src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_GeomP1P1_c.c

   1 /*
   2  * Note: this file was generated by the Gromacs c kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  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
  13  *
  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
  17  * later version.
  18  *
  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.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 /*
  34  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomP1P1_VF_c
  35  * Electrostatics interaction: Ewald
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecEw_VdwCSTab_GeomP1P1_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)
  49 {
  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;
  56     int              vdwioffset0;
  57     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  58     int              vdwjidx0;
  59     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  60     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  61     real             velec,felec,velecsum,facel,crf,krf,krf2;
  62     real             *charge;
  63     int              nvdwtype;
  64     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  65     int              *vdwtype;
  66     real             *vdwparam;
  67     int              vfitab;
  68     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  69     real             *vftab;
  70     int              ewitab;
  71     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  72     real             *ewtab;
  73
  74     x                = xx[0];
  75     f                = ff[0];
  76
  77     nri              = nlist->nri;
  78     iinr             = nlist->iinr;
  79     jindex           = nlist->jindex;
  80     jjnr             = nlist->jjnr;
  81     shiftidx         = nlist->shift;
  82     gid              = nlist->gid;
  83     shiftvec         = fr->shift_vec[0];
  84     fshift           = fr->fshift[0];
  85     facel            = fr->epsfac;
  86     charge           = mdatoms->chargeA;
  87     nvdwtype         = fr->ntype;
  88     vdwparam         = fr->nbfp;
  89     vdwtype          = mdatoms->typeA;
  90
  91     vftab            = kernel_data->table_vdw->data;
  92     vftabscale       = kernel_data->table_vdw->scale;
  93
  94     sh_ewald         = fr->ic->sh_ewald;
  95     ewtab            = fr->ic->tabq_coul_FDV0;
  96     ewtabscale       = fr->ic->tabq_scale;
  97     ewtabhalfspace   = 0.5/ewtabscale;
  98
  99     outeriter        = 0;
 100     inneriter        = 0;
 101
 102     /* Start outer loop over neighborlists */
 103     for(iidx=0; iidx<nri; iidx++)
 104     {
 105         /* Load shift vector for this list */
 106         i_shift_offset   = DIM*shiftidx[iidx];
 107         shX              = shiftvec[i_shift_offset+XX];
 108         shY              = shiftvec[i_shift_offset+YY];
 109         shZ              = shiftvec[i_shift_offset+ZZ];
 110
 111         /* Load limits for loop over neighbors */
 112         j_index_start    = jindex[iidx];
 113         j_index_end      = jindex[iidx+1];
 114
 115         /* Get outer coordinate index */
 116         inr              = iinr[iidx];
 117         i_coord_offset   = DIM*inr;
 118
 119         /* Load i particle coords and add shift vector */
 120         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 121         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 122         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 123
 124         fix0             = 0.0;
 125         fiy0             = 0.0;
 126         fiz0             = 0.0;
 127
 128         /* Load parameters for i particles */
 129         iq0              = facel*charge[inr+0];
 130         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 131
 132         /* Reset potential sums */
 133         velecsum         = 0.0;
 134         vvdwsum          = 0.0;
 135
 136         /* Start inner kernel loop */
 137         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 138         {
 139             /* Get j neighbor index, and coordinate index */
 140             jnr              = jjnr[jidx];
 141             j_coord_offset   = DIM*jnr;
 142
 143             /* load j atom coordinates */
 144             jx0              = x[j_coord_offset+DIM*0+XX];
 145             jy0              = x[j_coord_offset+DIM*0+YY];
 146             jz0              = x[j_coord_offset+DIM*0+ZZ];
 147
 148             /* Calculate displacement vector */
 149             dx00             = ix0 - jx0;
 150             dy00             = iy0 - jy0;
 151             dz00             = iz0 - jz0;
 152
 153             /* Calculate squared distance and things based on it */
 154             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 155
 156             rinv00           = gmx_invsqrt(rsq00);
 157
 158             rinvsq00         = rinv00*rinv00;
 159
 160             /* Load parameters for j particles */
 161             jq0              = charge[jnr+0];
 162             vdwjidx0         = 2*vdwtype[jnr+0];
 163
 164             /**************************
 165              * CALCULATE INTERACTIONS *
 166              **************************/
 167
 168             r00              = rsq00*rinv00;
 169
 170             qq00             = iq0*jq0;
 171             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 172             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 173
 174             /* Calculate table index by multiplying r with table scale and truncate to integer */
 175             rt               = r00*vftabscale;
 176             vfitab           = rt;
 177             vfeps            = rt-vfitab;
 178             vfitab           = 2*4*vfitab;
 179
 180             /* EWALD ELECTROSTATICS */
 181
 182             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 183             ewrt             = r00*ewtabscale;
 184             ewitab           = ewrt;
 185             eweps            = ewrt-ewitab;
 186             ewitab           = 4*ewitab;
 187             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 188             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 189             felec            = qq00*rinv00*(rinvsq00-felec);
 190
 191             /* CUBIC SPLINE TABLE DISPERSION */
 192             vfitab          += 0;
 193             Y                = vftab[vfitab];
 194             F                = vftab[vfitab+1];
 195             Geps             = vfeps*vftab[vfitab+2];
 196             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 197             Fp               = F+Geps+Heps2;
 198             VV               = Y+vfeps*Fp;
 199             vvdw6            = c6_00*VV;
 200             FF               = Fp+Geps+2.0*Heps2;
 201             fvdw6            = c6_00*FF;
 202
 203             /* CUBIC SPLINE TABLE REPULSION */
 204             Y                = vftab[vfitab+4];
 205             F                = vftab[vfitab+5];
 206             Geps             = vfeps*vftab[vfitab+6];
 207             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 208             Fp               = F+Geps+Heps2;
 209             VV               = Y+vfeps*Fp;
 210             vvdw12           = c12_00*VV;
 211             FF               = Fp+Geps+2.0*Heps2;
 212             fvdw12           = c12_00*FF;
 213             vvdw             = vvdw12+vvdw6;
 214             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 215
 216             /* Update potential sums from outer loop */
 217             velecsum        += velec;
 218             vvdwsum         += vvdw;
 219
 220             fscal            = felec+fvdw;
 221
 222             /* Calculate temporary vectorial force */
 223             tx               = fscal*dx00;
 224             ty               = fscal*dy00;
 225             tz               = fscal*dz00;
 226
 227             /* Update vectorial force */
 228             fix0            += tx;
 229             fiy0            += ty;
 230             fiz0            += tz;
 231             f[j_coord_offset+DIM*0+XX] -= tx;
 232             f[j_coord_offset+DIM*0+YY] -= ty;
 233             f[j_coord_offset+DIM*0+ZZ] -= tz;
 234
 235             /* Inner loop uses 74 flops */
 236         }
 237         /* End of innermost loop */
 238
 239         tx = ty = tz = 0;
 240         f[i_coord_offset+DIM*0+XX] += fix0;
 241         f[i_coord_offset+DIM*0+YY] += fiy0;
 242         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 243         tx                         += fix0;
 244         ty                         += fiy0;
 245         tz                         += fiz0;
 246         fshift[i_shift_offset+XX]  += tx;
 247         fshift[i_shift_offset+YY]  += ty;
 248         fshift[i_shift_offset+ZZ]  += tz;
 249
 250         ggid                        = gid[iidx];
 251         /* Update potential energies */
 252         kernel_data->energygrp_elec[ggid] += velecsum;
 253         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 254
 255         /* Increment number of inner iterations */
 256         inneriter                  += j_index_end - j_index_start;
 257
 258         /* Outer loop uses 15 flops */
 259     }
 260
 261     /* Increment number of outer iterations */
 262     outeriter        += nri;
 263
 264     /* Update outer/inner flops */
 265
 266     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*74);
 267 }
 268 /*
 269  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 270  * Electrostatics interaction: Ewald
 271  * VdW interaction:            CubicSplineTable
 272  * Geometry:                   Particle-Particle
 273  * Calculate force/pot:        Force
 274  */
 275 void
 276 nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 277                     (t_nblist * gmx_restrict                nlist,
 278                      rvec * gmx_restrict                    xx,
 279                      rvec * gmx_restrict                    ff,
 280                      t_forcerec * gmx_restrict              fr,
 281                      t_mdatoms * gmx_restrict               mdatoms,
 282                      nb_kernel_data_t * gmx_restrict        kernel_data,
 283                      t_nrnb * gmx_restrict                  nrnb)
 284 {
 285     int              i_shift_offset,i_coord_offset,j_coord_offset;
 286     int              j_index_start,j_index_end;
 287     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 288     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 289     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 290     real             *shiftvec,*fshift,*x,*f;
 291     int              vdwioffset0;
 292     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 293     int              vdwjidx0;
 294     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 295     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 296     real             velec,felec,velecsum,facel,crf,krf,krf2;
 297     real             *charge;
 298     int              nvdwtype;
 299     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 300     int              *vdwtype;
 301     real             *vdwparam;
 302     int              vfitab;
 303     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 304     real             *vftab;
 305     int              ewitab;
 306     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 307     real             *ewtab;
 308
 309     x                = xx[0];
 310     f                = ff[0];
 311
 312     nri              = nlist->nri;
 313     iinr             = nlist->iinr;
 314     jindex           = nlist->jindex;
 315     jjnr             = nlist->jjnr;
 316     shiftidx         = nlist->shift;
 317     gid              = nlist->gid;
 318     shiftvec         = fr->shift_vec[0];
 319     fshift           = fr->fshift[0];
 320     facel            = fr->epsfac;
 321     charge           = mdatoms->chargeA;
 322     nvdwtype         = fr->ntype;
 323     vdwparam         = fr->nbfp;
 324     vdwtype          = mdatoms->typeA;
 325
 326     vftab            = kernel_data->table_vdw->data;
 327     vftabscale       = kernel_data->table_vdw->scale;
 328
 329     sh_ewald         = fr->ic->sh_ewald;
 330     ewtab            = fr->ic->tabq_coul_F;
 331     ewtabscale       = fr->ic->tabq_scale;
 332     ewtabhalfspace   = 0.5/ewtabscale;
 333
 334     outeriter        = 0;
 335     inneriter        = 0;
 336
 337     /* Start outer loop over neighborlists */
 338     for(iidx=0; iidx<nri; iidx++)
 339     {
 340         /* Load shift vector for this list */
 341         i_shift_offset   = DIM*shiftidx[iidx];
 342         shX              = shiftvec[i_shift_offset+XX];
 343         shY              = shiftvec[i_shift_offset+YY];
 344         shZ              = shiftvec[i_shift_offset+ZZ];
 345
 346         /* Load limits for loop over neighbors */
 347         j_index_start    = jindex[iidx];
 348         j_index_end      = jindex[iidx+1];
 349
 350         /* Get outer coordinate index */
 351         inr              = iinr[iidx];
 352         i_coord_offset   = DIM*inr;
 353
 354         /* Load i particle coords and add shift vector */
 355         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 356         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 357         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 358
 359         fix0             = 0.0;
 360         fiy0             = 0.0;
 361         fiz0             = 0.0;
 362
 363         /* Load parameters for i particles */
 364         iq0              = facel*charge[inr+0];
 365         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 366
 367         /* Start inner kernel loop */
 368         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 369         {
 370             /* Get j neighbor index, and coordinate index */
 371             jnr              = jjnr[jidx];
 372             j_coord_offset   = DIM*jnr;
 373
 374             /* load j atom coordinates */
 375             jx0              = x[j_coord_offset+DIM*0+XX];
 376             jy0              = x[j_coord_offset+DIM*0+YY];
 377             jz0              = x[j_coord_offset+DIM*0+ZZ];
 378
 379             /* Calculate displacement vector */
 380             dx00             = ix0 - jx0;
 381             dy00             = iy0 - jy0;
 382             dz00             = iz0 - jz0;
 383
 384             /* Calculate squared distance and things based on it */
 385             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 386
 387             rinv00           = gmx_invsqrt(rsq00);
 388
 389             rinvsq00         = rinv00*rinv00;
 390
 391             /* Load parameters for j particles */
 392             jq0              = charge[jnr+0];
 393             vdwjidx0         = 2*vdwtype[jnr+0];
 394
 395             /**************************
 396              * CALCULATE INTERACTIONS *
 397              **************************/
 398
 399             r00              = rsq00*rinv00;
 400
 401             qq00             = iq0*jq0;
 402             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 403             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 404
 405             /* Calculate table index by multiplying r with table scale and truncate to integer */
 406             rt               = r00*vftabscale;
 407             vfitab           = rt;
 408             vfeps            = rt-vfitab;
 409             vfitab           = 2*4*vfitab;
 410
 411             /* EWALD ELECTROSTATICS */
 412
 413             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 414             ewrt             = r00*ewtabscale;
 415             ewitab           = ewrt;
 416             eweps            = ewrt-ewitab;
 417             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 418             felec            = qq00*rinv00*(rinvsq00-felec);
 419
 420             /* CUBIC SPLINE TABLE DISPERSION */
 421             vfitab          += 0;
 422             Y                = vftab[vfitab];
 423             F                = vftab[vfitab+1];
 424             Geps             = vfeps*vftab[vfitab+2];
 425             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 426             Fp               = F+Geps+Heps2;
 427             FF               = Fp+Geps+2.0*Heps2;
 428             fvdw6            = c6_00*FF;
 429
 430             /* CUBIC SPLINE TABLE REPULSION */
 431             Y                = vftab[vfitab+4];
 432             F                = vftab[vfitab+5];
 433             Geps             = vfeps*vftab[vfitab+6];
 434             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 435             Fp               = F+Geps+Heps2;
 436             FF               = Fp+Geps+2.0*Heps2;
 437             fvdw12           = c12_00*FF;
 438             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 439
 440             fscal            = felec+fvdw;
 441
 442             /* Calculate temporary vectorial force */
 443             tx               = fscal*dx00;
 444             ty               = fscal*dy00;
 445             tz               = fscal*dz00;
 446
 447             /* Update vectorial force */
 448             fix0            += tx;
 449             fiy0            += ty;
 450             fiz0            += tz;
 451             f[j_coord_offset+DIM*0+XX] -= tx;
 452             f[j_coord_offset+DIM*0+YY] -= ty;
 453             f[j_coord_offset+DIM*0+ZZ] -= tz;
 454
 455             /* Inner loop uses 59 flops */
 456         }
 457         /* End of innermost loop */
 458
 459         tx = ty = tz = 0;
 460         f[i_coord_offset+DIM*0+XX] += fix0;
 461         f[i_coord_offset+DIM*0+YY] += fiy0;
 462         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 463         tx                         += fix0;
 464         ty                         += fiy0;
 465         tz                         += fiz0;
 466         fshift[i_shift_offset+XX]  += tx;
 467         fshift[i_shift_offset+YY]  += ty;
 468         fshift[i_shift_offset+ZZ]  += tz;
 469
 470         /* Increment number of inner iterations */
 471         inneriter                  += j_index_end - j_index_start;
 472
 473         /* Outer loop uses 13 flops */
 474     }
 475
 476     /* Increment number of outer iterations */
 477     outeriter        += nri;
 478
 479     /* Update outer/inner flops */
 480
 481     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*59);
 482 }