src/gromacs/gmxlib/nonbonded/nb_kernel_adress_c/nb_kernel111_c_adress.c

   1 /*
   2  * Copyright (c) Erik Lindahl, David van der Spoel 2003
   3  *
   4  * This file is generated automatically at compile time
   5  * by the program mknb in the Gromacs distribution.
   6  *
   7  * Options used when generation this file:
   8  * Language:         c
   9  * Precision:        single
  10  * Threads:          yes
  11  * Software invsqrt: no
  12  * PowerPC invsqrt:  no
  13  * Prefetch forces:  no
  14  * Adress kernel:  yes
  15  * Comments:         no
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include<config.h>
  19 #endif
  20 #ifdef GMX_THREAD_SHM_FDECOMP
  21 #include<thread_mpi.h>
  22 #endif
  23 #define ALMOST_ZERO 1e-30
  24 #define ALMOST_ONE 1-(1e-30)
  25 #include<math.h>
  26
  27 #include "nb_kernel111_adress.h"
  28
  29
  30
  31 /*
  32  * Gromacs nonbonded kernel nb_kernel111_adress_cg
  33  * Coulomb interaction:     Normal Coulomb
  34  * VdW interaction:         Lennard-Jones
  35  * water optimization:      SPC/TIP3P - other atoms
  36  * Calculate forces:        yes
  37  */
  38 void nb_kernel111_adress_cg(
  39                     int *           p_nri,
  40                     int *           iinr,
  41                     int *           jindex,
  42                     int *           jjnr,
  43                     int *           shift,
  44                     real *         shiftvec,
  45                     real *         fshift,
  46                     int *           gid,
  47                     real *         pos,
  48                     real *         faction,
  49                     real *         charge,
  50                     real *         p_facel,
  51                     real *         p_krf,
  52                     real *         p_crf,
  53                     real *         Vc,
  54                     int *           type,
  55                     int *           p_ntype,
  56                     real *         vdwparam,
  57                     real *         Vvdw,
  58                     real *         p_tabscale,
  59                     real *         VFtab,
  60                     real *         invsqrta,
  61                     real *         dvda,
  62                     real *         p_gbtabscale,
  63                     real *         GBtab,
  64                     int *           p_nthreads,
  65                     int *           count,
  66                     void *          mtx,
  67                     int *           outeriter,
  68                     int *           inneriter,
  69                     real           force_cap,
  70                     real *         wf)
  71 {
  72     int           nri,ntype,nthreads;
  73     real         facel,krf,crf,tabscale,gbtabscale;
  74     int           n,ii,is3,ii3,k,nj0,nj1,jnr,j3,ggid;
  75     int           nn0,nn1,nouter,ninner;
  76     real         shX,shY,shZ;
  77     real         fscal,tx,ty,tz;
  78     real         rinvsq;
  79     real         jq;
  80     real         qq,vcoul,vctot;
  81     int           nti;
  82     int           tj;
  83     real         rinvsix;
  84     real         Vvdw6,Vvdwtot;
  85     real         Vvdw12;
  86     real         ix1,iy1,iz1,fix1,fiy1,fiz1;
  87     real         ix2,iy2,iz2,fix2,fiy2,fiz2;
  88     real         ix3,iy3,iz3,fix3,fiy3,fiz3;
  89     real         jx1,jy1,jz1,fjx1,fjy1,fjz1;
  90     real         dx11,dy11,dz11,rsq11,rinv11;
  91     real         dx21,dy21,dz21,rsq21,rinv21;
  92     real         dx31,dy31,dz31,rsq31,rinv31;
  93     real         qO,qH;
  94     real         c6,c12;
  95     real         weight_cg1, weight_cg2, weight_product;
  96     real         hybscal;
  97
  98     nri              = *p_nri;
  99     ntype            = *p_ntype;
 100     nthreads         = *p_nthreads;
 101     facel            = *p_facel;
 102     krf              = *p_krf;
 103     crf              = *p_crf;
 104     tabscale         = *p_tabscale;
 105     ii               = iinr[0];
 106     qO               = facel*charge[ii];
 107     qH               = facel*charge[ii+1];
 108     nti              = 2*ntype*type[ii];
 109
 110     nouter           = 0;
 111     ninner           = 0;
 112
 113     do
 114     {
 115         #ifdef GMX_THREAD_SHM_FDECOMP
 116         tMPI_Thread_mutex_lock((tMPI_Thread_mutex_t *)mtx);
 117         nn0              = *count;
 118         nn1              = nn0+(nri-nn0)/(2*nthreads)+10;
 119         *count           = nn1;
 120         tMPI_Thread_mutex_unlock((tMPI_Thread_mutex_t *)mtx);
 121         if(nn1>nri) nn1=nri;
 122         #else
 123         nn0 = 0;
 124         nn1 = nri;
 125         #endif
 126
 127         for(n=nn0; (n<nn1); n++)
 128         {
 129             is3              = 3*shift[n];
 130             shX              = shiftvec[is3];
 131             shY              = shiftvec[is3+1];
 132             shZ              = shiftvec[is3+2];
 133             nj0              = jindex[n];
 134             nj1              = jindex[n+1];
 135             ii               = iinr[n];
 136             ii3              = 3*ii;
 137             ix1              = shX + pos[ii3+0];
 138             iy1              = shY + pos[ii3+1];
 139             iz1              = shZ + pos[ii3+2];
 140             ix2              = shX + pos[ii3+3];
 141             iy2              = shY + pos[ii3+4];
 142             iz2              = shZ + pos[ii3+5];
 143             ix3              = shX + pos[ii3+6];
 144             iy3              = shY + pos[ii3+7];
 145             iz3              = shZ + pos[ii3+8];
 146             weight_cg1       = wf[ii];
 147             vctot            = 0;
 148             Vvdwtot          = 0;
 149             fix1             = 0;
 150             fiy1             = 0;
 151             fiz1             = 0;
 152             fix2             = 0;
 153             fiy2             = 0;
 154             fiz2             = 0;
 155             fix3             = 0;
 156             fiy3             = 0;
 157             fiz3             = 0;
 158
 159             for(k=nj0; (k<nj1); k++)
 160             {
 161                 jnr              = jjnr[k];
 162                 weight_cg2       = wf[jnr];
 163                 weight_product   = weight_cg1*weight_cg2;
 164                 if (weight_product < ALMOST_ZERO) {
 165                        hybscal = 1.0;
 166                 }
 167                 else if (weight_product >= ALMOST_ONE)
 168                 {
 169                   /* force is zero, skip this molecule */
 170                        continue;
 171                 }
 172                 else
 173                 {
 174                    hybscal = 1.0 - weight_product;
 175                 }
 176                 j3               = 3*jnr;
 177                 jx1              = pos[j3+0];
 178                 jy1              = pos[j3+1];
 179                 jz1              = pos[j3+2];
 180                 dx11             = ix1 - jx1;
 181                 dy11             = iy1 - jy1;
 182                 dz11             = iz1 - jz1;
 183                 rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 184                 dx21             = ix2 - jx1;
 185                 dy21             = iy2 - jy1;
 186                 dz21             = iz2 - jz1;
 187                 rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 188                 dx31             = ix3 - jx1;
 189                 dy31             = iy3 - jy1;
 190                 dz31             = iz3 - jz1;
 191                 rsq31            = dx31*dx31+dy31*dy31+dz31*dz31;
 192                 rinv11           = 1.0/sqrt(rsq11);
 193                 rinv21           = 1.0/sqrt(rsq21);
 194                 rinv31           = 1.0/sqrt(rsq31);
 195                 jq               = charge[jnr+0];
 196                 qq               = qO*jq;
 197                 tj               = nti+2*type[jnr];
 198                 c6               = vdwparam[tj];
 199                 c12              = vdwparam[tj+1];
 200                 rinvsq           = rinv11*rinv11;
 201                 vcoul            = qq*rinv11;
 202                 vctot            = vctot+vcoul;
 203                 rinvsix          = rinvsq*rinvsq*rinvsq;
 204                 Vvdw6            = c6*rinvsix;
 205                 Vvdw12           = c12*rinvsix*rinvsix;
 206                 Vvdwtot          = Vvdwtot+Vvdw12-Vvdw6;
 207                 fscal            = (vcoul+12.0*Vvdw12-6.0*Vvdw6)*rinvsq;
 208                 fscal *= hybscal;
 209                 tx               = fscal*dx11;
 210                 ty               = fscal*dy11;
 211                 tz               = fscal*dz11;
 212                 fix1             = fix1 + tx;
 213                 fiy1             = fiy1 + ty;
 214                 fiz1             = fiz1 + tz;
 215                 fjx1             = faction[j3+0] - tx;
 216                 fjy1             = faction[j3+1] - ty;
 217                 fjz1             = faction[j3+2] - tz;
 218                 qq               = qH*jq;
 219                 rinvsq           = rinv21*rinv21;
 220                 vcoul            = qq*rinv21;
 221                 vctot            = vctot+vcoul;
 222                 fscal            = (vcoul)*rinvsq;
 223                 fscal *= hybscal;
 224                 tx               = fscal*dx21;
 225                 ty               = fscal*dy21;
 226                 tz               = fscal*dz21;
 227                 fix2             = fix2 + tx;
 228                 fiy2             = fiy2 + ty;
 229                 fiz2             = fiz2 + tz;
 230                 fjx1             = fjx1 - tx;
 231                 fjy1             = fjy1 - ty;
 232                 fjz1             = fjz1 - tz;
 233                 rinvsq           = rinv31*rinv31;
 234                 vcoul            = qq*rinv31;
 235                 vctot            = vctot+vcoul;
 236                 fscal            = (vcoul)*rinvsq;
 237                 fscal *= hybscal;
 238                 tx               = fscal*dx31;
 239                 ty               = fscal*dy31;
 240                 tz               = fscal*dz31;
 241                 fix3             = fix3 + tx;
 242                 fiy3             = fiy3 + ty;
 243                 fiz3             = fiz3 + tz;
 244                 faction[j3+0]    = fjx1 - tx;
 245                 faction[j3+1]    = fjy1 - ty;
 246                 faction[j3+2]    = fjz1 - tz;
 247             }
 248
 249             faction[ii3+0]   = faction[ii3+0] + fix1;
 250             faction[ii3+1]   = faction[ii3+1] + fiy1;
 251             faction[ii3+2]   = faction[ii3+2] + fiz1;
 252             faction[ii3+3]   = faction[ii3+3] + fix2;
 253             faction[ii3+4]   = faction[ii3+4] + fiy2;
 254             faction[ii3+5]   = faction[ii3+5] + fiz2;
 255             faction[ii3+6]   = faction[ii3+6] + fix3;
 256             faction[ii3+7]   = faction[ii3+7] + fiy3;
 257             faction[ii3+8]   = faction[ii3+8] + fiz3;
 258             fshift[is3]      = fshift[is3]+fix1+fix2+fix3;
 259             fshift[is3+1]    = fshift[is3+1]+fiy1+fiy2+fiy3;
 260             fshift[is3+2]    = fshift[is3+2]+fiz1+fiz2+fiz3;
 261             ggid             = gid[n];
 262             Vc[ggid]         = Vc[ggid] + vctot;
 263             Vvdw[ggid]       = Vvdw[ggid] + Vvdwtot;
 264             ninner           = ninner + nj1 - nj0;
 265         }
 266
 267         nouter           = nouter + nn1 - nn0;
 268     }
 269     while (nn1<nri);
 270
 271     *outeriter       = nouter;
 272     *inneriter       = ninner;
 273 }
 274
 275
 276
 277
 278
 279 /*
 280  * Gromacs nonbonded kernel nb_kernel111_adress_ex
 281  * Coulomb interaction:     Normal Coulomb
 282  * VdW interaction:         Lennard-Jones
 283  * water optimization:      SPC/TIP3P - other atoms
 284  * Calculate forces:        yes
 285  */
 286 void nb_kernel111_adress_ex(
 287                     int *           p_nri,
 288                     int *           iinr,
 289                     int *           jindex,
 290                     int *           jjnr,
 291                     int *           shift,
 292                     real *         shiftvec,
 293                     real *         fshift,
 294                     int *           gid,
 295                     real *         pos,
 296                     real *         faction,
 297                     real *         charge,
 298                     real *         p_facel,
 299                     real *         p_krf,
 300                     real *         p_crf,
 301                     real *         Vc,
 302                     int *           type,
 303                     int *           p_ntype,
 304                     real *         vdwparam,
 305                     real *         Vvdw,
 306                     real *         p_tabscale,
 307                     real *         VFtab,
 308                     real *         invsqrta,
 309                     real *         dvda,
 310                     real *         p_gbtabscale,
 311                     real *         GBtab,
 312                     int *           p_nthreads,
 313                     int *           count,
 314                     void *          mtx,
 315                     int *           outeriter,
 316                     int *           inneriter,
 317                     real           force_cap,
 318                     real *         wf)
 319 {
 320     int           nri,ntype,nthreads;
 321     real         facel,krf,crf,tabscale,gbtabscale;
 322     int           n,ii,is3,ii3,k,nj0,nj1,jnr,j3,ggid;
 323     int           nn0,nn1,nouter,ninner;
 324     real         shX,shY,shZ;
 325     real         fscal,tx,ty,tz;
 326     real         rinvsq;
 327     real         jq;
 328     real         qq,vcoul,vctot;
 329     int           nti;
 330     int           tj;
 331     real         rinvsix;
 332     real         Vvdw6,Vvdwtot;
 333     real         Vvdw12;
 334     real         ix1,iy1,iz1,fix1,fiy1,fiz1;
 335     real         ix2,iy2,iz2,fix2,fiy2,fiz2;
 336     real         ix3,iy3,iz3,fix3,fiy3,fiz3;
 337     real         jx1,jy1,jz1,fjx1,fjy1,fjz1;
 338     real         dx11,dy11,dz11,rsq11,rinv11;
 339     real         dx21,dy21,dz21,rsq21,rinv21;
 340     real         dx31,dy31,dz31,rsq31,rinv31;
 341     real         qO,qH;
 342     real         c6,c12;
 343     real         weight_cg1, weight_cg2, weight_product;
 344     real         hybscal;
 345
 346     nri              = *p_nri;
 347     ntype            = *p_ntype;
 348     nthreads         = *p_nthreads;
 349     facel            = *p_facel;
 350     krf              = *p_krf;
 351     crf              = *p_crf;
 352     tabscale         = *p_tabscale;
 353     ii               = iinr[0];
 354     qO               = facel*charge[ii];
 355     qH               = facel*charge[ii+1];
 356     nti              = 2*ntype*type[ii];
 357
 358     nouter           = 0;
 359     ninner           = 0;
 360
 361     do
 362     {
 363         #ifdef GMX_THREAD_SHM_FDECOMP
 364         tMPI_Thread_mutex_lock((tMPI_Thread_mutex_t *)mtx);
 365         nn0              = *count;
 366         nn1              = nn0+(nri-nn0)/(2*nthreads)+10;
 367         *count           = nn1;
 368         tMPI_Thread_mutex_unlock((tMPI_Thread_mutex_t *)mtx);
 369         if(nn1>nri) nn1=nri;
 370         #else
 371         nn0 = 0;
 372         nn1 = nri;
 373         #endif
 374
 375         for(n=nn0; (n<nn1); n++)
 376         {
 377             is3              = 3*shift[n];
 378             shX              = shiftvec[is3];
 379             shY              = shiftvec[is3+1];
 380             shZ              = shiftvec[is3+2];
 381             nj0              = jindex[n];
 382             nj1              = jindex[n+1];
 383             ii               = iinr[n];
 384             ii3              = 3*ii;
 385             ix1              = shX + pos[ii3+0];
 386             iy1              = shY + pos[ii3+1];
 387             iz1              = shZ + pos[ii3+2];
 388             ix2              = shX + pos[ii3+3];
 389             iy2              = shY + pos[ii3+4];
 390             iz2              = shZ + pos[ii3+5];
 391             ix3              = shX + pos[ii3+6];
 392             iy3              = shY + pos[ii3+7];
 393             iz3              = shZ + pos[ii3+8];
 394             weight_cg1       = wf[ii];
 395             vctot            = 0;
 396             Vvdwtot          = 0;
 397             fix1             = 0;
 398             fiy1             = 0;
 399             fiz1             = 0;
 400             fix2             = 0;
 401             fiy2             = 0;
 402             fiz2             = 0;
 403             fix3             = 0;
 404             fiy3             = 0;
 405             fiz3             = 0;
 406
 407             for(k=nj0; (k<nj1); k++)
 408             {
 409                 jnr              = jjnr[k];
 410                 weight_cg2       = wf[jnr];
 411                 weight_product   = weight_cg1*weight_cg2;
 412                 if (weight_product < ALMOST_ZERO) {
 413                 /* force is zero, skip this molecule */
 414                  continue;
 415                 }
 416                 else if (weight_product >= ALMOST_ONE)
 417                 {
 418                        hybscal = 1.0;
 419                 }
 420                 else
 421                 {
 422                    hybscal = weight_product;
 423                 }
 424                 j3               = 3*jnr;
 425                 jx1              = pos[j3+0];
 426                 jy1              = pos[j3+1];
 427                 jz1              = pos[j3+2];
 428                 dx11             = ix1 - jx1;
 429                 dy11             = iy1 - jy1;
 430                 dz11             = iz1 - jz1;
 431                 rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 432                 dx21             = ix2 - jx1;
 433                 dy21             = iy2 - jy1;
 434                 dz21             = iz2 - jz1;
 435                 rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 436                 dx31             = ix3 - jx1;
 437                 dy31             = iy3 - jy1;
 438                 dz31             = iz3 - jz1;
 439                 rsq31            = dx31*dx31+dy31*dy31+dz31*dz31;
 440                 rinv11           = 1.0/sqrt(rsq11);
 441                 rinv21           = 1.0/sqrt(rsq21);
 442                 rinv31           = 1.0/sqrt(rsq31);
 443                 jq               = charge[jnr+0];
 444                 qq               = qO*jq;
 445                 tj               = nti+2*type[jnr];
 446                 c6               = vdwparam[tj];
 447                 c12              = vdwparam[tj+1];
 448                 rinvsq           = rinv11*rinv11;
 449                 vcoul            = qq*rinv11;
 450                 vctot            = vctot+vcoul;
 451                 rinvsix          = rinvsq*rinvsq*rinvsq;
 452                 Vvdw6            = c6*rinvsix;
 453                 Vvdw12           = c12*rinvsix*rinvsix;
 454                 Vvdwtot          = Vvdwtot+Vvdw12-Vvdw6;
 455                 fscal            = (vcoul+12.0*Vvdw12-6.0*Vvdw6)*rinvsq;
 456                 fscal *= hybscal;
 457                 if(force_cap>0 && (fabs(fscal)> force_cap)){
 458                 fscal=force_cap*fscal/fabs(fscal);
 459                 }
 460                 tx               = fscal*dx11;
 461                 ty               = fscal*dy11;
 462                 tz               = fscal*dz11;
 463                 fix1             = fix1 + tx;
 464                 fiy1             = fiy1 + ty;
 465                 fiz1             = fiz1 + tz;
 466                 fjx1             = faction[j3+0] - tx;
 467                 fjy1             = faction[j3+1] - ty;
 468                 fjz1             = faction[j3+2] - tz;
 469                 qq               = qH*jq;
 470                 rinvsq           = rinv21*rinv21;
 471                 vcoul            = qq*rinv21;
 472                 vctot            = vctot+vcoul;
 473                 fscal            = (vcoul)*rinvsq;
 474                 fscal *= hybscal;
 475                 if(force_cap>0 && (fabs(fscal)> force_cap)){
 476                 fscal=force_cap*fscal/fabs(fscal);
 477                 }
 478                 tx               = fscal*dx21;
 479                 ty               = fscal*dy21;
 480                 tz               = fscal*dz21;
 481                 fix2             = fix2 + tx;
 482                 fiy2             = fiy2 + ty;
 483                 fiz2             = fiz2 + tz;
 484                 fjx1             = fjx1 - tx;
 485                 fjy1             = fjy1 - ty;
 486                 fjz1             = fjz1 - tz;
 487                 rinvsq           = rinv31*rinv31;
 488                 vcoul            = qq*rinv31;
 489                 vctot            = vctot+vcoul;
 490                 fscal            = (vcoul)*rinvsq;
 491                 fscal *= hybscal;
 492                 if(force_cap>0 && (fabs(fscal)> force_cap)){
 493                 fscal=force_cap*fscal/fabs(fscal);
 494                 }
 495                 tx               = fscal*dx31;
 496                 ty               = fscal*dy31;
 497                 tz               = fscal*dz31;
 498                 fix3             = fix3 + tx;
 499                 fiy3             = fiy3 + ty;
 500                 fiz3             = fiz3 + tz;
 501                 faction[j3+0]    = fjx1 - tx;
 502                 faction[j3+1]    = fjy1 - ty;
 503                 faction[j3+2]    = fjz1 - tz;
 504             }
 505
 506             faction[ii3+0]   = faction[ii3+0] + fix1;
 507             faction[ii3+1]   = faction[ii3+1] + fiy1;
 508             faction[ii3+2]   = faction[ii3+2] + fiz1;
 509             faction[ii3+3]   = faction[ii3+3] + fix2;
 510             faction[ii3+4]   = faction[ii3+4] + fiy2;
 511             faction[ii3+5]   = faction[ii3+5] + fiz2;
 512             faction[ii3+6]   = faction[ii3+6] + fix3;
 513             faction[ii3+7]   = faction[ii3+7] + fiy3;
 514             faction[ii3+8]   = faction[ii3+8] + fiz3;
 515             fshift[is3]      = fshift[is3]+fix1+fix2+fix3;
 516             fshift[is3+1]    = fshift[is3+1]+fiy1+fiy2+fiy3;
 517             fshift[is3+2]    = fshift[is3+2]+fiz1+fiz2+fiz3;
 518             ggid             = gid[n];
 519             Vc[ggid]         = Vc[ggid] + vctot;
 520             Vvdw[ggid]       = Vvdw[ggid] + Vvdwtot;
 521             ninner           = ninner + nj1 - nj0;
 522         }
 523
 524         nouter           = nouter + nn1 - nn0;
 525     }
 526     while (nn1<nri);
 527
 528     *outeriter       = nouter;
 529     *inneriter       = ninner;
 530 }
 531
 532