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_VdwBhamSh_GeomW4W4_VF_c
49 * Electrostatics interaction: Ewald
50 * VdW interaction: Buckingham
51 * Geometry: Water4-Water4
52 * Calculate force/pot: PotentialAndForce
55 nb_kernel_ElecEwSh_VdwBhamSh_GeomW4W4_VF_c
56 (t_nblist * gmx_restrict nlist,
57 rvec * gmx_restrict xx,
58 rvec * gmx_restrict ff,
59 t_forcerec * gmx_restrict fr,
60 t_mdatoms * gmx_restrict mdatoms,
61 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
62 t_nrnb * gmx_restrict nrnb)
64 int i_shift_offset,i_coord_offset,j_coord_offset;
65 int j_index_start,j_index_end;
66 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
67 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
68 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
69 real *shiftvec,*fshift,*x,*f;
71 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
73 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
75 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
77 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
79 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
81 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
83 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
85 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
86 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
87 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
88 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
89 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
90 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
91 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
92 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
93 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
94 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
95 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
96 real velec,felec,velecsum,facel,crf,krf,krf2;
99 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
103 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
111 jindex = nlist->jindex;
113 shiftidx = nlist->shift;
115 shiftvec = fr->shift_vec[0];
116 fshift = fr->fshift[0];
118 charge = mdatoms->chargeA;
119 nvdwtype = fr->ntype;
121 vdwtype = mdatoms->typeA;
123 sh_ewald = fr->ic->sh_ewald;
124 ewtab = fr->ic->tabq_coul_FDV0;
125 ewtabscale = fr->ic->tabq_scale;
126 ewtabhalfspace = 0.5/ewtabscale;
128 /* Setup water-specific parameters */
129 inr = nlist->iinr[0];
130 iq1 = facel*charge[inr+1];
131 iq2 = facel*charge[inr+2];
132 iq3 = facel*charge[inr+3];
133 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
138 vdwjidx0 = 3*vdwtype[inr+0];
139 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
140 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
141 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
152 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
153 rcutoff = fr->rcoulomb;
154 rcutoff2 = rcutoff*rcutoff;
156 sh_vdw_invrcut6 = fr->ic->sh_invrc6;
162 /* Start outer loop over neighborlists */
163 for(iidx=0; iidx<nri; iidx++)
165 /* Load shift vector for this list */
166 i_shift_offset = DIM*shiftidx[iidx];
167 shX = shiftvec[i_shift_offset+XX];
168 shY = shiftvec[i_shift_offset+YY];
169 shZ = shiftvec[i_shift_offset+ZZ];
171 /* Load limits for loop over neighbors */
172 j_index_start = jindex[iidx];
173 j_index_end = jindex[iidx+1];
175 /* Get outer coordinate index */
177 i_coord_offset = DIM*inr;
179 /* Load i particle coords and add shift vector */
180 ix0 = shX + x[i_coord_offset+DIM*0+XX];
181 iy0 = shY + x[i_coord_offset+DIM*0+YY];
182 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
183 ix1 = shX + x[i_coord_offset+DIM*1+XX];
184 iy1 = shY + x[i_coord_offset+DIM*1+YY];
185 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
186 ix2 = shX + x[i_coord_offset+DIM*2+XX];
187 iy2 = shY + x[i_coord_offset+DIM*2+YY];
188 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
189 ix3 = shX + x[i_coord_offset+DIM*3+XX];
190 iy3 = shY + x[i_coord_offset+DIM*3+YY];
191 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
206 /* Reset potential sums */
210 /* Start inner kernel loop */
211 for(jidx=j_index_start; jidx<j_index_end; jidx++)
213 /* Get j neighbor index, and coordinate index */
215 j_coord_offset = DIM*jnr;
217 /* load j atom coordinates */
218 jx0 = x[j_coord_offset+DIM*0+XX];
219 jy0 = x[j_coord_offset+DIM*0+YY];
220 jz0 = x[j_coord_offset+DIM*0+ZZ];
221 jx1 = x[j_coord_offset+DIM*1+XX];
222 jy1 = x[j_coord_offset+DIM*1+YY];
223 jz1 = x[j_coord_offset+DIM*1+ZZ];
224 jx2 = x[j_coord_offset+DIM*2+XX];
225 jy2 = x[j_coord_offset+DIM*2+YY];
226 jz2 = x[j_coord_offset+DIM*2+ZZ];
227 jx3 = x[j_coord_offset+DIM*3+XX];
228 jy3 = x[j_coord_offset+DIM*3+YY];
229 jz3 = x[j_coord_offset+DIM*3+ZZ];
231 /* Calculate displacement vector */
263 /* Calculate squared distance and things based on it */
264 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
265 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
266 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
267 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
268 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
269 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
270 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
271 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
272 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
273 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
275 rinv00 = gmx_invsqrt(rsq00);
276 rinv11 = gmx_invsqrt(rsq11);
277 rinv12 = gmx_invsqrt(rsq12);
278 rinv13 = gmx_invsqrt(rsq13);
279 rinv21 = gmx_invsqrt(rsq21);
280 rinv22 = gmx_invsqrt(rsq22);
281 rinv23 = gmx_invsqrt(rsq23);
282 rinv31 = gmx_invsqrt(rsq31);
283 rinv32 = gmx_invsqrt(rsq32);
284 rinv33 = gmx_invsqrt(rsq33);
286 rinvsq00 = rinv00*rinv00;
287 rinvsq11 = rinv11*rinv11;
288 rinvsq12 = rinv12*rinv12;
289 rinvsq13 = rinv13*rinv13;
290 rinvsq21 = rinv21*rinv21;
291 rinvsq22 = rinv22*rinv22;
292 rinvsq23 = rinv23*rinv23;
293 rinvsq31 = rinv31*rinv31;
294 rinvsq32 = rinv32*rinv32;
295 rinvsq33 = rinv33*rinv33;
297 /**************************
298 * CALCULATE INTERACTIONS *
299 **************************/
306 /* BUCKINGHAM DISPERSION/REPULSION */
307 rinvsix = rinvsq00*rinvsq00*rinvsq00;
308 vvdw6 = c6_00*rinvsix;
310 vvdwexp = cexp1_00*exp(-br);
311 vvdw = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
312 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
314 /* Update potential sums from outer loop */
319 /* Calculate temporary vectorial force */
324 /* Update vectorial force */
328 f[j_coord_offset+DIM*0+XX] -= tx;
329 f[j_coord_offset+DIM*0+YY] -= ty;
330 f[j_coord_offset+DIM*0+ZZ] -= tz;
334 /**************************
335 * CALCULATE INTERACTIONS *
336 **************************/
343 /* EWALD ELECTROSTATICS */
345 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
346 ewrt = r11*ewtabscale;
350 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
351 velec = qq11*((rinv11-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
352 felec = qq11*rinv11*(rinvsq11-felec);
354 /* Update potential sums from outer loop */
359 /* Calculate temporary vectorial force */
364 /* Update vectorial force */
368 f[j_coord_offset+DIM*1+XX] -= tx;
369 f[j_coord_offset+DIM*1+YY] -= ty;
370 f[j_coord_offset+DIM*1+ZZ] -= tz;
374 /**************************
375 * CALCULATE INTERACTIONS *
376 **************************/
383 /* EWALD ELECTROSTATICS */
385 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
386 ewrt = r12*ewtabscale;
390 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
391 velec = qq12*((rinv12-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
392 felec = qq12*rinv12*(rinvsq12-felec);
394 /* Update potential sums from outer loop */
399 /* Calculate temporary vectorial force */
404 /* Update vectorial force */
408 f[j_coord_offset+DIM*2+XX] -= tx;
409 f[j_coord_offset+DIM*2+YY] -= ty;
410 f[j_coord_offset+DIM*2+ZZ] -= tz;
414 /**************************
415 * CALCULATE INTERACTIONS *
416 **************************/
423 /* EWALD ELECTROSTATICS */
425 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
426 ewrt = r13*ewtabscale;
430 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
431 velec = qq13*((rinv13-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
432 felec = qq13*rinv13*(rinvsq13-felec);
434 /* Update potential sums from outer loop */
439 /* Calculate temporary vectorial force */
444 /* Update vectorial force */
448 f[j_coord_offset+DIM*3+XX] -= tx;
449 f[j_coord_offset+DIM*3+YY] -= ty;
450 f[j_coord_offset+DIM*3+ZZ] -= tz;
454 /**************************
455 * CALCULATE INTERACTIONS *
456 **************************/
463 /* EWALD ELECTROSTATICS */
465 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
466 ewrt = r21*ewtabscale;
470 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
471 velec = qq21*((rinv21-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
472 felec = qq21*rinv21*(rinvsq21-felec);
474 /* Update potential sums from outer loop */
479 /* Calculate temporary vectorial force */
484 /* Update vectorial force */
488 f[j_coord_offset+DIM*1+XX] -= tx;
489 f[j_coord_offset+DIM*1+YY] -= ty;
490 f[j_coord_offset+DIM*1+ZZ] -= tz;
494 /**************************
495 * CALCULATE INTERACTIONS *
496 **************************/
503 /* EWALD ELECTROSTATICS */
505 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
506 ewrt = r22*ewtabscale;
510 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
511 velec = qq22*((rinv22-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
512 felec = qq22*rinv22*(rinvsq22-felec);
514 /* Update potential sums from outer loop */
519 /* Calculate temporary vectorial force */
524 /* Update vectorial force */
528 f[j_coord_offset+DIM*2+XX] -= tx;
529 f[j_coord_offset+DIM*2+YY] -= ty;
530 f[j_coord_offset+DIM*2+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 = r23*ewtabscale;
550 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
551 velec = qq23*((rinv23-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
552 felec = qq23*rinv23*(rinvsq23-felec);
554 /* Update potential sums from outer loop */
559 /* Calculate temporary vectorial force */
564 /* Update vectorial force */
568 f[j_coord_offset+DIM*3+XX] -= tx;
569 f[j_coord_offset+DIM*3+YY] -= ty;
570 f[j_coord_offset+DIM*3+ZZ] -= tz;
574 /**************************
575 * CALCULATE INTERACTIONS *
576 **************************/
583 /* EWALD ELECTROSTATICS */
585 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
586 ewrt = r31*ewtabscale;
590 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
591 velec = qq31*((rinv31-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
592 felec = qq31*rinv31*(rinvsq31-felec);
594 /* Update potential sums from outer loop */
599 /* Calculate temporary vectorial force */
604 /* Update vectorial force */
608 f[j_coord_offset+DIM*1+XX] -= tx;
609 f[j_coord_offset+DIM*1+YY] -= ty;
610 f[j_coord_offset+DIM*1+ZZ] -= tz;
614 /**************************
615 * CALCULATE INTERACTIONS *
616 **************************/
623 /* EWALD ELECTROSTATICS */
625 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
626 ewrt = r32*ewtabscale;
630 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
631 velec = qq32*((rinv32-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
632 felec = qq32*rinv32*(rinvsq32-felec);
634 /* Update potential sums from outer loop */
639 /* Calculate temporary vectorial force */
644 /* Update vectorial force */
648 f[j_coord_offset+DIM*2+XX] -= tx;
649 f[j_coord_offset+DIM*2+YY] -= ty;
650 f[j_coord_offset+DIM*2+ZZ] -= tz;
654 /**************************
655 * CALCULATE INTERACTIONS *
656 **************************/
663 /* EWALD ELECTROSTATICS */
665 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
666 ewrt = r33*ewtabscale;
670 felec = ewtab[ewitab]+eweps*ewtab[ewitab+1];
671 velec = qq33*((rinv33-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
672 felec = qq33*rinv33*(rinvsq33-felec);
674 /* Update potential sums from outer loop */
679 /* Calculate temporary vectorial force */
684 /* Update vectorial force */
688 f[j_coord_offset+DIM*3+XX] -= tx;
689 f[j_coord_offset+DIM*3+YY] -= ty;
690 f[j_coord_offset+DIM*3+ZZ] -= tz;
694 /* Inner loop uses 461 flops */
696 /* End of innermost loop */
699 f[i_coord_offset+DIM*0+XX] += fix0;
700 f[i_coord_offset+DIM*0+YY] += fiy0;
701 f[i_coord_offset+DIM*0+ZZ] += fiz0;
705 f[i_coord_offset+DIM*1+XX] += fix1;
706 f[i_coord_offset+DIM*1+YY] += fiy1;
707 f[i_coord_offset+DIM*1+ZZ] += fiz1;
711 f[i_coord_offset+DIM*2+XX] += fix2;
712 f[i_coord_offset+DIM*2+YY] += fiy2;
713 f[i_coord_offset+DIM*2+ZZ] += fiz2;
717 f[i_coord_offset+DIM*3+XX] += fix3;
718 f[i_coord_offset+DIM*3+YY] += fiy3;
719 f[i_coord_offset+DIM*3+ZZ] += fiz3;
723 fshift[i_shift_offset+XX] += tx;
724 fshift[i_shift_offset+YY] += ty;
725 fshift[i_shift_offset+ZZ] += tz;
728 /* Update potential energies */
729 kernel_data->energygrp_elec[ggid] += velecsum;
730 kernel_data->energygrp_vdw[ggid] += vvdwsum;
732 /* Increment number of inner iterations */
733 inneriter += j_index_end - j_index_start;
735 /* Outer loop uses 41 flops */
738 /* Increment number of outer iterations */
741 /* Update outer/inner flops */
743 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*41 + inneriter*461);
746 * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwBhamSh_GeomW4W4_F_c
747 * Electrostatics interaction: Ewald
748 * VdW interaction: Buckingham
749 * Geometry: Water4-Water4
750 * Calculate force/pot: Force
753 nb_kernel_ElecEwSh_VdwBhamSh_GeomW4W4_F_c
754 (t_nblist * gmx_restrict nlist,
755 rvec * gmx_restrict xx,
756 rvec * gmx_restrict ff,
757 t_forcerec * gmx_restrict fr,
758 t_mdatoms * gmx_restrict mdatoms,
759 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
760 t_nrnb * gmx_restrict nrnb)
762 int i_shift_offset,i_coord_offset,j_coord_offset;
763 int j_index_start,j_index_end;
764 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
765 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
766 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
767 real *shiftvec,*fshift,*x,*f;
769 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
771 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
773 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
775 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
777 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
779 real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
781 real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
783 real jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
784 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
785 real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
786 real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
787 real dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13,cexp1_13,cexp2_13;
788 real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
789 real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
790 real dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23,cexp1_23,cexp2_23;
791 real dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31,cexp1_31,cexp2_31;
792 real dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32,cexp1_32,cexp2_32;
793 real dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33,cexp1_33,cexp2_33;
794 real velec,felec,velecsum,facel,crf,krf,krf2;
797 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
801 real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
809 jindex = nlist->jindex;
811 shiftidx = nlist->shift;
813 shiftvec = fr->shift_vec[0];
814 fshift = fr->fshift[0];
816 charge = mdatoms->chargeA;
817 nvdwtype = fr->ntype;
819 vdwtype = mdatoms->typeA;
821 sh_ewald = fr->ic->sh_ewald;
822 ewtab = fr->ic->tabq_coul_F;
823 ewtabscale = fr->ic->tabq_scale;
824 ewtabhalfspace = 0.5/ewtabscale;
826 /* Setup water-specific parameters */
827 inr = nlist->iinr[0];
828 iq1 = facel*charge[inr+1];
829 iq2 = facel*charge[inr+2];
830 iq3 = facel*charge[inr+3];
831 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
836 vdwjidx0 = 3*vdwtype[inr+0];
837 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
838 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
839 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
850 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
851 rcutoff = fr->rcoulomb;
852 rcutoff2 = rcutoff*rcutoff;
854 sh_vdw_invrcut6 = fr->ic->sh_invrc6;
860 /* Start outer loop over neighborlists */
861 for(iidx=0; iidx<nri; iidx++)
863 /* Load shift vector for this list */
864 i_shift_offset = DIM*shiftidx[iidx];
865 shX = shiftvec[i_shift_offset+XX];
866 shY = shiftvec[i_shift_offset+YY];
867 shZ = shiftvec[i_shift_offset+ZZ];
869 /* Load limits for loop over neighbors */
870 j_index_start = jindex[iidx];
871 j_index_end = jindex[iidx+1];
873 /* Get outer coordinate index */
875 i_coord_offset = DIM*inr;
877 /* Load i particle coords and add shift vector */
878 ix0 = shX + x[i_coord_offset+DIM*0+XX];
879 iy0 = shY + x[i_coord_offset+DIM*0+YY];
880 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
881 ix1 = shX + x[i_coord_offset+DIM*1+XX];
882 iy1 = shY + x[i_coord_offset+DIM*1+YY];
883 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
884 ix2 = shX + x[i_coord_offset+DIM*2+XX];
885 iy2 = shY + x[i_coord_offset+DIM*2+YY];
886 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
887 ix3 = shX + x[i_coord_offset+DIM*3+XX];
888 iy3 = shY + x[i_coord_offset+DIM*3+YY];
889 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
904 /* Start inner kernel loop */
905 for(jidx=j_index_start; jidx<j_index_end; jidx++)
907 /* Get j neighbor index, and coordinate index */
909 j_coord_offset = DIM*jnr;
911 /* load j atom coordinates */
912 jx0 = x[j_coord_offset+DIM*0+XX];
913 jy0 = x[j_coord_offset+DIM*0+YY];
914 jz0 = x[j_coord_offset+DIM*0+ZZ];
915 jx1 = x[j_coord_offset+DIM*1+XX];
916 jy1 = x[j_coord_offset+DIM*1+YY];
917 jz1 = x[j_coord_offset+DIM*1+ZZ];
918 jx2 = x[j_coord_offset+DIM*2+XX];
919 jy2 = x[j_coord_offset+DIM*2+YY];
920 jz2 = x[j_coord_offset+DIM*2+ZZ];
921 jx3 = x[j_coord_offset+DIM*3+XX];
922 jy3 = x[j_coord_offset+DIM*3+YY];
923 jz3 = x[j_coord_offset+DIM*3+ZZ];
925 /* Calculate displacement vector */
957 /* Calculate squared distance and things based on it */
958 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
959 rsq11 = dx11*dx11+dy11*dy11+dz11*dz11;
960 rsq12 = dx12*dx12+dy12*dy12+dz12*dz12;
961 rsq13 = dx13*dx13+dy13*dy13+dz13*dz13;
962 rsq21 = dx21*dx21+dy21*dy21+dz21*dz21;
963 rsq22 = dx22*dx22+dy22*dy22+dz22*dz22;
964 rsq23 = dx23*dx23+dy23*dy23+dz23*dz23;
965 rsq31 = dx31*dx31+dy31*dy31+dz31*dz31;
966 rsq32 = dx32*dx32+dy32*dy32+dz32*dz32;
967 rsq33 = dx33*dx33+dy33*dy33+dz33*dz33;
969 rinv00 = gmx_invsqrt(rsq00);
970 rinv11 = gmx_invsqrt(rsq11);
971 rinv12 = gmx_invsqrt(rsq12);
972 rinv13 = gmx_invsqrt(rsq13);
973 rinv21 = gmx_invsqrt(rsq21);
974 rinv22 = gmx_invsqrt(rsq22);
975 rinv23 = gmx_invsqrt(rsq23);
976 rinv31 = gmx_invsqrt(rsq31);
977 rinv32 = gmx_invsqrt(rsq32);
978 rinv33 = gmx_invsqrt(rsq33);
980 rinvsq00 = rinv00*rinv00;
981 rinvsq11 = rinv11*rinv11;
982 rinvsq12 = rinv12*rinv12;
983 rinvsq13 = rinv13*rinv13;
984 rinvsq21 = rinv21*rinv21;
985 rinvsq22 = rinv22*rinv22;
986 rinvsq23 = rinv23*rinv23;
987 rinvsq31 = rinv31*rinv31;
988 rinvsq32 = rinv32*rinv32;
989 rinvsq33 = rinv33*rinv33;
991 /**************************
992 * CALCULATE INTERACTIONS *
993 **************************/
1000 /* BUCKINGHAM DISPERSION/REPULSION */
1001 rinvsix = rinvsq00*rinvsq00*rinvsq00;
1002 vvdw6 = c6_00*rinvsix;
1004 vvdwexp = cexp1_00*exp(-br);
1005 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
1009 /* Calculate temporary vectorial force */
1014 /* Update vectorial force */
1018 f[j_coord_offset+DIM*0+XX] -= tx;
1019 f[j_coord_offset+DIM*0+YY] -= ty;
1020 f[j_coord_offset+DIM*0+ZZ] -= tz;
1024 /**************************
1025 * CALCULATE INTERACTIONS *
1026 **************************/
1033 /* EWALD ELECTROSTATICS */
1035 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1036 ewrt = r11*ewtabscale;
1038 eweps = ewrt-ewitab;
1039 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1040 felec = qq11*rinv11*(rinvsq11-felec);
1044 /* Calculate temporary vectorial force */
1049 /* Update vectorial force */
1053 f[j_coord_offset+DIM*1+XX] -= tx;
1054 f[j_coord_offset+DIM*1+YY] -= ty;
1055 f[j_coord_offset+DIM*1+ZZ] -= tz;
1059 /**************************
1060 * CALCULATE INTERACTIONS *
1061 **************************/
1068 /* EWALD ELECTROSTATICS */
1070 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1071 ewrt = r12*ewtabscale;
1073 eweps = ewrt-ewitab;
1074 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1075 felec = qq12*rinv12*(rinvsq12-felec);
1079 /* Calculate temporary vectorial force */
1084 /* Update vectorial force */
1088 f[j_coord_offset+DIM*2+XX] -= tx;
1089 f[j_coord_offset+DIM*2+YY] -= ty;
1090 f[j_coord_offset+DIM*2+ZZ] -= tz;
1094 /**************************
1095 * CALCULATE INTERACTIONS *
1096 **************************/
1103 /* EWALD ELECTROSTATICS */
1105 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1106 ewrt = r13*ewtabscale;
1108 eweps = ewrt-ewitab;
1109 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1110 felec = qq13*rinv13*(rinvsq13-felec);
1114 /* Calculate temporary vectorial force */
1119 /* Update vectorial force */
1123 f[j_coord_offset+DIM*3+XX] -= tx;
1124 f[j_coord_offset+DIM*3+YY] -= ty;
1125 f[j_coord_offset+DIM*3+ZZ] -= tz;
1129 /**************************
1130 * CALCULATE INTERACTIONS *
1131 **************************/
1138 /* EWALD ELECTROSTATICS */
1140 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1141 ewrt = r21*ewtabscale;
1143 eweps = ewrt-ewitab;
1144 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1145 felec = qq21*rinv21*(rinvsq21-felec);
1149 /* Calculate temporary vectorial force */
1154 /* Update vectorial force */
1158 f[j_coord_offset+DIM*1+XX] -= tx;
1159 f[j_coord_offset+DIM*1+YY] -= ty;
1160 f[j_coord_offset+DIM*1+ZZ] -= tz;
1164 /**************************
1165 * CALCULATE INTERACTIONS *
1166 **************************/
1173 /* EWALD ELECTROSTATICS */
1175 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1176 ewrt = r22*ewtabscale;
1178 eweps = ewrt-ewitab;
1179 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1180 felec = qq22*rinv22*(rinvsq22-felec);
1184 /* Calculate temporary vectorial force */
1189 /* Update vectorial force */
1193 f[j_coord_offset+DIM*2+XX] -= tx;
1194 f[j_coord_offset+DIM*2+YY] -= ty;
1195 f[j_coord_offset+DIM*2+ZZ] -= tz;
1199 /**************************
1200 * CALCULATE INTERACTIONS *
1201 **************************/
1208 /* EWALD ELECTROSTATICS */
1210 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1211 ewrt = r23*ewtabscale;
1213 eweps = ewrt-ewitab;
1214 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1215 felec = qq23*rinv23*(rinvsq23-felec);
1219 /* Calculate temporary vectorial force */
1224 /* Update vectorial force */
1228 f[j_coord_offset+DIM*3+XX] -= tx;
1229 f[j_coord_offset+DIM*3+YY] -= ty;
1230 f[j_coord_offset+DIM*3+ZZ] -= tz;
1234 /**************************
1235 * CALCULATE INTERACTIONS *
1236 **************************/
1243 /* EWALD ELECTROSTATICS */
1245 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1246 ewrt = r31*ewtabscale;
1248 eweps = ewrt-ewitab;
1249 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1250 felec = qq31*rinv31*(rinvsq31-felec);
1254 /* Calculate temporary vectorial force */
1259 /* Update vectorial force */
1263 f[j_coord_offset+DIM*1+XX] -= tx;
1264 f[j_coord_offset+DIM*1+YY] -= ty;
1265 f[j_coord_offset+DIM*1+ZZ] -= tz;
1269 /**************************
1270 * CALCULATE INTERACTIONS *
1271 **************************/
1278 /* EWALD ELECTROSTATICS */
1280 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1281 ewrt = r32*ewtabscale;
1283 eweps = ewrt-ewitab;
1284 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1285 felec = qq32*rinv32*(rinvsq32-felec);
1289 /* Calculate temporary vectorial force */
1294 /* Update vectorial force */
1298 f[j_coord_offset+DIM*2+XX] -= tx;
1299 f[j_coord_offset+DIM*2+YY] -= ty;
1300 f[j_coord_offset+DIM*2+ZZ] -= tz;
1304 /**************************
1305 * CALCULATE INTERACTIONS *
1306 **************************/
1313 /* EWALD ELECTROSTATICS */
1315 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1316 ewrt = r33*ewtabscale;
1318 eweps = ewrt-ewitab;
1319 felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
1320 felec = qq33*rinv33*(rinvsq33-felec);
1324 /* Calculate temporary vectorial force */
1329 /* Update vectorial force */
1333 f[j_coord_offset+DIM*3+XX] -= tx;
1334 f[j_coord_offset+DIM*3+YY] -= ty;
1335 f[j_coord_offset+DIM*3+ZZ] -= tz;
1339 /* Inner loop uses 355 flops */
1341 /* End of innermost loop */
1344 f[i_coord_offset+DIM*0+XX] += fix0;
1345 f[i_coord_offset+DIM*0+YY] += fiy0;
1346 f[i_coord_offset+DIM*0+ZZ] += fiz0;
1350 f[i_coord_offset+DIM*1+XX] += fix1;
1351 f[i_coord_offset+DIM*1+YY] += fiy1;
1352 f[i_coord_offset+DIM*1+ZZ] += fiz1;
1356 f[i_coord_offset+DIM*2+XX] += fix2;
1357 f[i_coord_offset+DIM*2+YY] += fiy2;
1358 f[i_coord_offset+DIM*2+ZZ] += fiz2;
1362 f[i_coord_offset+DIM*3+XX] += fix3;
1363 f[i_coord_offset+DIM*3+YY] += fiy3;
1364 f[i_coord_offset+DIM*3+ZZ] += fiz3;
1368 fshift[i_shift_offset+XX] += tx;
1369 fshift[i_shift_offset+YY] += ty;
1370 fshift[i_shift_offset+ZZ] += tz;
1372 /* Increment number of inner iterations */
1373 inneriter += j_index_end - j_index_start;
1375 /* Outer loop uses 39 flops */
1378 /* Increment number of outer iterations */
1381 /* Update outer/inner flops */
1383 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*39 + inneriter*355);