2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
36 * Note: this file was generated by the GROMACS c kernel generator.
42 #include "../nb_kernel.h"
43 #include "gromacs/legacyheaders/types/simple.h"
44 #include "gromacs/math/vec.h"
45 #include "gromacs/legacyheaders/nrnb.h"
48 * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_VF_c
49 * Electrostatics interaction: Ewald
50 * VdW interaction: LennardJones
51 * Geometry: Water4-Particle
52 * Calculate force/pot: PotentialAndForce
55 nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_VF_c
56 (t_nblist * gmx_restrict nlist,
57 rvec * gmx_restrict xx,
58 rvec * gmx_restrict ff,
59 t_forcerec * gmx_restrict fr,
60 t_mdatoms * gmx_restrict mdatoms,
61 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
62 t_nrnb * gmx_restrict nrnb)
64 int i_shift_offset,i_coord_offset,j_coord_offset;
65 int j_index_start,j_index_end;
66 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
67 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
68 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
69 real *shiftvec,*fshift,*x,*f;
71 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
73 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
75 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
77 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
79 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
80 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
81 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
82 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
83 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
84 real velec,felec,velecsum,facel,crf,krf,krf2;
87 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
91 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
99 jindex = nlist->jindex;
101 shiftidx = nlist->shift;
103 shiftvec = fr->shift_vec[0];
104 fshift = fr->fshift[0];
106 charge = mdatoms->chargeA;
107 nvdwtype = fr->ntype;
109 vdwtype = mdatoms->typeA;
111 sh_ewald = fr->ic->sh_ewald;
112 ewtab = fr->ic->tabq_coul_FDV0;
113 ewtabscale = fr->ic->tabq_scale;
114 ewtabhalfspace = 0.5/ewtabscale;
116 /* Setup water-specific parameters */
117 inr = nlist->iinr[0];
118 iq1 = facel*charge[inr+1];
119 iq2 = facel*charge[inr+2];
120 iq3 = facel*charge[inr+3];
121 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
123 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
124 rcutoff = fr->rcoulomb;
125 rcutoff2 = rcutoff*rcutoff;
127 sh_vdw_invrcut6 = fr->ic->sh_invrc6;
133 /* Start outer loop over neighborlists */
134 for(iidx=0; iidx<nri; iidx++)
136 /* Load shift vector for this list */
137 i_shift_offset = DIM*shiftidx[iidx];
138 shX = shiftvec[i_shift_offset+XX];
139 shY = shiftvec[i_shift_offset+YY];
140 shZ = shiftvec[i_shift_offset+ZZ];
142 /* Load limits for loop over neighbors */
143 j_index_start = jindex[iidx];
144 j_index_end = jindex[iidx+1];
146 /* Get outer coordinate index */
148 i_coord_offset = DIM*inr;
150 /* Load i particle coords and add shift vector */
151 ix0 = shX + x[i_coord_offset+DIM*0+XX];
152 iy0 = shY + x[i_coord_offset+DIM*0+YY];
153 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
154 ix1 = shX + x[i_coord_offset+DIM*1+XX];
155 iy1 = shY + x[i_coord_offset+DIM*1+YY];
156 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
157 ix2 = shX + x[i_coord_offset+DIM*2+XX];
158 iy2 = shY + x[i_coord_offset+DIM*2+YY];
159 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
160 ix3 = shX + x[i_coord_offset+DIM*3+XX];
161 iy3 = shY + x[i_coord_offset+DIM*3+YY];
162 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
177 /* Reset potential sums */
181 /* Start inner kernel loop */
182 for(jidx=j_index_start; jidx<j_index_end; jidx++)
184 /* Get j neighbor index, and coordinate index */
186 j_coord_offset = DIM*jnr;
188 /* load j atom coordinates */
189 jx0 = x[j_coord_offset+DIM*0+XX];
190 jy0 = x[j_coord_offset+DIM*0+YY];
191 jz0 = x[j_coord_offset+DIM*0+ZZ];
193 /* Calculate displacement vector */
207 /* Calculate squared distance and things based on it */
208 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
209 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
210 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
211 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
213 rinv10 = gmx_invsqrt(rsq10);
214 rinv20 = gmx_invsqrt(rsq20);
215 rinv30 = gmx_invsqrt(rsq30);
217 rinvsq00 = 1.0/rsq00;
218 rinvsq10 = rinv10*rinv10;
219 rinvsq20 = rinv20*rinv20;
220 rinvsq30 = rinv30*rinv30;
222 /* Load parameters for j particles */
224 vdwjidx0 = 2*vdwtype[jnr+0];
226 /**************************
227 * CALCULATE INTERACTIONS *
228 **************************/
233 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
234 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
236 /* LENNARD-JONES DISPERSION/REPULSION */
238 rinvsix = rinvsq00*rinvsq00*rinvsq00;
239 vvdw6 = c6_00*rinvsix;
240 vvdw12 = c12_00*rinvsix*rinvsix;
241 vvdw = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
242 fvdw = (vvdw12-vvdw6)*rinvsq00;
244 /* Update potential sums from outer loop */
249 /* Calculate temporary vectorial force */
254 /* Update vectorial force */
258 f[j_coord_offset+DIM*0+XX] -= tx;
259 f[j_coord_offset+DIM*0+YY] -= ty;
260 f[j_coord_offset+DIM*0+ZZ] -= tz;
264 /**************************
265 * CALCULATE INTERACTIONS *
266 **************************/
275 /* EWALD ELECTROSTATICS */
277 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
278 ewrt = r10*ewtabscale;
282 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
283 velec = qq10*((rinv10-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
284 felec = qq10*rinv10*(rinvsq10-felec);
286 /* Update potential sums from outer loop */
291 /* Calculate temporary vectorial force */
296 /* Update vectorial force */
300 f[j_coord_offset+DIM*0+XX] -= tx;
301 f[j_coord_offset+DIM*0+YY] -= ty;
302 f[j_coord_offset+DIM*0+ZZ] -= tz;
306 /**************************
307 * CALCULATE INTERACTIONS *
308 **************************/
317 /* EWALD ELECTROSTATICS */
319 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
320 ewrt = r20*ewtabscale;
324 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
325 velec = qq20*((rinv20-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
326 felec = qq20*rinv20*(rinvsq20-felec);
328 /* Update potential sums from outer loop */
333 /* Calculate temporary vectorial force */
338 /* Update vectorial force */
342 f[j_coord_offset+DIM*0+XX] -= tx;
343 f[j_coord_offset+DIM*0+YY] -= ty;
344 f[j_coord_offset+DIM*0+ZZ] -= tz;
348 /**************************
349 * CALCULATE INTERACTIONS *
350 **************************/
359 /* EWALD ELECTROSTATICS */
361 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
362 ewrt = r30*ewtabscale;
366 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
367 velec = qq30*((rinv30-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
368 felec = qq30*rinv30*(rinvsq30-felec);
370 /* Update potential sums from outer loop */
375 /* Calculate temporary vectorial force */
380 /* Update vectorial force */
384 f[j_coord_offset+DIM*0+XX] -= tx;
385 f[j_coord_offset+DIM*0+YY] -= ty;
386 f[j_coord_offset+DIM*0+ZZ] -= tz;
390 /* Inner loop uses 163 flops */
392 /* End of innermost loop */
395 f[i_coord_offset+DIM*0+XX] += fix0;
396 f[i_coord_offset+DIM*0+YY] += fiy0;
397 f[i_coord_offset+DIM*0+ZZ] += fiz0;
401 f[i_coord_offset+DIM*1+XX] += fix1;
402 f[i_coord_offset+DIM*1+YY] += fiy1;
403 f[i_coord_offset+DIM*1+ZZ] += fiz1;
407 f[i_coord_offset+DIM*2+XX] += fix2;
408 f[i_coord_offset+DIM*2+YY] += fiy2;
409 f[i_coord_offset+DIM*2+ZZ] += fiz2;
413 f[i_coord_offset+DIM*3+XX] += fix3;
414 f[i_coord_offset+DIM*3+YY] += fiy3;
415 f[i_coord_offset+DIM*3+ZZ] += fiz3;
419 fshift[i_shift_offset+XX] += tx;
420 fshift[i_shift_offset+YY] += ty;
421 fshift[i_shift_offset+ZZ] += tz;
424 /* Update potential energies */
425 kernel_data->energygrp_elec[ggid] += velecsum;
426 kernel_data->energygrp_vdw[ggid] += vvdwsum;
428 /* Increment number of inner iterations */
429 inneriter += j_index_end - j_index_start;
431 /* Outer loop uses 41 flops */
434 /* Increment number of outer iterations */
437 /* Update outer/inner flops */
439 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*163);
442 * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_F_c
443 * Electrostatics interaction: Ewald
444 * VdW interaction: LennardJones
445 * Geometry: Water4-Particle
446 * Calculate force/pot: Force
449 nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_F_c
450 (t_nblist * gmx_restrict nlist,
451 rvec * gmx_restrict xx,
452 rvec * gmx_restrict ff,
453 t_forcerec * gmx_restrict fr,
454 t_mdatoms * gmx_restrict mdatoms,
455 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
456 t_nrnb * gmx_restrict nrnb)
458 int i_shift_offset,i_coord_offset,j_coord_offset;
459 int j_index_start,j_index_end;
460 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
461 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
462 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
463 real *shiftvec,*fshift,*x,*f;
465 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
467 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
469 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
471 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
473 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
474 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
475 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
476 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
477 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
478 real velec,felec,velecsum,facel,crf,krf,krf2;
481 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
485 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
493 jindex = nlist->jindex;
495 shiftidx = nlist->shift;
497 shiftvec = fr->shift_vec[0];
498 fshift = fr->fshift[0];
500 charge = mdatoms->chargeA;
501 nvdwtype = fr->ntype;
503 vdwtype = mdatoms->typeA;
505 sh_ewald = fr->ic->sh_ewald;
506 ewtab = fr->ic->tabq_coul_F;
507 ewtabscale = fr->ic->tabq_scale;
508 ewtabhalfspace = 0.5/ewtabscale;
510 /* Setup water-specific parameters */
511 inr = nlist->iinr[0];
512 iq1 = facel*charge[inr+1];
513 iq2 = facel*charge[inr+2];
514 iq3 = facel*charge[inr+3];
515 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
517 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
518 rcutoff = fr->rcoulomb;
519 rcutoff2 = rcutoff*rcutoff;
521 sh_vdw_invrcut6 = fr->ic->sh_invrc6;
527 /* Start outer loop over neighborlists */
528 for(iidx=0; iidx<nri; iidx++)
530 /* Load shift vector for this list */
531 i_shift_offset = DIM*shiftidx[iidx];
532 shX = shiftvec[i_shift_offset+XX];
533 shY = shiftvec[i_shift_offset+YY];
534 shZ = shiftvec[i_shift_offset+ZZ];
536 /* Load limits for loop over neighbors */
537 j_index_start = jindex[iidx];
538 j_index_end = jindex[iidx+1];
540 /* Get outer coordinate index */
542 i_coord_offset = DIM*inr;
544 /* Load i particle coords and add shift vector */
545 ix0 = shX + x[i_coord_offset+DIM*0+XX];
546 iy0 = shY + x[i_coord_offset+DIM*0+YY];
547 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
548 ix1 = shX + x[i_coord_offset+DIM*1+XX];
549 iy1 = shY + x[i_coord_offset+DIM*1+YY];
550 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
551 ix2 = shX + x[i_coord_offset+DIM*2+XX];
552 iy2 = shY + x[i_coord_offset+DIM*2+YY];
553 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
554 ix3 = shX + x[i_coord_offset+DIM*3+XX];
555 iy3 = shY + x[i_coord_offset+DIM*3+YY];
556 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
571 /* Start inner kernel loop */
572 for(jidx=j_index_start; jidx<j_index_end; jidx++)
574 /* Get j neighbor index, and coordinate index */
576 j_coord_offset = DIM*jnr;
578 /* load j atom coordinates */
579 jx0 = x[j_coord_offset+DIM*0+XX];
580 jy0 = x[j_coord_offset+DIM*0+YY];
581 jz0 = x[j_coord_offset+DIM*0+ZZ];
583 /* Calculate displacement vector */
597 /* Calculate squared distance and things based on it */
598 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
599 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
600 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
601 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
603 rinv10 = gmx_invsqrt(rsq10);
604 rinv20 = gmx_invsqrt(rsq20);
605 rinv30 = gmx_invsqrt(rsq30);
607 rinvsq00 = 1.0/rsq00;
608 rinvsq10 = rinv10*rinv10;
609 rinvsq20 = rinv20*rinv20;
610 rinvsq30 = rinv30*rinv30;
612 /* Load parameters for j particles */
614 vdwjidx0 = 2*vdwtype[jnr+0];
616 /**************************
617 * CALCULATE INTERACTIONS *
618 **************************/
623 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
624 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
626 /* LENNARD-JONES DISPERSION/REPULSION */
628 rinvsix = rinvsq00*rinvsq00*rinvsq00;
629 fvdw = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
633 /* Calculate temporary vectorial force */
638 /* Update vectorial force */
642 f[j_coord_offset+DIM*0+XX] -= tx;
643 f[j_coord_offset+DIM*0+YY] -= ty;
644 f[j_coord_offset+DIM*0+ZZ] -= tz;
648 /**************************
649 * CALCULATE INTERACTIONS *
650 **************************/
659 /* EWALD ELECTROSTATICS */
661 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
662 ewrt = r10*ewtabscale;
665 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
666 felec = qq10*rinv10*(rinvsq10-felec);
670 /* Calculate temporary vectorial force */
675 /* Update vectorial force */
679 f[j_coord_offset+DIM*0+XX] -= tx;
680 f[j_coord_offset+DIM*0+YY] -= ty;
681 f[j_coord_offset+DIM*0+ZZ] -= tz;
685 /**************************
686 * CALCULATE INTERACTIONS *
687 **************************/
696 /* EWALD ELECTROSTATICS */
698 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
699 ewrt = r20*ewtabscale;
702 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
703 felec = qq20*rinv20*(rinvsq20-felec);
707 /* Calculate temporary vectorial force */
712 /* Update vectorial force */
716 f[j_coord_offset+DIM*0+XX] -= tx;
717 f[j_coord_offset+DIM*0+YY] -= ty;
718 f[j_coord_offset+DIM*0+ZZ] -= tz;
722 /**************************
723 * CALCULATE INTERACTIONS *
724 **************************/
733 /* EWALD ELECTROSTATICS */
735 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
736 ewrt = r30*ewtabscale;
739 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
740 felec = qq30*rinv30*(rinvsq30-felec);
744 /* Calculate temporary vectorial force */
749 /* Update vectorial force */
753 f[j_coord_offset+DIM*0+XX] -= tx;
754 f[j_coord_offset+DIM*0+YY] -= ty;
755 f[j_coord_offset+DIM*0+ZZ] -= tz;
759 /* Inner loop uses 129 flops */
761 /* End of innermost loop */
764 f[i_coord_offset+DIM*0+XX] += fix0;
765 f[i_coord_offset+DIM*0+YY] += fiy0;
766 f[i_coord_offset+DIM*0+ZZ] += fiz0;
770 f[i_coord_offset+DIM*1+XX] += fix1;
771 f[i_coord_offset+DIM*1+YY] += fiy1;
772 f[i_coord_offset+DIM*1+ZZ] += fiz1;
776 f[i_coord_offset+DIM*2+XX] += fix2;
777 f[i_coord_offset+DIM*2+YY] += fiy2;
778 f[i_coord_offset+DIM*2+ZZ] += fiz2;
782 f[i_coord_offset+DIM*3+XX] += fix3;
783 f[i_coord_offset+DIM*3+YY] += fiy3;
784 f[i_coord_offset+DIM*3+ZZ] += fiz3;
788 fshift[i_shift_offset+XX] += tx;
789 fshift[i_shift_offset+YY] += ty;
790 fshift[i_shift_offset+ZZ] += tz;
792 /* Increment number of inner iterations */
793 inneriter += j_index_end - j_index_start;
795 /* Outer loop uses 39 flops */
798 /* Increment number of outer iterations */
801 /* Update outer/inner flops */
803 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*129);