src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEw_VdwLJEw_GeomP1P1_avx_256_single.c

   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  * Note: this file was generated by the GROMACS avx_256_single kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_single.h"
  48 #include "kernelutil_x86_avx_256_single.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomP1P1_VF_avx_256_single
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            LJEwald
  54  * Geometry:                   Particle-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEw_VdwLJEw_GeomP1P1_VF_avx_256_single
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB,jnrC,jnrD;
  75     int              jnrE,jnrF,jnrG,jnrH;
  76     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  77     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  78     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  79     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  80     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  81     real             rcutoff_scalar;
  82     real             *shiftvec,*fshift,*x,*f;
  83     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  84     real             scratch[4*DIM];
  85     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  86     real *           vdwioffsetptr0;
  87     real *           vdwgridioffsetptr0;
  88     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  89     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  90     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  91     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  92     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
  93     real             *charge;
  94     int              nvdwtype;
  95     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  96     int              *vdwtype;
  97     real             *vdwparam;
  98     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
  99     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 100     __m256           c6grid_00;
 101     real             *vdwgridparam;
 102     __m256           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 103     __m256           one_half  = _mm256_set1_ps(0.5);
 104     __m256           minus_one = _mm256_set1_ps(-1.0);
 105     __m256i          ewitab;
 106     __m128i          ewitab_lo,ewitab_hi;
 107     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 108     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 109     real             *ewtab;
 110     __m256           dummy_mask,cutoff_mask;
 111     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 112     __m256           one     = _mm256_set1_ps(1.0);
 113     __m256           two     = _mm256_set1_ps(2.0);
 114     x                = xx[0];
 115     f                = ff[0];
 116
 117     nri              = nlist->nri;
 118     iinr             = nlist->iinr;
 119     jindex           = nlist->jindex;
 120     jjnr             = nlist->jjnr;
 121     shiftidx         = nlist->shift;
 122     gid              = nlist->gid;
 123     shiftvec         = fr->shift_vec[0];
 124     fshift           = fr->fshift[0];
 125     facel            = _mm256_set1_ps(fr->epsfac);
 126     charge           = mdatoms->chargeA;
 127     nvdwtype         = fr->ntype;
 128     vdwparam         = fr->nbfp;
 129     vdwtype          = mdatoms->typeA;
 130     vdwgridparam     = fr->ljpme_c6grid;
 131     sh_lj_ewald      = _mm256_set1_ps(fr->ic->sh_lj_ewald);
 132     ewclj            = _mm256_set1_ps(fr->ewaldcoeff_lj);
 133     ewclj2           = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
 134
 135     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 136     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 137     beta2            = _mm256_mul_ps(beta,beta);
 138     beta3            = _mm256_mul_ps(beta,beta2);
 139
 140     ewtab            = fr->ic->tabq_coul_FDV0;
 141     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 142     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 143
 144     /* Avoid stupid compiler warnings */
 145     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 146     j_coord_offsetA = 0;
 147     j_coord_offsetB = 0;
 148     j_coord_offsetC = 0;
 149     j_coord_offsetD = 0;
 150     j_coord_offsetE = 0;
 151     j_coord_offsetF = 0;
 152     j_coord_offsetG = 0;
 153     j_coord_offsetH = 0;
 154
 155     outeriter        = 0;
 156     inneriter        = 0;
 157
 158     for(iidx=0;iidx<4*DIM;iidx++)
 159     {
 160         scratch[iidx] = 0.0;
 161     }
 162
 163     /* Start outer loop over neighborlists */
 164     for(iidx=0; iidx<nri; iidx++)
 165     {
 166         /* Load shift vector for this list */
 167         i_shift_offset   = DIM*shiftidx[iidx];
 168
 169         /* Load limits for loop over neighbors */
 170         j_index_start    = jindex[iidx];
 171         j_index_end      = jindex[iidx+1];
 172
 173         /* Get outer coordinate index */
 174         inr              = iinr[iidx];
 175         i_coord_offset   = DIM*inr;
 176
 177         /* Load i particle coords and add shift vector */
 178         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 179
 180         fix0             = _mm256_setzero_ps();
 181         fiy0             = _mm256_setzero_ps();
 182         fiz0             = _mm256_setzero_ps();
 183
 184         /* Load parameters for i particles */
 185         iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 186         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 187         vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
 188
 189         /* Reset potential sums */
 190         velecsum         = _mm256_setzero_ps();
 191         vvdwsum          = _mm256_setzero_ps();
 192
 193         /* Start inner kernel loop */
 194         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 195         {
 196
 197             /* Get j neighbor index, and coordinate index */
 198             jnrA             = jjnr[jidx];
 199             jnrB             = jjnr[jidx+1];
 200             jnrC             = jjnr[jidx+2];
 201             jnrD             = jjnr[jidx+3];
 202             jnrE             = jjnr[jidx+4];
 203             jnrF             = jjnr[jidx+5];
 204             jnrG             = jjnr[jidx+6];
 205             jnrH             = jjnr[jidx+7];
 206             j_coord_offsetA  = DIM*jnrA;
 207             j_coord_offsetB  = DIM*jnrB;
 208             j_coord_offsetC  = DIM*jnrC;
 209             j_coord_offsetD  = DIM*jnrD;
 210             j_coord_offsetE  = DIM*jnrE;
 211             j_coord_offsetF  = DIM*jnrF;
 212             j_coord_offsetG  = DIM*jnrG;
 213             j_coord_offsetH  = DIM*jnrH;
 214
 215             /* load j atom coordinates */
 216             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 217                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 218                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 219                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 220                                                  &jx0,&jy0,&jz0);
 221
 222             /* Calculate displacement vector */
 223             dx00             = _mm256_sub_ps(ix0,jx0);
 224             dy00             = _mm256_sub_ps(iy0,jy0);
 225             dz00             = _mm256_sub_ps(iz0,jz0);
 226
 227             /* Calculate squared distance and things based on it */
 228             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 229
 230             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 231
 232             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 233
 234             /* Load parameters for j particles */
 235             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 236                                                                  charge+jnrC+0,charge+jnrD+0,
 237                                                                  charge+jnrE+0,charge+jnrF+0,
 238                                                                  charge+jnrG+0,charge+jnrH+0);
 239             vdwjidx0A        = 2*vdwtype[jnrA+0];
 240             vdwjidx0B        = 2*vdwtype[jnrB+0];
 241             vdwjidx0C        = 2*vdwtype[jnrC+0];
 242             vdwjidx0D        = 2*vdwtype[jnrD+0];
 243             vdwjidx0E        = 2*vdwtype[jnrE+0];
 244             vdwjidx0F        = 2*vdwtype[jnrF+0];
 245             vdwjidx0G        = 2*vdwtype[jnrG+0];
 246             vdwjidx0H        = 2*vdwtype[jnrH+0];
 247
 248             /**************************
 249              * CALCULATE INTERACTIONS *
 250              **************************/
 251
 252             r00              = _mm256_mul_ps(rsq00,rinv00);
 253
 254             /* Compute parameters for interactions between i and j atoms */
 255             qq00             = _mm256_mul_ps(iq0,jq0);
 256             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 257                                             vdwioffsetptr0+vdwjidx0B,
 258                                             vdwioffsetptr0+vdwjidx0C,
 259                                             vdwioffsetptr0+vdwjidx0D,
 260                                             vdwioffsetptr0+vdwjidx0E,
 261                                             vdwioffsetptr0+vdwjidx0F,
 262                                             vdwioffsetptr0+vdwjidx0G,
 263                                             vdwioffsetptr0+vdwjidx0H,
 264                                             &c6_00,&c12_00);
 265
 266             c6grid_00       = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
 267                                                                   vdwgridioffsetptr0+vdwjidx0B,
 268                                                                   vdwgridioffsetptr0+vdwjidx0C,
 269                                                                   vdwgridioffsetptr0+vdwjidx0D,
 270                                                                   vdwgridioffsetptr0+vdwjidx0E,
 271                                                                   vdwgridioffsetptr0+vdwjidx0F,
 272                                                                   vdwgridioffsetptr0+vdwjidx0G,
 273                                                                   vdwgridioffsetptr0+vdwjidx0H);
 274
 275             /* EWALD ELECTROSTATICS */
 276
 277             /* Analytical PME correction */
 278             zeta2            = _mm256_mul_ps(beta2,rsq00);
 279             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 280             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 281             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 282             felec            = _mm256_mul_ps(qq00,felec);
 283             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 284             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 285             velec            = _mm256_sub_ps(rinv00,pmecorrV);
 286             velec            = _mm256_mul_ps(qq00,velec);
 287
 288             /* Analytical LJ-PME */
 289             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 290             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 291             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 292             exponent         = gmx_simd_exp_r(ewcljrsq);
 293             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 294             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 295             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 296             vvdw6            = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
 297             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 298             vvdw             = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
 299             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 300             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
 301
 302             /* Update potential sum for this i atom from the interaction with this j atom. */
 303             velecsum         = _mm256_add_ps(velecsum,velec);
 304             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 305
 306             fscal            = _mm256_add_ps(felec,fvdw);
 307
 308             /* Calculate temporary vectorial force */
 309             tx               = _mm256_mul_ps(fscal,dx00);
 310             ty               = _mm256_mul_ps(fscal,dy00);
 311             tz               = _mm256_mul_ps(fscal,dz00);
 312
 313             /* Update vectorial force */
 314             fix0             = _mm256_add_ps(fix0,tx);
 315             fiy0             = _mm256_add_ps(fiy0,ty);
 316             fiz0             = _mm256_add_ps(fiz0,tz);
 317
 318             fjptrA             = f+j_coord_offsetA;
 319             fjptrB             = f+j_coord_offsetB;
 320             fjptrC             = f+j_coord_offsetC;
 321             fjptrD             = f+j_coord_offsetD;
 322             fjptrE             = f+j_coord_offsetE;
 323             fjptrF             = f+j_coord_offsetF;
 324             fjptrG             = f+j_coord_offsetG;
 325             fjptrH             = f+j_coord_offsetH;
 326             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 327
 328             /* Inner loop uses 112 flops */
 329         }
 330
 331         if(jidx<j_index_end)
 332         {
 333
 334             /* Get j neighbor index, and coordinate index */
 335             jnrlistA         = jjnr[jidx];
 336             jnrlistB         = jjnr[jidx+1];
 337             jnrlistC         = jjnr[jidx+2];
 338             jnrlistD         = jjnr[jidx+3];
 339             jnrlistE         = jjnr[jidx+4];
 340             jnrlistF         = jjnr[jidx+5];
 341             jnrlistG         = jjnr[jidx+6];
 342             jnrlistH         = jjnr[jidx+7];
 343             /* Sign of each element will be negative for non-real atoms.
 344              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 345              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 346              */
 347             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 348                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 349
 350             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 351             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 352             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 353             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 354             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 355             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 356             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 357             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 358             j_coord_offsetA  = DIM*jnrA;
 359             j_coord_offsetB  = DIM*jnrB;
 360             j_coord_offsetC  = DIM*jnrC;
 361             j_coord_offsetD  = DIM*jnrD;
 362             j_coord_offsetE  = DIM*jnrE;
 363             j_coord_offsetF  = DIM*jnrF;
 364             j_coord_offsetG  = DIM*jnrG;
 365             j_coord_offsetH  = DIM*jnrH;
 366
 367             /* load j atom coordinates */
 368             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 369                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 370                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 371                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 372                                                  &jx0,&jy0,&jz0);
 373
 374             /* Calculate displacement vector */
 375             dx00             = _mm256_sub_ps(ix0,jx0);
 376             dy00             = _mm256_sub_ps(iy0,jy0);
 377             dz00             = _mm256_sub_ps(iz0,jz0);
 378
 379             /* Calculate squared distance and things based on it */
 380             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 381
 382             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 383
 384             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 385
 386             /* Load parameters for j particles */
 387             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 388                                                                  charge+jnrC+0,charge+jnrD+0,
 389                                                                  charge+jnrE+0,charge+jnrF+0,
 390                                                                  charge+jnrG+0,charge+jnrH+0);
 391             vdwjidx0A        = 2*vdwtype[jnrA+0];
 392             vdwjidx0B        = 2*vdwtype[jnrB+0];
 393             vdwjidx0C        = 2*vdwtype[jnrC+0];
 394             vdwjidx0D        = 2*vdwtype[jnrD+0];
 395             vdwjidx0E        = 2*vdwtype[jnrE+0];
 396             vdwjidx0F        = 2*vdwtype[jnrF+0];
 397             vdwjidx0G        = 2*vdwtype[jnrG+0];
 398             vdwjidx0H        = 2*vdwtype[jnrH+0];
 399
 400             /**************************
 401              * CALCULATE INTERACTIONS *
 402              **************************/
 403
 404             r00              = _mm256_mul_ps(rsq00,rinv00);
 405             r00              = _mm256_andnot_ps(dummy_mask,r00);
 406
 407             /* Compute parameters for interactions between i and j atoms */
 408             qq00             = _mm256_mul_ps(iq0,jq0);
 409             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 410                                             vdwioffsetptr0+vdwjidx0B,
 411                                             vdwioffsetptr0+vdwjidx0C,
 412                                             vdwioffsetptr0+vdwjidx0D,
 413                                             vdwioffsetptr0+vdwjidx0E,
 414                                             vdwioffsetptr0+vdwjidx0F,
 415                                             vdwioffsetptr0+vdwjidx0G,
 416                                             vdwioffsetptr0+vdwjidx0H,
 417                                             &c6_00,&c12_00);
 418
 419             c6grid_00       = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
 420                                                                   vdwgridioffsetptr0+vdwjidx0B,
 421                                                                   vdwgridioffsetptr0+vdwjidx0C,
 422                                                                   vdwgridioffsetptr0+vdwjidx0D,
 423                                                                   vdwgridioffsetptr0+vdwjidx0E,
 424                                                                   vdwgridioffsetptr0+vdwjidx0F,
 425                                                                   vdwgridioffsetptr0+vdwjidx0G,
 426                                                                   vdwgridioffsetptr0+vdwjidx0H);
 427
 428             /* EWALD ELECTROSTATICS */
 429
 430             /* Analytical PME correction */
 431             zeta2            = _mm256_mul_ps(beta2,rsq00);
 432             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 433             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 434             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 435             felec            = _mm256_mul_ps(qq00,felec);
 436             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 437             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 438             velec            = _mm256_sub_ps(rinv00,pmecorrV);
 439             velec            = _mm256_mul_ps(qq00,velec);
 440
 441             /* Analytical LJ-PME */
 442             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 443             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 444             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 445             exponent         = gmx_simd_exp_r(ewcljrsq);
 446             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 447             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 448             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 449             vvdw6            = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
 450             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 451             vvdw             = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
 452             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 453             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
 454
 455             /* Update potential sum for this i atom from the interaction with this j atom. */
 456             velec            = _mm256_andnot_ps(dummy_mask,velec);
 457             velecsum         = _mm256_add_ps(velecsum,velec);
 458             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 459             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 460
 461             fscal            = _mm256_add_ps(felec,fvdw);
 462
 463             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 464
 465             /* Calculate temporary vectorial force */
 466             tx               = _mm256_mul_ps(fscal,dx00);
 467             ty               = _mm256_mul_ps(fscal,dy00);
 468             tz               = _mm256_mul_ps(fscal,dz00);
 469
 470             /* Update vectorial force */
 471             fix0             = _mm256_add_ps(fix0,tx);
 472             fiy0             = _mm256_add_ps(fiy0,ty);
 473             fiz0             = _mm256_add_ps(fiz0,tz);
 474
 475             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 476             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 477             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 478             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 479             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 480             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 481             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 482             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 483             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 484
 485             /* Inner loop uses 113 flops */
 486         }
 487
 488         /* End of innermost loop */
 489
 490         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 491                                                  f+i_coord_offset,fshift+i_shift_offset);
 492
 493         ggid                        = gid[iidx];
 494         /* Update potential energies */
 495         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 496         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 497
 498         /* Increment number of inner iterations */
 499         inneriter                  += j_index_end - j_index_start;
 500
 501         /* Outer loop uses 9 flops */
 502     }
 503
 504     /* Increment number of outer iterations */
 505     outeriter        += nri;
 506
 507     /* Update outer/inner flops */
 508
 509     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*9 + inneriter*113);
 510 }
 511 /*
 512  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_avx_256_single
 513  * Electrostatics interaction: Ewald
 514  * VdW interaction:            LJEwald
 515  * Geometry:                   Particle-Particle
 516  * Calculate force/pot:        Force
 517  */
 518 void
 519 nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_avx_256_single
 520                     (t_nblist                    * gmx_restrict       nlist,
 521                      rvec                        * gmx_restrict          xx,
 522                      rvec                        * gmx_restrict          ff,
 523                      t_forcerec                  * gmx_restrict          fr,
 524                      t_mdatoms                   * gmx_restrict     mdatoms,
 525                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 526                      t_nrnb                      * gmx_restrict        nrnb)
 527 {
 528     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 529      * just 0 for non-waters.
 530      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 531      * jnr indices corresponding to data put in the four positions in the SIMD register.
 532      */
 533     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 534     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 535     int              jnrA,jnrB,jnrC,jnrD;
 536     int              jnrE,jnrF,jnrG,jnrH;
 537     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 538     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 539     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 540     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 541     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 542     real             rcutoff_scalar;
 543     real             *shiftvec,*fshift,*x,*f;
 544     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 545     real             scratch[4*DIM];
 546     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 547     real *           vdwioffsetptr0;
 548     real *           vdwgridioffsetptr0;
 549     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 550     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
 551     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 552     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 553     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 554     real             *charge;
 555     int              nvdwtype;
 556     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 557     int              *vdwtype;
 558     real             *vdwparam;
 559     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 560     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 561     __m256           c6grid_00;
 562     real             *vdwgridparam;
 563     __m256           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 564     __m256           one_half  = _mm256_set1_ps(0.5);
 565     __m256           minus_one = _mm256_set1_ps(-1.0);
 566     __m256i          ewitab;
 567     __m128i          ewitab_lo,ewitab_hi;
 568     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 569     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 570     real             *ewtab;
 571     __m256           dummy_mask,cutoff_mask;
 572     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 573     __m256           one     = _mm256_set1_ps(1.0);
 574     __m256           two     = _mm256_set1_ps(2.0);
 575     x                = xx[0];
 576     f                = ff[0];
 577
 578     nri              = nlist->nri;
 579     iinr             = nlist->iinr;
 580     jindex           = nlist->jindex;
 581     jjnr             = nlist->jjnr;
 582     shiftidx         = nlist->shift;
 583     gid              = nlist->gid;
 584     shiftvec         = fr->shift_vec[0];
 585     fshift           = fr->fshift[0];
 586     facel            = _mm256_set1_ps(fr->epsfac);
 587     charge           = mdatoms->chargeA;
 588     nvdwtype         = fr->ntype;
 589     vdwparam         = fr->nbfp;
 590     vdwtype          = mdatoms->typeA;
 591     vdwgridparam     = fr->ljpme_c6grid;
 592     sh_lj_ewald      = _mm256_set1_ps(fr->ic->sh_lj_ewald);
 593     ewclj            = _mm256_set1_ps(fr->ewaldcoeff_lj);
 594     ewclj2           = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
 595
 596     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 597     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 598     beta2            = _mm256_mul_ps(beta,beta);
 599     beta3            = _mm256_mul_ps(beta,beta2);
 600
 601     ewtab            = fr->ic->tabq_coul_F;
 602     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 603     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 604
 605     /* Avoid stupid compiler warnings */
 606     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 607     j_coord_offsetA = 0;
 608     j_coord_offsetB = 0;
 609     j_coord_offsetC = 0;
 610     j_coord_offsetD = 0;
 611     j_coord_offsetE = 0;
 612     j_coord_offsetF = 0;
 613     j_coord_offsetG = 0;
 614     j_coord_offsetH = 0;
 615
 616     outeriter        = 0;
 617     inneriter        = 0;
 618
 619     for(iidx=0;iidx<4*DIM;iidx++)
 620     {
 621         scratch[iidx] = 0.0;
 622     }
 623
 624     /* Start outer loop over neighborlists */
 625     for(iidx=0; iidx<nri; iidx++)
 626     {
 627         /* Load shift vector for this list */
 628         i_shift_offset   = DIM*shiftidx[iidx];
 629
 630         /* Load limits for loop over neighbors */
 631         j_index_start    = jindex[iidx];
 632         j_index_end      = jindex[iidx+1];
 633
 634         /* Get outer coordinate index */
 635         inr              = iinr[iidx];
 636         i_coord_offset   = DIM*inr;
 637
 638         /* Load i particle coords and add shift vector */
 639         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 640
 641         fix0             = _mm256_setzero_ps();
 642         fiy0             = _mm256_setzero_ps();
 643         fiz0             = _mm256_setzero_ps();
 644
 645         /* Load parameters for i particles */
 646         iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 647         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 648         vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
 649
 650         /* Start inner kernel loop */
 651         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 652         {
 653
 654             /* Get j neighbor index, and coordinate index */
 655             jnrA             = jjnr[jidx];
 656             jnrB             = jjnr[jidx+1];
 657             jnrC             = jjnr[jidx+2];
 658             jnrD             = jjnr[jidx+3];
 659             jnrE             = jjnr[jidx+4];
 660             jnrF             = jjnr[jidx+5];
 661             jnrG             = jjnr[jidx+6];
 662             jnrH             = jjnr[jidx+7];
 663             j_coord_offsetA  = DIM*jnrA;
 664             j_coord_offsetB  = DIM*jnrB;
 665             j_coord_offsetC  = DIM*jnrC;
 666             j_coord_offsetD  = DIM*jnrD;
 667             j_coord_offsetE  = DIM*jnrE;
 668             j_coord_offsetF  = DIM*jnrF;
 669             j_coord_offsetG  = DIM*jnrG;
 670             j_coord_offsetH  = DIM*jnrH;
 671
 672             /* load j atom coordinates */
 673             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 674                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 675                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 676                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 677                                                  &jx0,&jy0,&jz0);
 678
 679             /* Calculate displacement vector */
 680             dx00             = _mm256_sub_ps(ix0,jx0);
 681             dy00             = _mm256_sub_ps(iy0,jy0);
 682             dz00             = _mm256_sub_ps(iz0,jz0);
 683
 684             /* Calculate squared distance and things based on it */
 685             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 686
 687             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 688
 689             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 690
 691             /* Load parameters for j particles */
 692             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 693                                                                  charge+jnrC+0,charge+jnrD+0,
 694                                                                  charge+jnrE+0,charge+jnrF+0,
 695                                                                  charge+jnrG+0,charge+jnrH+0);
 696             vdwjidx0A        = 2*vdwtype[jnrA+0];
 697             vdwjidx0B        = 2*vdwtype[jnrB+0];
 698             vdwjidx0C        = 2*vdwtype[jnrC+0];
 699             vdwjidx0D        = 2*vdwtype[jnrD+0];
 700             vdwjidx0E        = 2*vdwtype[jnrE+0];
 701             vdwjidx0F        = 2*vdwtype[jnrF+0];
 702             vdwjidx0G        = 2*vdwtype[jnrG+0];
 703             vdwjidx0H        = 2*vdwtype[jnrH+0];
 704
 705             /**************************
 706              * CALCULATE INTERACTIONS *
 707              **************************/
 708
 709             r00              = _mm256_mul_ps(rsq00,rinv00);
 710
 711             /* Compute parameters for interactions between i and j atoms */
 712             qq00             = _mm256_mul_ps(iq0,jq0);
 713             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 714                                             vdwioffsetptr0+vdwjidx0B,
 715                                             vdwioffsetptr0+vdwjidx0C,
 716                                             vdwioffsetptr0+vdwjidx0D,
 717                                             vdwioffsetptr0+vdwjidx0E,
 718                                             vdwioffsetptr0+vdwjidx0F,
 719                                             vdwioffsetptr0+vdwjidx0G,
 720                                             vdwioffsetptr0+vdwjidx0H,
 721                                             &c6_00,&c12_00);
 722
 723             c6grid_00       = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
 724                                                                   vdwgridioffsetptr0+vdwjidx0B,
 725                                                                   vdwgridioffsetptr0+vdwjidx0C,
 726                                                                   vdwgridioffsetptr0+vdwjidx0D,
 727                                                                   vdwgridioffsetptr0+vdwjidx0E,
 728                                                                   vdwgridioffsetptr0+vdwjidx0F,
 729                                                                   vdwgridioffsetptr0+vdwjidx0G,
 730                                                                   vdwgridioffsetptr0+vdwjidx0H);
 731
 732             /* EWALD ELECTROSTATICS */
 733
 734             /* Analytical PME correction */
 735             zeta2            = _mm256_mul_ps(beta2,rsq00);
 736             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 737             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 738             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 739             felec            = _mm256_mul_ps(qq00,felec);
 740
 741             /* Analytical LJ-PME */
 742             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 743             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 744             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 745             exponent         = gmx_simd_exp_r(ewcljrsq);
 746             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 747             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 748             /* f6A = 6 * C6grid * (1 - poly) */
 749             f6A              = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
 750             /* f6B = C6grid * exponent * beta^6 */
 751             f6B              = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
 752             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
 753             fvdw              = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
 754
 755             fscal            = _mm256_add_ps(felec,fvdw);
 756
 757             /* Calculate temporary vectorial force */
 758             tx               = _mm256_mul_ps(fscal,dx00);
 759             ty               = _mm256_mul_ps(fscal,dy00);
 760             tz               = _mm256_mul_ps(fscal,dz00);
 761
 762             /* Update vectorial force */
 763             fix0             = _mm256_add_ps(fix0,tx);
 764             fiy0             = _mm256_add_ps(fiy0,ty);
 765             fiz0             = _mm256_add_ps(fiz0,tz);
 766
 767             fjptrA             = f+j_coord_offsetA;
 768             fjptrB             = f+j_coord_offsetB;
 769             fjptrC             = f+j_coord_offsetC;
 770             fjptrD             = f+j_coord_offsetD;
 771             fjptrE             = f+j_coord_offsetE;
 772             fjptrF             = f+j_coord_offsetF;
 773             fjptrG             = f+j_coord_offsetG;
 774             fjptrH             = f+j_coord_offsetH;
 775             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 776
 777             /* Inner loop uses 79 flops */
 778         }
 779
 780         if(jidx<j_index_end)
 781         {
 782
 783             /* Get j neighbor index, and coordinate index */
 784             jnrlistA         = jjnr[jidx];
 785             jnrlistB         = jjnr[jidx+1];
 786             jnrlistC         = jjnr[jidx+2];
 787             jnrlistD         = jjnr[jidx+3];
 788             jnrlistE         = jjnr[jidx+4];
 789             jnrlistF         = jjnr[jidx+5];
 790             jnrlistG         = jjnr[jidx+6];
 791             jnrlistH         = jjnr[jidx+7];
 792             /* Sign of each element will be negative for non-real atoms.
 793              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 794              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 795              */
 796             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 797                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 798
 799             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 800             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 801             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 802             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 803             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 804             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 805             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 806             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 807             j_coord_offsetA  = DIM*jnrA;
 808             j_coord_offsetB  = DIM*jnrB;
 809             j_coord_offsetC  = DIM*jnrC;
 810             j_coord_offsetD  = DIM*jnrD;
 811             j_coord_offsetE  = DIM*jnrE;
 812             j_coord_offsetF  = DIM*jnrF;
 813             j_coord_offsetG  = DIM*jnrG;
 814             j_coord_offsetH  = DIM*jnrH;
 815
 816             /* load j atom coordinates */
 817             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 818                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 819                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 820                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 821                                                  &jx0,&jy0,&jz0);
 822
 823             /* Calculate displacement vector */
 824             dx00             = _mm256_sub_ps(ix0,jx0);
 825             dy00             = _mm256_sub_ps(iy0,jy0);
 826             dz00             = _mm256_sub_ps(iz0,jz0);
 827
 828             /* Calculate squared distance and things based on it */
 829             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 830
 831             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 832
 833             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 834
 835             /* Load parameters for j particles */
 836             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 837                                                                  charge+jnrC+0,charge+jnrD+0,
 838                                                                  charge+jnrE+0,charge+jnrF+0,
 839                                                                  charge+jnrG+0,charge+jnrH+0);
 840             vdwjidx0A        = 2*vdwtype[jnrA+0];
 841             vdwjidx0B        = 2*vdwtype[jnrB+0];
 842             vdwjidx0C        = 2*vdwtype[jnrC+0];
 843             vdwjidx0D        = 2*vdwtype[jnrD+0];
 844             vdwjidx0E        = 2*vdwtype[jnrE+0];
 845             vdwjidx0F        = 2*vdwtype[jnrF+0];
 846             vdwjidx0G        = 2*vdwtype[jnrG+0];
 847             vdwjidx0H        = 2*vdwtype[jnrH+0];
 848
 849             /**************************
 850              * CALCULATE INTERACTIONS *
 851              **************************/
 852
 853             r00              = _mm256_mul_ps(rsq00,rinv00);
 854             r00              = _mm256_andnot_ps(dummy_mask,r00);
 855
 856             /* Compute parameters for interactions between i and j atoms */
 857             qq00             = _mm256_mul_ps(iq0,jq0);
 858             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 859                                             vdwioffsetptr0+vdwjidx0B,
 860                                             vdwioffsetptr0+vdwjidx0C,
 861                                             vdwioffsetptr0+vdwjidx0D,
 862                                             vdwioffsetptr0+vdwjidx0E,
 863                                             vdwioffsetptr0+vdwjidx0F,
 864                                             vdwioffsetptr0+vdwjidx0G,
 865                                             vdwioffsetptr0+vdwjidx0H,
 866                                             &c6_00,&c12_00);
 867
 868             c6grid_00       = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
 869                                                                   vdwgridioffsetptr0+vdwjidx0B,
 870                                                                   vdwgridioffsetptr0+vdwjidx0C,
 871                                                                   vdwgridioffsetptr0+vdwjidx0D,
 872                                                                   vdwgridioffsetptr0+vdwjidx0E,
 873                                                                   vdwgridioffsetptr0+vdwjidx0F,
 874                                                                   vdwgridioffsetptr0+vdwjidx0G,
 875                                                                   vdwgridioffsetptr0+vdwjidx0H);
 876
 877             /* EWALD ELECTROSTATICS */
 878
 879             /* Analytical PME correction */
 880             zeta2            = _mm256_mul_ps(beta2,rsq00);
 881             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 882             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 883             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 884             felec            = _mm256_mul_ps(qq00,felec);
 885
 886             /* Analytical LJ-PME */
 887             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 888             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 889             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 890             exponent         = gmx_simd_exp_r(ewcljrsq);
 891             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 892             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 893             /* f6A = 6 * C6grid * (1 - poly) */
 894             f6A              = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
 895             /* f6B = C6grid * exponent * beta^6 */
 896             f6B              = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
 897             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
 898             fvdw              = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
 899
 900             fscal            = _mm256_add_ps(felec,fvdw);
 901
 902             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 903
 904             /* Calculate temporary vectorial force */
 905             tx               = _mm256_mul_ps(fscal,dx00);
 906             ty               = _mm256_mul_ps(fscal,dy00);
 907             tz               = _mm256_mul_ps(fscal,dz00);
 908
 909             /* Update vectorial force */
 910             fix0             = _mm256_add_ps(fix0,tx);
 911             fiy0             = _mm256_add_ps(fiy0,ty);
 912             fiz0             = _mm256_add_ps(fiz0,tz);
 913
 914             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 915             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 916             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 917             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 918             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 919             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 920             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 921             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 922             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 923
 924             /* Inner loop uses 80 flops */
 925         }
 926
 927         /* End of innermost loop */
 928
 929         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 930                                                  f+i_coord_offset,fshift+i_shift_offset);
 931
 932         /* Increment number of inner iterations */
 933         inneriter                  += j_index_end - j_index_start;
 934
 935         /* Outer loop uses 7 flops */
 936     }
 937
 938     /* Increment number of outer iterations */
 939     outeriter        += nri;
 940
 941     /* Update outer/inner flops */
 942
 943     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*7 + inneriter*80);
 944 }