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_ElecCSTab_VdwCSTab_GeomW4P1_VF_c
35 * Electrostatics interaction: CubicSplineTable
36 * VdW interaction: CubicSplineTable
37 * Geometry: Water4-Particle
38 * Calculate force/pot: PotentialAndForce
41 nb_kernel_ElecCSTab_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_elec_vdw->data;
98 vftabscale = kernel_data->table_elec_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 /* Load parameters for j particles */
197 vdwjidx0 = 2*vdwtype[jnr+0];
199 /**************************
200 * CALCULATE INTERACTIONS *
201 **************************/
205 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
206 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
208 /* Calculate table index by multiplying r with table scale and truncate to integer */
214 /* CUBIC SPLINE TABLE DISPERSION */
218 Geps = vfeps*vftab[vfitab+2];
219 Heps2 = vfeps*vfeps*vftab[vfitab+3];
223 FF = Fp+Geps+2.0*Heps2;
226 /* CUBIC SPLINE TABLE REPULSION */
229 Geps = vfeps*vftab[vfitab+6];
230 Heps2 = vfeps*vfeps*vftab[vfitab+7];
234 FF = Fp+Geps+2.0*Heps2;
237 fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00;
239 /* Update potential sums from outer loop */
244 /* Calculate temporary vectorial force */
249 /* Update vectorial force */
253 f[j_coord_offset+DIM*0+XX] -= tx;
254 f[j_coord_offset+DIM*0+YY] -= ty;
255 f[j_coord_offset+DIM*0+ZZ] -= tz;
257 /**************************
258 * CALCULATE INTERACTIONS *
259 **************************/
265 /* Calculate table index by multiplying r with table scale and truncate to integer */
271 /* CUBIC SPLINE TABLE ELECTROSTATICS */
274 Geps = vfeps*vftab[vfitab+2];
275 Heps2 = vfeps*vfeps*vftab[vfitab+3];
279 FF = Fp+Geps+2.0*Heps2;
280 felec = -qq10*FF*vftabscale*rinv10;
282 /* Update potential sums from outer loop */
287 /* Calculate temporary vectorial force */
292 /* Update vectorial force */
296 f[j_coord_offset+DIM*0+XX] -= tx;
297 f[j_coord_offset+DIM*0+YY] -= ty;
298 f[j_coord_offset+DIM*0+ZZ] -= tz;
300 /**************************
301 * CALCULATE INTERACTIONS *
302 **************************/
308 /* Calculate table index by multiplying r with table scale and truncate to integer */
314 /* CUBIC SPLINE TABLE ELECTROSTATICS */
317 Geps = vfeps*vftab[vfitab+2];
318 Heps2 = vfeps*vfeps*vftab[vfitab+3];
322 FF = Fp+Geps+2.0*Heps2;
323 felec = -qq20*FF*vftabscale*rinv20;
325 /* Update potential sums from outer loop */
330 /* Calculate temporary vectorial force */
335 /* Update vectorial force */
339 f[j_coord_offset+DIM*0+XX] -= tx;
340 f[j_coord_offset+DIM*0+YY] -= ty;
341 f[j_coord_offset+DIM*0+ZZ] -= tz;
343 /**************************
344 * CALCULATE INTERACTIONS *
345 **************************/
351 /* Calculate table index by multiplying r with table scale and truncate to integer */
357 /* CUBIC SPLINE TABLE ELECTROSTATICS */
360 Geps = vfeps*vftab[vfitab+2];
361 Heps2 = vfeps*vfeps*vftab[vfitab+3];
365 FF = Fp+Geps+2.0*Heps2;
366 felec = -qq30*FF*vftabscale*rinv30;
368 /* Update potential sums from outer loop */
373 /* Calculate temporary vectorial force */
378 /* Update vectorial force */
382 f[j_coord_offset+DIM*0+XX] -= tx;
383 f[j_coord_offset+DIM*0+YY] -= ty;
384 f[j_coord_offset+DIM*0+ZZ] -= tz;
386 /* Inner loop uses 181 flops */
388 /* End of innermost loop */
391 f[i_coord_offset+DIM*0+XX] += fix0;
392 f[i_coord_offset+DIM*0+YY] += fiy0;
393 f[i_coord_offset+DIM*0+ZZ] += fiz0;
397 f[i_coord_offset+DIM*1+XX] += fix1;
398 f[i_coord_offset+DIM*1+YY] += fiy1;
399 f[i_coord_offset+DIM*1+ZZ] += fiz1;
403 f[i_coord_offset+DIM*2+XX] += fix2;
404 f[i_coord_offset+DIM*2+YY] += fiy2;
405 f[i_coord_offset+DIM*2+ZZ] += fiz2;
409 f[i_coord_offset+DIM*3+XX] += fix3;
410 f[i_coord_offset+DIM*3+YY] += fiy3;
411 f[i_coord_offset+DIM*3+ZZ] += fiz3;
415 fshift[i_shift_offset+XX] += tx;
416 fshift[i_shift_offset+YY] += ty;
417 fshift[i_shift_offset+ZZ] += tz;
420 /* Update potential energies */
421 kernel_data->energygrp_elec[ggid] += velecsum;
422 kernel_data->energygrp_vdw[ggid] += vvdwsum;
424 /* Increment number of inner iterations */
425 inneriter += j_index_end - j_index_start;
427 /* Outer loop uses 41 flops */
430 /* Increment number of outer iterations */
433 /* Update outer/inner flops */
435 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*181);
438 * Gromacs nonbonded kernel: nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
439 * Electrostatics interaction: CubicSplineTable
440 * VdW interaction: CubicSplineTable
441 * Geometry: Water4-Particle
442 * Calculate force/pot: Force
445 nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
446 (t_nblist * gmx_restrict nlist,
447 rvec * gmx_restrict xx,
448 rvec * gmx_restrict ff,
449 t_forcerec * gmx_restrict fr,
450 t_mdatoms * gmx_restrict mdatoms,
451 nb_kernel_data_t * gmx_restrict kernel_data,
452 t_nrnb * gmx_restrict nrnb)
454 int i_shift_offset,i_coord_offset,j_coord_offset;
455 int j_index_start,j_index_end;
456 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
457 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
458 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
459 real *shiftvec,*fshift,*x,*f;
461 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
463 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
465 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
467 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
469 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
470 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
471 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
472 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
473 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
474 real velec,felec,velecsum,facel,crf,krf,krf2;
477 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
481 real rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
489 jindex = nlist->jindex;
491 shiftidx = nlist->shift;
493 shiftvec = fr->shift_vec[0];
494 fshift = fr->fshift[0];
496 charge = mdatoms->chargeA;
497 nvdwtype = fr->ntype;
499 vdwtype = mdatoms->typeA;
501 vftab = kernel_data->table_elec_vdw->data;
502 vftabscale = kernel_data->table_elec_vdw->scale;
504 /* Setup water-specific parameters */
505 inr = nlist->iinr[0];
506 iq1 = facel*charge[inr+1];
507 iq2 = facel*charge[inr+2];
508 iq3 = facel*charge[inr+3];
509 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
514 /* Start outer loop over neighborlists */
515 for(iidx=0; iidx<nri; iidx++)
517 /* Load shift vector for this list */
518 i_shift_offset = DIM*shiftidx[iidx];
519 shX = shiftvec[i_shift_offset+XX];
520 shY = shiftvec[i_shift_offset+YY];
521 shZ = shiftvec[i_shift_offset+ZZ];
523 /* Load limits for loop over neighbors */
524 j_index_start = jindex[iidx];
525 j_index_end = jindex[iidx+1];
527 /* Get outer coordinate index */
529 i_coord_offset = DIM*inr;
531 /* Load i particle coords and add shift vector */
532 ix0 = shX + x[i_coord_offset+DIM*0+XX];
533 iy0 = shY + x[i_coord_offset+DIM*0+YY];
534 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
535 ix1 = shX + x[i_coord_offset+DIM*1+XX];
536 iy1 = shY + x[i_coord_offset+DIM*1+YY];
537 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
538 ix2 = shX + x[i_coord_offset+DIM*2+XX];
539 iy2 = shY + x[i_coord_offset+DIM*2+YY];
540 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
541 ix3 = shX + x[i_coord_offset+DIM*3+XX];
542 iy3 = shY + x[i_coord_offset+DIM*3+YY];
543 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
558 /* Start inner kernel loop */
559 for(jidx=j_index_start; jidx<j_index_end; jidx++)
561 /* Get j neighbor index, and coordinate index */
563 j_coord_offset = DIM*jnr;
565 /* load j atom coordinates */
566 jx0 = x[j_coord_offset+DIM*0+XX];
567 jy0 = x[j_coord_offset+DIM*0+YY];
568 jz0 = x[j_coord_offset+DIM*0+ZZ];
570 /* Calculate displacement vector */
584 /* Calculate squared distance and things based on it */
585 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
586 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
587 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
588 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
590 rinv00 = gmx_invsqrt(rsq00);
591 rinv10 = gmx_invsqrt(rsq10);
592 rinv20 = gmx_invsqrt(rsq20);
593 rinv30 = gmx_invsqrt(rsq30);
595 /* Load parameters for j particles */
597 vdwjidx0 = 2*vdwtype[jnr+0];
599 /**************************
600 * CALCULATE INTERACTIONS *
601 **************************/
605 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
606 c12_00 = vdwparam[vdwioffset0+vdwjidx0+1];
608 /* Calculate table index by multiplying r with table scale and truncate to integer */
614 /* CUBIC SPLINE TABLE DISPERSION */
617 Geps = vfeps*vftab[vfitab+2];
618 Heps2 = vfeps*vfeps*vftab[vfitab+3];
620 FF = Fp+Geps+2.0*Heps2;
623 /* CUBIC SPLINE TABLE REPULSION */
625 Geps = vfeps*vftab[vfitab+6];
626 Heps2 = vfeps*vfeps*vftab[vfitab+7];
628 FF = Fp+Geps+2.0*Heps2;
630 fvdw = -(fvdw6+fvdw12)*vftabscale*rinv00;
634 /* Calculate temporary vectorial force */
639 /* Update vectorial force */
643 f[j_coord_offset+DIM*0+XX] -= tx;
644 f[j_coord_offset+DIM*0+YY] -= ty;
645 f[j_coord_offset+DIM*0+ZZ] -= tz;
647 /**************************
648 * CALCULATE INTERACTIONS *
649 **************************/
655 /* Calculate table index by multiplying r with table scale and truncate to integer */
661 /* CUBIC SPLINE TABLE ELECTROSTATICS */
663 Geps = vfeps*vftab[vfitab+2];
664 Heps2 = vfeps*vfeps*vftab[vfitab+3];
666 FF = Fp+Geps+2.0*Heps2;
667 felec = -qq10*FF*vftabscale*rinv10;
671 /* Calculate temporary vectorial force */
676 /* Update vectorial force */
680 f[j_coord_offset+DIM*0+XX] -= tx;
681 f[j_coord_offset+DIM*0+YY] -= ty;
682 f[j_coord_offset+DIM*0+ZZ] -= tz;
684 /**************************
685 * CALCULATE INTERACTIONS *
686 **************************/
692 /* Calculate table index by multiplying r with table scale and truncate to integer */
698 /* CUBIC SPLINE TABLE ELECTROSTATICS */
700 Geps = vfeps*vftab[vfitab+2];
701 Heps2 = vfeps*vfeps*vftab[vfitab+3];
703 FF = Fp+Geps+2.0*Heps2;
704 felec = -qq20*FF*vftabscale*rinv20;
708 /* Calculate temporary vectorial force */
713 /* Update vectorial force */
717 f[j_coord_offset+DIM*0+XX] -= tx;
718 f[j_coord_offset+DIM*0+YY] -= ty;
719 f[j_coord_offset+DIM*0+ZZ] -= tz;
721 /**************************
722 * CALCULATE INTERACTIONS *
723 **************************/
729 /* Calculate table index by multiplying r with table scale and truncate to integer */
735 /* CUBIC SPLINE TABLE ELECTROSTATICS */
737 Geps = vfeps*vftab[vfitab+2];
738 Heps2 = vfeps*vfeps*vftab[vfitab+3];
740 FF = Fp+Geps+2.0*Heps2;
741 felec = -qq30*FF*vftabscale*rinv30;
745 /* Calculate temporary vectorial force */
750 /* Update vectorial force */
754 f[j_coord_offset+DIM*0+XX] -= tx;
755 f[j_coord_offset+DIM*0+YY] -= ty;
756 f[j_coord_offset+DIM*0+ZZ] -= tz;
758 /* Inner loop uses 161 flops */
760 /* End of innermost loop */
763 f[i_coord_offset+DIM*0+XX] += fix0;
764 f[i_coord_offset+DIM*0+YY] += fiy0;
765 f[i_coord_offset+DIM*0+ZZ] += fiz0;
769 f[i_coord_offset+DIM*1+XX] += fix1;
770 f[i_coord_offset+DIM*1+YY] += fiy1;
771 f[i_coord_offset+DIM*1+ZZ] += fiz1;
775 f[i_coord_offset+DIM*2+XX] += fix2;
776 f[i_coord_offset+DIM*2+YY] += fiy2;
777 f[i_coord_offset+DIM*2+ZZ] += fiz2;
781 f[i_coord_offset+DIM*3+XX] += fix3;
782 f[i_coord_offset+DIM*3+YY] += fiy3;
783 f[i_coord_offset+DIM*3+ZZ] += fiz3;
787 fshift[i_shift_offset+XX] += tx;
788 fshift[i_shift_offset+YY] += ty;
789 fshift[i_shift_offset+ZZ] += tz;
791 /* Increment number of inner iterations */
792 inneriter += j_index_end - j_index_start;
794 /* Outer loop uses 39 flops */
797 /* Increment number of outer iterations */
800 /* Update outer/inner flops */
802 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*161);