate switch function */
666 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
667 felec = felec*sw - rinv21*velec*dsw;
670 /* Update potential sums from outer loop */
675 /* Calculate temporary vectorial force */
680 /* Update vectorial force */
684 f[j_coord_offset+DIM*1+XX] -= tx;
685 f[j_coord_offset+DIM*1+YY] -= ty;
686 f[j_coord_offset+DIM*1+ZZ] -= tz;
690 /**************************
691 * CALCULATE INTERACTIONS *
692 **************************/
699 /* EWALD ELECTROSTATICS */
701 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
702 ewrt = r22*ewtabscale;
706 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
707 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
708 felec = qq22*rinv22*(rinvsq22-felec);
711 d = (d>0.0) ? d : 0.0;
713 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
715 dsw = d2*(swF2+d*(swF3+d*swF4));
717 /* Evaluate switch function */
718 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
719 felec = felec*sw - rinv22*velec*dsw;
722 /* Update potential sums from outer loop */
727 /* Calculate temporary vectorial force */
732 /* Update vectorial force */
736 f[j_coord_offset+DIM*2+XX] -= tx;
737 f[j_coord_offset+DIM*2+YY] -= ty;
738 f[j_coord_offset+DIM*2+ZZ] -= tz;
742 /* Inner loop uses 522 flops */
744 /* End of innermost loop */
747 f[i_coord_offset+DIM*0+XX] += fix0;
748 f[i_coord_offset+DIM*0+YY] += fiy0;
749 f[i_coord_offset+DIM*0+ZZ] += fiz0;
753 f[i_coord_offset+DIM*1+XX] += fix1;
754 f[i_coord_offset+DIM*1+YY] += fiy1;
755 f[i_coord_offset+DIM*1+ZZ] += fiz1;
759 f[i_coord_offset+DIM*2+XX] += fix2;
760 f[i_coord_offset+DIM*2+YY] += fiy2;
761 f[i_coord_offset+DIM*2+ZZ] += fiz2;
765 fshift[i_shift_offset+XX] += tx;
766 fshift[i_shift_offset+YY] += ty;
767 fshift[i_shift_offset+ZZ] += tz;
770 /* Update potential energies */
771 kernel_data->energygrp_elec[ggid] += velecsum;
773 /* Increment number of inner iterations */
774 inneriter += j_index_end - j_index_start;
776 /* Outer loop uses 31 flops */
779 /* Increment number of outer iterations */
782 /* Update outer/inner flops */
784 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_VF,outeriter*31 + inneriter*522);
787 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_c
788 * Electrostatics interaction: Ewald
789 * VdW interaction: None
790 * Geometry: Water3-Water3
791 * Calculate force/pot: Force
794 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_c
795 (t_nblist * gmx_restrict nlist,
796 rvec * gmx_restrict xx,
797 rvec * gmx_restrict ff,
798 t_forcerec * gmx_restrict fr,
799 t_mdatoms * gmx_restrict mdatoms,
800 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
801 t_nrnb * gmx_restrict nrnb)
803 int i_shift_offset,i_coord_offset,j_coord_offset;
804 int j_index_start,j_index_end;
805 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
806 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
807 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
808 real *shiftvec,*fshift,*x,*f;
810 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
812 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
814 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
816 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
818 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
820 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
821 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
822 real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01;
823 real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02;
824 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
825 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
826 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
827 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
828 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
829 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
830 real velec,felec,velecsum,facel,crf,krf,krf2;
833 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
835 real rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
842 jindex = nlist->jindex;
844 shiftidx = nlist->shift;
846 shiftvec = fr->shift_vec[0];
847 fshift = fr->fshift[0];
849 charge = mdatoms->chargeA;
851 sh_ewald = fr->ic->sh_ewald;
852 ewtab = fr->ic->tabq_coul_FDV0;
853 ewtabscale = fr->ic->tabq_scale;
854 ewtabhalfspace = 0.5/ewtabscale;
856 /* Setup water-specific parameters */
857 inr = nlist->iinr[0];
858 iq0 = facel*charge[inr+0];
859 iq1 = facel*charge[inr+1];
860 iq2 = facel*charge[inr+2];
875 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
876 rcutoff = fr->rcoulomb;
877 rcutoff2 = rcutoff*rcutoff;
879 rswitch = fr->rcoulomb_switch;
880 /* Setup switch parameters */
882 swV3 = -10.0/(d*d*d);
883 swV4 = 15.0/(d*d*d*d);
884 swV5 = -6.0/(d*d*d*d*d);
885 swF2 = -30.0/(d*d*d);
886 swF3 = 60.0/(d*d*d*d);
887 swF4 = -30.0/(d*d*d*d*d);
892 /* Start outer loop over neighborlists */
893 for(iidx=0; iidx<nri; iidx++)
895 /* Load shift vector for this list */
896 i_shift_offset = DIM*shiftidx[iidx];
897 shX = shiftvec[i_shift_offset+XX];
898 shY = shiftvec[i_shift_offset+YY];
899 shZ = shiftvec[i_shift_offset+ZZ];
901 /* Load limits for loop over neighbors */
902 j_index_start = jindex[iidx];
903 j_index_end = jindex[iidx+1];
905 /* Get outer coordinate index */
907 i_coord_offset = DIM*inr;
909 /* Load i particle coords and add shift vector */
910 ix0 = shX + x[i_coord_offset+DIM*0+XX];
911 iy0 = shY + x[i_coord_offset+DIM*0+YY];
912 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
913 ix1 = shX + x[i_coord_offset+DIM*1+XX];
914 iy1 = shY + x[i_coord_offset+DIM*1+YY];
915 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
916 ix2 = shX + x[i_coord_offset+DIM*2+XX];
917 iy2 = shY + x[i_coord_offset+DIM*2+YY];
918 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
930 /* Start inner kernel loop */
931 for(jidx=j_index_start; jidx<j_index_end; jidx++)
933 /* Get j neighbor index, and coordinate index */
935 j_coord_offset = DIM*jnr;
937 /* load j atom coordinates */
938 jx0 = x[j_coord_offset+DIM*0+XX];
939 jy0 = x[j_coord_offset+DIM*0+YY];
940 jz0 = x[j_coord_offset+DIM*0+ZZ];
941 jx1 = x[j_coord_offset+DIM*1+XX];
942 jy1 = x[j_coord_offset+DIM*1+YY];
943 jz1 = x[j_coord_offset+DIM*1+ZZ];
944 jx2 = x[j_coord_offset+DIM*2+XX];
945 jy2 = x[j_coord_offset+DIM*2+YY];
946 jz2 = x[j_coord_offset+DIM*2+ZZ];
948 /* Calculate displacement vector */
977 /* Calculate squared distance and things based on it */
978 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
979 rsq01 = dx01*dx01+dy01*dy01+dz01*dz01;
980 rsq02 = dx02*dx02+dy02*dy02+dz02*dz02;
981 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
982 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
983 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
984 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
985 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
986 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
988 rinv00 = gmx_invsqrt(rsq00);
989 rinv01 = gmx_invsqrt(rsq01);
990 rinv02 = gmx_invsqrt(rsq02);
991 rinv10 = gmx_invsqrt(rsq10);
992 rinv11 = gmx_invsqrt(rsq11);
993 rinv12 = gmx_invsqrt(rsq12);
994 rinv20 = gmx_invsqrt(rsq20);
995 rinv21 = gmx_invsqrt(rsq21);
996 rinv22 = gmx_invsqrt(rsq22);
998 rinvsq00 = rinv00*rinv00;
999 rinvsq01 = rinv01*rinv01;
1000 rinvsq02 = rinv02*rinv02;
1001 rinvsq10 = rinv10*rinv10;
1002 rinvsq11 = rinv11*rinv11;
1003 rinvsq12 = rinv12*rinv12;
1004 rinvsq20 = rinv20*rinv20;
1005 rinvsq21 = rinv21*rinv21;
1006 rinvsq22 = rinv22*rinv22;
1008 /**************************
1009 * CALCULATE INTERACTIONS *
1010 **************************/
1017 /* EWALD ELECTROSTATICS */
1019 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1020 ewrt = r00*ewtabscale;
1022 eweps = ewrt-ewitab;
1024 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1025 velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1026 felec = qq00*rinv00*(rinvsq00-felec);
1029 d = (d>0.0) ? d : 0.0;
1031 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1033 dsw = d2*(swF2+d*(swF3+d*swF4));
1035 /* Evaluate switch function */
1036 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1037 felec = felec*sw - rinv00*velec*dsw;
1041 /* Calculate temporary vectorial force */
1046 /* Update vectorial force */
1050 f[j_coord_offset+DIM*0+XX] -= tx;
1051 f[j_coord_offset+DIM*0+YY] -= ty;
1052 f[j_coord_offset+DIM*0+ZZ] -= tz;
1056 /**************************
1057 * CALCULATE INTERACTIONS *
1058 **************************/
1065 /* EWALD ELECTROSTATICS */
1067 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1068 ewrt = r01*ewtabscale;
1070 eweps = ewrt-ewitab;
1072 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1073 velec = qq01*(rinv01-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1074 felec = qq01*rinv01*(rinvsq01-felec);
1077 d = (d>0.0) ? d : 0.0;
1079 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1081 dsw = d2*(swF2+d*(swF3+d*swF4));
1083 /* Evaluate switch function */
1084 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1085 felec = felec*sw - rinv01*velec*dsw;
1089 /* Calculate temporary vectorial force */
1094 /* Update vectorial force */
1098 f[j_coord_offset+DIM*1+XX] -= tx;
1099 f[j_coord_offset+DIM*1+YY] -= ty;
1100 f[j_coord_offset+DIM*1+ZZ] -= tz;
1104 /**************************
1105 * CALCULATE INTERACTIONS *
1106 **************************/
1113 /* EWALD ELECTROSTATICS */
1115 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1116 ewrt = r02*ewtabscale;
1118 eweps = ewrt-ewitab;
1120 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1121 velec = qq02*(rinv02-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1122 felec = qq02*rinv02*(rinvsq02-felec);
1125 d = (d>0.0) ? d : 0.0;
1127 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1129 dsw = d2*(swF2+d*(swF3+d*swF4));
1131 /* Evaluate switch function */
1132 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1133 felec = felec*sw - rinv02*velec*dsw;
1137 /* Calculate temporary vectorial force */
1142 /* Update vectorial force */
1146 f[j_coord_offset+DIM*2+XX] -= tx;
1147 f[j_coord_offset+DIM*2+YY] -= ty;
1148 f[j_coord_offset+DIM*2+ZZ] -= tz;
1152 /**************************
1153 * CALCULATE INTERACTIONS *
1154 **************************/
1161 /* EWALD ELECTROSTATICS */
1163 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1164 ewrt = r10*ewtabscale;
1166 eweps = ewrt-ewitab;
1168 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1169 velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1170 felec = qq10*rinv10*(rinvsq10-felec);
1173 d = (d>0.0) ? d : 0.0;
1175 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1177 dsw = d2*(swF2+d*(swF3+d*swF4));
1179 /* Evaluate switch function */
1180 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1181 felec = felec*sw - rinv10*velec*dsw;
1185 /* Calculate temporary vectorial force */
1190 /* Update vectorial force */
1194 f[j_coord_offset+DIM*0+XX] -= tx;
1195 f[j_coord_offset+DIM*0+YY] -= ty;
1196 f[j_coord_offset+DIM*0+ZZ] -= tz;
1200 /**************************
1201 * CALCULATE INTERACTIONS *
1202 **************************/
1209 /* EWALD ELECTROSTATICS */
1211 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1212 ewrt = r11*ewtabscale;
1214 eweps = ewrt-ewitab;
1216 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1217 velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1218 felec = qq11*rinv11*(rinvsq11-felec);
1221 d = (d>0.0) ? d : 0.0;
1223 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1225 dsw = d2*(swF2+d*(swF3+d*swF4));
1227 /* Evaluate switch function */
1228 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1229 felec = felec*sw - rinv11*velec*dsw;
1233 /* Calculate temporary vectorial force */
1238 /* Update vectorial force */
1242 f[j_coord_offset+DIM*1+XX] -= tx;
1243 f[j_coord_offset+DIM*1+YY] -= ty;
1244 f[j_coord_offset+DIM*1+ZZ] -= tz;
1248 /**************************
1249 * CALCULATE INTERACTIONS *
1250 **************************/
1257 /* EWALD ELECTROSTATICS */
1259 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1260 ewrt = r12*ewtabscale;
1262 eweps = ewrt-ewitab;
1264 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1265 velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1266 felec = qq12*rinv12*(rinvsq12-felec);
1269 d = (d>0.0) ? d : 0.0;
1271 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1273 dsw = d2*(swF2+d*(swF3+d*swF4));
1275 /* Evaluate switch function */
1276 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1277 felec = felec*sw - rinv12*velec*dsw;
1281 /* Calculate temporary vectorial force */
1286 /* Update vectorial force */
1290 f[j_coord_offset+DIM*2+XX] -= tx;
1291 f[j_coord_offset+DIM*2+YY] -= ty;
1292 f[j_coord_offset+DIM*2+ZZ] -= tz;
1296 /**************************
1297 * CALCULATE INTERACTIONS *
1298 **************************/
1305 /* EWALD ELECTROSTATICS */
1307 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1308 ewrt = r20*ewtabscale;
1310 eweps = ewrt-ewitab;
1312 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1313 velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1314 felec = qq20*rinv20*(rinvsq20-felec);
1317 d = (d>0.0) ? d : 0.0;
1319 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1321 dsw = d2*(swF2+d*(swF3+d*swF4));
1323 /* Evaluate switch function */
1324 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1325 felec = felec*sw - rinv20*velec*dsw;
1329 /* Calculate temporary vectorial force */
1334 /* Update vectorial force */
1338 f[j_coord_offset+DIM*0+XX] -= tx;
1339 f[j_coord_offset+DIM*0+YY] -= ty;
1340 f[j_coord_offset+DIM*0+ZZ] -= tz;
1344 /**************************
1345 * CALCULATE INTERACTIONS *
1346 **************************/
1353 /* EWALD ELECTROSTATICS */
1355 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1356 ewrt = r21*ewtabscale;
1358 eweps = ewrt-ewitab;
1360 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1361 velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1362 felec = qq21*rinv21*(rinvsq21-felec);
1365 d = (d>0.0) ? d : 0.0;
1367 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1369 dsw = d2*(swF2+d*(swF3+d*swF4));
1371 /* Evaluate switch function */
1372 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1373 felec = felec*sw - rinv21*velec*dsw;
1377 /* Calculate temporary vectorial force */
1382 /* Update vectorial force */
1386 f[j_coord_offset+DIM*1+XX] -= tx;
1387 f[j_coord_offset+DIM*1+YY] -= ty;
1388 f[j_coord_offset+DIM*1+ZZ] -= tz;
1392 /**************************
1393 * CALCULATE INTERACTIONS *
1394 **************************/
1401 /* EWALD ELECTROSTATICS */
1403 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1404 ewrt = r22*ewtabscale;
1406 eweps = ewrt-ewitab;
1408 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
1409 velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
1410 felec = qq22*rinv22*(rinvsq22-felec);
1413 d = (d>0.0) ? d : 0.0;
1415 sw = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
1417 dsw = d2*(swF2+d*(swF3+d*swF4));
1419 /* Evaluate switch function */
1420 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1421 felec = felec*sw - rinv22*velec*dsw;
1425 /* Calculate temporary vectorial force */
1430 /* Update vectorial force */
1434 f[j_coord_offset+DIM*2+XX] -= tx;
1435 f[j_coord_offset+DIM*2+YY] -= ty;
1436 f[j_coord_offset+DIM*2+ZZ] -= tz;
1440 /* Inner loop uses 504 flops */
1442 /* End of innermost loop */
1445 f[i_coord_offset+DIM*0+XX] += fix0;
1446 f[i_coord_offset+DIM*0+YY] += fiy0;
1447 f[i_coord_offset+DIM*0+ZZ] += fiz0;
1451 f[i_coord_offset+DIM*1+XX] += fix1;
1452 f[i_coord_offset+DIM*1+YY] += fiy1;
1453 f[i_coord_offset+DIM*1+ZZ] += fiz1;
1457 f[i_coord_offset+DIM*2+XX] += fix2;
1458 f[i_coord_offset+DIM*2+YY] += fiy2;
1459 f[i_coord_offset+DIM*2+ZZ] += fiz2;
1463 fshift[i_shift_offset+XX] += tx;
1464 fshift[i_shift_offset+YY] += ty;
1465 fshift[i_shift_offset+ZZ] += tz;
1467 /* Increment number of inner iterations */
1468 inneriter += j_index_end - j_index_start;
1470 /* Outer loop uses 30 flops */
1473 /* Increment number of outer iterations */
1476 /* Update outer/inner flops */
1478 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_F,outeriter*30 + inneriter*504);