src/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecRFCut_VdwCSTab_GeomP1P1_avx_256_double.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_256_double kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 #include "gmx_math_x86_avx_256_double.h"
  34 #include "kernelutil_x86_avx_256_double.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomP1P1_VF_avx_256_double
  38  * Electrostatics interaction: ReactionField
  39  * VdW interaction:            CubicSplineTable
  40  * Geometry:                   Particle-Particle
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecRFCut_VdwCSTab_GeomP1P1_VF_avx_256_double
  45                     (t_nblist * gmx_restrict                nlist,
  46                      rvec * gmx_restrict                    xx,
  47                      rvec * gmx_restrict                    ff,
  48                      t_forcerec * gmx_restrict              fr,
  49                      t_mdatoms * gmx_restrict               mdatoms,
  50                      nb_kernel_data_t * gmx_restrict        kernel_data,
  51                      t_nrnb * gmx_restrict                  nrnb)
  52 {
  53     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  54      * just 0 for non-waters.
  55      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  56      * jnr indices corresponding to data put in the four positions in the SIMD register.
  57      */
  58     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  59     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  60     int              jnrA,jnrB,jnrC,jnrD;
  61     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  62     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  63     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  64     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  65     real             rcutoff_scalar;
  66     real             *shiftvec,*fshift,*x,*f;
  67     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  68     real             scratch[4*DIM];
  69     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  70     real *           vdwioffsetptr0;
  71     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  73     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  74     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  75     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
  76     real             *charge;
  77     int              nvdwtype;
  78     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  79     int              *vdwtype;
  80     real             *vdwparam;
  81     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
  82     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
  83     __m128i          vfitab;
  84     __m128i          ifour       = _mm_set1_epi32(4);
  85     __m256d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
  86     real             *vftab;
  87     __m256d          dummy_mask,cutoff_mask;
  88     __m128           tmpmask0,tmpmask1;
  89     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
  90     __m256d          one     = _mm256_set1_pd(1.0);
  91     __m256d          two     = _mm256_set1_pd(2.0);
  92     x                = xx[0];
  93     f                = ff[0];
  94
  95     nri              = nlist->nri;
  96     iinr             = nlist->iinr;
  97     jindex           = nlist->jindex;
  98     jjnr             = nlist->jjnr;
  99     shiftidx         = nlist->shift;
 100     gid              = nlist->gid;
 101     shiftvec         = fr->shift_vec[0];
 102     fshift           = fr->fshift[0];
 103     facel            = _mm256_set1_pd(fr->epsfac);
 104     charge           = mdatoms->chargeA;
 105     krf              = _mm256_set1_pd(fr->ic->k_rf);
 106     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 107     crf              = _mm256_set1_pd(fr->ic->c_rf);
 108     nvdwtype         = fr->ntype;
 109     vdwparam         = fr->nbfp;
 110     vdwtype          = mdatoms->typeA;
 111
 112     vftab            = kernel_data->table_vdw->data;
 113     vftabscale       = _mm256_set1_pd(kernel_data->table_vdw->scale);
 114
 115     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 116     rcutoff_scalar   = fr->rcoulomb;
 117     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 118     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 119
 120     /* Avoid stupid compiler warnings */
 121     jnrA = jnrB = jnrC = jnrD = 0;
 122     j_coord_offsetA = 0;
 123     j_coord_offsetB = 0;
 124     j_coord_offsetC = 0;
 125     j_coord_offsetD = 0;
 126
 127     outeriter        = 0;
 128     inneriter        = 0;
 129
 130     for(iidx=0;iidx<4*DIM;iidx++)
 131     {
 132         scratch[iidx] = 0.0;
 133     }
 134
 135     /* Start outer loop over neighborlists */
 136     for(iidx=0; iidx<nri; iidx++)
 137     {
 138         /* Load shift vector for this list */
 139         i_shift_offset   = DIM*shiftidx[iidx];
 140
 141         /* Load limits for loop over neighbors */
 142         j_index_start    = jindex[iidx];
 143         j_index_end      = jindex[iidx+1];
 144
 145         /* Get outer coordinate index */
 146         inr              = iinr[iidx];
 147         i_coord_offset   = DIM*inr;
 148
 149         /* Load i particle coords and add shift vector */
 150         gmx_mm256_load_shift_and_1rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 151
 152         fix0             = _mm256_setzero_pd();
 153         fiy0             = _mm256_setzero_pd();
 154         fiz0             = _mm256_setzero_pd();
 155
 156         /* Load parameters for i particles */
 157         iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 158         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 159
 160         /* Reset potential sums */
 161         velecsum         = _mm256_setzero_pd();
 162         vvdwsum          = _mm256_setzero_pd();
 163
 164         /* Start inner kernel loop */
 165         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 166         {
 167
 168             /* Get j neighbor index, and coordinate index */
 169             jnrA             = jjnr[jidx];
 170             jnrB             = jjnr[jidx+1];
 171             jnrC             = jjnr[jidx+2];
 172             jnrD             = jjnr[jidx+3];
 173             j_coord_offsetA  = DIM*jnrA;
 174             j_coord_offsetB  = DIM*jnrB;
 175             j_coord_offsetC  = DIM*jnrC;
 176             j_coord_offsetD  = DIM*jnrD;
 177
 178             /* load j atom coordinates */
 179             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 180                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 181                                                  &jx0,&jy0,&jz0);
 182
 183             /* Calculate displacement vector */
 184             dx00             = _mm256_sub_pd(ix0,jx0);
 185             dy00             = _mm256_sub_pd(iy0,jy0);
 186             dz00             = _mm256_sub_pd(iz0,jz0);
 187
 188             /* Calculate squared distance and things based on it */
 189             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 190
 191             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 192
 193             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 194
 195             /* Load parameters for j particles */
 196             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 197                                                                  charge+jnrC+0,charge+jnrD+0);
 198             vdwjidx0A        = 2*vdwtype[jnrA+0];
 199             vdwjidx0B        = 2*vdwtype[jnrB+0];
 200             vdwjidx0C        = 2*vdwtype[jnrC+0];
 201             vdwjidx0D        = 2*vdwtype[jnrD+0];
 202
 203             /**************************
 204              * CALCULATE INTERACTIONS *
 205              **************************/
 206
 207             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 208             {
 209
 210             r00              = _mm256_mul_pd(rsq00,rinv00);
 211
 212             /* Compute parameters for interactions between i and j atoms */
 213             qq00             = _mm256_mul_pd(iq0,jq0);
 214             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 215                                             vdwioffsetptr0+vdwjidx0B,
 216                                             vdwioffsetptr0+vdwjidx0C,
 217                                             vdwioffsetptr0+vdwjidx0D,
 218                                             &c6_00,&c12_00);
 219
 220             /* Calculate table index by multiplying r with table scale and truncate to integer */
 221             rt               = _mm256_mul_pd(r00,vftabscale);
 222             vfitab           = _mm256_cvttpd_epi32(rt);
 223             vfeps            = _mm256_sub_pd(rt,_mm256_round_pd(rt, _MM_FROUND_FLOOR));
 224             vfitab           = _mm_slli_epi32(vfitab,3);
 225
 226             /* REACTION-FIELD ELECTROSTATICS */
 227             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 228             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 229
 230             /* CUBIC SPLINE TABLE DISPERSION */
 231             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 232             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 233             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 234             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 235             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 236             Heps             = _mm256_mul_pd(vfeps,H);
 237             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 238             VV               = _mm256_add_pd(Y,_mm256_mul_pd(vfeps,Fp));
 239             vvdw6            = _mm256_mul_pd(c6_00,VV);
 240             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 241             fvdw6            = _mm256_mul_pd(c6_00,FF);
 242
 243             /* CUBIC SPLINE TABLE REPULSION */
 244             vfitab           = _mm_add_epi32(vfitab,ifour);
 245             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 246             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 247             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 248             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 249             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 250             Heps             = _mm256_mul_pd(vfeps,H);
 251             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 252             VV               = _mm256_add_pd(Y,_mm256_mul_pd(vfeps,Fp));
 253             vvdw12           = _mm256_mul_pd(c12_00,VV);
 254             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 255             fvdw12           = _mm256_mul_pd(c12_00,FF);
 256             vvdw             = _mm256_add_pd(vvdw12,vvdw6);
 257             fvdw             = _mm256_xor_pd(signbit,_mm256_mul_pd(_mm256_add_pd(fvdw6,fvdw12),_mm256_mul_pd(vftabscale,rinv00)));
 258
 259             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 260
 261             /* Update potential sum for this i atom from the interaction with this j atom. */
 262             velec            = _mm256_and_pd(velec,cutoff_mask);
 263             velecsum         = _mm256_add_pd(velecsum,velec);
 264             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 265             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 266
 267             fscal            = _mm256_add_pd(felec,fvdw);
 268
 269             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 270
 271             /* Calculate temporary vectorial force */
 272             tx               = _mm256_mul_pd(fscal,dx00);
 273             ty               = _mm256_mul_pd(fscal,dy00);
 274             tz               = _mm256_mul_pd(fscal,dz00);
 275
 276             /* Update vectorial force */
 277             fix0             = _mm256_add_pd(fix0,tx);
 278             fiy0             = _mm256_add_pd(fiy0,ty);
 279             fiz0             = _mm256_add_pd(fiz0,tz);
 280
 281             fjptrA             = f+j_coord_offsetA;
 282             fjptrB             = f+j_coord_offsetB;
 283             fjptrC             = f+j_coord_offsetC;
 284             fjptrD             = f+j_coord_offsetD;
 285             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 286
 287             }
 288
 289             /* Inner loop uses 72 flops */
 290         }
 291
 292         if(jidx<j_index_end)
 293         {
 294
 295             /* Get j neighbor index, and coordinate index */
 296             jnrlistA         = jjnr[jidx];
 297             jnrlistB         = jjnr[jidx+1];
 298             jnrlistC         = jjnr[jidx+2];
 299             jnrlistD         = jjnr[jidx+3];
 300             /* Sign of each element will be negative for non-real atoms.
 301              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 302              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 303              */
 304             tmpmask0 = gmx_mm_castsi128_pd(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 305
 306             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 307             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 308             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 309
 310             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 311             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 312             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 313             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 314             j_coord_offsetA  = DIM*jnrA;
 315             j_coord_offsetB  = DIM*jnrB;
 316             j_coord_offsetC  = DIM*jnrC;
 317             j_coord_offsetD  = DIM*jnrD;
 318
 319             /* load j atom coordinates */
 320             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 321                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 322                                                  &jx0,&jy0,&jz0);
 323
 324             /* Calculate displacement vector */
 325             dx00             = _mm256_sub_pd(ix0,jx0);
 326             dy00             = _mm256_sub_pd(iy0,jy0);
 327             dz00             = _mm256_sub_pd(iz0,jz0);
 328
 329             /* Calculate squared distance and things based on it */
 330             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 331
 332             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 333
 334             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 335
 336             /* Load parameters for j particles */
 337             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 338                                                                  charge+jnrC+0,charge+jnrD+0);
 339             vdwjidx0A        = 2*vdwtype[jnrA+0];
 340             vdwjidx0B        = 2*vdwtype[jnrB+0];
 341             vdwjidx0C        = 2*vdwtype[jnrC+0];
 342             vdwjidx0D        = 2*vdwtype[jnrD+0];
 343
 344             /**************************
 345              * CALCULATE INTERACTIONS *
 346              **************************/
 347
 348             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 349             {
 350
 351             r00              = _mm256_mul_pd(rsq00,rinv00);
 352             r00              = _mm256_andnot_pd(dummy_mask,r00);
 353
 354             /* Compute parameters for interactions between i and j atoms */
 355             qq00             = _mm256_mul_pd(iq0,jq0);
 356             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 357                                             vdwioffsetptr0+vdwjidx0B,
 358                                             vdwioffsetptr0+vdwjidx0C,
 359                                             vdwioffsetptr0+vdwjidx0D,
 360                                             &c6_00,&c12_00);
 361
 362             /* Calculate table index by multiplying r with table scale and truncate to integer */
 363             rt               = _mm256_mul_pd(r00,vftabscale);
 364             vfitab           = _mm256_cvttpd_epi32(rt);
 365             vfeps            = _mm256_sub_pd(rt,_mm256_round_pd(rt, _MM_FROUND_FLOOR));
 366             vfitab           = _mm_slli_epi32(vfitab,3);
 367
 368             /* REACTION-FIELD ELECTROSTATICS */
 369             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 370             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 371
 372             /* CUBIC SPLINE TABLE DISPERSION */
 373             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 374             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 375             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 376             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 377             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 378             Heps             = _mm256_mul_pd(vfeps,H);
 379             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 380             VV               = _mm256_add_pd(Y,_mm256_mul_pd(vfeps,Fp));
 381             vvdw6            = _mm256_mul_pd(c6_00,VV);
 382             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 383             fvdw6            = _mm256_mul_pd(c6_00,FF);
 384
 385             /* CUBIC SPLINE TABLE REPULSION */
 386             vfitab           = _mm_add_epi32(vfitab,ifour);
 387             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 388             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 389             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 390             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 391             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 392             Heps             = _mm256_mul_pd(vfeps,H);
 393             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 394             VV               = _mm256_add_pd(Y,_mm256_mul_pd(vfeps,Fp));
 395             vvdw12           = _mm256_mul_pd(c12_00,VV);
 396             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 397             fvdw12           = _mm256_mul_pd(c12_00,FF);
 398             vvdw             = _mm256_add_pd(vvdw12,vvdw6);
 399             fvdw             = _mm256_xor_pd(signbit,_mm256_mul_pd(_mm256_add_pd(fvdw6,fvdw12),_mm256_mul_pd(vftabscale,rinv00)));
 400
 401             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 402
 403             /* Update potential sum for this i atom from the interaction with this j atom. */
 404             velec            = _mm256_and_pd(velec,cutoff_mask);
 405             velec            = _mm256_andnot_pd(dummy_mask,velec);
 406             velecsum         = _mm256_add_pd(velecsum,velec);
 407             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 408             vvdw             = _mm256_andnot_pd(dummy_mask,vvdw);
 409             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 410
 411             fscal            = _mm256_add_pd(felec,fvdw);
 412
 413             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 414
 415             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 416
 417             /* Calculate temporary vectorial force */
 418             tx               = _mm256_mul_pd(fscal,dx00);
 419             ty               = _mm256_mul_pd(fscal,dy00);
 420             tz               = _mm256_mul_pd(fscal,dz00);
 421
 422             /* Update vectorial force */
 423             fix0             = _mm256_add_pd(fix0,tx);
 424             fiy0             = _mm256_add_pd(fiy0,ty);
 425             fiz0             = _mm256_add_pd(fiz0,tz);
 426
 427             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 428             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 429             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 430             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 431             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 432
 433             }
 434
 435             /* Inner loop uses 73 flops */
 436         }
 437
 438         /* End of innermost loop */
 439
 440         gmx_mm256_update_iforce_1atom_swizzle_pd(fix0,fiy0,fiz0,
 441                                                  f+i_coord_offset,fshift+i_shift_offset);
 442
 443         ggid                        = gid[iidx];
 444         /* Update potential energies */
 445         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 446         gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 447
 448         /* Increment number of inner iterations */
 449         inneriter                  += j_index_end - j_index_start;
 450
 451         /* Outer loop uses 9 flops */
 452     }
 453
 454     /* Increment number of outer iterations */
 455     outeriter        += nri;
 456
 457     /* Update outer/inner flops */
 458
 459     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*9 + inneriter*73);
 460 }
 461 /*
 462  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomP1P1_F_avx_256_double
 463  * Electrostatics interaction: ReactionField
 464  * VdW interaction:            CubicSplineTable
 465  * Geometry:                   Particle-Particle
 466  * Calculate force/pot:        Force
 467  */
 468 void
 469 nb_kernel_ElecRFCut_VdwCSTab_GeomP1P1_F_avx_256_double
 470                     (t_nblist * gmx_restrict                nlist,
 471                      rvec * gmx_restrict                    xx,
 472                      rvec * gmx_restrict                    ff,
 473                      t_forcerec * gmx_restrict              fr,
 474                      t_mdatoms * gmx_restrict               mdatoms,
 475                      nb_kernel_data_t * gmx_restrict        kernel_data,
 476                      t_nrnb * gmx_restrict                  nrnb)
 477 {
 478     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 479      * just 0 for non-waters.
 480      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
 481      * jnr indices corresponding to data put in the four positions in the SIMD register.
 482      */
 483     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 484     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 485     int              jnrA,jnrB,jnrC,jnrD;
 486     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 487     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 488     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 489     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 490     real             rcutoff_scalar;
 491     real             *shiftvec,*fshift,*x,*f;
 492     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 493     real             scratch[4*DIM];
 494     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 495     real *           vdwioffsetptr0;
 496     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 497     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 498     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 499     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 500     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 501     real             *charge;
 502     int              nvdwtype;
 503     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 504     int              *vdwtype;
 505     real             *vdwparam;
 506     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 507     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 508     __m128i          vfitab;
 509     __m128i          ifour       = _mm_set1_epi32(4);
 510     __m256d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 511     real             *vftab;
 512     __m256d          dummy_mask,cutoff_mask;
 513     __m128           tmpmask0,tmpmask1;
 514     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 515     __m256d          one     = _mm256_set1_pd(1.0);
 516     __m256d          two     = _mm256_set1_pd(2.0);
 517     x                = xx[0];
 518     f                = ff[0];
 519
 520     nri              = nlist->nri;
 521     iinr             = nlist->iinr;
 522     jindex           = nlist->jindex;
 523     jjnr             = nlist->jjnr;
 524     shiftidx         = nlist->shift;
 525     gid              = nlist->gid;
 526     shiftvec         = fr->shift_vec[0];
 527     fshift           = fr->fshift[0];
 528     facel            = _mm256_set1_pd(fr->epsfac);
 529     charge           = mdatoms->chargeA;
 530     krf              = _mm256_set1_pd(fr->ic->k_rf);
 531     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 532     crf              = _mm256_set1_pd(fr->ic->c_rf);
 533     nvdwtype         = fr->ntype;
 534     vdwparam         = fr->nbfp;
 535     vdwtype          = mdatoms->typeA;
 536
 537     vftab            = kernel_data->table_vdw->data;
 538     vftabscale       = _mm256_set1_pd(kernel_data->table_vdw->scale);
 539
 540     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 541     rcutoff_scalar   = fr->rcoulomb;
 542     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 543     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 544
 545     /* Avoid stupid compiler warnings */
 546     jnrA = jnrB = jnrC = jnrD = 0;
 547     j_coord_offsetA = 0;
 548     j_coord_offsetB = 0;
 549     j_coord_offsetC = 0;
 550     j_coord_offsetD = 0;
 551
 552     outeriter        = 0;
 553     inneriter        = 0;
 554
 555     for(iidx=0;iidx<4*DIM;iidx++)
 556     {
 557         scratch[iidx] = 0.0;
 558     }
 559
 560     /* Start outer loop over neighborlists */
 561     for(iidx=0; iidx<nri; iidx++)
 562     {
 563         /* Load shift vector for this list */
 564         i_shift_offset   = DIM*shiftidx[iidx];
 565
 566         /* Load limits for loop over neighbors */
 567         j_index_start    = jindex[iidx];
 568         j_index_end      = jindex[iidx+1];
 569
 570         /* Get outer coordinate index */
 571         inr              = iinr[iidx];
 572         i_coord_offset   = DIM*inr;
 573
 574         /* Load i particle coords and add shift vector */
 575         gmx_mm256_load_shift_and_1rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 576
 577         fix0             = _mm256_setzero_pd();
 578         fiy0             = _mm256_setzero_pd();
 579         fiz0             = _mm256_setzero_pd();
 580
 581         /* Load parameters for i particles */
 582         iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 583         vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 584
 585         /* Start inner kernel loop */
 586         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 587         {
 588
 589             /* Get j neighbor index, and coordinate index */
 590             jnrA             = jjnr[jidx];
 591             jnrB             = jjnr[jidx+1];
 592             jnrC             = jjnr[jidx+2];
 593             jnrD             = jjnr[jidx+3];
 594             j_coord_offsetA  = DIM*jnrA;
 595             j_coord_offsetB  = DIM*jnrB;
 596             j_coord_offsetC  = DIM*jnrC;
 597             j_coord_offsetD  = DIM*jnrD;
 598
 599             /* load j atom coordinates */
 600             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 601                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 602                                                  &jx0,&jy0,&jz0);
 603
 604             /* Calculate displacement vector */
 605             dx00             = _mm256_sub_pd(ix0,jx0);
 606             dy00             = _mm256_sub_pd(iy0,jy0);
 607             dz00             = _mm256_sub_pd(iz0,jz0);
 608
 609             /* Calculate squared distance and things based on it */
 610             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 611
 612             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 613
 614             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 615
 616             /* Load parameters for j particles */
 617             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 618                                                                  charge+jnrC+0,charge+jnrD+0);
 619             vdwjidx0A        = 2*vdwtype[jnrA+0];
 620             vdwjidx0B        = 2*vdwtype[jnrB+0];
 621             vdwjidx0C        = 2*vdwtype[jnrC+0];
 622             vdwjidx0D        = 2*vdwtype[jnrD+0];
 623
 624             /**************************
 625              * CALCULATE INTERACTIONS *
 626              **************************/
 627
 628             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 629             {
 630
 631             r00              = _mm256_mul_pd(rsq00,rinv00);
 632
 633             /* Compute parameters for interactions between i and j atoms */
 634             qq00             = _mm256_mul_pd(iq0,jq0);
 635             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 636                                             vdwioffsetptr0+vdwjidx0B,
 637                                             vdwioffsetptr0+vdwjidx0C,
 638                                             vdwioffsetptr0+vdwjidx0D,
 639                                             &c6_00,&c12_00);
 640
 641             /* Calculate table index by multiplying r with table scale and truncate to integer */
 642             rt               = _mm256_mul_pd(r00,vftabscale);
 643             vfitab           = _mm256_cvttpd_epi32(rt);
 644             vfeps            = _mm256_sub_pd(rt,_mm256_round_pd(rt, _MM_FROUND_FLOOR));
 645             vfitab           = _mm_slli_epi32(vfitab,3);
 646
 647             /* REACTION-FIELD ELECTROSTATICS */
 648             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 649
 650             /* CUBIC SPLINE TABLE DISPERSION */
 651             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 652             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 653             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 654             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 655             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 656             Heps             = _mm256_mul_pd(vfeps,H);
 657             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 658             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 659             fvdw6            = _mm256_mul_pd(c6_00,FF);
 660
 661             /* CUBIC SPLINE TABLE REPULSION */
 662             vfitab           = _mm_add_epi32(vfitab,ifour);
 663             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 664             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 665             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 666             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 667             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 668             Heps             = _mm256_mul_pd(vfeps,H);
 669             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 670             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 671             fvdw12           = _mm256_mul_pd(c12_00,FF);
 672             fvdw             = _mm256_xor_pd(signbit,_mm256_mul_pd(_mm256_add_pd(fvdw6,fvdw12),_mm256_mul_pd(vftabscale,rinv00)));
 673
 674             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 675
 676             fscal            = _mm256_add_pd(felec,fvdw);
 677
 678             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 679
 680             /* Calculate temporary vectorial force */
 681             tx               = _mm256_mul_pd(fscal,dx00);
 682             ty               = _mm256_mul_pd(fscal,dy00);
 683             tz               = _mm256_mul_pd(fscal,dz00);
 684
 685             /* Update vectorial force */
 686             fix0             = _mm256_add_pd(fix0,tx);
 687             fiy0             = _mm256_add_pd(fiy0,ty);
 688             fiz0             = _mm256_add_pd(fiz0,tz);
 689
 690             fjptrA             = f+j_coord_offsetA;
 691             fjptrB             = f+j_coord_offsetB;
 692             fjptrC             = f+j_coord_offsetC;
 693             fjptrD             = f+j_coord_offsetD;
 694             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 695
 696             }
 697
 698             /* Inner loop uses 57 flops */
 699         }
 700
 701         if(jidx<j_index_end)
 702         {
 703
 704             /* Get j neighbor index, and coordinate index */
 705             jnrlistA         = jjnr[jidx];
 706             jnrlistB         = jjnr[jidx+1];
 707             jnrlistC         = jjnr[jidx+2];
 708             jnrlistD         = jjnr[jidx+3];
 709             /* Sign of each element will be negative for non-real atoms.
 710              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 711              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 712              */
 713             tmpmask0 = gmx_mm_castsi128_pd(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 714
 715             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 716             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 717             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 718
 719             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 720             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 721             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 722             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 723             j_coord_offsetA  = DIM*jnrA;
 724             j_coord_offsetB  = DIM*jnrB;
 725             j_coord_offsetC  = DIM*jnrC;
 726             j_coord_offsetD  = DIM*jnrD;
 727
 728             /* load j atom coordinates */
 729             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 730                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 731                                                  &jx0,&jy0,&jz0);
 732
 733             /* Calculate displacement vector */
 734             dx00             = _mm256_sub_pd(ix0,jx0);
 735             dy00             = _mm256_sub_pd(iy0,jy0);
 736             dz00             = _mm256_sub_pd(iz0,jz0);
 737
 738             /* Calculate squared distance and things based on it */
 739             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 740
 741             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 742
 743             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 744
 745             /* Load parameters for j particles */
 746             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 747                                                                  charge+jnrC+0,charge+jnrD+0);
 748             vdwjidx0A        = 2*vdwtype[jnrA+0];
 749             vdwjidx0B        = 2*vdwtype[jnrB+0];
 750             vdwjidx0C        = 2*vdwtype[jnrC+0];
 751             vdwjidx0D        = 2*vdwtype[jnrD+0];
 752
 753             /**************************
 754              * CALCULATE INTERACTIONS *
 755              **************************/
 756
 757             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 758             {
 759
 760             r00              = _mm256_mul_pd(rsq00,rinv00);
 761             r00              = _mm256_andnot_pd(dummy_mask,r00);
 762
 763             /* Compute parameters for interactions between i and j atoms */
 764             qq00             = _mm256_mul_pd(iq0,jq0);
 765             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 766                                             vdwioffsetptr0+vdwjidx0B,
 767                                             vdwioffsetptr0+vdwjidx0C,
 768                                             vdwioffsetptr0+vdwjidx0D,
 769                                             &c6_00,&c12_00);
 770
 771             /* Calculate table index by multiplying r with table scale and truncate to integer */
 772             rt               = _mm256_mul_pd(r00,vftabscale);
 773             vfitab           = _mm256_cvttpd_epi32(rt);
 774             vfeps            = _mm256_sub_pd(rt,_mm256_round_pd(rt, _MM_FROUND_FLOOR));
 775             vfitab           = _mm_slli_epi32(vfitab,3);
 776
 777             /* REACTION-FIELD ELECTROSTATICS */
 778             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 779
 780             /* CUBIC SPLINE TABLE DISPERSION */
 781             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 782             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 783             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 784             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 785             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 786             Heps             = _mm256_mul_pd(vfeps,H);
 787             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 788             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 789             fvdw6            = _mm256_mul_pd(c6_00,FF);
 790
 791             /* CUBIC SPLINE TABLE REPULSION */
 792             vfitab           = _mm_add_epi32(vfitab,ifour);
 793             Y                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 794             F                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 795             G                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,2) );
 796             H                = _mm256_load_pd( vftab + _mm_extract_epi32(vfitab,3) );
 797             GMX_MM256_FULLTRANSPOSE4_PD(Y,F,G,H);
 798             Heps             = _mm256_mul_pd(vfeps,H);
 799             Fp               = _mm256_add_pd(F,_mm256_mul_pd(vfeps,_mm256_add_pd(G,Heps)));
 800             FF               = _mm256_add_pd(Fp,_mm256_mul_pd(vfeps,_mm256_add_pd(G,_mm256_add_pd(Heps,Heps))));
 801             fvdw12           = _mm256_mul_pd(c12_00,FF);
 802             fvdw             = _mm256_xor_pd(signbit,_mm256_mul_pd(_mm256_add_pd(fvdw6,fvdw12),_mm256_mul_pd(vftabscale,rinv00)));
 803
 804             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 805
 806             fscal            = _mm256_add_pd(felec,fvdw);
 807
 808             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 809
 810             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 811
 812             /* Calculate temporary vectorial force */
 813             tx               = _mm256_mul_pd(fscal,dx00);
 814             ty               = _mm256_mul_pd(fscal,dy00);
 815             tz               = _mm256_mul_pd(fscal,dz00);
 816
 817             /* Update vectorial force */
 818             fix0             = _mm256_add_pd(fix0,tx);
 819             fiy0             = _mm256_add_pd(fiy0,ty);
 820             fiz0             = _mm256_add_pd(fiz0,tz);
 821
 822             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 823             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 824             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 825             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 826             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 827
 828             }
 829
 830             /* Inner loop uses 58 flops */
 831         }
 832
 833         /* End of innermost loop */
 834
 835         gmx_mm256_update_iforce_1atom_swizzle_pd(fix0,fiy0,fiz0,
 836                                                  f+i_coord_offset,fshift+i_shift_offset);
 837
 838         /* Increment number of inner iterations */
 839         inneriter                  += j_index_end - j_index_start;
 840
 841         /* Outer loop uses 7 flops */
 842     }
 843
 844     /* Increment number of outer iterations */
 845     outeriter        += nri;
 846
 847     /* Update outer/inner flops */
 848
 849     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*7 + inneriter*58);
 850 }