src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecRF_VdwCSTab_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_ElecRF_VdwCSTab_GeomP1P1_VF_avx_256_single
  52  * Electrostatics interaction: ReactionField
  53  * VdW interaction:            CubicSplineTable
  54  * Geometry:                   Particle-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecRF_VdwCSTab_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     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  88     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  89     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  90     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  91     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
  92     real             *charge;
  93     int              nvdwtype;
  94     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  95     int              *vdwtype;
  96     real             *vdwparam;
  97     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
  98     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
  99     __m256i          vfitab;
 100     __m128i          vfitab_lo,vfitab_hi;
 101     __m128i          ifour       = _mm_set1_epi32(4);
 102     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 103     real             *vftab;
 104     __m256           dummy_mask,cutoff_mask;
 105     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 106     __m256           one     = _mm256_set1_ps(1.0);
 107     __m256           two     = _mm256_set1_ps(2.0);
 108     x                = xx[0];
 109     f                = ff[0];
 110
 111     nri              = nlist->nri;
 112     iinr             = nlist->iinr;
 113     jindex           = nlist->jindex;
 114     jjnr             = nlist->jjnr;
 115     shiftidx         = nlist->shift;
 116     gid              = nlist->gid;
 117     shiftvec         = fr->shift_vec[0];
 118     fshift           = fr->fshift[0];
 119     facel            = _mm256_set1_ps(fr->epsfac);
 120     charge           = mdatoms->chargeA;
 121     krf              = _mm256_set1_ps(fr->ic->k_rf);
 122     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 123     crf              = _mm256_set1_ps(fr->ic->c_rf);
 124     nvdwtype         = fr->ntype;
 125     vdwparam         = fr->nbfp;
 126     vdwtype          = mdatoms->typeA;
 127
 128     vftab            = kernel_data->table_vdw->data;
 129     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 130
 131     /* Avoid stupid compiler warnings */
 132     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 133     j_coord_offsetA = 0;
 134     j_coord_offsetB = 0;
 135     j_coord_offsetC = 0;
 136     j_coord_offsetD = 0;
 137     j_coord_offsetE = 0;
 138     j_coord_offsetF = 0;
 139     j_coord_offsetG = 0;
 140     j_coord_offsetH = 0;
 141
 142     outeriter        = 0;
 143     inneriter        = 0;
 144
 145     for(iidx=0;iidx<4*DIM;iidx++)
 146     {
 147         scratch[iidx] = 0.0;
 148     }
 149
 150     /* Start outer loop over neighborlists */
 151     for(iidx=0; iidx<nri; iidx++)
 152     {
 153         /* Load shift vector for this list */
 154         i_shift_offset   = DIM*shiftidx[iidx];
 155
 156         /* Load limits for loop over neighbors */
 157         j_index_start    = jindex[iidx];
 158         j_index_end      = jindex[iidx+1];
 159
 160         /* Get outer coordinate index */
 161         inr              = iinr[iidx];
 162         i_coord_offset   = DIM*inr;
 163
 164         /* Load i particle coords and add shift vector */
 165         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 166
 167         fix0             = _mm256_setzero_ps();
 168         fiy0             = _mm256_setzero_ps();
 169         fiz0             = _mm256_setzero_ps();
 170
 171         /* Load parameters for i particles */
 172         iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 173         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 174
 175         /* Reset potential sums */
 176         velecsum         = _mm256_setzero_ps();
 177         vvdwsum          = _mm256_setzero_ps();
 178
 179         /* Start inner kernel loop */
 180         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 181         {
 182
 183             /* Get j neighbor index, and coordinate index */
 184             jnrA             = jjnr[jidx];
 185             jnrB             = jjnr[jidx+1];
 186             jnrC             = jjnr[jidx+2];
 187             jnrD             = jjnr[jidx+3];
 188             jnrE             = jjnr[jidx+4];
 189             jnrF             = jjnr[jidx+5];
 190             jnrG             = jjnr[jidx+6];
 191             jnrH             = jjnr[jidx+7];
 192             j_coord_offsetA  = DIM*jnrA;
 193             j_coord_offsetB  = DIM*jnrB;
 194             j_coord_offsetC  = DIM*jnrC;
 195             j_coord_offsetD  = DIM*jnrD;
 196             j_coord_offsetE  = DIM*jnrE;
 197             j_coord_offsetF  = DIM*jnrF;
 198             j_coord_offsetG  = DIM*jnrG;
 199             j_coord_offsetH  = DIM*jnrH;
 200
 201             /* load j atom coordinates */
 202             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 203                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 204                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 205                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 206                                                  &jx0,&jy0,&jz0);
 207
 208             /* Calculate displacement vector */
 209             dx00             = _mm256_sub_ps(ix0,jx0);
 210             dy00             = _mm256_sub_ps(iy0,jy0);
 211             dz00             = _mm256_sub_ps(iz0,jz0);
 212
 213             /* Calculate squared distance and things based on it */
 214             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 215
 216             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 217
 218             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 219
 220             /* Load parameters for j particles */
 221             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 222                                                                  charge+jnrC+0,charge+jnrD+0,
 223                                                                  charge+jnrE+0,charge+jnrF+0,
 224                                                                  charge+jnrG+0,charge+jnrH+0);
 225             vdwjidx0A        = 2*vdwtype[jnrA+0];
 226             vdwjidx0B        = 2*vdwtype[jnrB+0];
 227             vdwjidx0C        = 2*vdwtype[jnrC+0];
 228             vdwjidx0D        = 2*vdwtype[jnrD+0];
 229             vdwjidx0E        = 2*vdwtype[jnrE+0];
 230             vdwjidx0F        = 2*vdwtype[jnrF+0];
 231             vdwjidx0G        = 2*vdwtype[jnrG+0];
 232             vdwjidx0H        = 2*vdwtype[jnrH+0];
 233
 234             /**************************
 235              * CALCULATE INTERACTIONS *
 236              **************************/
 237
 238             r00              = _mm256_mul_ps(rsq00,rinv00);
 239
 240             /* Compute parameters for interactions between i and j atoms */
 241             qq00             = _mm256_mul_ps(iq0,jq0);
 242             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 243                                             vdwioffsetptr0+vdwjidx0B,
 244                                             vdwioffsetptr0+vdwjidx0C,
 245                                             vdwioffsetptr0+vdwjidx0D,
 246                                             vdwioffsetptr0+vdwjidx0E,
 247                                             vdwioffsetptr0+vdwjidx0F,
 248                                             vdwioffsetptr0+vdwjidx0G,
 249                                             vdwioffsetptr0+vdwjidx0H,
 250                                             &c6_00,&c12_00);
 251
 252             /* Calculate table index by multiplying r with table scale and truncate to integer */
 253             rt               = _mm256_mul_ps(r00,vftabscale);
 254             vfitab           = _mm256_cvttps_epi32(rt);
 255             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 256             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 257             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 258             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 259             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 260             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 261
 262             /* REACTION-FIELD ELECTROSTATICS */
 263             velec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_add_ps(rinv00,_mm256_mul_ps(krf,rsq00)),crf));
 264             felec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_mul_ps(rinv00,rinvsq00),krf2));
 265
 266             /* CUBIC SPLINE TABLE DISPERSION */
 267             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 268                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 269             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 270                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 271             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 272                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 273             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 274                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 275             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 276             Heps             = _mm256_mul_ps(vfeps,H);
 277             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 278             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 279             vvdw6            = _mm256_mul_ps(c6_00,VV);
 280             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 281             fvdw6            = _mm256_mul_ps(c6_00,FF);
 282
 283             /* CUBIC SPLINE TABLE REPULSION */
 284             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 285             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 286             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 287                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 288             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 289                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 290             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 291                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 292             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 293                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 294             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 295             Heps             = _mm256_mul_ps(vfeps,H);
 296             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 297             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 298             vvdw12           = _mm256_mul_ps(c12_00,VV);
 299             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 300             fvdw12           = _mm256_mul_ps(c12_00,FF);
 301             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 302             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 303
 304             /* Update potential sum for this i atom from the interaction with this j atom. */
 305             velecsum         = _mm256_add_ps(velecsum,velec);
 306             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 307
 308             fscal            = _mm256_add_ps(felec,fvdw);
 309
 310             /* Calculate temporary vectorial force */
 311             tx               = _mm256_mul_ps(fscal,dx00);
 312             ty               = _mm256_mul_ps(fscal,dy00);
 313             tz               = _mm256_mul_ps(fscal,dz00);
 314
 315             /* Update vectorial force */
 316             fix0             = _mm256_add_ps(fix0,tx);
 317             fiy0             = _mm256_add_ps(fiy0,ty);
 318             fiz0             = _mm256_add_ps(fiz0,tz);
 319
 320             fjptrA             = f+j_coord_offsetA;
 321             fjptrB             = f+j_coord_offsetB;
 322             fjptrC             = f+j_coord_offsetC;
 323             fjptrD             = f+j_coord_offsetD;
 324             fjptrE             = f+j_coord_offsetE;
 325             fjptrF             = f+j_coord_offsetF;
 326             fjptrG             = f+j_coord_offsetG;
 327             fjptrH             = f+j_coord_offsetH;
 328             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 329
 330             /* Inner loop uses 67 flops */
 331         }
 332
 333         if(jidx<j_index_end)
 334         {
 335
 336             /* Get j neighbor index, and coordinate index */
 337             jnrlistA         = jjnr[jidx];
 338             jnrlistB         = jjnr[jidx+1];
 339             jnrlistC         = jjnr[jidx+2];
 340             jnrlistD         = jjnr[jidx+3];
 341             jnrlistE         = jjnr[jidx+4];
 342             jnrlistF         = jjnr[jidx+5];
 343             jnrlistG         = jjnr[jidx+6];
 344             jnrlistH         = jjnr[jidx+7];
 345             /* Sign of each element will be negative for non-real atoms.
 346              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 347              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 348              */
 349             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 350                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 351
 352             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 353             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 354             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 355             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 356             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 357             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 358             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 359             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 360             j_coord_offsetA  = DIM*jnrA;
 361             j_coord_offsetB  = DIM*jnrB;
 362             j_coord_offsetC  = DIM*jnrC;
 363             j_coord_offsetD  = DIM*jnrD;
 364             j_coord_offsetE  = DIM*jnrE;
 365             j_coord_offsetF  = DIM*jnrF;
 366             j_coord_offsetG  = DIM*jnrG;
 367             j_coord_offsetH  = DIM*jnrH;
 368
 369             /* load j atom coordinates */
 370             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 371                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 372                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 373                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 374                                                  &jx0,&jy0,&jz0);
 375
 376             /* Calculate displacement vector */
 377             dx00             = _mm256_sub_ps(ix0,jx0);
 378             dy00             = _mm256_sub_ps(iy0,jy0);
 379             dz00             = _mm256_sub_ps(iz0,jz0);
 380
 381             /* Calculate squared distance and things based on it */
 382             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 383
 384             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 385
 386             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 387
 388             /* Load parameters for j particles */
 389             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 390                                                                  charge+jnrC+0,charge+jnrD+0,
 391                                                                  charge+jnrE+0,charge+jnrF+0,
 392                                                                  charge+jnrG+0,charge+jnrH+0);
 393             vdwjidx0A        = 2*vdwtype[jnrA+0];
 394             vdwjidx0B        = 2*vdwtype[jnrB+0];
 395             vdwjidx0C        = 2*vdwtype[jnrC+0];
 396             vdwjidx0D        = 2*vdwtype[jnrD+0];
 397             vdwjidx0E        = 2*vdwtype[jnrE+0];
 398             vdwjidx0F        = 2*vdwtype[jnrF+0];
 399             vdwjidx0G        = 2*vdwtype[jnrG+0];
 400             vdwjidx0H        = 2*vdwtype[jnrH+0];
 401
 402             /**************************
 403              * CALCULATE INTERACTIONS *
 404              **************************/
 405
 406             r00              = _mm256_mul_ps(rsq00,rinv00);
 407             r00              = _mm256_andnot_ps(dummy_mask,r00);
 408
 409             /* Compute parameters for interactions between i and j atoms */
 410             qq00             = _mm256_mul_ps(iq0,jq0);
 411             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 412                                             vdwioffsetptr0+vdwjidx0B,
 413                                             vdwioffsetptr0+vdwjidx0C,
 414                                             vdwioffsetptr0+vdwjidx0D,
 415                                             vdwioffsetptr0+vdwjidx0E,
 416                                             vdwioffsetptr0+vdwjidx0F,
 417                                             vdwioffsetptr0+vdwjidx0G,
 418                                             vdwioffsetptr0+vdwjidx0H,
 419                                             &c6_00,&c12_00);
 420
 421             /* Calculate table index by multiplying r with table scale and truncate to integer */
 422             rt               = _mm256_mul_ps(r00,vftabscale);
 423             vfitab           = _mm256_cvttps_epi32(rt);
 424             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 425             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 426             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 427             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 428             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 429             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 430
 431             /* REACTION-FIELD ELECTROSTATICS */
 432             velec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_add_ps(rinv00,_mm256_mul_ps(krf,rsq00)),crf));
 433             felec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_mul_ps(rinv00,rinvsq00),krf2));
 434
 435             /* CUBIC SPLINE TABLE DISPERSION */
 436             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 437                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 438             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 439                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 440             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 441                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 442             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 443                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 444             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 445             Heps             = _mm256_mul_ps(vfeps,H);
 446             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 447             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 448             vvdw6            = _mm256_mul_ps(c6_00,VV);
 449             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 450             fvdw6            = _mm256_mul_ps(c6_00,FF);
 451
 452             /* CUBIC SPLINE TABLE REPULSION */
 453             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 454             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 455             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 456                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 457             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 458                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 459             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 460                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 461             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 462                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 463             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 464             Heps             = _mm256_mul_ps(vfeps,H);
 465             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 466             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 467             vvdw12           = _mm256_mul_ps(c12_00,VV);
 468             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 469             fvdw12           = _mm256_mul_ps(c12_00,FF);
 470             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 471             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 472
 473             /* Update potential sum for this i atom from the interaction with this j atom. */
 474             velec            = _mm256_andnot_ps(dummy_mask,velec);
 475             velecsum         = _mm256_add_ps(velecsum,velec);
 476             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 477             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 478
 479             fscal            = _mm256_add_ps(felec,fvdw);
 480
 481             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 482
 483             /* Calculate temporary vectorial force */
 484             tx               = _mm256_mul_ps(fscal,dx00);
 485             ty               = _mm256_mul_ps(fscal,dy00);
 486             tz               = _mm256_mul_ps(fscal,dz00);
 487
 488             /* Update vectorial force */
 489             fix0             = _mm256_add_ps(fix0,tx);
 490             fiy0             = _mm256_add_ps(fiy0,ty);
 491             fiz0             = _mm256_add_ps(fiz0,tz);
 492
 493             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 494             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 495             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 496             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 497             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 498             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 499             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 500             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 501             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 502
 503             /* Inner loop uses 68 flops */
 504         }
 505
 506         /* End of innermost loop */
 507
 508         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 509                                                  f+i_coord_offset,fshift+i_shift_offset);
 510
 511         ggid                        = gid[iidx];
 512         /* Update potential energies */
 513         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 514         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 515
 516         /* Increment number of inner iterations */
 517         inneriter                  += j_index_end - j_index_start;
 518
 519         /* Outer loop uses 9 flops */
 520     }
 521
 522     /* Increment number of outer iterations */
 523     outeriter        += nri;
 524
 525     /* Update outer/inner flops */
 526
 527     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*9 + inneriter*68);
 528 }
 529 /*
 530  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomP1P1_F_avx_256_single
 531  * Electrostatics interaction: ReactionField
 532  * VdW interaction:            CubicSplineTable
 533  * Geometry:                   Particle-Particle
 534  * Calculate force/pot:        Force
 535  */
 536 void
 537 nb_kernel_ElecRF_VdwCSTab_GeomP1P1_F_avx_256_single
 538                     (t_nblist                    * gmx_restrict       nlist,
 539                      rvec                        * gmx_restrict          xx,
 540                      rvec                        * gmx_restrict          ff,
 541                      t_forcerec                  * gmx_restrict          fr,
 542                      t_mdatoms                   * gmx_restrict     mdatoms,
 543                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 544                      t_nrnb                      * gmx_restrict        nrnb)
 545 {
 546     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 547      * just 0 for non-waters.
 548      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 549      * jnr indices corresponding to data put in the four positions in the SIMD register.
 550      */
 551     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 552     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 553     int              jnrA,jnrB,jnrC,jnrD;
 554     int              jnrE,jnrF,jnrG,jnrH;
 555     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 556     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 557     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 558     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 559     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 560     real             rcutoff_scalar;
 561     real             *shiftvec,*fshift,*x,*f;
 562     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 563     real             scratch[4*DIM];
 564     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 565     real *           vdwioffsetptr0;
 566     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 567     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
 568     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 569     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 570     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 571     real             *charge;
 572     int              nvdwtype;
 573     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 574     int              *vdwtype;
 575     real             *vdwparam;
 576     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 577     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 578     __m256i          vfitab;
 579     __m128i          vfitab_lo,vfitab_hi;
 580     __m128i          ifour       = _mm_set1_epi32(4);
 581     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 582     real             *vftab;
 583     __m256           dummy_mask,cutoff_mask;
 584     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 585     __m256           one     = _mm256_set1_ps(1.0);
 586     __m256           two     = _mm256_set1_ps(2.0);
 587     x                = xx[0];
 588     f                = ff[0];
 589
 590     nri              = nlist->nri;
 591     iinr             = nlist->iinr;
 592     jindex           = nlist->jindex;
 593     jjnr             = nlist->jjnr;
 594     shiftidx         = nlist->shift;
 595     gid              = nlist->gid;
 596     shiftvec         = fr->shift_vec[0];
 597     fshift           = fr->fshift[0];
 598     facel            = _mm256_set1_ps(fr->epsfac);
 599     charge           = mdatoms->chargeA;
 600     krf              = _mm256_set1_ps(fr->ic->k_rf);
 601     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 602     crf              = _mm256_set1_ps(fr->ic->c_rf);
 603     nvdwtype         = fr->ntype;
 604     vdwparam         = fr->nbfp;
 605     vdwtype          = mdatoms->typeA;
 606
 607     vftab            = kernel_data->table_vdw->data;
 608     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 609
 610     /* Avoid stupid compiler warnings */
 611     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 612     j_coord_offsetA = 0;
 613     j_coord_offsetB = 0;
 614     j_coord_offsetC = 0;
 615     j_coord_offsetD = 0;
 616     j_coord_offsetE = 0;
 617     j_coord_offsetF = 0;
 618     j_coord_offsetG = 0;
 619     j_coord_offsetH = 0;
 620
 621     outeriter        = 0;
 622     inneriter        = 0;
 623
 624     for(iidx=0;iidx<4*DIM;iidx++)
 625     {
 626         scratch[iidx] = 0.0;
 627     }
 628
 629     /* Start outer loop over neighborlists */
 630     for(iidx=0; iidx<nri; iidx++)
 631     {
 632         /* Load shift vector for this list */
 633         i_shift_offset   = DIM*shiftidx[iidx];
 634
 635         /* Load limits for loop over neighbors */
 636         j_index_start    = jindex[iidx];
 637         j_index_end      = jindex[iidx+1];
 638
 639         /* Get outer coordinate index */
 640         inr              = iinr[iidx];
 641         i_coord_offset   = DIM*inr;
 642
 643         /* Load i particle coords and add shift vector */
 644         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 645
 646         fix0             = _mm256_setzero_ps();
 647         fiy0             = _mm256_setzero_ps();
 648         fiz0             = _mm256_setzero_ps();
 649
 650         /* Load parameters for i particles */
 651         iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 652         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 653
 654         /* Start inner kernel loop */
 655         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 656         {
 657
 658             /* Get j neighbor index, and coordinate index */
 659             jnrA             = jjnr[jidx];
 660             jnrB             = jjnr[jidx+1];
 661             jnrC             = jjnr[jidx+2];
 662             jnrD             = jjnr[jidx+3];
 663             jnrE             = jjnr[jidx+4];
 664             jnrF             = jjnr[jidx+5];
 665             jnrG             = jjnr[jidx+6];
 666             jnrH             = jjnr[jidx+7];
 667             j_coord_offsetA  = DIM*jnrA;
 668             j_coord_offsetB  = DIM*jnrB;
 669             j_coord_offsetC  = DIM*jnrC;
 670             j_coord_offsetD  = DIM*jnrD;
 671             j_coord_offsetE  = DIM*jnrE;
 672             j_coord_offsetF  = DIM*jnrF;
 673             j_coord_offsetG  = DIM*jnrG;
 674             j_coord_offsetH  = DIM*jnrH;
 675
 676             /* load j atom coordinates */
 677             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 678                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 679                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 680                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 681                                                  &jx0,&jy0,&jz0);
 682
 683             /* Calculate displacement vector */
 684             dx00             = _mm256_sub_ps(ix0,jx0);
 685             dy00             = _mm256_sub_ps(iy0,jy0);
 686             dz00             = _mm256_sub_ps(iz0,jz0);
 687
 688             /* Calculate squared distance and things based on it */
 689             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 690
 691             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 692
 693             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 694
 695             /* Load parameters for j particles */
 696             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 697                                                                  charge+jnrC+0,charge+jnrD+0,
 698                                                                  charge+jnrE+0,charge+jnrF+0,
 699                                                                  charge+jnrG+0,charge+jnrH+0);
 700             vdwjidx0A        = 2*vdwtype[jnrA+0];
 701             vdwjidx0B        = 2*vdwtype[jnrB+0];
 702             vdwjidx0C        = 2*vdwtype[jnrC+0];
 703             vdwjidx0D        = 2*vdwtype[jnrD+0];
 704             vdwjidx0E        = 2*vdwtype[jnrE+0];
 705             vdwjidx0F        = 2*vdwtype[jnrF+0];
 706             vdwjidx0G        = 2*vdwtype[jnrG+0];
 707             vdwjidx0H        = 2*vdwtype[jnrH+0];
 708
 709             /**************************
 710              * CALCULATE INTERACTIONS *
 711              **************************/
 712
 713             r00              = _mm256_mul_ps(rsq00,rinv00);
 714
 715             /* Compute parameters for interactions between i and j atoms */
 716             qq00             = _mm256_mul_ps(iq0,jq0);
 717             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 718                                             vdwioffsetptr0+vdwjidx0B,
 719                                             vdwioffsetptr0+vdwjidx0C,
 720                                             vdwioffsetptr0+vdwjidx0D,
 721                                             vdwioffsetptr0+vdwjidx0E,
 722                                             vdwioffsetptr0+vdwjidx0F,
 723                                             vdwioffsetptr0+vdwjidx0G,
 724                                             vdwioffsetptr0+vdwjidx0H,
 725                                             &c6_00,&c12_00);
 726
 727             /* Calculate table index by multiplying r with table scale and truncate to integer */
 728             rt               = _mm256_mul_ps(r00,vftabscale);
 729             vfitab           = _mm256_cvttps_epi32(rt);
 730             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 731             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 732             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 733             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 734             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 735             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 736
 737             /* REACTION-FIELD ELECTROSTATICS */
 738             felec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_mul_ps(rinv00,rinvsq00),krf2));
 739
 740             /* CUBIC SPLINE TABLE DISPERSION */
 741             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 742                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 743             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 744                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 745             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 746                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 747             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 748                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 749             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 750             Heps             = _mm256_mul_ps(vfeps,H);
 751             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 752             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 753             fvdw6            = _mm256_mul_ps(c6_00,FF);
 754
 755             /* CUBIC SPLINE TABLE REPULSION */
 756             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 757             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 758             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 759                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 760             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 761                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 762             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 763                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 764             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 765                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 766             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 767             Heps             = _mm256_mul_ps(vfeps,H);
 768             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 769             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 770             fvdw12           = _mm256_mul_ps(c12_00,FF);
 771             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 772
 773             fscal            = _mm256_add_ps(felec,fvdw);
 774
 775             /* Calculate temporary vectorial force */
 776             tx               = _mm256_mul_ps(fscal,dx00);
 777             ty               = _mm256_mul_ps(fscal,dy00);
 778             tz               = _mm256_mul_ps(fscal,dz00);
 779
 780             /* Update vectorial force */
 781             fix0             = _mm256_add_ps(fix0,tx);
 782             fiy0             = _mm256_add_ps(fiy0,ty);
 783             fiz0             = _mm256_add_ps(fiz0,tz);
 784
 785             fjptrA             = f+j_coord_offsetA;
 786             fjptrB             = f+j_coord_offsetB;
 787             fjptrC             = f+j_coord_offsetC;
 788             fjptrD             = f+j_coord_offsetD;
 789             fjptrE             = f+j_coord_offsetE;
 790             fjptrF             = f+j_coord_offsetF;
 791             fjptrG             = f+j_coord_offsetG;
 792             fjptrH             = f+j_coord_offsetH;
 793             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 794
 795             /* Inner loop uses 54 flops */
 796         }
 797
 798         if(jidx<j_index_end)
 799         {
 800
 801             /* Get j neighbor index, and coordinate index */
 802             jnrlistA         = jjnr[jidx];
 803             jnrlistB         = jjnr[jidx+1];
 804             jnrlistC         = jjnr[jidx+2];
 805             jnrlistD         = jjnr[jidx+3];
 806             jnrlistE         = jjnr[jidx+4];
 807             jnrlistF         = jjnr[jidx+5];
 808             jnrlistG         = jjnr[jidx+6];
 809             jnrlistH         = jjnr[jidx+7];
 810             /* Sign of each element will be negative for non-real atoms.
 811              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 812              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 813              */
 814             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 815                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 816
 817             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 818             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 819             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 820             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 821             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 822             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 823             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 824             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 825             j_coord_offsetA  = DIM*jnrA;
 826             j_coord_offsetB  = DIM*jnrB;
 827             j_coord_offsetC  = DIM*jnrC;
 828             j_coord_offsetD  = DIM*jnrD;
 829             j_coord_offsetE  = DIM*jnrE;
 830             j_coord_offsetF  = DIM*jnrF;
 831             j_coord_offsetG  = DIM*jnrG;
 832             j_coord_offsetH  = DIM*jnrH;
 833
 834             /* load j atom coordinates */
 835             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 836                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 837                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 838                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 839                                                  &jx0,&jy0,&jz0);
 840
 841             /* Calculate displacement vector */
 842             dx00             = _mm256_sub_ps(ix0,jx0);
 843             dy00             = _mm256_sub_ps(iy0,jy0);
 844             dz00             = _mm256_sub_ps(iz0,jz0);
 845
 846             /* Calculate squared distance and things based on it */
 847             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 848
 849             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 850
 851             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 852
 853             /* Load parameters for j particles */
 854             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 855                                                                  charge+jnrC+0,charge+jnrD+0,
 856                                                                  charge+jnrE+0,charge+jnrF+0,
 857                                                                  charge+jnrG+0,charge+jnrH+0);
 858             vdwjidx0A        = 2*vdwtype[jnrA+0];
 859             vdwjidx0B        = 2*vdwtype[jnrB+0];
 860             vdwjidx0C        = 2*vdwtype[jnrC+0];
 861             vdwjidx0D        = 2*vdwtype[jnrD+0];
 862             vdwjidx0E        = 2*vdwtype[jnrE+0];
 863             vdwjidx0F        = 2*vdwtype[jnrF+0];
 864             vdwjidx0G        = 2*vdwtype[jnrG+0];
 865             vdwjidx0H        = 2*vdwtype[jnrH+0];
 866
 867             /**************************
 868              * CALCULATE INTERACTIONS *
 869              **************************/
 870
 871             r00              = _mm256_mul_ps(rsq00,rinv00);
 872             r00              = _mm256_andnot_ps(dummy_mask,r00);
 873
 874             /* Compute parameters for interactions between i and j atoms */
 875             qq00             = _mm256_mul_ps(iq0,jq0);
 876             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 877                                             vdwioffsetptr0+vdwjidx0B,
 878                                             vdwioffsetptr0+vdwjidx0C,
 879                                             vdwioffsetptr0+vdwjidx0D,
 880                                             vdwioffsetptr0+vdwjidx0E,
 881                                             vdwioffsetptr0+vdwjidx0F,
 882                                             vdwioffsetptr0+vdwjidx0G,
 883                                             vdwioffsetptr0+vdwjidx0H,
 884                                             &c6_00,&c12_00);
 885
 886             /* Calculate table index by multiplying r with table scale and truncate to integer */
 887             rt               = _mm256_mul_ps(r00,vftabscale);
 888             vfitab           = _mm256_cvttps_epi32(rt);
 889             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 890             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 891             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 892             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 893             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 894             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 895
 896             /* REACTION-FIELD ELECTROSTATICS */
 897             felec            = _mm256_mul_ps(qq00,_mm256_sub_ps(_mm256_mul_ps(rinv00,rinvsq00),krf2));
 898
 899             /* CUBIC SPLINE TABLE DISPERSION */
 900             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 901                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 902             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 903                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 904             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 905                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 906             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 907                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 908             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 909             Heps             = _mm256_mul_ps(vfeps,H);
 910             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 911             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 912             fvdw6            = _mm256_mul_ps(c6_00,FF);
 913
 914             /* CUBIC SPLINE TABLE REPULSION */
 915             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 916             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 917             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 918                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 919             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 920                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 921             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 922                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 923             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 924                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 925             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 926             Heps             = _mm256_mul_ps(vfeps,H);
 927             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 928             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 929             fvdw12           = _mm256_mul_ps(c12_00,FF);
 930             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 931
 932             fscal            = _mm256_add_ps(felec,fvdw);
 933
 934             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 935
 936             /* Calculate temporary vectorial force */
 937             tx               = _mm256_mul_ps(fscal,dx00);
 938             ty               = _mm256_mul_ps(fscal,dy00);
 939             tz               = _mm256_mul_ps(fscal,dz00);
 940
 941             /* Update vectorial force */
 942             fix0             = _mm256_add_ps(fix0,tx);
 943             fiy0             = _mm256_add_ps(fiy0,ty);
 944             fiz0             = _mm256_add_ps(fiz0,tz);
 945
 946             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 947             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 948             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 949             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 950             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 951             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 952             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 953             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 954             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 955
 956             /* Inner loop uses 55 flops */
 957         }
 958
 959         /* End of innermost loop */
 960
 961         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 962                                                  f+i_coord_offset,fshift+i_shift_offset);
 963
 964         /* Increment number of inner iterations */
 965         inneriter                  += j_index_end - j_index_start;
 966
 967         /* Outer loop uses 7 flops */
 968     }
 969
 970     /* Increment number of outer iterations */
 971     outeriter        += nri;
 972
 973     /* Update outer/inner flops */
 974
 975     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*7 + inneriter*55);
 976 }