2 * Note: this file was generated by the Gromacs c kernel generator.
4 * This source code is part of
8 * Copyright (c) 2001-2012, The GROMACS Development Team
10 * Gromacs is a library for molecular simulation and trajectory analysis,
11 * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
12 * a full list of developers and information, check out http://www.gromacs.org
14 * This program is free software; you can redistribute it and/or modify it under
15 * the terms of the GNU Lesser General Public License as published by the Free
16 * Software Foundation; either version 2 of the License, or (at your option) any
19 * To help fund GROMACS development, we humbly ask that you cite
20 * the papers people have written on it - you can find them on the website.
28 #include "../nb_kernel.h"
29 #include "types/simple.h"
34 * Gromacs nonbonded kernel: nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_VF_c
35 * Electrostatics interaction: Coulomb
36 * VdW interaction: CubicSplineTable
37 * Geometry: Water4-Particle
38 * Calculate force/pot: PotentialAndForce
41 nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_VF_c
42 (t_nblist * gmx_restrict nlist,
43 rvec * gmx_restrict xx,
44 rvec * gmx_restrict ff,
45 t_forcerec * gmx_restrict fr,
46 t_mdatoms * gmx_restrict mdatoms,
47 nb_kernel_data_t * gmx_restrict kernel_data,
48 t_nrnb * gmx_restrict nrnb)
50 int i_shift_offset,i_coord_offset,j_coord_offset;
51 int j_index_start,j_index_end;
52 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
53 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
54 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
55 real *shiftvec,*fshift,*x,*f;
57 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
59 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
61 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
63 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
65 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
66 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
67 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
68 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
69 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
70 real velec,felec,velecsum,facel,crf,krf,krf2;
73 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
77 real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
85 jindex = nlist->jindex;
87 shiftidx = nlist->shift;
89 shiftvec = fr->shift_vec[0];
90 fshift = fr->fshift[0];
92 charge = mdatoms->chargeA;
95 vdwtype = mdatoms->typeA;
97 vftab = kernel_data->table_vdw->data;
98 vftabscale = kernel_data->table_vdw->scale;
100 /* Setup water-specific parameters */
101 inr = nlist->iinr[0];
102 iq1 = facel*charge[inr+1];
103 iq2 = facel*charge[inr+2];
104 iq3 = facel*charge[inr+3];
105 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
110 /* Start outer loop over neighborlists */
111 for(iidx=0; iidx<nri; iidx++)
113 /* Load shift vector for this list */
114 i_shift_offset = DIM*shiftidx[iidx];
115 shX = shiftvec[i_shift_offset+XX];
116 shY = shiftvec[i_shift_offset+YY];
117 shZ = shiftvec[i_shift_offset+ZZ];
119 /* Load limits for loop over neighbors */
120 j_index_start = jindex[iidx];
121 j_index_end = jindex[iidx+1];
123 /* Get outer coordinate index */
125 i_coord_offset = DIM*inr;
127 /* Load i particle coords and add shift vector */
128 ix0 = shX + x[i_coord_offset+DIM*0+XX];
129 iy0 = shY + x[i_coord_offset+DIM*0+YY];
130 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
131 ix1 = shX + x[i_coord_offset+DIM*1+XX];
132 iy1 = shY + x[i_coord_offset+DIM*1+YY];
133 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
134 ix2 = shX + x[i_coord_offset+DIM*2+XX];
135 iy2 = shY + x[i_coord_offset+DIM*2+YY];
136 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
137 ix3 = shX + x[i_coord_offset+DIM*3+XX];
138 iy3 = shY + x[i_coord_offset+DIM*3+YY];
139 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
154 /* Reset potential sums */
158 /* Start inner kernel loop */
159 for(jidx=j_index_start; jidx<j_index_end; jidx++)
161 /* Get j neighbor index, and coordinate index */
163 j_coord_offset = DIM*jnr;
165 /* load j atom coordinates */
166 jx0 = x[j_coord_offset+DIM*0+XX];
167 jy0 = x[j_coord_offset+DIM*0+YY];
168 jz0 = x[j_coord_offset+DIM*0+ZZ];
170 /* Calculate displacement vector */
184 /* Calculate squared distance and things based on it */
185 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
186 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
187 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
188 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
190 rinv00 = gmx_invsqrt(rsq00);
191 rinv10 = gmx_invsqrt(rsq10);
192 rinv20 = gmx_invsqrt(rsq20);
193 rinv30 = gmx_invsqrt(rsq30);
195 rinvsq10 = rinv10*rinv10;
196 rinvsq20 = rinv20*rinv20;
197 rinvsq30 = rinv30*rinv30;
199 /* Load parameters for j particles */
201 vdwjidx0 = 2*vdwtype[jnr+0];
203 /**************************
204 * CALCULATE INTERACTIONS *
205 **************************/
209 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
210 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
212 /* Calculate table index by multiplying r with table scale and truncate to integer */
218 /* CUBIC SPLINE TABLE DISPERSION */
222 Geps = vfeps*vftab[vfitab+2];
223 Heps2 = vfeps*vfeps*vftab[vfitab+3];
227 FF = Fp+Geps+2.0*Heps2;
230 /* CUBIC SPLINE TABLE REPULSION */
233 Geps = vfeps*vftab[vfitab+6];
234 Heps2 = vfeps*vfeps*vftab[vfitab+7];
238 FF = Fp+Geps+2.0*Heps2;
241 fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00;
243 /* Update potential sums from outer loop */
248 /* Calculate temporary vectorial force */
253 /* Update vectorial force */
257 f[j_coord_offset+DIM*0+XX] -= tx;
258 f[j_coord_offset+DIM*0+YY] -= ty;
259 f[j_coord_offset+DIM*0+ZZ] -= tz;
261 /**************************
262 * CALCULATE INTERACTIONS *
263 **************************/
267 /* COULOMB ELECTROSTATICS */
269 felec = velec*rinvsq10;
271 /* Update potential sums from outer loop */
276 /* Calculate temporary vectorial force */
281 /* Update vectorial force */
285 f[j_coord_offset+DIM*0+XX] -= tx;
286 f[j_coord_offset+DIM*0+YY] -= ty;
287 f[j_coord_offset+DIM*0+ZZ] -= tz;
289 /**************************
290 * CALCULATE INTERACTIONS *
291 **************************/
295 /* COULOMB ELECTROSTATICS */
297 felec = velec*rinvsq20;
299 /* Update potential sums from outer loop */
304 /* Calculate temporary vectorial force */
309 /* Update vectorial force */
313 f[j_coord_offset+DIM*0+XX] -= tx;
314 f[j_coord_offset+DIM*0+YY] -= ty;
315 f[j_coord_offset+DIM*0+ZZ] -= tz;
317 /**************************
318 * CALCULATE INTERACTIONS *
319 **************************/
323 /* COULOMB ELECTROSTATICS */
325 felec = velec*rinvsq30;
327 /* Update potential sums from outer loop */
332 /* Calculate temporary vectorial force */
337 /* Update vectorial force */
341 f[j_coord_offset+DIM*0+XX] -= tx;
342 f[j_coord_offset+DIM*0+YY] -= ty;
343 f[j_coord_offset+DIM*0+ZZ] -= tz;
345 /* Inner loop uses 139 flops */
347 /* End of innermost loop */
350 f[i_coord_offset+DIM*0+XX] += fix0;
351 f[i_coord_offset+DIM*0+YY] += fiy0;
352 f[i_coord_offset+DIM*0+ZZ] += fiz0;
356 f[i_coord_offset+DIM*1+XX] += fix1;
357 f[i_coord_offset+DIM*1+YY] += fiy1;
358 f[i_coord_offset+DIM*1+ZZ] += fiz1;
362 f[i_coord_offset+DIM*2+XX] += fix2;
363 f[i_coord_offset+DIM*2+YY] += fiy2;
364 f[i_coord_offset+DIM*2+ZZ] += fiz2;
368 f[i_coord_offset+DIM*3+XX] += fix3;
369 f[i_coord_offset+DIM*3+YY] += fiy3;
370 f[i_coord_offset+DIM*3+ZZ] += fiz3;
374 fshift[i_shift_offset+XX] += tx;
375 fshift[i_shift_offset+YY] += ty;
376 fshift[i_shift_offset+ZZ] += tz;
379 /* Update potential energies */
380 kernel_data->energygrp_elec[ggid] += velecsum;
381 kernel_data->energygrp_vdw[ggid] += vvdwsum;
383 /* Increment number of inner iterations */
384 inneriter += j_index_end - j_index_start;
386 /* Outer loop uses 41 flops */
389 /* Increment number of outer iterations */
392 /* Update outer/inner flops */
394 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*139);
397 * Gromacs nonbonded kernel: nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
398 * Electrostatics interaction: Coulomb
399 * VdW interaction: CubicSplineTable
400 * Geometry: Water4-Particle
401 * Calculate force/pot: Force
404 nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
405 (t_nblist * gmx_restrict nlist,
406 rvec * gmx_restrict xx,
407 rvec * gmx_restrict ff,
408 t_forcerec * gmx_restrict fr,
409 t_mdatoms * gmx_restrict mdatoms,
410 nb_kernel_data_t * gmx_restrict kernel_data,
411 t_nrnb * gmx_restrict nrnb)
413 int i_shift_offset,i_coord_offset,j_coord_offset;
414 int j_index_start,j_index_end;
415 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
416 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
417 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
418 real *shiftvec,*fshift,*x,*f;
420 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
422 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
424 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
426 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
428 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
429 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
430 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
431 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
432 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
433 real velec,felec,velecsum,facel,crf,krf,krf2;
436 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
440 real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
448 jindex = nlist->jindex;
450 shiftidx = nlist->shift;
452 shiftvec = fr->shift_vec[0];
453 fshift = fr->fshift[0];
455 charge = mdatoms->chargeA;
456 nvdwtype = fr->ntype;
458 vdwtype = mdatoms->typeA;
460 vftab = kernel_data->table_vdw->data;
461 vftabscale = kernel_data->table_vdw->scale;
463 /* Setup water-specific parameters */
464 inr = nlist->iinr[0];
465 iq1 = facel*charge[inr+1];
466 iq2 = facel*charge[inr+2];
467 iq3 = facel*charge[inr+3];
468 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
473 /* Start outer loop over neighborlists */
474 for(iidx=0; iidx<nri; iidx++)
476 /* Load shift vector for this list */
477 i_shift_offset = DIM*shiftidx[iidx];
478 shX = shiftvec[i_shift_offset+XX];
479 shY = shiftvec[i_shift_offset+YY];
480 shZ = shiftvec[i_shift_offset+ZZ];
482 /* Load limits for loop over neighbors */
483 j_index_start = jindex[iidx];
484 j_index_end = jindex[iidx+1];
486 /* Get outer coordinate index */
488 i_coord_offset = DIM*inr;
490 /* Load i particle coords and add shift vector */
491 ix0 = shX + x[i_coord_offset+DIM*0+XX];
492 iy0 = shY + x[i_coord_offset+DIM*0+YY];
493 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
494 ix1 = shX + x[i_coord_offset+DIM*1+XX];
495 iy1 = shY + x[i_coord_offset+DIM*1+YY];
496 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
497 ix2 = shX + x[i_coord_offset+DIM*2+XX];
498 iy2 = shY + x[i_coord_offset+DIM*2+YY];
499 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
500 ix3 = shX + x[i_coord_offset+DIM*3+XX];
501 iy3 = shY + x[i_coord_offset+DIM*3+YY];
502 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
517 /* Start inner kernel loop */
518 for(jidx=j_index_start; jidx<j_index_end; jidx++)
520 /* Get j neighbor index, and coordinate index */
522 j_coord_offset = DIM*jnr;
524 /* load j atom coordinates */
525 jx0 = x[j_coord_offset+DIM*0+XX];
526 jy0 = x[j_coord_offset+DIM*0+YY];
527 jz0 = x[j_coord_offset+DIM*0+ZZ];
529 /* Calculate displacement vector */
543 /* Calculate squared distance and things based on it */
544 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
545 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
546 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
547 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
549 rinv00 = gmx_invsqrt(rsq00);
550 rinv10 = gmx_invsqrt(rsq10);
551 rinv20 = gmx_invsqrt(rsq20);
552 rinv30 = gmx_invsqrt(rsq30);
554 rinvsq10 = rinv10*rinv10;
555 rinvsq20 = rinv20*rinv20;
556 rinvsq30 = rinv30*rinv30;
558 /* Load parameters for j particles */
560 vdwjidx0 = 2*vdwtype[jnr+0];
562 /**************************
563 * CALCULATE INTERACTIONS *
564 **************************/
568 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
569 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
571 /* Calculate table index by multiplying r with table scale and truncate to integer */
577 /* CUBIC SPLINE TABLE DISPERSION */
581 Geps = vfeps*vftab[vfitab+2];
582 Heps2 = vfeps*vfeps*vftab[vfitab+3];
584 FF = Fp+Geps+2.0*Heps2;
587 /* CUBIC SPLINE TABLE REPULSION */
590 Geps = vfeps*vftab[vfitab+6];
591 Heps2 = vfeps*vfeps*vftab[vfitab+7];
593 FF = Fp+Geps+2.0*Heps2;
595 fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00;
599 /* Calculate temporary vectorial force */
604 /* Update vectorial force */
608 f[j_coord_offset+DIM*0+XX] -= tx;
609 f[j_coord_offset+DIM*0+YY] -= ty;
610 f[j_coord_offset+DIM*0+ZZ] -= tz;
612 /**************************
613 * CALCULATE INTERACTIONS *
614 **************************/
618 /* COULOMB ELECTROSTATICS */
620 felec = velec*rinvsq10;
624 /* Calculate temporary vectorial force */
629 /* Update vectorial force */
633 f[j_coord_offset+DIM*0+XX] -= tx;
634 f[j_coord_offset+DIM*0+YY] -= ty;
635 f[j_coord_offset+DIM*0+ZZ] -= tz;
637 /**************************
638 * CALCULATE INTERACTIONS *
639 **************************/
643 /* COULOMB ELECTROSTATICS */
645 felec = velec*rinvsq20;
649 /* Calculate temporary vectorial force */
654 /* Update vectorial force */
658 f[j_coord_offset+DIM*0+XX] -= tx;
659 f[j_coord_offset+DIM*0+YY] -= ty;
660 f[j_coord_offset+DIM*0+ZZ] -= tz;
662 /**************************
663 * CALCULATE INTERACTIONS *
664 **************************/
668 /* COULOMB ELECTROSTATICS */
670 felec = velec*rinvsq30;
674 /* Calculate temporary vectorial force */
679 /* Update vectorial force */
683 f[j_coord_offset+DIM*0+XX] -= tx;
684 f[j_coord_offset+DIM*0+YY] -= ty;
685 f[j_coord_offset+DIM*0+ZZ] -= tz;
687 /* Inner loop uses 128 flops */
689 /* End of innermost loop */
692 f[i_coord_offset+DIM*0+XX] += fix0;
693 f[i_coord_offset+DIM*0+YY] += fiy0;
694 f[i_coord_offset+DIM*0+ZZ] += fiz0;
698 f[i_coord_offset+DIM*1+XX] += fix1;
699 f[i_coord_offset+DIM*1+YY] += fiy1;
700 f[i_coord_offset+DIM*1+ZZ] += fiz1;
704 f[i_coord_offset+DIM*2+XX] += fix2;
705 f[i_coord_offset+DIM*2+YY] += fiy2;
706 f[i_coord_offset+DIM*2+ZZ] += fiz2;
710 f[i_coord_offset+DIM*3+XX] += fix3;
711 f[i_coord_offset+DIM*3+YY] += fiy3;
712 f[i_coord_offset+DIM*3+ZZ] += fiz3;
716 fshift[i_shift_offset+XX] += tx;
717 fshift[i_shift_offset+YY] += ty;
718 fshift[i_shift_offset+ZZ] += tz;
720 /* Increment number of inner iterations */
721 inneriter += j_index_end - j_index_start;
723 /* Outer loop uses 39 flops */
726 /* Increment number of outer iterations */
729 /* Update outer/inner flops */
731 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*128);