src/gromacs/gmxlib/nonbonded/nb_kernel_sparc64_hpc_ace_double/nb_kernel_ElecRF_VdwCSTab_GeomW4W4_sparc64_hpc_ace_double.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 sparc64_hpc_ace_double 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 "kernelutil_sparc64_hpc_ace_double.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomW4W4_VF_sparc64_hpc_ace_double
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            CubicSplineTable
  53  * Geometry:                   Water4-Water4
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRF_VdwCSTab_GeomW4W4_VF_sparc64_hpc_ace_double
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB;
  74     int              j_coord_offsetA,j_coord_offsetB;
  75     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  76     real             rcutoff_scalar;
  77     real             *shiftvec,*fshift,*x,*f;
  78     _fjsp_v2r8       tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  79     int              vdwioffset0;
  80     _fjsp_v2r8       ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  81     int              vdwioffset1;
  82     _fjsp_v2r8       ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  83     int              vdwioffset2;
  84     _fjsp_v2r8       ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  85     int              vdwioffset3;
  86     _fjsp_v2r8       ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  87     int              vdwjidx0A,vdwjidx0B;
  88     _fjsp_v2r8       jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  89     int              vdwjidx1A,vdwjidx1B;
  90     _fjsp_v2r8       jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  91     int              vdwjidx2A,vdwjidx2B;
  92     _fjsp_v2r8       jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  93     int              vdwjidx3A,vdwjidx3B;
  94     _fjsp_v2r8       jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
  95     _fjsp_v2r8       dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  96     _fjsp_v2r8       dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  97     _fjsp_v2r8       dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  98     _fjsp_v2r8       dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
  99     _fjsp_v2r8       dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 100     _fjsp_v2r8       dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 101     _fjsp_v2r8       dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
 102     _fjsp_v2r8       dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
 103     _fjsp_v2r8       dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
 104     _fjsp_v2r8       dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
 105     _fjsp_v2r8       velec,felec,velecsum,facel,crf,krf,krf2;
 106     real             *charge;
 107     int              nvdwtype;
 108     _fjsp_v2r8       rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 109     int              *vdwtype;
 110     real             *vdwparam;
 111     _fjsp_v2r8       one_sixth   = gmx_fjsp_set1_v2r8(1.0/6.0);
 112     _fjsp_v2r8       one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
 113     _fjsp_v2r8       rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 114     real             *vftab;
 115     _fjsp_v2r8       itab_tmp;
 116     _fjsp_v2r8       dummy_mask,cutoff_mask;
 117     _fjsp_v2r8       one     = gmx_fjsp_set1_v2r8(1.0);
 118     _fjsp_v2r8       two     = gmx_fjsp_set1_v2r8(2.0);
 119     union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
 120
 121     x                = xx[0];
 122     f                = ff[0];
 123
 124     nri              = nlist->nri;
 125     iinr             = nlist->iinr;
 126     jindex           = nlist->jindex;
 127     jjnr             = nlist->jjnr;
 128     shiftidx         = nlist->shift;
 129     gid              = nlist->gid;
 130     shiftvec         = fr->shift_vec[0];
 131     fshift           = fr->fshift[0];
 132     facel            = gmx_fjsp_set1_v2r8(fr->epsfac);
 133     charge           = mdatoms->chargeA;
 134     krf              = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
 135     krf2             = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
 136     crf              = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
 137     nvdwtype         = fr->ntype;
 138     vdwparam         = fr->nbfp;
 139     vdwtype          = mdatoms->typeA;
 140
 141     vftab            = kernel_data->table_vdw->data;
 142     vftabscale       = gmx_fjsp_set1_v2r8(kernel_data->table_vdw->scale);
 143
 144     /* Setup water-specific parameters */
 145     inr              = nlist->iinr[0];
 146     iq1              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+1]));
 147     iq2              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+2]));
 148     iq3              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+3]));
 149     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 150
 151     jq1              = gmx_fjsp_set1_v2r8(charge[inr+1]);
 152     jq2              = gmx_fjsp_set1_v2r8(charge[inr+2]);
 153     jq3              = gmx_fjsp_set1_v2r8(charge[inr+3]);
 154     vdwjidx0A        = 2*vdwtype[inr+0];
 155     c6_00            = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset0+vdwjidx0A]);
 156     c12_00           = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset0+vdwjidx0A+1]);
 157     qq11             = _fjsp_mul_v2r8(iq1,jq1);
 158     qq12             = _fjsp_mul_v2r8(iq1,jq2);
 159     qq13             = _fjsp_mul_v2r8(iq1,jq3);
 160     qq21             = _fjsp_mul_v2r8(iq2,jq1);
 161     qq22             = _fjsp_mul_v2r8(iq2,jq2);
 162     qq23             = _fjsp_mul_v2r8(iq2,jq3);
 163     qq31             = _fjsp_mul_v2r8(iq3,jq1);
 164     qq32             = _fjsp_mul_v2r8(iq3,jq2);
 165     qq33             = _fjsp_mul_v2r8(iq3,jq3);
 166
 167     /* Avoid stupid compiler warnings */
 168     jnrA = jnrB = 0;
 169     j_coord_offsetA = 0;
 170     j_coord_offsetB = 0;
 171
 172     outeriter        = 0;
 173     inneriter        = 0;
 174
 175     /* Start outer loop over neighborlists */
 176     for(iidx=0; iidx<nri; iidx++)
 177     {
 178         /* Load shift vector for this list */
 179         i_shift_offset   = DIM*shiftidx[iidx];
 180
 181         /* Load limits for loop over neighbors */
 182         j_index_start    = jindex[iidx];
 183         j_index_end      = jindex[iidx+1];
 184
 185         /* Get outer coordinate index */
 186         inr              = iinr[iidx];
 187         i_coord_offset   = DIM*inr;
 188
 189         /* Load i particle coords and add shift vector */
 190         gmx_fjsp_load_shift_and_4rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
 191                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 192
 193         fix0             = _fjsp_setzero_v2r8();
 194         fiy0             = _fjsp_setzero_v2r8();
 195         fiz0             = _fjsp_setzero_v2r8();
 196         fix1             = _fjsp_setzero_v2r8();
 197         fiy1             = _fjsp_setzero_v2r8();
 198         fiz1             = _fjsp_setzero_v2r8();
 199         fix2             = _fjsp_setzero_v2r8();
 200         fiy2             = _fjsp_setzero_v2r8();
 201         fiz2             = _fjsp_setzero_v2r8();
 202         fix3             = _fjsp_setzero_v2r8();
 203         fiy3             = _fjsp_setzero_v2r8();
 204         fiz3             = _fjsp_setzero_v2r8();
 205
 206         /* Reset potential sums */
 207         velecsum         = _fjsp_setzero_v2r8();
 208         vvdwsum          = _fjsp_setzero_v2r8();
 209
 210         /* Start inner kernel loop */
 211         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 212         {
 213
 214             /* Get j neighbor index, and coordinate index */
 215             jnrA             = jjnr[jidx];
 216             jnrB             = jjnr[jidx+1];
 217             j_coord_offsetA  = DIM*jnrA;
 218             j_coord_offsetB  = DIM*jnrB;
 219
 220             /* load j atom coordinates */
 221             gmx_fjsp_load_4rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
 222                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 223                                               &jy2,&jz2,&jx3,&jy3,&jz3);
 224
 225             /* Calculate displacement vector */
 226             dx00             = _fjsp_sub_v2r8(ix0,jx0);
 227             dy00             = _fjsp_sub_v2r8(iy0,jy0);
 228             dz00             = _fjsp_sub_v2r8(iz0,jz0);
 229             dx11             = _fjsp_sub_v2r8(ix1,jx1);
 230             dy11             = _fjsp_sub_v2r8(iy1,jy1);
 231             dz11             = _fjsp_sub_v2r8(iz1,jz1);
 232             dx12             = _fjsp_sub_v2r8(ix1,jx2);
 233             dy12             = _fjsp_sub_v2r8(iy1,jy2);
 234             dz12             = _fjsp_sub_v2r8(iz1,jz2);
 235             dx13             = _fjsp_sub_v2r8(ix1,jx3);
 236             dy13             = _fjsp_sub_v2r8(iy1,jy3);
 237             dz13             = _fjsp_sub_v2r8(iz1,jz3);
 238             dx21             = _fjsp_sub_v2r8(ix2,jx1);
 239             dy21             = _fjsp_sub_v2r8(iy2,jy1);
 240             dz21             = _fjsp_sub_v2r8(iz2,jz1);
 241             dx22             = _fjsp_sub_v2r8(ix2,jx2);
 242             dy22             = _fjsp_sub_v2r8(iy2,jy2);
 243             dz22             = _fjsp_sub_v2r8(iz2,jz2);
 244             dx23             = _fjsp_sub_v2r8(ix2,jx3);
 245             dy23             = _fjsp_sub_v2r8(iy2,jy3);
 246             dz23             = _fjsp_sub_v2r8(iz2,jz3);
 247             dx31             = _fjsp_sub_v2r8(ix3,jx1);
 248             dy31             = _fjsp_sub_v2r8(iy3,jy1);
 249             dz31             = _fjsp_sub_v2r8(iz3,jz1);
 250             dx32             = _fjsp_sub_v2r8(ix3,jx2);
 251             dy32             = _fjsp_sub_v2r8(iy3,jy2);
 252             dz32             = _fjsp_sub_v2r8(iz3,jz2);
 253             dx33             = _fjsp_sub_v2r8(ix3,jx3);
 254             dy33             = _fjsp_sub_v2r8(iy3,jy3);
 255             dz33             = _fjsp_sub_v2r8(iz3,jz3);
 256
 257             /* Calculate squared distance and things based on it */
 258             rsq00            = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
 259             rsq11            = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
 260             rsq12            = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
 261             rsq13            = gmx_fjsp_calc_rsq_v2r8(dx13,dy13,dz13);
 262             rsq21            = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
 263             rsq22            = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
 264             rsq23            = gmx_fjsp_calc_rsq_v2r8(dx23,dy23,dz23);
 265             rsq31            = gmx_fjsp_calc_rsq_v2r8(dx31,dy31,dz31);
 266             rsq32            = gmx_fjsp_calc_rsq_v2r8(dx32,dy32,dz32);
 267             rsq33            = gmx_fjsp_calc_rsq_v2r8(dx33,dy33,dz33);
 268
 269             rinv00           = gmx_fjsp_invsqrt_v2r8(rsq00);
 270             rinv11           = gmx_fjsp_invsqrt_v2r8(rsq11);
 271             rinv12           = gmx_fjsp_invsqrt_v2r8(rsq12);
 272             rinv13           = gmx_fjsp_invsqrt_v2r8(rsq13);
 273             rinv21           = gmx_fjsp_invsqrt_v2r8(rsq21);
 274             rinv22           = gmx_fjsp_invsqrt_v2r8(rsq22);
 275             rinv23           = gmx_fjsp_invsqrt_v2r8(rsq23);
 276             rinv31           = gmx_fjsp_invsqrt_v2r8(rsq31);
 277             rinv32           = gmx_fjsp_invsqrt_v2r8(rsq32);
 278             rinv33           = gmx_fjsp_invsqrt_v2r8(rsq33);
 279
 280             rinvsq11         = _fjsp_mul_v2r8(rinv11,rinv11);
 281             rinvsq12         = _fjsp_mul_v2r8(rinv12,rinv12);
 282             rinvsq13         = _fjsp_mul_v2r8(rinv13,rinv13);
 283             rinvsq21         = _fjsp_mul_v2r8(rinv21,rinv21);
 284             rinvsq22         = _fjsp_mul_v2r8(rinv22,rinv22);
 285             rinvsq23         = _fjsp_mul_v2r8(rinv23,rinv23);
 286             rinvsq31         = _fjsp_mul_v2r8(rinv31,rinv31);
 287             rinvsq32         = _fjsp_mul_v2r8(rinv32,rinv32);
 288             rinvsq33         = _fjsp_mul_v2r8(rinv33,rinv33);
 289
 290             fjx0             = _fjsp_setzero_v2r8();
 291             fjy0             = _fjsp_setzero_v2r8();
 292             fjz0             = _fjsp_setzero_v2r8();
 293             fjx1             = _fjsp_setzero_v2r8();
 294             fjy1             = _fjsp_setzero_v2r8();
 295             fjz1             = _fjsp_setzero_v2r8();
 296             fjx2             = _fjsp_setzero_v2r8();
 297             fjy2             = _fjsp_setzero_v2r8();
 298             fjz2             = _fjsp_setzero_v2r8();
 299             fjx3             = _fjsp_setzero_v2r8();
 300             fjy3             = _fjsp_setzero_v2r8();
 301             fjz3             = _fjsp_setzero_v2r8();
 302
 303             /**************************
 304              * CALCULATE INTERACTIONS *
 305              **************************/
 306
 307             r00              = _fjsp_mul_v2r8(rsq00,rinv00);
 308
 309             /* Calculate table index by multiplying r with table scale and truncate to integer */
 310             rt               = _fjsp_mul_v2r8(r00,vftabscale);
 311             itab_tmp         = _fjsp_dtox_v2r8(rt);
 312             vfeps            = _fjsp_sub_v2r8(rt, _fjsp_xtod_v2r8(itab_tmp));
 313             twovfeps         = _fjsp_add_v2r8(vfeps,vfeps);
 314             _fjsp_store_v2r8(&vfconv.simd,itab_tmp);
 315
 316             vfconv.i[0]     *= 8;
 317             vfconv.i[1]     *= 8;
 318
 319             /* CUBIC SPLINE TABLE DISPERSION */
 320             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
 321             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] );
 322             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 323             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 2 );
 324             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 2 );
 325             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 326             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 327             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 328             vvdw6            = _fjsp_mul_v2r8(c6_00,VV);
 329             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 330             fvdw6            = _fjsp_mul_v2r8(c6_00,FF);
 331
 332             /* CUBIC SPLINE TABLE REPULSION */
 333             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 4 );
 334             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 4 );
 335             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 336             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 6 );
 337             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 6 );
 338             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 339             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 340             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 341             vvdw12           = _fjsp_mul_v2r8(c12_00,VV);
 342             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 343             fvdw12           = _fjsp_mul_v2r8(c12_00,FF);
 344             vvdw             = _fjsp_add_v2r8(vvdw12,vvdw6);
 345             fvdw             = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_add_v2r8(fvdw6,fvdw12),_fjsp_mul_v2r8(vftabscale,rinv00)));
 346
 347             /* Update potential sum for this i atom from the interaction with this j atom. */
 348             vvdwsum          = _fjsp_add_v2r8(vvdwsum,vvdw);
 349
 350             fscal            = fvdw;
 351
 352             /* Update vectorial force */
 353             fix0             = _fjsp_madd_v2r8(dx00,fscal,fix0);
 354             fiy0             = _fjsp_madd_v2r8(dy00,fscal,fiy0);
 355             fiz0             = _fjsp_madd_v2r8(dz00,fscal,fiz0);
 356
 357             fjx0             = _fjsp_madd_v2r8(dx00,fscal,fjx0);
 358             fjy0             = _fjsp_madd_v2r8(dy00,fscal,fjy0);
 359             fjz0             = _fjsp_madd_v2r8(dz00,fscal,fjz0);
 360
 361             /**************************
 362              * CALCULATE INTERACTIONS *
 363              **************************/
 364
 365             /* REACTION-FIELD ELECTROSTATICS */
 366             velec            = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq11,rinv11),crf));
 367             felec            = _fjsp_mul_v2r8(qq11,_fjsp_msub_v2r8(rinv11,rinvsq11,krf2));
 368
 369             /* Update potential sum for this i atom from the interaction with this j atom. */
 370             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 371
 372             fscal            = felec;
 373
 374             /* Update vectorial force */
 375             fix1             = _fjsp_madd_v2r8(dx11,fscal,fix1);
 376             fiy1             = _fjsp_madd_v2r8(dy11,fscal,fiy1);
 377             fiz1             = _fjsp_madd_v2r8(dz11,fscal,fiz1);
 378
 379             fjx1             = _fjsp_madd_v2r8(dx11,fscal,fjx1);
 380             fjy1             = _fjsp_madd_v2r8(dy11,fscal,fjy1);
 381             fjz1             = _fjsp_madd_v2r8(dz11,fscal,fjz1);
 382
 383             /**************************
 384              * CALCULATE INTERACTIONS *
 385              **************************/
 386
 387             /* REACTION-FIELD ELECTROSTATICS */
 388             velec            = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq12,rinv12),crf));
 389             felec            = _fjsp_mul_v2r8(qq12,_fjsp_msub_v2r8(rinv12,rinvsq12,krf2));
 390
 391             /* Update potential sum for this i atom from the interaction with this j atom. */
 392             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 393
 394             fscal            = felec;
 395
 396             /* Update vectorial force */
 397             fix1             = _fjsp_madd_v2r8(dx12,fscal,fix1);
 398             fiy1             = _fjsp_madd_v2r8(dy12,fscal,fiy1);
 399             fiz1             = _fjsp_madd_v2r8(dz12,fscal,fiz1);
 400
 401             fjx2             = _fjsp_madd_v2r8(dx12,fscal,fjx2);
 402             fjy2             = _fjsp_madd_v2r8(dy12,fscal,fjy2);
 403             fjz2             = _fjsp_madd_v2r8(dz12,fscal,fjz2);
 404
 405             /**************************
 406              * CALCULATE INTERACTIONS *
 407              **************************/
 408
 409             /* REACTION-FIELD ELECTROSTATICS */
 410             velec            = _fjsp_mul_v2r8(qq13,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq13,rinv13),crf));
 411             felec            = _fjsp_mul_v2r8(qq13,_fjsp_msub_v2r8(rinv13,rinvsq13,krf2));
 412
 413             /* Update potential sum for this i atom from the interaction with this j atom. */
 414             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 415
 416             fscal            = felec;
 417
 418             /* Update vectorial force */
 419             fix1             = _fjsp_madd_v2r8(dx13,fscal,fix1);
 420             fiy1             = _fjsp_madd_v2r8(dy13,fscal,fiy1);
 421             fiz1             = _fjsp_madd_v2r8(dz13,fscal,fiz1);
 422
 423             fjx3             = _fjsp_madd_v2r8(dx13,fscal,fjx3);
 424             fjy3             = _fjsp_madd_v2r8(dy13,fscal,fjy3);
 425             fjz3             = _fjsp_madd_v2r8(dz13,fscal,fjz3);
 426
 427             /**************************
 428              * CALCULATE INTERACTIONS *
 429              **************************/
 430
 431             /* REACTION-FIELD ELECTROSTATICS */
 432             velec            = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq21,rinv21),crf));
 433             felec            = _fjsp_mul_v2r8(qq21,_fjsp_msub_v2r8(rinv21,rinvsq21,krf2));
 434
 435             /* Update potential sum for this i atom from the interaction with this j atom. */
 436             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 437
 438             fscal            = felec;
 439
 440             /* Update vectorial force */
 441             fix2             = _fjsp_madd_v2r8(dx21,fscal,fix2);
 442             fiy2             = _fjsp_madd_v2r8(dy21,fscal,fiy2);
 443             fiz2             = _fjsp_madd_v2r8(dz21,fscal,fiz2);
 444
 445             fjx1             = _fjsp_madd_v2r8(dx21,fscal,fjx1);
 446             fjy1             = _fjsp_madd_v2r8(dy21,fscal,fjy1);
 447             fjz1             = _fjsp_madd_v2r8(dz21,fscal,fjz1);
 448
 449             /**************************
 450              * CALCULATE INTERACTIONS *
 451              **************************/
 452
 453             /* REACTION-FIELD ELECTROSTATICS */
 454             velec            = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq22,rinv22),crf));
 455             felec            = _fjsp_mul_v2r8(qq22,_fjsp_msub_v2r8(rinv22,rinvsq22,krf2));
 456
 457             /* Update potential sum for this i atom from the interaction with this j atom. */
 458             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 459
 460             fscal            = felec;
 461
 462             /* Update vectorial force */
 463             fix2             = _fjsp_madd_v2r8(dx22,fscal,fix2);
 464             fiy2             = _fjsp_madd_v2r8(dy22,fscal,fiy2);
 465             fiz2             = _fjsp_madd_v2r8(dz22,fscal,fiz2);
 466
 467             fjx2             = _fjsp_madd_v2r8(dx22,fscal,fjx2);
 468             fjy2             = _fjsp_madd_v2r8(dy22,fscal,fjy2);
 469             fjz2             = _fjsp_madd_v2r8(dz22,fscal,fjz2);
 470
 471             /**************************
 472              * CALCULATE INTERACTIONS *
 473              **************************/
 474
 475             /* REACTION-FIELD ELECTROSTATICS */
 476             velec            = _fjsp_mul_v2r8(qq23,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq23,rinv23),crf));
 477             felec            = _fjsp_mul_v2r8(qq23,_fjsp_msub_v2r8(rinv23,rinvsq23,krf2));
 478
 479             /* Update potential sum for this i atom from the interaction with this j atom. */
 480             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 481
 482             fscal            = felec;
 483
 484             /* Update vectorial force */
 485             fix2             = _fjsp_madd_v2r8(dx23,fscal,fix2);
 486             fiy2             = _fjsp_madd_v2r8(dy23,fscal,fiy2);
 487             fiz2             = _fjsp_madd_v2r8(dz23,fscal,fiz2);
 488
 489             fjx3             = _fjsp_madd_v2r8(dx23,fscal,fjx3);
 490             fjy3             = _fjsp_madd_v2r8(dy23,fscal,fjy3);
 491             fjz3             = _fjsp_madd_v2r8(dz23,fscal,fjz3);
 492
 493             /**************************
 494              * CALCULATE INTERACTIONS *
 495              **************************/
 496
 497             /* REACTION-FIELD ELECTROSTATICS */
 498             velec            = _fjsp_mul_v2r8(qq31,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq31,rinv31),crf));
 499             felec            = _fjsp_mul_v2r8(qq31,_fjsp_msub_v2r8(rinv31,rinvsq31,krf2));
 500
 501             /* Update potential sum for this i atom from the interaction with this j atom. */
 502             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 503
 504             fscal            = felec;
 505
 506             /* Update vectorial force */
 507             fix3             = _fjsp_madd_v2r8(dx31,fscal,fix3);
 508             fiy3             = _fjsp_madd_v2r8(dy31,fscal,fiy3);
 509             fiz3             = _fjsp_madd_v2r8(dz31,fscal,fiz3);
 510
 511             fjx1             = _fjsp_madd_v2r8(dx31,fscal,fjx1);
 512             fjy1             = _fjsp_madd_v2r8(dy31,fscal,fjy1);
 513             fjz1             = _fjsp_madd_v2r8(dz31,fscal,fjz1);
 514
 515             /**************************
 516              * CALCULATE INTERACTIONS *
 517              **************************/
 518
 519             /* REACTION-FIELD ELECTROSTATICS */
 520             velec            = _fjsp_mul_v2r8(qq32,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq32,rinv32),crf));
 521             felec            = _fjsp_mul_v2r8(qq32,_fjsp_msub_v2r8(rinv32,rinvsq32,krf2));
 522
 523             /* Update potential sum for this i atom from the interaction with this j atom. */
 524             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 525
 526             fscal            = felec;
 527
 528             /* Update vectorial force */
 529             fix3             = _fjsp_madd_v2r8(dx32,fscal,fix3);
 530             fiy3             = _fjsp_madd_v2r8(dy32,fscal,fiy3);
 531             fiz3             = _fjsp_madd_v2r8(dz32,fscal,fiz3);
 532
 533             fjx2             = _fjsp_madd_v2r8(dx32,fscal,fjx2);
 534             fjy2             = _fjsp_madd_v2r8(dy32,fscal,fjy2);
 535             fjz2             = _fjsp_madd_v2r8(dz32,fscal,fjz2);
 536
 537             /**************************
 538              * CALCULATE INTERACTIONS *
 539              **************************/
 540
 541             /* REACTION-FIELD ELECTROSTATICS */
 542             velec            = _fjsp_mul_v2r8(qq33,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq33,rinv33),crf));
 543             felec            = _fjsp_mul_v2r8(qq33,_fjsp_msub_v2r8(rinv33,rinvsq33,krf2));
 544
 545             /* Update potential sum for this i atom from the interaction with this j atom. */
 546             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 547
 548             fscal            = felec;
 549
 550             /* Update vectorial force */
 551             fix3             = _fjsp_madd_v2r8(dx33,fscal,fix3);
 552             fiy3             = _fjsp_madd_v2r8(dy33,fscal,fiy3);
 553             fiz3             = _fjsp_madd_v2r8(dz33,fscal,fiz3);
 554
 555             fjx3             = _fjsp_madd_v2r8(dx33,fscal,fjx3);
 556             fjy3             = _fjsp_madd_v2r8(dy33,fscal,fjy3);
 557             fjz3             = _fjsp_madd_v2r8(dz33,fscal,fjz3);
 558
 559             gmx_fjsp_decrement_4rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 560
 561             /* Inner loop uses 377 flops */
 562         }
 563
 564         if(jidx<j_index_end)
 565         {
 566
 567             jnrA             = jjnr[jidx];
 568             j_coord_offsetA  = DIM*jnrA;
 569
 570             /* load j atom coordinates */
 571             gmx_fjsp_load_4rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
 572                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 573                                               &jy2,&jz2,&jx3,&jy3,&jz3);
 574
 575             /* Calculate displacement vector */
 576             dx00             = _fjsp_sub_v2r8(ix0,jx0);
 577             dy00             = _fjsp_sub_v2r8(iy0,jy0);
 578             dz00             = _fjsp_sub_v2r8(iz0,jz0);
 579             dx11             = _fjsp_sub_v2r8(ix1,jx1);
 580             dy11             = _fjsp_sub_v2r8(iy1,jy1);
 581             dz11             = _fjsp_sub_v2r8(iz1,jz1);
 582             dx12             = _fjsp_sub_v2r8(ix1,jx2);
 583             dy12             = _fjsp_sub_v2r8(iy1,jy2);
 584             dz12             = _fjsp_sub_v2r8(iz1,jz2);
 585             dx13             = _fjsp_sub_v2r8(ix1,jx3);
 586             dy13             = _fjsp_sub_v2r8(iy1,jy3);
 587             dz13             = _fjsp_sub_v2r8(iz1,jz3);
 588             dx21             = _fjsp_sub_v2r8(ix2,jx1);
 589             dy21             = _fjsp_sub_v2r8(iy2,jy1);
 590             dz21             = _fjsp_sub_v2r8(iz2,jz1);
 591             dx22             = _fjsp_sub_v2r8(ix2,jx2);
 592             dy22             = _fjsp_sub_v2r8(iy2,jy2);
 593             dz22             = _fjsp_sub_v2r8(iz2,jz2);
 594             dx23             = _fjsp_sub_v2r8(ix2,jx3);
 595             dy23             = _fjsp_sub_v2r8(iy2,jy3);
 596             dz23             = _fjsp_sub_v2r8(iz2,jz3);
 597             dx31             = _fjsp_sub_v2r8(ix3,jx1);
 598             dy31             = _fjsp_sub_v2r8(iy3,jy1);
 599             dz31             = _fjsp_sub_v2r8(iz3,jz1);
 600             dx32             = _fjsp_sub_v2r8(ix3,jx2);
 601             dy32             = _fjsp_sub_v2r8(iy3,jy2);
 602             dz32             = _fjsp_sub_v2r8(iz3,jz2);
 603             dx33             = _fjsp_sub_v2r8(ix3,jx3);
 604             dy33             = _fjsp_sub_v2r8(iy3,jy3);
 605             dz33             = _fjsp_sub_v2r8(iz3,jz3);
 606
 607             /* Calculate squared distance and things based on it */
 608             rsq00            = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
 609             rsq11            = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
 610             rsq12            = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
 611             rsq13            = gmx_fjsp_calc_rsq_v2r8(dx13,dy13,dz13);
 612             rsq21            = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
 613             rsq22            = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
 614             rsq23            = gmx_fjsp_calc_rsq_v2r8(dx23,dy23,dz23);
 615             rsq31            = gmx_fjsp_calc_rsq_v2r8(dx31,dy31,dz31);
 616             rsq32            = gmx_fjsp_calc_rsq_v2r8(dx32,dy32,dz32);
 617             rsq33            = gmx_fjsp_calc_rsq_v2r8(dx33,dy33,dz33);
 618
 619             rinv00           = gmx_fjsp_invsqrt_v2r8(rsq00);
 620             rinv11           = gmx_fjsp_invsqrt_v2r8(rsq11);
 621             rinv12           = gmx_fjsp_invsqrt_v2r8(rsq12);
 622             rinv13           = gmx_fjsp_invsqrt_v2r8(rsq13);
 623             rinv21           = gmx_fjsp_invsqrt_v2r8(rsq21);
 624             rinv22           = gmx_fjsp_invsqrt_v2r8(rsq22);
 625             rinv23           = gmx_fjsp_invsqrt_v2r8(rsq23);
 626             rinv31           = gmx_fjsp_invsqrt_v2r8(rsq31);
 627             rinv32           = gmx_fjsp_invsqrt_v2r8(rsq32);
 628             rinv33           = gmx_fjsp_invsqrt_v2r8(rsq33);
 629
 630             rinvsq11         = _fjsp_mul_v2r8(rinv11,rinv11);
 631             rinvsq12         = _fjsp_mul_v2r8(rinv12,rinv12);
 632             rinvsq13         = _fjsp_mul_v2r8(rinv13,rinv13);
 633             rinvsq21         = _fjsp_mul_v2r8(rinv21,rinv21);
 634             rinvsq22         = _fjsp_mul_v2r8(rinv22,rinv22);
 635             rinvsq23         = _fjsp_mul_v2r8(rinv23,rinv23);
 636             rinvsq31         = _fjsp_mul_v2r8(rinv31,rinv31);
 637             rinvsq32         = _fjsp_mul_v2r8(rinv32,rinv32);
 638             rinvsq33         = _fjsp_mul_v2r8(rinv33,rinv33);
 639
 640             fjx0             = _fjsp_setzero_v2r8();
 641             fjy0             = _fjsp_setzero_v2r8();
 642             fjz0             = _fjsp_setzero_v2r8();
 643             fjx1             = _fjsp_setzero_v2r8();
 644             fjy1             = _fjsp_setzero_v2r8();
 645             fjz1             = _fjsp_setzero_v2r8();
 646             fjx2             = _fjsp_setzero_v2r8();
 647             fjy2             = _fjsp_setzero_v2r8();
 648             fjz2             = _fjsp_setzero_v2r8();
 649             fjx3             = _fjsp_setzero_v2r8();
 650             fjy3             = _fjsp_setzero_v2r8();
 651             fjz3             = _fjsp_setzero_v2r8();
 652
 653             /**************************
 654              * CALCULATE INTERACTIONS *
 655              **************************/
 656
 657             r00              = _fjsp_mul_v2r8(rsq00,rinv00);
 658
 659             /* Calculate table index by multiplying r with table scale and truncate to integer */
 660             rt               = _fjsp_mul_v2r8(r00,vftabscale);
 661             itab_tmp         = _fjsp_dtox_v2r8(rt);
 662             vfeps            = _fjsp_sub_v2r8(rt, _fjsp_xtod_v2r8(itab_tmp));
 663             twovfeps         = _fjsp_add_v2r8(vfeps,vfeps);
 664             _fjsp_store_v2r8(&vfconv.simd,itab_tmp);
 665
 666             vfconv.i[0]     *= 8;
 667             vfconv.i[1]     *= 8;
 668
 669             /* CUBIC SPLINE TABLE DISPERSION */
 670             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
 671             F                = _fjsp_setzero_v2r8();
 672             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 673             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 2 );
 674             H                = _fjsp_setzero_v2r8();
 675             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 676             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 677             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 678             vvdw6            = _fjsp_mul_v2r8(c6_00,VV);
 679             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 680             fvdw6            = _fjsp_mul_v2r8(c6_00,FF);
 681
 682             /* CUBIC SPLINE TABLE REPULSION */
 683             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 4 );
 684             F                = _fjsp_setzero_v2r8();
 685             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 686             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 6 );
 687             H                = _fjsp_setzero_v2r8();
 688             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 689             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 690             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 691             vvdw12           = _fjsp_mul_v2r8(c12_00,VV);
 692             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 693             fvdw12           = _fjsp_mul_v2r8(c12_00,FF);
 694             vvdw             = _fjsp_add_v2r8(vvdw12,vvdw6);
 695             fvdw             = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_add_v2r8(fvdw6,fvdw12),_fjsp_mul_v2r8(vftabscale,rinv00)));
 696
 697             /* Update potential sum for this i atom from the interaction with this j atom. */
 698             vvdw             = _fjsp_unpacklo_v2r8(vvdw,_fjsp_setzero_v2r8());
 699             vvdwsum          = _fjsp_add_v2r8(vvdwsum,vvdw);
 700
 701             fscal            = fvdw;
 702
 703             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 704
 705             /* Update vectorial force */
 706             fix0             = _fjsp_madd_v2r8(dx00,fscal,fix0);
 707             fiy0             = _fjsp_madd_v2r8(dy00,fscal,fiy0);
 708             fiz0             = _fjsp_madd_v2r8(dz00,fscal,fiz0);
 709
 710             fjx0             = _fjsp_madd_v2r8(dx00,fscal,fjx0);
 711             fjy0             = _fjsp_madd_v2r8(dy00,fscal,fjy0);
 712             fjz0             = _fjsp_madd_v2r8(dz00,fscal,fjz0);
 713
 714             /**************************
 715              * CALCULATE INTERACTIONS *
 716              **************************/
 717
 718             /* REACTION-FIELD ELECTROSTATICS */
 719             velec            = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq11,rinv11),crf));
 720             felec            = _fjsp_mul_v2r8(qq11,_fjsp_msub_v2r8(rinv11,rinvsq11,krf2));
 721
 722             /* Update potential sum for this i atom from the interaction with this j atom. */
 723             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 724             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 725
 726             fscal            = felec;
 727
 728             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 729
 730             /* Update vectorial force */
 731             fix1             = _fjsp_madd_v2r8(dx11,fscal,fix1);
 732             fiy1             = _fjsp_madd_v2r8(dy11,fscal,fiy1);
 733             fiz1             = _fjsp_madd_v2r8(dz11,fscal,fiz1);
 734
 735             fjx1             = _fjsp_madd_v2r8(dx11,fscal,fjx1);
 736             fjy1             = _fjsp_madd_v2r8(dy11,fscal,fjy1);
 737             fjz1             = _fjsp_madd_v2r8(dz11,fscal,fjz1);
 738
 739             /**************************
 740              * CALCULATE INTERACTIONS *
 741              **************************/
 742
 743             /* REACTION-FIELD ELECTROSTATICS */
 744             velec            = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq12,rinv12),crf));
 745             felec            = _fjsp_mul_v2r8(qq12,_fjsp_msub_v2r8(rinv12,rinvsq12,krf2));
 746
 747             /* Update potential sum for this i atom from the interaction with this j atom. */
 748             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 749             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 750
 751             fscal            = felec;
 752
 753             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 754
 755             /* Update vectorial force */
 756             fix1             = _fjsp_madd_v2r8(dx12,fscal,fix1);
 757             fiy1             = _fjsp_madd_v2r8(dy12,fscal,fiy1);
 758             fiz1             = _fjsp_madd_v2r8(dz12,fscal,fiz1);
 759
 760             fjx2             = _fjsp_madd_v2r8(dx12,fscal,fjx2);
 761             fjy2             = _fjsp_madd_v2r8(dy12,fscal,fjy2);
 762             fjz2             = _fjsp_madd_v2r8(dz12,fscal,fjz2);
 763
 764             /**************************
 765              * CALCULATE INTERACTIONS *
 766              **************************/
 767
 768             /* REACTION-FIELD ELECTROSTATICS */
 769             velec            = _fjsp_mul_v2r8(qq13,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq13,rinv13),crf));
 770             felec            = _fjsp_mul_v2r8(qq13,_fjsp_msub_v2r8(rinv13,rinvsq13,krf2));
 771
 772             /* Update potential sum for this i atom from the interaction with this j atom. */
 773             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 774             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 775
 776             fscal            = felec;
 777
 778             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 779
 780             /* Update vectorial force */
 781             fix1             = _fjsp_madd_v2r8(dx13,fscal,fix1);
 782             fiy1             = _fjsp_madd_v2r8(dy13,fscal,fiy1);
 783             fiz1             = _fjsp_madd_v2r8(dz13,fscal,fiz1);
 784
 785             fjx3             = _fjsp_madd_v2r8(dx13,fscal,fjx3);
 786             fjy3             = _fjsp_madd_v2r8(dy13,fscal,fjy3);
 787             fjz3             = _fjsp_madd_v2r8(dz13,fscal,fjz3);
 788
 789             /**************************
 790              * CALCULATE INTERACTIONS *
 791              **************************/
 792
 793             /* REACTION-FIELD ELECTROSTATICS */
 794             velec            = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq21,rinv21),crf));
 795             felec            = _fjsp_mul_v2r8(qq21,_fjsp_msub_v2r8(rinv21,rinvsq21,krf2));
 796
 797             /* Update potential sum for this i atom from the interaction with this j atom. */
 798             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 799             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 800
 801             fscal            = felec;
 802
 803             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 804
 805             /* Update vectorial force */
 806             fix2             = _fjsp_madd_v2r8(dx21,fscal,fix2);
 807             fiy2             = _fjsp_madd_v2r8(dy21,fscal,fiy2);
 808             fiz2             = _fjsp_madd_v2r8(dz21,fscal,fiz2);
 809
 810             fjx1             = _fjsp_madd_v2r8(dx21,fscal,fjx1);
 811             fjy1             = _fjsp_madd_v2r8(dy21,fscal,fjy1);
 812             fjz1             = _fjsp_madd_v2r8(dz21,fscal,fjz1);
 813
 814             /**************************
 815              * CALCULATE INTERACTIONS *
 816              **************************/
 817
 818             /* REACTION-FIELD ELECTROSTATICS */
 819             velec            = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq22,rinv22),crf));
 820             felec            = _fjsp_mul_v2r8(qq22,_fjsp_msub_v2r8(rinv22,rinvsq22,krf2));
 821
 822             /* Update potential sum for this i atom from the interaction with this j atom. */
 823             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 824             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 825
 826             fscal            = felec;
 827
 828             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 829
 830             /* Update vectorial force */
 831             fix2             = _fjsp_madd_v2r8(dx22,fscal,fix2);
 832             fiy2             = _fjsp_madd_v2r8(dy22,fscal,fiy2);
 833             fiz2             = _fjsp_madd_v2r8(dz22,fscal,fiz2);
 834
 835             fjx2             = _fjsp_madd_v2r8(dx22,fscal,fjx2);
 836             fjy2             = _fjsp_madd_v2r8(dy22,fscal,fjy2);
 837             fjz2             = _fjsp_madd_v2r8(dz22,fscal,fjz2);
 838
 839             /**************************
 840              * CALCULATE INTERACTIONS *
 841              **************************/
 842
 843             /* REACTION-FIELD ELECTROSTATICS */
 844             velec            = _fjsp_mul_v2r8(qq23,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq23,rinv23),crf));
 845             felec            = _fjsp_mul_v2r8(qq23,_fjsp_msub_v2r8(rinv23,rinvsq23,krf2));
 846
 847             /* Update potential sum for this i atom from the interaction with this j atom. */
 848             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 849             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 850
 851             fscal            = felec;
 852
 853             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 854
 855             /* Update vectorial force */
 856             fix2             = _fjsp_madd_v2r8(dx23,fscal,fix2);
 857             fiy2             = _fjsp_madd_v2r8(dy23,fscal,fiy2);
 858             fiz2             = _fjsp_madd_v2r8(dz23,fscal,fiz2);
 859
 860             fjx3             = _fjsp_madd_v2r8(dx23,fscal,fjx3);
 861             fjy3             = _fjsp_madd_v2r8(dy23,fscal,fjy3);
 862             fjz3             = _fjsp_madd_v2r8(dz23,fscal,fjz3);
 863
 864             /**************************
 865              * CALCULATE INTERACTIONS *
 866              **************************/
 867
 868             /* REACTION-FIELD ELECTROSTATICS */
 869             velec            = _fjsp_mul_v2r8(qq31,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq31,rinv31),crf));
 870             felec            = _fjsp_mul_v2r8(qq31,_fjsp_msub_v2r8(rinv31,rinvsq31,krf2));
 871
 872             /* Update potential sum for this i atom from the interaction with this j atom. */
 873             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 874             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 875
 876             fscal            = felec;
 877
 878             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 879
 880             /* Update vectorial force */
 881             fix3             = _fjsp_madd_v2r8(dx31,fscal,fix3);
 882             fiy3             = _fjsp_madd_v2r8(dy31,fscal,fiy3);
 883             fiz3             = _fjsp_madd_v2r8(dz31,fscal,fiz3);
 884
 885             fjx1             = _fjsp_madd_v2r8(dx31,fscal,fjx1);
 886             fjy1             = _fjsp_madd_v2r8(dy31,fscal,fjy1);
 887             fjz1             = _fjsp_madd_v2r8(dz31,fscal,fjz1);
 888
 889             /**************************
 890              * CALCULATE INTERACTIONS *
 891              **************************/
 892
 893             /* REACTION-FIELD ELECTROSTATICS */
 894             velec            = _fjsp_mul_v2r8(qq32,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq32,rinv32),crf));
 895             felec            = _fjsp_mul_v2r8(qq32,_fjsp_msub_v2r8(rinv32,rinvsq32,krf2));
 896
 897             /* Update potential sum for this i atom from the interaction with this j atom. */
 898             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 899             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 900
 901             fscal            = felec;
 902
 903             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 904
 905             /* Update vectorial force */
 906             fix3             = _fjsp_madd_v2r8(dx32,fscal,fix3);
 907             fiy3             = _fjsp_madd_v2r8(dy32,fscal,fiy3);
 908             fiz3             = _fjsp_madd_v2r8(dz32,fscal,fiz3);
 909
 910             fjx2             = _fjsp_madd_v2r8(dx32,fscal,fjx2);
 911             fjy2             = _fjsp_madd_v2r8(dy32,fscal,fjy2);
 912             fjz2             = _fjsp_madd_v2r8(dz32,fscal,fjz2);
 913
 914             /**************************
 915              * CALCULATE INTERACTIONS *
 916              **************************/
 917
 918             /* REACTION-FIELD ELECTROSTATICS */
 919             velec            = _fjsp_mul_v2r8(qq33,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq33,rinv33),crf));
 920             felec            = _fjsp_mul_v2r8(qq33,_fjsp_msub_v2r8(rinv33,rinvsq33,krf2));
 921
 922             /* Update potential sum for this i atom from the interaction with this j atom. */
 923             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 924             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 925
 926             fscal            = felec;
 927
 928             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 929
 930             /* Update vectorial force */
 931             fix3             = _fjsp_madd_v2r8(dx33,fscal,fix3);
 932             fiy3             = _fjsp_madd_v2r8(dy33,fscal,fiy3);
 933             fiz3             = _fjsp_madd_v2r8(dz33,fscal,fiz3);
 934
 935             fjx3             = _fjsp_madd_v2r8(dx33,fscal,fjx3);
 936             fjy3             = _fjsp_madd_v2r8(dy33,fscal,fjy3);
 937             fjz3             = _fjsp_madd_v2r8(dz33,fscal,fjz3);
 938
 939             gmx_fjsp_decrement_4rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 940
 941             /* Inner loop uses 377 flops */
 942         }
 943
 944         /* End of innermost loop */
 945
 946         gmx_fjsp_update_iforce_4atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 947                                               f+i_coord_offset,fshift+i_shift_offset);
 948
 949         ggid                        = gid[iidx];
 950         /* Update potential energies */
 951         gmx_fjsp_update_1pot_v2r8(velecsum,kernel_data->energygrp_elec+ggid);
 952         gmx_fjsp_update_1pot_v2r8(vvdwsum,kernel_data->energygrp_vdw+ggid);
 953
 954         /* Increment number of inner iterations */
 955         inneriter                  += j_index_end - j_index_start;
 956
 957         /* Outer loop uses 26 flops */
 958     }
 959
 960     /* Increment number of outer iterations */
 961     outeriter        += nri;
 962
 963     /* Update outer/inner flops */
 964
 965     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*26 + inneriter*377);
 966 }
 967 /*
 968  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomW4W4_F_sparc64_hpc_ace_double
 969  * Electrostatics interaction: ReactionField
 970  * VdW interaction:            CubicSplineTable
 971  * Geometry:                   Water4-Water4
 972  * Calculate force/pot:        Force
 973  */
 974 void
 975 nb_kernel_ElecRF_VdwCSTab_GeomW4W4_F_sparc64_hpc_ace_double
 976                     (t_nblist                    * gmx_restrict       nlist,
 977                      rvec                        * gmx_restrict          xx,
 978                      rvec                        * gmx_restrict          ff,
 979                      t_forcerec                  * gmx_restrict          fr,
 980                      t_mdatoms                   * gmx_restrict     mdatoms,
 981                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 982                      t_nrnb                      * gmx_restrict        nrnb)
 983 {
 984     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 985      * just 0 for non-waters.
 986      * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
 987      * jnr indices corresponding to data put in the four positions in the SIMD register.
 988      */
 989     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 990     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 991     int              jnrA,jnrB;
 992     int              j_coord_offsetA,j_coord_offsetB;
 993     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 994     real             rcutoff_scalar;
 995     real             *shiftvec,*fshift,*x,*f;
 996     _fjsp_v2r8       tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 997     int              vdwioffset0;
 998     _fjsp_v2r8       ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 999     int              vdwioffset1;
1000     _fjsp_v2r8       ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1001     int              vdwioffset2;
1002     _fjsp_v2r8       ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1003     int              vdwioffset3;
1004     _fjsp_v2r8       ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1005     int              vdwjidx0A,vdwjidx0B;
1006     _fjsp_v2r8       jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1007     int              vdwjidx1A,vdwjidx1B;
1008     _fjsp_v2r8       jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1009     int              vdwjidx2A,vdwjidx2B;
1010     _fjsp_v2r8       jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1011     int              vdwjidx3A,vdwjidx3B;
1012     _fjsp_v2r8       jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1013     _fjsp_v2r8       dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1014     _fjsp_v2r8       dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1015     _fjsp_v2r8       dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1016     _fjsp_v2r8       dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1017     _fjsp_v2r8       dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1018     _fjsp_v2r8       dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1019     _fjsp_v2r8       dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1020     _fjsp_v2r8       dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1021     _fjsp_v2r8       dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1022     _fjsp_v2r8       dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1023     _fjsp_v2r8       velec,felec,velecsum,facel,crf,krf,krf2;
1024     real             *charge;
1025     int              nvdwtype;
1026     _fjsp_v2r8       rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1027     int              *vdwtype;
1028     real             *vdwparam;
1029     _fjsp_v2r8       one_sixth   = gmx_fjsp_set1_v2r8(1.0/6.0);
1030     _fjsp_v2r8       one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
1031     _fjsp_v2r8       rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
1032     real             *vftab;
1033     _fjsp_v2r8       itab_tmp;
1034     _fjsp_v2r8       dummy_mask,cutoff_mask;
1035     _fjsp_v2r8       one     = gmx_fjsp_set1_v2r8(1.0);
1036     _fjsp_v2r8       two     = gmx_fjsp_set1_v2r8(2.0);
1037     union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
1038
1039     x                = xx[0];
1040     f                = ff[0];
1041
1042     nri              = nlist->nri;
1043     iinr             = nlist->iinr;
1044     jindex           = nlist->jindex;
1045     jjnr             = nlist->jjnr;
1046     shiftidx         = nlist->shift;
1047     gid              = nlist->gid;
1048     shiftvec         = fr->shift_vec[0];
1049     fshift           = fr->fshift[0];
1050     facel            = gmx_fjsp_set1_v2r8(fr->epsfac);
1051     charge           = mdatoms->chargeA;
1052     krf              = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
1053     krf2             = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
1054     crf              = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
1055     nvdwtype         = fr->ntype;
1056     vdwparam         = fr->nbfp;
1057     vdwtype          = mdatoms->typeA;
1058
1059     vftab            = kernel_data->table_vdw->data;
1060     vftabscale       = gmx_fjsp_set1_v2r8(kernel_data->table_vdw->scale);
1061
1062     /* Setup water-specific parameters */
1063     inr              = nlist->iinr[0];
1064     iq1              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+1]));
1065     iq2              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+2]));
1066     iq3              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+3]));
1067     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
1068
1069     jq1              = gmx_fjsp_set1_v2r8(charge[inr+1]);
1070     jq2              = gmx_fjsp_set1_v2r8(charge[inr+2]);
1071     jq3              = gmx_fjsp_set1_v2r8(charge[inr+3]);
1072     vdwjidx0A        = 2*vdwtype[inr+0];
1073     c6_00            = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset0+vdwjidx0A]);
1074     c12_00           = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset0+vdwjidx0A+1]);
1075     qq11             = _fjsp_mul_v2r8(iq1,jq1);
1076     qq12             = _fjsp_mul_v2r8(iq1,jq2);
1077     qq13             = _fjsp_mul_v2r8(iq1,jq3);
1078     qq21             = _fjsp_mul_v2r8(iq2,jq1);
1079     qq22             = _fjsp_mul_v2r8(iq2,jq2);
1080     qq23             = _fjsp_mul_v2r8(iq2,jq3);
1081     qq31             = _fjsp_mul_v2r8(iq3,jq1);
1082     qq32             = _fjsp_mul_v2r8(iq3,jq2);
1083     qq33             = _fjsp_mul_v2r8(iq3,jq3);
1084
1085     /* Avoid stupid compiler warnings */
1086     jnrA = jnrB = 0;
1087     j_coord_offsetA = 0;
1088     j_coord_offsetB = 0;
1089
1090     outeriter        = 0;
1091     inneriter        = 0;
1092
1093     /* Start outer loop over neighborlists */
1094     for(iidx=0; iidx<nri; iidx++)
1095     {
1096         /* Load shift vector for this list */
1097         i_shift_offset   = DIM*shiftidx[iidx];
1098
1099         /* Load limits for loop over neighbors */
1100         j_index_start    = jindex[iidx];
1101         j_index_end      = jindex[iidx+1];
1102
1103         /* Get outer coordinate index */
1104         inr              = iinr[iidx];
1105         i_coord_offset   = DIM*inr;
1106
1107         /* Load i particle coords and add shift vector */
1108         gmx_fjsp_load_shift_and_4rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
1109                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1110
1111         fix0             = _fjsp_setzero_v2r8();
1112         fiy0             = _fjsp_setzero_v2r8();
1113         fiz0             = _fjsp_setzero_v2r8();
1114         fix1             = _fjsp_setzero_v2r8();
1115         fiy1             = _fjsp_setzero_v2r8();
1116         fiz1             = _fjsp_setzero_v2r8();
1117         fix2             = _fjsp_setzero_v2r8();
1118         fiy2             = _fjsp_setzero_v2r8();
1119         fiz2             = _fjsp_setzero_v2r8();
1120         fix3             = _fjsp_setzero_v2r8();
1121         fiy3             = _fjsp_setzero_v2r8();
1122         fiz3             = _fjsp_setzero_v2r8();
1123
1124         /* Start inner kernel loop */
1125         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
1126         {
1127
1128             /* Get j neighbor index, and coordinate index */
1129             jnrA             = jjnr[jidx];
1130             jnrB             = jjnr[jidx+1];
1131             j_coord_offsetA  = DIM*jnrA;
1132             j_coord_offsetB  = DIM*jnrB;
1133
1134             /* load j atom coordinates */
1135             gmx_fjsp_load_4rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
1136                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1137                                               &jy2,&jz2,&jx3,&jy3,&jz3);
1138
1139             /* Calculate displacement vector */
1140             dx00             = _fjsp_sub_v2r8(ix0,jx0);
1141             dy00             = _fjsp_sub_v2r8(iy0,jy0);
1142             dz00             = _fjsp_sub_v2r8(iz0,jz0);
1143             dx11             = _fjsp_sub_v2r8(ix1,jx1);
1144             dy11             = _fjsp_sub_v2r8(iy1,jy1);
1145             dz11             = _fjsp_sub_v2r8(iz1,jz1);
1146             dx12             = _fjsp_sub_v2r8(ix1,jx2);
1147             dy12             = _fjsp_sub_v2r8(iy1,jy2);
1148             dz12             = _fjsp_sub_v2r8(iz1,jz2);
1149             dx13             = _fjsp_sub_v2r8(ix1,jx3);
1150             dy13             = _fjsp_sub_v2r8(iy1,jy3);
1151             dz13             = _fjsp_sub_v2r8(iz1,jz3);
1152             dx21             = _fjsp_sub_v2r8(ix2,jx1);
1153             dy21             = _fjsp_sub_v2r8(iy2,jy1);
1154             dz21             = _fjsp_sub_v2r8(iz2,jz1);
1155             dx22             = _fjsp_sub_v2r8(ix2,jx2);
1156             dy22             = _fjsp_sub_v2r8(iy2,jy2);
1157             dz22             = _fjsp_sub_v2r8(iz2,jz2);
1158             dx23             = _fjsp_sub_v2r8(ix2,jx3);
1159             dy23             = _fjsp_sub_v2r8(iy2,jy3);
1160             dz23             = _fjsp_sub_v2r8(iz2,jz3);
1161             dx31             = _fjsp_sub_v2r8(ix3,jx1);
1162             dy31             = _fjsp_sub_v2r8(iy3,jy1);
1163             dz31             = _fjsp_sub_v2r8(iz3,jz1);
1164             dx32             = _fjsp_sub_v2r8(ix3,jx2);
1165             dy32             = _fjsp_sub_v2r8(iy3,jy2);
1166             dz32             = _fjsp_sub_v2r8(iz3,jz2);
1167             dx33             = _fjsp_sub_v2r8(ix3,jx3);
1168             dy33             = _fjsp_sub_v2r8(iy3,jy3);
1169             dz33             = _fjsp_sub_v2r8(iz3,jz3);
1170
1171             /* Calculate squared distance and things based on it */
1172             rsq00            = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
1173             rsq11            = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
1174             rsq12            = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
1175             rsq13            = gmx_fjsp_calc_rsq_v2r8(dx13,dy13,dz13);
1176             rsq21            = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
1177             rsq22            = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
1178             rsq23            = gmx_fjsp_calc_rsq_v2r8(dx23,dy23,dz23);
1179             rsq31            = gmx_fjsp_calc_rsq_v2r8(dx31,dy31,dz31);
1180             rsq32            = gmx_fjsp_calc_rsq_v2r8(dx32,dy32,dz32);
1181             rsq33            = gmx_fjsp_calc_rsq_v2r8(dx33,dy33,dz33);
1182
1183             rinv00           = gmx_fjsp_invsqrt_v2r8(rsq00);
1184             rinv11           = gmx_fjsp_invsqrt_v2r8(rsq11);
1185             rinv12           = gmx_fjsp_invsqrt_v2r8(rsq12);
1186             rinv13           = gmx_fjsp_invsqrt_v2r8(rsq13);
1187             rinv21           = gmx_fjsp_invsqrt_v2r8(rsq21);
1188             rinv22           = gmx_fjsp_invsqrt_v2r8(rsq22);
1189             rinv23           = gmx_fjsp_invsqrt_v2r8(rsq23);
1190             rinv31           = gmx_fjsp_invsqrt_v2r8(rsq31);
1191             rinv32           = gmx_fjsp_invsqrt_v2r8(rsq32);
1192             rinv33           = gmx_fjsp_invsqrt_v2r8(rsq33);
1193
1194             rinvsq11         = _fjsp_mul_v2r8(rinv11,rinv11);
1195             rinvsq12         = _fjsp_mul_v2r8(rinv12,rinv12);
1196             rinvsq13         = _fjsp_mul_v2r8(rinv13,rinv13);
1197             rinvsq21         = _fjsp_mul_v2r8(rinv21,rinv21);
1198             rinvsq22         = _fjsp_mul_v2r8(rinv22,rinv22);
1199             rinvsq23         = _fjsp_mul_v2r8(rinv23,rinv23);
1200             rinvsq31         = _fjsp_mul_v2r8(rinv31,rinv31);
1201             rinvsq32         = _fjsp_mul_v2r8(rinv32,rinv32);
1202             rinvsq33         = _fjsp_mul_v2r8(rinv33,rinv33);
1203
1204             fjx0             = _fjsp_setzero_v2r8();
1205             fjy0             = _fjsp_setzero_v2r8();
1206             fjz0             = _fjsp_setzero_v2r8();
1207             fjx1             = _fjsp_setzero_v2r8();
1208             fjy1             = _fjsp_setzero_v2r8();
1209             fjz1             = _fjsp_setzero_v2r8();
1210             fjx2             = _fjsp_setzero_v2r8();
1211             fjy2             = _fjsp_setzero_v2r8();
1212             fjz2             = _fjsp_setzero_v2r8();
1213             fjx3             = _fjsp_setzero_v2r8();
1214             fjy3             = _fjsp_setzero_v2r8();
1215             fjz3             = _fjsp_setzero_v2r8();
1216
1217             /**************************
1218              * CALCULATE INTERACTIONS *
1219              **************************/
1220
1221             r00              = _fjsp_mul_v2r8(rsq00,rinv00);
1222
1223             /* Calculate table index by multiplying r with table scale and truncate to integer */
1224             rt               = _fjsp_mul_v2r8(r00,vftabscale);
1225             itab_tmp         = _fjsp_dtox_v2r8(rt);
1226             vfeps            = _fjsp_sub_v2r8(rt, _fjsp_xtod_v2r8(itab_tmp));
1227             twovfeps         = _fjsp_add_v2r8(vfeps,vfeps);
1228             _fjsp_store_v2r8(&vfconv.simd,itab_tmp);
1229
1230             vfconv.i[0]     *= 8;
1231             vfconv.i[1]     *= 8;
1232
1233             /* CUBIC SPLINE TABLE DISPERSION */
1234             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
1235             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] );
1236             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
1237             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 2 );
1238             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 2 );
1239             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
1240             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
1241             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
1242             fvdw6            = _fjsp_mul_v2r8(c6_00,FF);
1243
1244             /* CUBIC SPLINE TABLE REPULSION */
1245             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 4 );
1246             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 4 );
1247             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
1248             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 6 );
1249             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 6 );
1250             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
1251             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
1252             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
1253             fvdw12           = _fjsp_mul_v2r8(c12_00,FF);
1254             fvdw             = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_add_v2r8(fvdw6,fvdw12),_fjsp_mul_v2r8(vftabscale,rinv00)));
1255
1256             fscal            = fvdw;
1257
1258             /* Update vectorial force */
1259             fix0             = _fjsp_madd_v2r8(dx00,fscal,fix0);
1260             fiy0             = _fjsp_madd_v2r8(dy00,fscal,fiy0);
1261             fiz0             = _fjsp_madd_v2r8(dz00,fscal,fiz0);
1262
1263             fjx0             = _fjsp_madd_v2r8(dx00,fscal,fjx0);
1264             fjy0             = _fjsp_madd_v2r8(dy00,fscal,fjy0);
1265             fjz0             = _fjsp_madd_v2r8(dz00,fscal,fjz0);
1266
1267             /**************************
1268              * CALCULATE INTERACTIONS *
1269              **************************/
1270
1271             /* REACTION-FIELD ELECTROSTATICS */
1272             felec            = _fjsp_mul_v2r8(qq11,_fjsp_msub_v2r8(rinv11,rinvsq11,krf2));
1273
1274             fscal            = felec;
1275
1276             /* Update vectorial force */
1277             fix1             = _fjsp_madd_v2r8(dx11,fscal,fix1);
1278             fiy1             = _fjsp_madd_v2r8(dy11,fscal,fiy1);
1279             fiz1             = _fjsp_madd_v2r8(dz11,fscal,fiz1);
1280
1281             fjx1             = _fjsp_madd_v2r8(dx11,fscal,fjx1);
1282             fjy1             = _fjsp_madd_v2r8(dy11,fscal,fjy1);
1283             fjz1             = _fjsp_madd_v2r8(dz11,fscal,fjz1);
1284
1285             /**************************
1286              * CALCULATE INTERACTIONS *
1287              **************************/
1288
1289             /* REACTION-FIELD ELECTROSTATICS */
1290             felec            = _fjsp_mul_v2r8(qq12,_fjsp_msub_v2r8(rinv12,rinvsq12,krf2));
1291
1292             fscal            = felec;
1293
1294             /* Update vectorial force */
1295             fix1             = _fjsp_madd_v2r8(dx12,fscal,fix1);
1296             fiy1             = _fjsp_madd_v2r8(dy12,fscal,fiy1);
1297             fiz1             = _fjsp_madd_v2r8(dz12,fscal,fiz1);
1298
1299             fjx2             = _fjsp_madd_v2r8(dx12,fscal,fjx2);
1300             fjy2             = _fjsp_madd_v2r8(dy12,fscal,fjy2);
1301             fjz2             = _fjsp_madd_v2r8(dz12,fscal,fjz2);
1302
1303             /**************************
1304              * CALCULATE INTERACTIONS *
1305              **************************/
1306
1307             /* REACTION-FIELD ELECTROSTATICS */
1308             felec            = _fjsp_mul_v2r8(qq13,_fjsp_msub_v2r8(rinv13,rinvsq13,krf2));
1309
1310             fscal            = felec;
1311
1312             /* Update vectorial force */
1313             fix1             = _fjsp_madd_v2r8(dx13,fscal,fix1);
1314             fiy1             = _fjsp_madd_v2r8(dy13,fscal,fiy1);
1315             fiz1             = _fjsp_madd_v2r8(dz13,fscal,fiz1);
1316
1317             fjx3             = _fjsp_madd_v2r8(dx13,fscal,fjx3);
1318             fjy3             = _fjsp_madd_v2r8(dy13,fscal,fjy3);
1319             fjz3             = _fjsp_madd_v2r8(dz13,fscal,fjz3);
1320
1321             /**************************
1322              * CALCULATE INTERACTIONS *
1323              **************************/
1324
1325             /* REACTION-FIELD ELECTROSTATICS */
1326             felec            = _fjsp_mul_v2r8(qq21,_fjsp_msub_v2r8(rinv21,rinvsq21,krf2));
1327
1328             fscal            = felec;
1329
1330             /* Update vectorial force */
1331             fix2             = _fjsp_madd_v2r8(dx21,fscal,fix2);
1332             fiy2             = _fjsp_madd_v2r8(dy21,fscal,fiy2);
1333             fiz2             = _fjsp_madd_v2r8(dz21,fscal,fiz2);
1334
1335             fjx1             = _fjsp_madd_v2r8(dx21,fscal,fjx1);
1336             fjy1             = _fjsp_madd_v2r8(dy21,fscal,fjy1);
1337             fjz1             = _fjsp_madd_v2r8(dz21,fscal,fjz1);
1338
1339             /**************************
1340              * CALCULATE INTERACTIONS *
1341              **************************/
1342
1343             /* REACTION-FIELD ELECTROSTATICS */
1344             felec            = _fjsp_mul_v2r8(qq22,_fjsp_msub_v2r8(rinv22,rinvsq22,krf2));
1345
1346             fscal            = felec;
1347
1348             /* Update vectorial force */
1349             fix2             = _fjsp_madd_v2r8(dx22,fscal,fix2);
1350             fiy2             = _fjsp_madd_v2r8(dy22,fscal,fiy2);
1351             fiz2             = _fjsp_madd_v2r8(dz22,fscal,fiz2);
1352
1353             fjx2             = _fjsp_madd_v2r8(dx22,fscal,fjx2);
1354             fjy2             = _fjsp_madd_v2r8(dy22,fscal,fjy2);
1355             fjz2             = _fjsp_madd_v2r8(dz22,fscal,fjz2);
1356
1357             /**************************
1358              * CALCULATE INTERACTIONS *
1359              **************************/
1360
1361             /* REACTION-FIELD ELECTROSTATICS */
1362             felec            = _fjsp_mul_v2r8(qq23,_fjsp_msub_v2r8(rinv23,rinvsq23,krf2));
1363
1364             fscal            = felec;
1365
1366             /* Update vectorial force */
1367             fix2             = _fjsp_madd_v2r8(dx23,fscal,fix2);
1368             fiy2             = _fjsp_madd_v2r8(dy23,fscal,fiy2);
1369             fiz2             = _fjsp_madd_v2r8(dz23,fscal,fiz2);
1370
1371             fjx3             = _fjsp_madd_v2r8(dx23,fscal,fjx3);
1372             fjy3             = _fjsp_madd_v2r8(dy23,fscal,fjy3);
1373             fjz3             = _fjsp_madd_v2r8(dz23,fscal,fjz3);
1374
1375             /**************************
1376              * CALCULATE INTERACTIONS *
1377              **************************/
1378
1379             /* REACTION-FIELD ELECTROSTATICS */
1380             felec            = _fjsp_mul_v2r8(qq31,_fjsp_msub_v2r8(rinv31,rinvsq31,krf2));
1381
1382             fscal            = felec;
1383
1384             /* Update vectorial force */
1385             fix3             = _fjsp_madd_v2r8(dx31,fscal,fix3);
1386             fiy3             = _fjsp_madd_v2r8(dy31,fscal,fiy3);
1387             fiz3             = _fjsp_madd_v2r8(dz31,fscal,fiz3);
1388
1389             fjx1             = _fjsp_madd_v2r8(dx31,fscal,fjx1);
1390             fjy1             = _fjsp_madd_v2r8(dy31,fscal,fjy1);
1391             fjz1             = _fjsp_madd_v2r8(dz31,fscal,fjz1);
1392
1393             /**************************
1394              * CALCULATE INTERACTIONS *
1395              **************************/
1396
1397             /* REACTION-FIELD ELECTROSTATICS */
1398             felec            = _fjsp_mul_v2r8(qq32,_fjsp_msub_v2r8(rinv32,rinvsq32,krf2));
1399
1400             fscal            = felec;
1401
1402             /* Update vectorial force */
1403             fix3             = _fjsp_madd_v2r8(dx32,fscal,fix3);
1404             fiy3             = _fjsp_madd_v2r8(dy32,fscal,fiy3);
1405             fiz3             = _fjsp_madd_v2r8(dz32,fscal,fiz3);
1406
1407             fjx2             = _fjsp_madd_v2r8(dx32,fscal,fjx2);
1408             fjy2             = _fjsp_madd_v2r8(dy32,fscal,fjy2);
1409             fjz2             = _fjsp_madd_v2r8(dz32,fscal,fjz2);
1410
1411             /**************************
1412              * CALCULATE INTERACTIONS *
1413              **************************/
1414
1415             /* REACTION-FIELD ELECTROSTATICS */
1416             felec            = _fjsp_mul_v2r8(qq33,_fjsp_msub_v2r8(rinv33,rinvsq33,krf2));
1417
1418             fscal            = felec;
1419
1420             /* Update vectorial force */
1421             fix3             = _fjsp_madd_v2r8(dx33,fscal,fix3);
1422             fiy3             = _fjsp_madd_v2r8(dy33,fscal,fiy3);
1423             fiz3             = _fjsp_madd_v2r8(dz33,fscal,fiz3);
1424
1425             fjx3             = _fjsp_madd_v2r8(dx33,fscal,fjx3);
1426             fjy3             = _fjsp_madd_v2r8(dy33,fscal,fjy3);
1427             fjz3             = _fjsp_madd_v2r8(dz33,fscal,fjz3);
1428
1429             gmx_fjsp_decrement_4rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1430
1431             /* Inner loop uses 324 flops */
1432         }
1433
1434         if(jidx<j_index_end)
1435         {
1436
1437             jnrA             = jjnr[jidx];
1438             j_coord_offsetA  = DIM*jnrA;
1439
1440             /* load j atom coordinates */
1441             gmx_fjsp_load_4rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
1442                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1443                                               &jy2,&jz2,&jx3,&jy3,&jz3);
1444
1445             /* Calculate displacement vector */
1446             dx00             = _fjsp_sub_v2r8(ix0,jx0);
1447             dy00             = _fjsp_sub_v2r8(iy0,jy0);
1448             dz00             = _fjsp_sub_v2r8(iz0,jz0);
1449             dx11             = _fjsp_sub_v2r8(ix1,jx1);
1450             dy11             = _fjsp_sub_v2r8(iy1,jy1);
1451             dz11             = _fjsp_sub_v2r8(iz1,jz1);
1452             dx12             = _fjsp_sub_v2r8(ix1,jx2);
1453             dy12             = _fjsp_sub_v2r8(iy1,jy2);
1454             dz12             = _fjsp_sub_v2r8(iz1,jz2);
1455             dx13             = _fjsp_sub_v2r8(ix1,jx3);
1456             dy13             = _fjsp_sub_v2r8(iy1,jy3);
1457             dz13             = _fjsp_sub_v2r8(iz1,jz3);
1458             dx21             = _fjsp_sub_v2r8(ix2,jx1);
1459             dy21             = _fjsp_sub_v2r8(iy2,jy1);
1460             dz21             = _fjsp_sub_v2r8(iz2,jz1);
1461             dx22             = _fjsp_sub_v2r8(ix2,jx2);
1462             dy22             = _fjsp_sub_v2r8(iy2,jy2);
1463             dz22             = _fjsp_sub_v2r8(iz2,jz2);
1464             dx23             = _fjsp_sub_v2r8(ix2,jx3);
1465             dy23             = _fjsp_sub_v2r8(iy2,jy3);
1466             dz23             = _fjsp_sub_v2r8(iz2,jz3);
1467             dx31             = _fjsp_sub_v2r8(ix3,jx1);
1468             dy31             = _fjsp_sub_v2r8(iy3,jy1);
1469             dz31             = _fjsp_sub_v2r8(iz3,jz1);
1470             dx32             = _fjsp_sub_v2r8(ix3,jx2);
1471             dy32             = _fjsp_sub_v2r8(iy3,jy2);
1472             dz32             = _fjsp_sub_v2r8(iz3,jz2);
1473             dx33             = _fjsp_sub_v2r8(ix3,jx3);
1474             dy33             = _fjsp_sub_v2r8(iy3,jy3);
1475             dz33             = _fjsp_sub_v2r8(iz3,jz3);
1476
1477             /* Calculate squared distance and things based on it */
1478             rsq00            = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
1479             rsq11            = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
1480             rsq12            = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
1481             rsq13            = gmx_fjsp_calc_rsq_v2r8(dx13,dy13,dz13);
1482             rsq21            = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
1483             rsq22            = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
1484             rsq23            = gmx_fjsp_calc_rsq_v2r8(dx23,dy23,dz23);
1485             rsq31            = gmx_fjsp_calc_rsq_v2r8(dx31,dy31,dz31);
1486             rsq32            = gmx_fjsp_calc_rsq_v2r8(dx32,dy32,dz32);
1487             rsq33            = gmx_fjsp_calc_rsq_v2r8(dx33,dy33,dz33);
1488
1489             rinv00           = gmx_fjsp_invsqrt_v2r8(rsq00);
1490             rinv11           = gmx_fjsp_invsqrt_v2r8(rsq11);
1491             rinv12           = gmx_fjsp_invsqrt_v2r8(rsq12);
1492             rinv13           = gmx_fjsp_invsqrt_v2r8(rsq13);
1493             rinv21           = gmx_fjsp_invsqrt_v2r8(rsq21);
1494             rinv22           = gmx_fjsp_invsqrt_v2r8(rsq22);
1495             rinv23           = gmx_fjsp_invsqrt_v2r8(rsq23);
1496             rinv31           = gmx_fjsp_invsqrt_v2r8(rsq31);
1497             rinv32           = gmx_fjsp_invsqrt_v2r8(rsq32);
1498             rinv33           = gmx_fjsp_invsqrt_v2r8(rsq33);
1499
1500             rinvsq11         = _fjsp_mul_v2r8(rinv11,rinv11);
1501             rinvsq12         = _fjsp_mul_v2r8(rinv12,rinv12);
1502             rinvsq13         = _fjsp_mul_v2r8(rinv13,rinv13);
1503             rinvsq21         = _fjsp_mul_v2r8(rinv21,rinv21);
1504             rinvsq22         = _fjsp_mul_v2r8(rinv22,rinv22);
1505             rinvsq23         = _fjsp_mul_v2r8(rinv23,rinv23);
1506             rinvsq31         = _fjsp_mul_v2r8(rinv31,rinv31);
1507             rinvsq32         = _fjsp_mul_v2r8(rinv32,rinv32);
1508             rinvsq33         = _fjsp_mul_v2r8(rinv33,rinv33);
1509
1510             fjx0             = _fjsp_setzero_v2r8();
1511             fjy0             = _fjsp_setzero_v2r8();
1512             fjz0             = _fjsp_setzero_v2r8();
1513             fjx1             = _fjsp_setzero_v2r8();
1514             fjy1             = _fjsp_setzero_v2r8();
1515             fjz1             = _fjsp_setzero_v2r8();
1516             fjx2             = _fjsp_setzero_v2r8();
1517             fjy2             = _fjsp_setzero_v2r8();
1518             fjz2             = _fjsp_setzero_v2r8();
1519             fjx3             = _fjsp_setzero_v2r8();
1520             fjy3             = _fjsp_setzero_v2r8();
1521             fjz3             = _fjsp_setzero_v2r8();
1522
1523             /**************************
1524              * CALCULATE INTERACTIONS *
1525              **************************/
1526
1527             r00              = _fjsp_mul_v2r8(rsq00,rinv00);
1528
1529             /* Calculate table index by multiplying r with table scale and truncate to integer */
1530             rt               = _fjsp_mul_v2r8(r00,vftabscale);
1531             itab_tmp         = _fjsp_dtox_v2r8(rt);
1532             vfeps            = _fjsp_sub_v2r8(rt, _fjsp_xtod_v2r8(itab_tmp));
1533             twovfeps         = _fjsp_add_v2r8(vfeps,vfeps);
1534             _fjsp_store_v2r8(&vfconv.simd,itab_tmp);
1535
1536             vfconv.i[0]     *= 8;
1537             vfconv.i[1]     *= 8;
1538
1539             /* CUBIC SPLINE TABLE DISPERSION */
1540             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
1541             F                = _fjsp_setzero_v2r8();
1542             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
1543             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 2 );
1544             H                = _fjsp_setzero_v2r8();
1545             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
1546             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
1547             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
1548             fvdw6            = _fjsp_mul_v2r8(c6_00,FF);
1549
1550             /* CUBIC SPLINE TABLE REPULSION */
1551             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 4 );
1552             F                = _fjsp_setzero_v2r8();
1553             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
1554             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 6 );
1555             H                = _fjsp_setzero_v2r8();
1556             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
1557             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
1558             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
1559             fvdw12           = _fjsp_mul_v2r8(c12_00,FF);
1560             fvdw             = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_add_v2r8(fvdw6,fvdw12),_fjsp_mul_v2r8(vftabscale,rinv00)));
1561
1562             fscal            = fvdw;
1563
1564             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1565
1566             /* Update vectorial force */
1567             fix0             = _fjsp_madd_v2r8(dx00,fscal,fix0);
1568             fiy0             = _fjsp_madd_v2r8(dy00,fscal,fiy0);
1569             fiz0             = _fjsp_madd_v2r8(dz00,fscal,fiz0);
1570
1571             fjx0             = _fjsp_madd_v2r8(dx00,fscal,fjx0);
1572             fjy0             = _fjsp_madd_v2r8(dy00,fscal,fjy0);
1573             fjz0             = _fjsp_madd_v2r8(dz00,fscal,fjz0);
1574
1575             /**************************
1576              * CALCULATE INTERACTIONS *
1577              **************************/
1578
1579             /* REACTION-FIELD ELECTROSTATICS */
1580             felec            = _fjsp_mul_v2r8(qq11,_fjsp_msub_v2r8(rinv11,rinvsq11,krf2));
1581
1582             fscal            = felec;
1583
1584             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1585
1586             /* Update vectorial force */
1587             fix1             = _fjsp_madd_v2r8(dx11,fscal,fix1);
1588             fiy1             = _fjsp_madd_v2r8(dy11,fscal,fiy1);
1589             fiz1             = _fjsp_madd_v2r8(dz11,fscal,fiz1);
1590
1591             fjx1             = _fjsp_madd_v2r8(dx11,fscal,fjx1);
1592             fjy1             = _fjsp_madd_v2r8(dy11,fscal,fjy1);
1593             fjz1             = _fjsp_madd_v2r8(dz11,fscal,fjz1);
1594
1595             /**************************
1596              * CALCULATE INTERACTIONS *
1597              **************************/
1598
1599             /* REACTION-FIELD ELECTROSTATICS */
1600             felec            = _fjsp_mul_v2r8(qq12,_fjsp_msub_v2r8(rinv12,rinvsq12,krf2));
1601
1602             fscal            = felec;
1603
1604             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1605
1606             /* Update vectorial force */
1607             fix1             = _fjsp_madd_v2r8(dx12,fscal,fix1);
1608             fiy1             = _fjsp_madd_v2r8(dy12,fscal,fiy1);
1609             fiz1             = _fjsp_madd_v2r8(dz12,fscal,fiz1);
1610
1611             fjx2             = _fjsp_madd_v2r8(dx12,fscal,fjx2);
1612             fjy2             = _fjsp_madd_v2r8(dy12,fscal,fjy2);
1613             fjz2             = _fjsp_madd_v2r8(dz12,fscal,fjz2);
1614
1615             /**************************
1616              * CALCULATE INTERACTIONS *
1617              **************************/
1618
1619             /* REACTION-FIELD ELECTROSTATICS */
1620             felec            = _fjsp_mul_v2r8(qq13,_fjsp_msub_v2r8(rinv13,rinvsq13,krf2));
1621
1622             fscal            = felec;
1623
1624             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1625
1626             /* Update vectorial force */
1627             fix1             = _fjsp_madd_v2r8(dx13,fscal,fix1);
1628             fiy1             = _fjsp_madd_v2r8(dy13,fscal,fiy1);
1629             fiz1             = _fjsp_madd_v2r8(dz13,fscal,fiz1);
1630
1631             fjx3             = _fjsp_madd_v2r8(dx13,fscal,fjx3);
1632             fjy3             = _fjsp_madd_v2r8(dy13,fscal,fjy3);
1633             fjz3             = _fjsp_madd_v2r8(dz13,fscal,fjz3);
1634
1635             /**************************
1636              * CALCULATE INTERACTIONS *
1637              **************************/
1638
1639             /* REACTION-FIELD ELECTROSTATICS */
1640             felec            = _fjsp_mul_v2r8(qq21,_fjsp_msub_v2r8(rinv21,rinvsq21,krf2));
1641
1642             fscal            = felec;
1643
1644             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1645
1646             /* Update vectorial force */
1647             fix2             = _fjsp_madd_v2r8(dx21,fscal,fix2);
1648             fiy2             = _fjsp_madd_v2r8(dy21,fscal,fiy2);
1649             fiz2             = _fjsp_madd_v2r8(dz21,fscal,fiz2);
1650
1651             fjx1             = _fjsp_madd_v2r8(dx21,fscal,fjx1);
1652             fjy1             = _fjsp_madd_v2r8(dy21,fscal,fjy1);
1653             fjz1             = _fjsp_madd_v2r8(dz21,fscal,fjz1);
1654
1655             /**************************
1656              * CALCULATE INTERACTIONS *
1657              **************************/
1658
1659             /* REACTION-FIELD ELECTROSTATICS */
1660             felec            = _fjsp_mul_v2r8(qq22,_fjsp_msub_v2r8(rinv22,rinvsq22,krf2));
1661
1662             fscal            = felec;
1663
1664             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1665
1666             /* Update vectorial force */
1667             fix2             = _fjsp_madd_v2r8(dx22,fscal,fix2);
1668             fiy2             = _fjsp_madd_v2r8(dy22,fscal,fiy2);
1669             fiz2             = _fjsp_madd_v2r8(dz22,fscal,fiz2);
1670
1671             fjx2             = _fjsp_madd_v2r8(dx22,fscal,fjx2);
1672             fjy2             = _fjsp_madd_v2r8(dy22,fscal,fjy2);
1673             fjz2             = _fjsp_madd_v2r8(dz22,fscal,fjz2);
1674
1675             /**************************
1676              * CALCULATE INTERACTIONS *
1677              **************************/
1678
1679             /* REACTION-FIELD ELECTROSTATICS */
1680             felec            = _fjsp_mul_v2r8(qq23,_fjsp_msub_v2r8(rinv23,rinvsq23,krf2));
1681
1682             fscal            = felec;
1683
1684             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1685
1686             /* Update vectorial force */
1687             fix2             = _fjsp_madd_v2r8(dx23,fscal,fix2);
1688             fiy2             = _fjsp_madd_v2r8(dy23,fscal,fiy2);
1689             fiz2             = _fjsp_madd_v2r8(dz23,fscal,fiz2);
1690
1691             fjx3             = _fjsp_madd_v2r8(dx23,fscal,fjx3);
1692             fjy3             = _fjsp_madd_v2r8(dy23,fscal,fjy3);
1693             fjz3             = _fjsp_madd_v2r8(dz23,fscal,fjz3);
1694
1695             /**************************
1696              * CALCULATE INTERACTIONS *
1697              **************************/
1698
1699             /* REACTION-FIELD ELECTROSTATICS */
1700             felec            = _fjsp_mul_v2r8(qq31,_fjsp_msub_v2r8(rinv31,rinvsq31,krf2));
1701
1702             fscal            = felec;
1703
1704             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1705
1706             /* Update vectorial force */
1707             fix3             = _fjsp_madd_v2r8(dx31,fscal,fix3);
1708             fiy3             = _fjsp_madd_v2r8(dy31,fscal,fiy3);
1709             fiz3             = _fjsp_madd_v2r8(dz31,fscal,fiz3);
1710
1711             fjx1             = _fjsp_madd_v2r8(dx31,fscal,fjx1);
1712             fjy1             = _fjsp_madd_v2r8(dy31,fscal,fjy1);
1713             fjz1             = _fjsp_madd_v2r8(dz31,fscal,fjz1);
1714
1715             /**************************
1716              * CALCULATE INTERACTIONS *
1717              **************************/
1718
1719             /* REACTION-FIELD ELECTROSTATICS */
1720             felec            = _fjsp_mul_v2r8(qq32,_fjsp_msub_v2r8(rinv32,rinvsq32,krf2));
1721
1722             fscal            = felec;
1723
1724             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1725
1726             /* Update vectorial force */
1727             fix3             = _fjsp_madd_v2r8(dx32,fscal,fix3);
1728             fiy3             = _fjsp_madd_v2r8(dy32,fscal,fiy3);
1729             fiz3             = _fjsp_madd_v2r8(dz32,fscal,fiz3);
1730
1731             fjx2             = _fjsp_madd_v2r8(dx32,fscal,fjx2);
1732             fjy2             = _fjsp_madd_v2r8(dy32,fscal,fjy2);
1733             fjz2             = _fjsp_madd_v2r8(dz32,fscal,fjz2);
1734
1735             /**************************
1736              * CALCULATE INTERACTIONS *
1737              **************************/
1738
1739             /* REACTION-FIELD ELECTROSTATICS */
1740             felec            = _fjsp_mul_v2r8(qq33,_fjsp_msub_v2r8(rinv33,rinvsq33,krf2));
1741
1742             fscal            = felec;
1743
1744             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1745
1746             /* Update vectorial force */
1747             fix3             = _fjsp_madd_v2r8(dx33,fscal,fix3);
1748             fiy3             = _fjsp_madd_v2r8(dy33,fscal,fiy3);
1749             fiz3             = _fjsp_madd_v2r8(dz33,fscal,fiz3);
1750
1751             fjx3             = _fjsp_madd_v2r8(dx33,fscal,fjx3);
1752             fjy3             = _fjsp_madd_v2r8(dy33,fscal,fjy3);
1753             fjz3             = _fjsp_madd_v2r8(dz33,fscal,fjz3);
1754
1755             gmx_fjsp_decrement_4rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1756
1757             /* Inner loop uses 324 flops */
1758         }
1759
1760         /* End of innermost loop */
1761
1762         gmx_fjsp_update_iforce_4atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1763                                               f+i_coord_offset,fshift+i_shift_offset);
1764
1765         /* Increment number of inner iterations */
1766         inneriter                  += j_index_end - j_index_start;
1767
1768         /* Outer loop uses 24 flops */
1769     }
1770
1771     /* Increment number of outer iterations */
1772     outeriter        += nri;
1773
1774     /* Update outer/inner flops */
1775
1776     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*24 + inneriter*324);
1777 }