src/gromacs/mdlib/nbnxn_kernels/simd_4xn/nbnxn_kernel_simd_4xn_outer.h

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   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.
   8  *
   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.
  13  *
  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.
  18  *
  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.
  23  *
  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.
  31  *
  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.
  34  */
  35
  36
  37 {
  38     const nbnxn_ci_t   *nbln;
  39     const nbnxn_cj_t   *l_cj;
  40     const int          *type;
  41     const real         *q;
  42     const real         *shiftvec;
  43     const real         *x;
  44     const real         *nbfp0, *nbfp1, *nbfp2 = NULL, *nbfp3 = NULL;
  45     real                facel;
  46     real               *nbfp_ptr;
  47     int                 n, ci, ci_sh;
  48     int                 ish, ish3;
  49     gmx_bool            do_LJ, half_LJ, do_coul;
  50     int                 sci, scix, sciy, sciz, sci2;
  51     int                 cjind0, cjind1, cjind;
  52     int                 ip, jp;
  53
  54 #ifdef ENERGY_GROUPS
  55     int         Vstride_i;
  56     int         egps_ishift, egps_imask;
  57     int         egps_jshift, egps_jmask, egps_jstride;
  58     int         egps_i;
  59     real       *vvdwtp[UNROLLI];
  60     real       *vctp[UNROLLI];
  61 #endif
  62
  63     gmx_simd_real_t  shX_S;
  64     gmx_simd_real_t  shY_S;
  65     gmx_simd_real_t  shZ_S;
  66     gmx_simd_real_t  ix_S0, iy_S0, iz_S0;
  67     gmx_simd_real_t  ix_S1, iy_S1, iz_S1;
  68     gmx_simd_real_t  ix_S2, iy_S2, iz_S2;
  69     gmx_simd_real_t  ix_S3, iy_S3, iz_S3;
  70     gmx_simd_real_t  fix_S0, fiy_S0, fiz_S0;
  71     gmx_simd_real_t  fix_S1, fiy_S1, fiz_S1;
  72     gmx_simd_real_t  fix_S2, fiy_S2, fiz_S2;
  73     gmx_simd_real_t  fix_S3, fiy_S3, fiz_S3;
  74 #if UNROLLJ >= 4
  75     /* We use an i-force SIMD register width of 4 */
  76     gmx_simd4_real_t fix_S, fiy_S, fiz_S;
  77 #else
  78     /* We use an i-force SIMD register width of 2 */
  79     gmx_simd_real_t  fix0_S, fiy0_S, fiz0_S;
  80     gmx_simd_real_t  fix2_S, fiy2_S, fiz2_S;
  81 #endif
  82
  83     gmx_simd_real_t  diagonal_jmi_S;
  84 #if UNROLLI == UNROLLJ
  85     gmx_simd_bool_t  diagonal_mask_S0, diagonal_mask_S1, diagonal_mask_S2, diagonal_mask_S3;
  86 #else
  87     gmx_simd_bool_t  diagonal_mask0_S0, diagonal_mask0_S1, diagonal_mask0_S2, diagonal_mask0_S3;
  88     gmx_simd_bool_t  diagonal_mask1_S0, diagonal_mask1_S1, diagonal_mask1_S2, diagonal_mask1_S3;
  89 #endif
  90
  91     unsigned            *exclusion_filter;
  92     gmx_exclfilter       filter_S0, filter_S1, filter_S2, filter_S3;
  93
  94     gmx_simd_real_t      zero_S = gmx_simd_set1_r(0.0);
  95
  96     gmx_simd_real_t      one_S  = gmx_simd_set1_r(1.0);
  97     gmx_simd_real_t      iq_S0  = gmx_simd_setzero_r();
  98     gmx_simd_real_t      iq_S1  = gmx_simd_setzero_r();
  99     gmx_simd_real_t      iq_S2  = gmx_simd_setzero_r();
 100     gmx_simd_real_t      iq_S3  = gmx_simd_setzero_r();
 101
 102 #ifdef CALC_COUL_RF
 103     gmx_simd_real_t      mrc_3_S;
 104 #ifdef CALC_ENERGIES
 105     gmx_simd_real_t      hrc_3_S, moh_rc_S;
 106 #endif
 107 #endif
 108
 109 #ifdef CALC_COUL_TAB
 110     /* Coulomb table variables */
 111     gmx_simd_real_t   invtsp_S;
 112     const real       *tab_coul_F;
 113 #ifndef TAB_FDV0
 114     const real       *tab_coul_V;
 115 #endif
 116     /* Thread-local working buffers for force and potential lookups */
 117     int               ti0_array[2*GMX_SIMD_REAL_WIDTH], *ti0 = NULL;
 118     int               ti1_array[2*GMX_SIMD_REAL_WIDTH], *ti1 = NULL;
 119     int               ti2_array[2*GMX_SIMD_REAL_WIDTH], *ti2 = NULL;
 120     int               ti3_array[2*GMX_SIMD_REAL_WIDTH], *ti3 = NULL;
 121 #ifdef CALC_ENERGIES
 122     gmx_simd_real_t   mhalfsp_S;
 123 #endif
 124 #endif
 125
 126 #ifdef CALC_COUL_EWALD
 127     gmx_simd_real_t beta2_S, beta_S;
 128 #endif
 129
 130 #if defined CALC_ENERGIES && (defined CALC_COUL_EWALD || defined CALC_COUL_TAB)
 131     gmx_simd_real_t  sh_ewald_S;
 132 #endif
 133
 134 #ifdef LJ_POT_SWITCH
 135     gmx_simd_real_t   rswitch_S;
 136     gmx_simd_real_t   swV3_S, swV4_S, swV5_S;
 137     gmx_simd_real_t   swF2_S, swF3_S, swF4_S;
 138 #else
 139 #ifdef LJ_FORCE_SWITCH
 140     gmx_simd_real_t   rswitch_S;
 141     gmx_simd_real_t   p6_fc2_S, p6_fc3_S;
 142     gmx_simd_real_t   p12_fc2_S, p12_fc3_S;
 143 #ifdef CALC_ENERGIES
 144     gmx_simd_real_t   p6_vc3_S, p6_vc4_S;
 145     gmx_simd_real_t   p12_vc3_S, p12_vc4_S;
 146     gmx_simd_real_t   p6_6cpot_S, p12_12cpot_S;
 147 #endif
 148 #else
 149 #ifdef CALC_ENERGIES
 150     gmx_simd_real_t  p6_cpot_S, p12_cpot_S;
 151 #endif
 152 #endif
 153 #endif
 154
 155 #ifdef LJ_COMB_LB
 156     const real       *ljc;
 157
 158     gmx_simd_real_t   hsig_i_S0, seps_i_S0;
 159     gmx_simd_real_t   hsig_i_S1, seps_i_S1;
 160     gmx_simd_real_t   hsig_i_S2, seps_i_S2;
 161     gmx_simd_real_t   hsig_i_S3, seps_i_S3;
 162 #else
 163 #ifdef FIX_LJ_C
 164     real              pvdw_array[2*UNROLLI*UNROLLJ+3];
 165     real             *pvdw_c6, *pvdw_c12;
 166     gmx_simd_real_t   c6_S0, c12_S0;
 167     gmx_simd_real_t   c6_S1, c12_S1;
 168     gmx_simd_real_t   c6_S2, c12_S2;
 169     gmx_simd_real_t   c6_S3, c12_S3;
 170 #endif
 171
 172 #ifdef LJ_COMB_GEOM
 173     const real       *ljc;
 174
 175     gmx_simd_real_t   c6s_S0, c12s_S0;
 176     gmx_simd_real_t   c6s_S1, c12s_S1;
 177     gmx_simd_real_t   c6s_S2  = gmx_simd_setzero_r();
 178     gmx_simd_real_t   c12s_S2 = gmx_simd_setzero_r();
 179     gmx_simd_real_t   c6s_S3  = gmx_simd_setzero_r();
 180     gmx_simd_real_t   c12s_S3 = gmx_simd_setzero_r();
 181 #endif
 182 #endif /* LJ_COMB_LB */
 183
 184     gmx_simd_real_t  vctot_S, Vvdwtot_S;
 185     gmx_simd_real_t  sixth_S, twelveth_S;
 186
 187     gmx_simd_real_t  avoid_sing_S;
 188     gmx_simd_real_t  rc2_S;
 189 #ifdef VDW_CUTOFF_CHECK
 190     gmx_simd_real_t  rcvdw2_S;
 191 #endif
 192
 193 #ifdef CALC_ENERGIES
 194     /* cppcheck-suppress unassignedVariable */
 195     real       tmpsum_array[GMX_SIMD_REAL_WIDTH*2], *tmpsum;
 196 #endif
 197 #ifdef CALC_SHIFTFORCES
 198     /* cppcheck-suppress unassignedVariable */
 199     real       shf_array[GMX_SIMD_REAL_WIDTH*2], *shf;
 200 #endif
 201
 202     int ninner;
 203
 204 #ifdef COUNT_PAIRS
 205     int npair = 0;
 206 #endif
 207
 208 #if defined LJ_COMB_GEOM || defined LJ_COMB_LB
 209     ljc = nbat->lj_comb;
 210 #else
 211     /* No combination rule used */
 212     nbfp_ptr    = (4 == nbfp_stride) ? nbat->nbfp_s4 : nbat->nbfp;
 213 #endif
 214
 215     /* Load j-i for the first i */
 216     diagonal_jmi_S    = gmx_simd_load_r(nbat->simd_4xn_diagonal_j_minus_i);
 217     /* Generate all the diagonal masks as comparison results */
 218 #if UNROLLI == UNROLLJ
 219     diagonal_mask_S0  = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 220     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 221     diagonal_mask_S1  = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 222     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 223     diagonal_mask_S2  = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 224     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 225     diagonal_mask_S3  = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 226 #else
 227 #if UNROLLI == 2*UNROLLJ || 2*UNROLLI == UNROLLJ
 228     diagonal_mask0_S0 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 229     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 230     diagonal_mask0_S1 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 231     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 232     diagonal_mask0_S2 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 233     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 234     diagonal_mask0_S3 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 235     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 236
 237 #if UNROLLI == 2*UNROLLJ
 238     /* Load j-i for the second half of the j-cluster */
 239     diagonal_jmi_S    = gmx_simd_load_r(nbat->simd_4xn_diagonal_j_minus_i + UNROLLJ);
 240 #endif
 241
 242     diagonal_mask1_S0 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 243     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 244     diagonal_mask1_S1 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 245     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 246     diagonal_mask1_S2 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 247     diagonal_jmi_S    = gmx_simd_sub_r(diagonal_jmi_S, one_S);
 248     diagonal_mask1_S3 = gmx_simd_cmplt_r(zero_S, diagonal_jmi_S);
 249 #endif
 250 #endif
 251
 252     /* Load masks for topology exclusion masking. filter_stride is
 253        static const, so the conditional will be optimized away. */
 254     if (1 == filter_stride)
 255     {
 256         exclusion_filter = nbat->simd_exclusion_filter1;
 257     }
 258     else /* (2 == filter_stride) */
 259     {
 260         exclusion_filter = nbat->simd_exclusion_filter2;
 261     }
 262
 263     /* Here we cast the exclusion filters from unsigned * to int * or real *.
 264      * Since we only check bits, the actual value they represent does not
 265      * matter, as long as both filter and mask data are treated the same way.
 266      */
 267     filter_S0 = gmx_load_exclusion_filter(exclusion_filter + 0*UNROLLJ*filter_stride);
 268     filter_S1 = gmx_load_exclusion_filter(exclusion_filter + 1*UNROLLJ*filter_stride);
 269     filter_S2 = gmx_load_exclusion_filter(exclusion_filter + 2*UNROLLJ*filter_stride);
 270     filter_S3 = gmx_load_exclusion_filter(exclusion_filter + 3*UNROLLJ*filter_stride);
 271
 272 #ifdef CALC_COUL_RF
 273     /* Reaction-field constants */
 274     mrc_3_S  = gmx_simd_set1_r(-2*ic->k_rf);
 275 #ifdef CALC_ENERGIES
 276     hrc_3_S  = gmx_simd_set1_r(ic->k_rf);
 277     moh_rc_S = gmx_simd_set1_r(-ic->c_rf);
 278 #endif
 279 #endif
 280
 281 #ifdef CALC_COUL_TAB
 282     /* Generate aligned table index pointers */
 283     ti0 = prepare_table_load_buffer(ti0_array);
 284     ti1 = prepare_table_load_buffer(ti1_array);
 285     ti2 = prepare_table_load_buffer(ti2_array);
 286     ti3 = prepare_table_load_buffer(ti3_array);
 287
 288     invtsp_S  = gmx_simd_set1_r(ic->tabq_scale);
 289 #ifdef CALC_ENERGIES
 290     mhalfsp_S = gmx_simd_set1_r(-0.5/ic->tabq_scale);
 291 #endif
 292
 293 #ifdef TAB_FDV0
 294     tab_coul_F = ic->tabq_coul_FDV0;
 295 #else
 296     tab_coul_F = ic->tabq_coul_F;
 297     tab_coul_V = ic->tabq_coul_V;
 298 #endif
 299 #endif /* CALC_COUL_TAB */
 300
 301 #ifdef CALC_COUL_EWALD
 302     beta2_S = gmx_simd_set1_r(ic->ewaldcoeff_q*ic->ewaldcoeff_q);
 303     beta_S  = gmx_simd_set1_r(ic->ewaldcoeff_q);
 304 #endif
 305
 306 #if (defined CALC_COUL_TAB || defined CALC_COUL_EWALD) && defined CALC_ENERGIES
 307     sh_ewald_S = gmx_simd_set1_r(ic->sh_ewald);
 308 #endif
 309
 310     /* LJ function constants */
 311 #if defined CALC_ENERGIES || defined LJ_POT_SWITCH
 312     sixth_S      = gmx_simd_set1_r(1.0/6.0);
 313     twelveth_S   = gmx_simd_set1_r(1.0/12.0);
 314 #endif
 315
 316 #ifdef LJ_POT_SWITCH
 317     rswitch_S = gmx_simd_set1_r(ic->rvdw_switch);
 318     swV3_S    = gmx_simd_set1_r(ic->vdw_switch.c3);
 319     swV4_S    = gmx_simd_set1_r(ic->vdw_switch.c4);
 320     swV5_S    = gmx_simd_set1_r(ic->vdw_switch.c5);
 321     swF2_S    = gmx_simd_set1_r(3*ic->vdw_switch.c3);
 322     swF3_S    = gmx_simd_set1_r(4*ic->vdw_switch.c4);
 323     swF4_S    = gmx_simd_set1_r(5*ic->vdw_switch.c5);
 324 #else
 325     sixth_S      = gmx_simd_set1_r(1.0/6.0);
 326     twelveth_S   = gmx_simd_set1_r(1.0/12.0);
 327 #ifdef LJ_FORCE_SWITCH
 328     rswitch_S = gmx_simd_set1_r(ic->rvdw_switch);
 329     p6_fc2_S  = gmx_simd_set1_r(ic->dispersion_shift.c2);
 330     p6_fc3_S  = gmx_simd_set1_r(ic->dispersion_shift.c3);
 331     p12_fc2_S = gmx_simd_set1_r(ic->repulsion_shift.c2);
 332     p12_fc3_S = gmx_simd_set1_r(ic->repulsion_shift.c3);
 333 #ifdef CALC_ENERGIES
 334     {
 335         gmx_simd_real_t mthird_S  = gmx_simd_set1_r(-1.0/3.0);
 336         gmx_simd_real_t mfourth_S = gmx_simd_set1_r(-1.0/4.0);
 337
 338         p6_vc3_S     = gmx_simd_mul_r(mthird_S,  p6_fc2_S);
 339         p6_vc4_S     = gmx_simd_mul_r(mfourth_S, p6_fc3_S);
 340         p6_6cpot_S   = gmx_simd_set1_r(ic->dispersion_shift.cpot/6);
 341         p12_vc3_S    = gmx_simd_mul_r(mthird_S,  p12_fc2_S);
 342         p12_vc4_S    = gmx_simd_mul_r(mfourth_S, p12_fc3_S);
 343         p12_12cpot_S = gmx_simd_set1_r(ic->repulsion_shift.cpot/12);
 344     }
 345 #endif
 346 #else
 347     /* Plain LJ cut-off, with potential shift cpot, which can be 0 */
 348 #ifdef CALC_ENERGIES
 349     p6_cpot_S    = gmx_simd_set1_r(ic->dispersion_shift.cpot);
 350     p12_cpot_S   = gmx_simd_set1_r(ic->repulsion_shift.cpot);
 351 #endif
 352 #endif
 353 #endif /* LJ_POT_SWITCH */
 354
 355     /* The kernel either supports rcoulomb = rvdw or rcoulomb >= rvdw */
 356     rc2_S    = gmx_simd_set1_r(ic->rcoulomb*ic->rcoulomb);
 357 #ifdef VDW_CUTOFF_CHECK
 358     rcvdw2_S = gmx_simd_set1_r(ic->rvdw*ic->rvdw);
 359 #endif
 360
 361     avoid_sing_S = gmx_simd_set1_r(NBNXN_AVOID_SING_R2_INC);
 362
 363     q                   = nbat->q;
 364     type                = nbat->type;
 365     facel               = ic->epsfac;
 366     shiftvec            = shift_vec[0];
 367     x                   = nbat->x;
 368
 369 #ifdef CALC_ENERGIES
 370     tmpsum   = gmx_simd_align_r(tmpsum_array);
 371 #endif
 372 #ifdef CALC_SHIFTFORCES
 373     shf      = gmx_simd_align_r(shf_array);
 374 #endif
 375
 376 #ifdef FIX_LJ_C
 377     pvdw_c6  = gmx_simd_align_real(pvdw_array);
 378     pvdw_c12 = pvdw_c6 + UNROLLI*UNROLLJ;
 379
 380     for (jp = 0; jp < UNROLLJ; jp++)
 381     {
 382         pvdw_c6 [0*UNROLLJ+jp] = nbat->nbfp[0*2];
 383         pvdw_c6 [1*UNROLLJ+jp] = nbat->nbfp[0*2];
 384         pvdw_c6 [2*UNROLLJ+jp] = nbat->nbfp[0*2];
 385         pvdw_c6 [3*UNROLLJ+jp] = nbat->nbfp[0*2];
 386
 387         pvdw_c12[0*UNROLLJ+jp] = nbat->nbfp[0*2+1];
 388         pvdw_c12[1*UNROLLJ+jp] = nbat->nbfp[0*2+1];
 389         pvdw_c12[2*UNROLLJ+jp] = nbat->nbfp[0*2+1];
 390         pvdw_c12[3*UNROLLJ+jp] = nbat->nbfp[0*2+1];
 391     }
 392     c6_S0            = gmx_simd_load_r(pvdw_c6 +0*UNROLLJ);
 393     c6_S1            = gmx_simd_load_r(pvdw_c6 +1*UNROLLJ);
 394     c6_S2            = gmx_simd_load_r(pvdw_c6 +2*UNROLLJ);
 395     c6_S3            = gmx_simd_load_r(pvdw_c6 +3*UNROLLJ);
 396
 397     c12_S0           = gmx_simd_load_r(pvdw_c12+0*UNROLLJ);
 398     c12_S1           = gmx_simd_load_r(pvdw_c12+1*UNROLLJ);
 399     c12_S2           = gmx_simd_load_r(pvdw_c12+2*UNROLLJ);
 400     c12_S3           = gmx_simd_load_r(pvdw_c12+3*UNROLLJ);
 401 #endif /* FIX_LJ_C */
 402
 403 #ifdef ENERGY_GROUPS
 404     egps_ishift  = nbat->neg_2log;
 405     egps_imask   = (1<<egps_ishift) - 1;
 406     egps_jshift  = 2*nbat->neg_2log;
 407     egps_jmask   = (1<<egps_jshift) - 1;
 408     egps_jstride = (UNROLLJ>>1)*UNROLLJ;
 409     /* Major division is over i-particle energy groups, determine the stride */
 410     Vstride_i    = nbat->nenergrp*(1<<nbat->neg_2log)*egps_jstride;
 411 #endif
 412
 413     l_cj = nbl->cj;
 414
 415     ninner = 0;
 416     for (n = 0; n < nbl->nci; n++)
 417     {
 418         nbln = &nbl->ci[n];
 419
 420         ish              = (nbln->shift & NBNXN_CI_SHIFT);
 421         ish3             = ish*3;
 422         cjind0           = nbln->cj_ind_start;
 423         cjind1           = nbln->cj_ind_end;
 424         ci               = nbln->ci;
 425         ci_sh            = (ish == CENTRAL ? ci : -1);
 426
 427         shX_S = gmx_simd_load1_r(shiftvec+ish3);
 428         shY_S = gmx_simd_load1_r(shiftvec+ish3+1);
 429         shZ_S = gmx_simd_load1_r(shiftvec+ish3+2);
 430
 431 #if UNROLLJ <= 4
 432         sci              = ci*STRIDE;
 433         scix             = sci*DIM;
 434         sci2             = sci*2;
 435 #else
 436         sci              = (ci>>1)*STRIDE;
 437         scix             = sci*DIM + (ci & 1)*(STRIDE>>1);
 438         sci2             = sci*2 + (ci & 1)*(STRIDE>>1);
 439         sci             += (ci & 1)*(STRIDE>>1);
 440 #endif
 441
 442         /* We have 5 LJ/C combinations, but use only three inner loops,
 443          * as the other combinations are unlikely and/or not much faster:
 444          * inner half-LJ + C for half-LJ + C / no-LJ + C
 445          * inner LJ + C      for full-LJ + C
 446          * inner LJ          for full-LJ + no-C / half-LJ + no-C
 447          */
 448         do_LJ   = (nbln->shift & NBNXN_CI_DO_LJ(0));
 449         do_coul = (nbln->shift & NBNXN_CI_DO_COUL(0));
 450         half_LJ = ((nbln->shift & NBNXN_CI_HALF_LJ(0)) || !do_LJ) && do_coul;
 451
 452 #ifdef ENERGY_GROUPS
 453         egps_i = nbat->energrp[ci];
 454         {
 455             int ia, egp_ia;
 456
 457             for (ia = 0; ia < UNROLLI; ia++)
 458             {
 459                 egp_ia     = (egps_i >> (ia*egps_ishift)) & egps_imask;
 460                 vvdwtp[ia] = Vvdw + egp_ia*Vstride_i;
 461                 vctp[ia]   = Vc   + egp_ia*Vstride_i;
 462             }
 463         }
 464 #endif
 465 #if defined CALC_ENERGIES
 466 #if UNROLLJ == 4
 467         if (do_coul && l_cj[nbln->cj_ind_start].cj == ci_sh)
 468 #endif
 469 #if UNROLLJ == 2
 470         if (do_coul && l_cj[nbln->cj_ind_start].cj == (ci_sh<<1))
 471 #endif
 472 #if UNROLLJ == 8
 473         if (do_coul && l_cj[nbln->cj_ind_start].cj == (ci_sh>>1))
 474 #endif
 475         {
 476             int  ia;
 477             real Vc_sub_self;
 478
 479 #ifdef CALC_COUL_RF
 480             Vc_sub_self = 0.5*ic->c_rf;
 481 #endif
 482 #ifdef CALC_COUL_TAB
 483 #ifdef TAB_FDV0
 484             Vc_sub_self = 0.5*tab_coul_F[2];
 485 #else
 486             Vc_sub_self = 0.5*tab_coul_V[0];
 487 #endif
 488 #endif
 489 #ifdef CALC_COUL_EWALD
 490             /* beta/sqrt(pi) */
 491             Vc_sub_self = 0.5*ic->ewaldcoeff_q*M_2_SQRTPI;
 492 #endif
 493
 494             for (ia = 0; ia < UNROLLI; ia++)
 495             {
 496                 real qi;
 497
 498                 qi = q[sci+ia];
 499 #ifdef ENERGY_GROUPS
 500                 vctp[ia][((egps_i>>(ia*egps_ishift)) & egps_imask)*egps_jstride]
 501 #else
 502                 Vc[0]
 503 #endif
 504                     -= facel*qi*qi*Vc_sub_self;
 505             }
 506         }
 507 #endif
 508
 509         /* Load i atom data */
 510         sciy             = scix + STRIDE;
 511         sciz             = sciy + STRIDE;
 512         ix_S0            = gmx_simd_add_r(gmx_simd_load1_r(x+scix), shX_S);
 513         ix_S1            = gmx_simd_add_r(gmx_simd_load1_r(x+scix+1), shX_S);
 514         ix_S2            = gmx_simd_add_r(gmx_simd_load1_r(x+scix+2), shX_S);
 515         ix_S3            = gmx_simd_add_r(gmx_simd_load1_r(x+scix+3), shX_S);
 516         iy_S0            = gmx_simd_add_r(gmx_simd_load1_r(x+sciy), shY_S);
 517         iy_S1            = gmx_simd_add_r(gmx_simd_load1_r(x+sciy+1), shY_S);
 518         iy_S2            = gmx_simd_add_r(gmx_simd_load1_r(x+sciy+2), shY_S);
 519         iy_S3            = gmx_simd_add_r(gmx_simd_load1_r(x+sciy+3), shY_S);
 520         iz_S0            = gmx_simd_add_r(gmx_simd_load1_r(x+sciz), shZ_S);
 521         iz_S1            = gmx_simd_add_r(gmx_simd_load1_r(x+sciz+1), shZ_S);
 522         iz_S2            = gmx_simd_add_r(gmx_simd_load1_r(x+sciz+2), shZ_S);
 523         iz_S3            = gmx_simd_add_r(gmx_simd_load1_r(x+sciz+3), shZ_S);
 524
 525         if (do_coul)
 526         {
 527             iq_S0      = gmx_simd_set1_r(facel*q[sci]);
 528             iq_S1      = gmx_simd_set1_r(facel*q[sci+1]);
 529             iq_S2      = gmx_simd_set1_r(facel*q[sci+2]);
 530             iq_S3      = gmx_simd_set1_r(facel*q[sci+3]);
 531         }
 532
 533 #ifdef LJ_COMB_LB
 534         hsig_i_S0      = gmx_simd_load1_r(ljc+sci2+0);
 535         hsig_i_S1      = gmx_simd_load1_r(ljc+sci2+1);
 536         hsig_i_S2      = gmx_simd_load1_r(ljc+sci2+2);
 537         hsig_i_S3      = gmx_simd_load1_r(ljc+sci2+3);
 538         seps_i_S0      = gmx_simd_load1_r(ljc+sci2+STRIDE+0);
 539         seps_i_S1      = gmx_simd_load1_r(ljc+sci2+STRIDE+1);
 540         seps_i_S2      = gmx_simd_load1_r(ljc+sci2+STRIDE+2);
 541         seps_i_S3      = gmx_simd_load1_r(ljc+sci2+STRIDE+3);
 542 #else
 543 #ifdef LJ_COMB_GEOM
 544         c6s_S0         = gmx_simd_load1_r(ljc+sci2+0);
 545         c6s_S1         = gmx_simd_load1_r(ljc+sci2+1);
 546         if (!half_LJ)
 547         {
 548             c6s_S2     = gmx_simd_load1_r(ljc+sci2+2);
 549             c6s_S3     = gmx_simd_load1_r(ljc+sci2+3);
 550         }
 551         c12s_S0        = gmx_simd_load1_r(ljc+sci2+STRIDE+0);
 552         c12s_S1        = gmx_simd_load1_r(ljc+sci2+STRIDE+1);
 553         if (!half_LJ)
 554         {
 555             c12s_S2    = gmx_simd_load1_r(ljc+sci2+STRIDE+2);
 556             c12s_S3    = gmx_simd_load1_r(ljc+sci2+STRIDE+3);
 557         }
 558 #else
 559         nbfp0     = nbfp_ptr + type[sci  ]*nbat->ntype*nbfp_stride;
 560         nbfp1     = nbfp_ptr + type[sci+1]*nbat->ntype*nbfp_stride;
 561         if (!half_LJ)
 562         {
 563             nbfp2 = nbfp_ptr + type[sci+2]*nbat->ntype*nbfp_stride;
 564             nbfp3 = nbfp_ptr + type[sci+3]*nbat->ntype*nbfp_stride;
 565         }
 566 #endif
 567 #endif
 568
 569         /* Zero the potential energy for this list */
 570         Vvdwtot_S        = gmx_simd_setzero_r();
 571         vctot_S          = gmx_simd_setzero_r();
 572
 573         /* Clear i atom forces */
 574         fix_S0           = gmx_simd_setzero_r();
 575         fix_S1           = gmx_simd_setzero_r();
 576         fix_S2           = gmx_simd_setzero_r();
 577         fix_S3           = gmx_simd_setzero_r();
 578         fiy_S0           = gmx_simd_setzero_r();
 579         fiy_S1           = gmx_simd_setzero_r();
 580         fiy_S2           = gmx_simd_setzero_r();
 581         fiy_S3           = gmx_simd_setzero_r();
 582         fiz_S0           = gmx_simd_setzero_r();
 583         fiz_S1           = gmx_simd_setzero_r();
 584         fiz_S2           = gmx_simd_setzero_r();
 585         fiz_S3           = gmx_simd_setzero_r();
 586
 587         cjind = cjind0;
 588
 589         /* Currently all kernels use (at least half) LJ */
 590 #define CALC_LJ
 591         if (half_LJ)
 592         {
 593             /* Coulomb: all i-atoms, LJ: first half i-atoms */
 594 #define CALC_COULOMB
 595 #define HALF_LJ
 596 #define CHECK_EXCLS
 597             while (cjind < cjind1 && nbl->cj[cjind].excl != NBNXN_INTERACTION_MASK_ALL)
 598             {
 599 #include "nbnxn_kernel_simd_4xn_inner.h"
 600                 cjind++;
 601             }
 602 #undef CHECK_EXCLS
 603             for (; (cjind < cjind1); cjind++)
 604             {
 605 #include "nbnxn_kernel_simd_4xn_inner.h"
 606             }
 607 #undef HALF_LJ
 608 #undef CALC_COULOMB
 609         }
 610         else if (do_coul)
 611         {
 612             /* Coulomb: all i-atoms, LJ: all i-atoms */
 613 #define CALC_COULOMB
 614 #define CHECK_EXCLS
 615             while (cjind < cjind1 && nbl->cj[cjind].excl != NBNXN_INTERACTION_MASK_ALL)
 616             {
 617 #include "nbnxn_kernel_simd_4xn_inner.h"
 618                 cjind++;
 619             }
 620 #undef CHECK_EXCLS
 621             for (; (cjind < cjind1); cjind++)
 622             {
 623 #include "nbnxn_kernel_simd_4xn_inner.h"
 624             }
 625 #undef CALC_COULOMB
 626         }
 627         else
 628         {
 629             /* Coulomb: none, LJ: all i-atoms */
 630 #define CHECK_EXCLS
 631             while (cjind < cjind1 && nbl->cj[cjind].excl != NBNXN_INTERACTION_MASK_ALL)
 632             {
 633 #include "nbnxn_kernel_simd_4xn_inner.h"
 634                 cjind++;
 635             }
 636 #undef CHECK_EXCLS
 637             for (; (cjind < cjind1); cjind++)
 638             {
 639 #include "nbnxn_kernel_simd_4xn_inner.h"
 640             }
 641         }
 642 #undef CALC_LJ
 643         ninner += cjind1 - cjind0;
 644
 645         /* Add accumulated i-forces to the force array */
 646 #if UNROLLJ >= 4
 647         fix_S = gmx_mm_transpose_sum4_pr(fix_S0, fix_S1, fix_S2, fix_S3);
 648         gmx_simd4_store_r(f+scix, gmx_add_pr4(fix_S, gmx_load_pr4(f+scix)));
 649
 650         fiy_S = gmx_mm_transpose_sum4_pr(fiy_S0, fiy_S1, fiy_S2, fiy_S3);
 651         gmx_simd4_store_r(f+sciy, gmx_add_pr4(fiy_S, gmx_load_pr4(f+sciy)));
 652
 653         fiz_S = gmx_mm_transpose_sum4_pr(fiz_S0, fiz_S1, fiz_S2, fiz_S3);
 654         gmx_simd4_store_r(f+sciz, gmx_add_pr4(fiz_S, gmx_load_pr4(f+sciz)));
 655
 656 #ifdef CALC_SHIFTFORCES
 657         fshift[ish3+0] += gmx_sum_simd4(fix_S, shf);
 658         fshift[ish3+1] += gmx_sum_simd4(fiy_S, shf);
 659         fshift[ish3+2] += gmx_sum_simd4(fiz_S, shf);
 660 #endif
 661 #else
 662         fix0_S = gmx_mm_transpose_sum2_pr(fix_S0, fix_S1);
 663         gmx_simd_store_r(f+scix, gmx_simd_add_r(fix0_S, gmx_simd_load_r(f+scix)));
 664         fix2_S = gmx_mm_transpose_sum2_pr(fix_S2, fix_S3);
 665         gmx_simd_store_r(f+scix+2, gmx_simd_add_r(fix2_S, gmx_simd_load_r(f+scix+2)));
 666
 667         fiy0_S = gmx_mm_transpose_sum2_pr(fiy_S0, fiy_S1);
 668         gmx_simd_store_r(f+sciy, gmx_simd_add_r(fiy0_S, gmx_simd_load_r(f+sciy)));
 669         fiy2_S = gmx_mm_transpose_sum2_pr(fiy_S2, fiy_S3);
 670         gmx_simd_store_r(f+sciy+2, gmx_simd_add_r(fiy2_S, gmx_simd_load_r(f+sciy+2)));
 671
 672         fiz0_S = gmx_mm_transpose_sum2_pr(fiz_S0, fiz_S1);
 673         gmx_simd_store_r(f+sciz, gmx_simd_add_r(fiz0_S, gmx_simd_load_r(f+sciz)));
 674         fiz2_S = gmx_mm_transpose_sum2_pr(fiz_S2, fiz_S3);
 675         gmx_simd_store_r(f+sciz+2, gmx_simd_add_r(fiz2_S, gmx_simd_load_r(f+sciz+2)));
 676
 677 #ifdef CALC_SHIFTFORCES
 678         fshift[ish3+0] += gmx_sum_simd2(gmx_simd_add_r(fix0_S, fix2_S), shf);
 679         fshift[ish3+1] += gmx_sum_simd2(gmx_simd_add_r(fiy0_S, fiy2_S), shf);
 680         fshift[ish3+2] += gmx_sum_simd2(gmx_simd_add_r(fiz0_S, fiz2_S), shf);
 681 #endif
 682 #endif
 683
 684 #ifdef CALC_ENERGIES
 685         if (do_coul)
 686         {
 687             *Vc += gmx_sum_simd(vctot_S, tmpsum);
 688         }
 689
 690         *Vvdw += gmx_sum_simd(Vvdwtot_S, tmpsum);
 691 #endif
 692
 693         /* Outer loop uses 6 flops/iteration */
 694     }
 695
 696 #ifdef COUNT_PAIRS
 697     printf("atom pairs %d\n", npair);
 698 #endif
 699 }