src/gmxlib/nonbonded/nb_kernel_avx_128_fma_single/nb_kernel_ElecCSTab_VdwNone_GeomW4W4_avx_128_fma_single.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_128_fma_single 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_128_fma_single.h"
  34 #include "kernelutil_x86_avx_128_fma_single.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwNone_GeomW4W4_VF_avx_128_fma_single
  38  * Electrostatics interaction: CubicSplineTable
  39  * VdW interaction:            None
  40  * Geometry:                   Water4-Water4
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecCSTab_VdwNone_GeomW4W4_VF_avx_128_fma_single
  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_128, 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              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  63     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  64     real             rcutoff_scalar;
  65     real             *shiftvec,*fshift,*x,*f;
  66     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  67     real             scratch[4*DIM];
  68     __m128           fscal,rcutoff,rcutoff2,jidxall;
  69     int              vdwioffset1;
  70     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  71     int              vdwioffset2;
  72     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  73     int              vdwioffset3;
  74     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  75     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
  76     __m128           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  77     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
  78     __m128           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  79     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
  80     __m128           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
  81     __m128           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  82     __m128           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  83     __m128           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
  84     __m128           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
  85     __m128           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
  86     __m128           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
  87     __m128           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
  88     __m128           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
  89     __m128           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
  90     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
  91     real             *charge;
  92     __m128i          vfitab;
  93     __m128i          ifour       = _mm_set1_epi32(4);
  94     __m128           rt,vfeps,twovfeps,vftabscale,Y,F,G,H,Fp,VV,FF;
  95     real             *vftab;
  96     __m128           dummy_mask,cutoff_mask;
  97     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
  98     __m128           one     = _mm_set1_ps(1.0);
  99     __m128           two     = _mm_set1_ps(2.0);
 100     x                = xx[0];
 101     f                = ff[0];
 102
 103     nri              = nlist->nri;
 104     iinr             = nlist->iinr;
 105     jindex           = nlist->jindex;
 106     jjnr             = nlist->jjnr;
 107     shiftidx         = nlist->shift;
 108     gid              = nlist->gid;
 109     shiftvec         = fr->shift_vec[0];
 110     fshift           = fr->fshift[0];
 111     facel            = _mm_set1_ps(fr->epsfac);
 112     charge           = mdatoms->chargeA;
 113
 114     vftab            = kernel_data->table_elec->data;
 115     vftabscale       = _mm_set1_ps(kernel_data->table_elec->scale);
 116
 117     /* Setup water-specific parameters */
 118     inr              = nlist->iinr[0];
 119     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 120     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 121     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 122
 123     jq1              = _mm_set1_ps(charge[inr+1]);
 124     jq2              = _mm_set1_ps(charge[inr+2]);
 125     jq3              = _mm_set1_ps(charge[inr+3]);
 126     qq11             = _mm_mul_ps(iq1,jq1);
 127     qq12             = _mm_mul_ps(iq1,jq2);
 128     qq13             = _mm_mul_ps(iq1,jq3);
 129     qq21             = _mm_mul_ps(iq2,jq1);
 130     qq22             = _mm_mul_ps(iq2,jq2);
 131     qq23             = _mm_mul_ps(iq2,jq3);
 132     qq31             = _mm_mul_ps(iq3,jq1);
 133     qq32             = _mm_mul_ps(iq3,jq2);
 134     qq33             = _mm_mul_ps(iq3,jq3);
 135
 136     /* Avoid stupid compiler warnings */
 137     jnrA = jnrB = jnrC = jnrD = 0;
 138     j_coord_offsetA = 0;
 139     j_coord_offsetB = 0;
 140     j_coord_offsetC = 0;
 141     j_coord_offsetD = 0;
 142
 143     outeriter        = 0;
 144     inneriter        = 0;
 145
 146     for(iidx=0;iidx<4*DIM;iidx++)
 147     {
 148         scratch[iidx] = 0.0;
 149     }
 150
 151     /* Start outer loop over neighborlists */
 152     for(iidx=0; iidx<nri; iidx++)
 153     {
 154         /* Load shift vector for this list */
 155         i_shift_offset   = DIM*shiftidx[iidx];
 156
 157         /* Load limits for loop over neighbors */
 158         j_index_start    = jindex[iidx];
 159         j_index_end      = jindex[iidx+1];
 160
 161         /* Get outer coordinate index */
 162         inr              = iinr[iidx];
 163         i_coord_offset   = DIM*inr;
 164
 165         /* Load i particle coords and add shift vector */
 166         gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 167                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 168
 169         fix1             = _mm_setzero_ps();
 170         fiy1             = _mm_setzero_ps();
 171         fiz1             = _mm_setzero_ps();
 172         fix2             = _mm_setzero_ps();
 173         fiy2             = _mm_setzero_ps();
 174         fiz2             = _mm_setzero_ps();
 175         fix3             = _mm_setzero_ps();
 176         fiy3             = _mm_setzero_ps();
 177         fiz3             = _mm_setzero_ps();
 178
 179         /* Reset potential sums */
 180         velecsum         = _mm_setzero_ps();
 181
 182         /* Start inner kernel loop */
 183         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 184         {
 185
 186             /* Get j neighbor index, and coordinate index */
 187             jnrA             = jjnr[jidx];
 188             jnrB             = jjnr[jidx+1];
 189             jnrC             = jjnr[jidx+2];
 190             jnrD             = jjnr[jidx+3];
 191             j_coord_offsetA  = DIM*jnrA;
 192             j_coord_offsetB  = DIM*jnrB;
 193             j_coord_offsetC  = DIM*jnrC;
 194             j_coord_offsetD  = DIM*jnrD;
 195
 196             /* load j atom coordinates */
 197             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 198                                               x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 199                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 200
 201             /* Calculate displacement vector */
 202             dx11             = _mm_sub_ps(ix1,jx1);
 203             dy11             = _mm_sub_ps(iy1,jy1);
 204             dz11             = _mm_sub_ps(iz1,jz1);
 205             dx12             = _mm_sub_ps(ix1,jx2);
 206             dy12             = _mm_sub_ps(iy1,jy2);
 207             dz12             = _mm_sub_ps(iz1,jz2);
 208             dx13             = _mm_sub_ps(ix1,jx3);
 209             dy13             = _mm_sub_ps(iy1,jy3);
 210             dz13             = _mm_sub_ps(iz1,jz3);
 211             dx21             = _mm_sub_ps(ix2,jx1);
 212             dy21             = _mm_sub_ps(iy2,jy1);
 213             dz21             = _mm_sub_ps(iz2,jz1);
 214             dx22             = _mm_sub_ps(ix2,jx2);
 215             dy22             = _mm_sub_ps(iy2,jy2);
 216             dz22             = _mm_sub_ps(iz2,jz2);
 217             dx23             = _mm_sub_ps(ix2,jx3);
 218             dy23             = _mm_sub_ps(iy2,jy3);
 219             dz23             = _mm_sub_ps(iz2,jz3);
 220             dx31             = _mm_sub_ps(ix3,jx1);
 221             dy31             = _mm_sub_ps(iy3,jy1);
 222             dz31             = _mm_sub_ps(iz3,jz1);
 223             dx32             = _mm_sub_ps(ix3,jx2);
 224             dy32             = _mm_sub_ps(iy3,jy2);
 225             dz32             = _mm_sub_ps(iz3,jz2);
 226             dx33             = _mm_sub_ps(ix3,jx3);
 227             dy33             = _mm_sub_ps(iy3,jy3);
 228             dz33             = _mm_sub_ps(iz3,jz3);
 229
 230             /* Calculate squared distance and things based on it */
 231             rsq11            = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
 232             rsq12            = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
 233             rsq13            = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
 234             rsq21            = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
 235             rsq22            = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
 236             rsq23            = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
 237             rsq31            = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
 238             rsq32            = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
 239             rsq33            = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
 240
 241             rinv11           = gmx_mm_invsqrt_ps(rsq11);
 242             rinv12           = gmx_mm_invsqrt_ps(rsq12);
 243             rinv13           = gmx_mm_invsqrt_ps(rsq13);
 244             rinv21           = gmx_mm_invsqrt_ps(rsq21);
 245             rinv22           = gmx_mm_invsqrt_ps(rsq22);
 246             rinv23           = gmx_mm_invsqrt_ps(rsq23);
 247             rinv31           = gmx_mm_invsqrt_ps(rsq31);
 248             rinv32           = gmx_mm_invsqrt_ps(rsq32);
 249             rinv33           = gmx_mm_invsqrt_ps(rsq33);
 250
 251             fjx1             = _mm_setzero_ps();
 252             fjy1             = _mm_setzero_ps();
 253             fjz1             = _mm_setzero_ps();
 254             fjx2             = _mm_setzero_ps();
 255             fjy2             = _mm_setzero_ps();
 256             fjz2             = _mm_setzero_ps();
 257             fjx3             = _mm_setzero_ps();
 258             fjy3             = _mm_setzero_ps();
 259             fjz3             = _mm_setzero_ps();
 260
 261             /**************************
 262              * CALCULATE INTERACTIONS *
 263              **************************/
 264
 265             r11              = _mm_mul_ps(rsq11,rinv11);
 266
 267             /* Calculate table index by multiplying r with table scale and truncate to integer */
 268             rt               = _mm_mul_ps(r11,vftabscale);
 269             vfitab           = _mm_cvttps_epi32(rt);
 270 #ifdef __XOP__
 271             vfeps            = _mm_frcz_ps(rt);
 272 #else
 273             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 274 #endif
 275             twovfeps         = _mm_add_ps(vfeps,vfeps);
 276             vfitab           = _mm_slli_epi32(vfitab,2);
 277
 278             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 279             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 280             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 281             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 282             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 283             _MM_TRANSPOSE4_PS(Y,F,G,H);
 284             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 285             VV               = _mm_macc_ps(vfeps,Fp,Y);
 286             velec            = _mm_mul_ps(qq11,VV);
 287             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 288             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq11,FF),_mm_mul_ps(vftabscale,rinv11)));
 289
 290             /* Update potential sum for this i atom from the interaction with this j atom. */
 291             velecsum         = _mm_add_ps(velecsum,velec);
 292
 293             fscal            = felec;
 294
 295              /* Update vectorial force */
 296             fix1             = _mm_macc_ps(dx11,fscal,fix1);
 297             fiy1             = _mm_macc_ps(dy11,fscal,fiy1);
 298             fiz1             = _mm_macc_ps(dz11,fscal,fiz1);
 299
 300             fjx1             = _mm_macc_ps(dx11,fscal,fjx1);
 301             fjy1             = _mm_macc_ps(dy11,fscal,fjy1);
 302             fjz1             = _mm_macc_ps(dz11,fscal,fjz1);
 303
 304             /**************************
 305              * CALCULATE INTERACTIONS *
 306              **************************/
 307
 308             r12              = _mm_mul_ps(rsq12,rinv12);
 309
 310             /* Calculate table index by multiplying r with table scale and truncate to integer */
 311             rt               = _mm_mul_ps(r12,vftabscale);
 312             vfitab           = _mm_cvttps_epi32(rt);
 313 #ifdef __XOP__
 314             vfeps            = _mm_frcz_ps(rt);
 315 #else
 316             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 317 #endif
 318             twovfeps         = _mm_add_ps(vfeps,vfeps);
 319             vfitab           = _mm_slli_epi32(vfitab,2);
 320
 321             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 322             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 323             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 324             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 325             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 326             _MM_TRANSPOSE4_PS(Y,F,G,H);
 327             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 328             VV               = _mm_macc_ps(vfeps,Fp,Y);
 329             velec            = _mm_mul_ps(qq12,VV);
 330             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 331             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq12,FF),_mm_mul_ps(vftabscale,rinv12)));
 332
 333             /* Update potential sum for this i atom from the interaction with this j atom. */
 334             velecsum         = _mm_add_ps(velecsum,velec);
 335
 336             fscal            = felec;
 337
 338              /* Update vectorial force */
 339             fix1             = _mm_macc_ps(dx12,fscal,fix1);
 340             fiy1             = _mm_macc_ps(dy12,fscal,fiy1);
 341             fiz1             = _mm_macc_ps(dz12,fscal,fiz1);
 342
 343             fjx2             = _mm_macc_ps(dx12,fscal,fjx2);
 344             fjy2             = _mm_macc_ps(dy12,fscal,fjy2);
 345             fjz2             = _mm_macc_ps(dz12,fscal,fjz2);
 346
 347             /**************************
 348              * CALCULATE INTERACTIONS *
 349              **************************/
 350
 351             r13              = _mm_mul_ps(rsq13,rinv13);
 352
 353             /* Calculate table index by multiplying r with table scale and truncate to integer */
 354             rt               = _mm_mul_ps(r13,vftabscale);
 355             vfitab           = _mm_cvttps_epi32(rt);
 356 #ifdef __XOP__
 357             vfeps            = _mm_frcz_ps(rt);
 358 #else
 359             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 360 #endif
 361             twovfeps         = _mm_add_ps(vfeps,vfeps);
 362             vfitab           = _mm_slli_epi32(vfitab,2);
 363
 364             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 365             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 366             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 367             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 368             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 369             _MM_TRANSPOSE4_PS(Y,F,G,H);
 370             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 371             VV               = _mm_macc_ps(vfeps,Fp,Y);
 372             velec            = _mm_mul_ps(qq13,VV);
 373             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 374             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq13,FF),_mm_mul_ps(vftabscale,rinv13)));
 375
 376             /* Update potential sum for this i atom from the interaction with this j atom. */
 377             velecsum         = _mm_add_ps(velecsum,velec);
 378
 379             fscal            = felec;
 380
 381              /* Update vectorial force */
 382             fix1             = _mm_macc_ps(dx13,fscal,fix1);
 383             fiy1             = _mm_macc_ps(dy13,fscal,fiy1);
 384             fiz1             = _mm_macc_ps(dz13,fscal,fiz1);
 385
 386             fjx3             = _mm_macc_ps(dx13,fscal,fjx3);
 387             fjy3             = _mm_macc_ps(dy13,fscal,fjy3);
 388             fjz3             = _mm_macc_ps(dz13,fscal,fjz3);
 389
 390             /**************************
 391              * CALCULATE INTERACTIONS *
 392              **************************/
 393
 394             r21              = _mm_mul_ps(rsq21,rinv21);
 395
 396             /* Calculate table index by multiplying r with table scale and truncate to integer */
 397             rt               = _mm_mul_ps(r21,vftabscale);
 398             vfitab           = _mm_cvttps_epi32(rt);
 399 #ifdef __XOP__
 400             vfeps            = _mm_frcz_ps(rt);
 401 #else
 402             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 403 #endif
 404             twovfeps         = _mm_add_ps(vfeps,vfeps);
 405             vfitab           = _mm_slli_epi32(vfitab,2);
 406
 407             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 408             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 409             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 410             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 411             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 412             _MM_TRANSPOSE4_PS(Y,F,G,H);
 413             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 414             VV               = _mm_macc_ps(vfeps,Fp,Y);
 415             velec            = _mm_mul_ps(qq21,VV);
 416             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 417             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq21,FF),_mm_mul_ps(vftabscale,rinv21)));
 418
 419             /* Update potential sum for this i atom from the interaction with this j atom. */
 420             velecsum         = _mm_add_ps(velecsum,velec);
 421
 422             fscal            = felec;
 423
 424              /* Update vectorial force */
 425             fix2             = _mm_macc_ps(dx21,fscal,fix2);
 426             fiy2             = _mm_macc_ps(dy21,fscal,fiy2);
 427             fiz2             = _mm_macc_ps(dz21,fscal,fiz2);
 428
 429             fjx1             = _mm_macc_ps(dx21,fscal,fjx1);
 430             fjy1             = _mm_macc_ps(dy21,fscal,fjy1);
 431             fjz1             = _mm_macc_ps(dz21,fscal,fjz1);
 432
 433             /**************************
 434              * CALCULATE INTERACTIONS *
 435              **************************/
 436
 437             r22              = _mm_mul_ps(rsq22,rinv22);
 438
 439             /* Calculate table index by multiplying r with table scale and truncate to integer */
 440             rt               = _mm_mul_ps(r22,vftabscale);
 441             vfitab           = _mm_cvttps_epi32(rt);
 442 #ifdef __XOP__
 443             vfeps            = _mm_frcz_ps(rt);
 444 #else
 445             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 446 #endif
 447             twovfeps         = _mm_add_ps(vfeps,vfeps);
 448             vfitab           = _mm_slli_epi32(vfitab,2);
 449
 450             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 451             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 452             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 453             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 454             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 455             _MM_TRANSPOSE4_PS(Y,F,G,H);
 456             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 457             VV               = _mm_macc_ps(vfeps,Fp,Y);
 458             velec            = _mm_mul_ps(qq22,VV);
 459             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 460             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq22,FF),_mm_mul_ps(vftabscale,rinv22)));
 461
 462             /* Update potential sum for this i atom from the interaction with this j atom. */
 463             velecsum         = _mm_add_ps(velecsum,velec);
 464
 465             fscal            = felec;
 466
 467              /* Update vectorial force */
 468             fix2             = _mm_macc_ps(dx22,fscal,fix2);
 469             fiy2             = _mm_macc_ps(dy22,fscal,fiy2);
 470             fiz2             = _mm_macc_ps(dz22,fscal,fiz2);
 471
 472             fjx2             = _mm_macc_ps(dx22,fscal,fjx2);
 473             fjy2             = _mm_macc_ps(dy22,fscal,fjy2);
 474             fjz2             = _mm_macc_ps(dz22,fscal,fjz2);
 475
 476             /**************************
 477              * CALCULATE INTERACTIONS *
 478              **************************/
 479
 480             r23              = _mm_mul_ps(rsq23,rinv23);
 481
 482             /* Calculate table index by multiplying r with table scale and truncate to integer */
 483             rt               = _mm_mul_ps(r23,vftabscale);
 484             vfitab           = _mm_cvttps_epi32(rt);
 485 #ifdef __XOP__
 486             vfeps            = _mm_frcz_ps(rt);
 487 #else
 488             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 489 #endif
 490             twovfeps         = _mm_add_ps(vfeps,vfeps);
 491             vfitab           = _mm_slli_epi32(vfitab,2);
 492
 493             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 494             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 495             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 496             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 497             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 498             _MM_TRANSPOSE4_PS(Y,F,G,H);
 499             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 500             VV               = _mm_macc_ps(vfeps,Fp,Y);
 501             velec            = _mm_mul_ps(qq23,VV);
 502             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 503             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq23,FF),_mm_mul_ps(vftabscale,rinv23)));
 504
 505             /* Update potential sum for this i atom from the interaction with this j atom. */
 506             velecsum         = _mm_add_ps(velecsum,velec);
 507
 508             fscal            = felec;
 509
 510              /* Update vectorial force */
 511             fix2             = _mm_macc_ps(dx23,fscal,fix2);
 512             fiy2             = _mm_macc_ps(dy23,fscal,fiy2);
 513             fiz2             = _mm_macc_ps(dz23,fscal,fiz2);
 514
 515             fjx3             = _mm_macc_ps(dx23,fscal,fjx3);
 516             fjy3             = _mm_macc_ps(dy23,fscal,fjy3);
 517             fjz3             = _mm_macc_ps(dz23,fscal,fjz3);
 518
 519             /**************************
 520              * CALCULATE INTERACTIONS *
 521              **************************/
 522
 523             r31              = _mm_mul_ps(rsq31,rinv31);
 524
 525             /* Calculate table index by multiplying r with table scale and truncate to integer */
 526             rt               = _mm_mul_ps(r31,vftabscale);
 527             vfitab           = _mm_cvttps_epi32(rt);
 528 #ifdef __XOP__
 529             vfeps            = _mm_frcz_ps(rt);
 530 #else
 531             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 532 #endif
 533             twovfeps         = _mm_add_ps(vfeps,vfeps);
 534             vfitab           = _mm_slli_epi32(vfitab,2);
 535
 536             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 537             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 538             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 539             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 540             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 541             _MM_TRANSPOSE4_PS(Y,F,G,H);
 542             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 543             VV               = _mm_macc_ps(vfeps,Fp,Y);
 544             velec            = _mm_mul_ps(qq31,VV);
 545             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 546             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq31,FF),_mm_mul_ps(vftabscale,rinv31)));
 547
 548             /* Update potential sum for this i atom from the interaction with this j atom. */
 549             velecsum         = _mm_add_ps(velecsum,velec);
 550
 551             fscal            = felec;
 552
 553              /* Update vectorial force */
 554             fix3             = _mm_macc_ps(dx31,fscal,fix3);
 555             fiy3             = _mm_macc_ps(dy31,fscal,fiy3);
 556             fiz3             = _mm_macc_ps(dz31,fscal,fiz3);
 557
 558             fjx1             = _mm_macc_ps(dx31,fscal,fjx1);
 559             fjy1             = _mm_macc_ps(dy31,fscal,fjy1);
 560             fjz1             = _mm_macc_ps(dz31,fscal,fjz1);
 561
 562             /**************************
 563              * CALCULATE INTERACTIONS *
 564              **************************/
 565
 566             r32              = _mm_mul_ps(rsq32,rinv32);
 567
 568             /* Calculate table index by multiplying r with table scale and truncate to integer */
 569             rt               = _mm_mul_ps(r32,vftabscale);
 570             vfitab           = _mm_cvttps_epi32(rt);
 571 #ifdef __XOP__
 572             vfeps            = _mm_frcz_ps(rt);
 573 #else
 574             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 575 #endif
 576             twovfeps         = _mm_add_ps(vfeps,vfeps);
 577             vfitab           = _mm_slli_epi32(vfitab,2);
 578
 579             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 580             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 581             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 582             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 583             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 584             _MM_TRANSPOSE4_PS(Y,F,G,H);
 585             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 586             VV               = _mm_macc_ps(vfeps,Fp,Y);
 587             velec            = _mm_mul_ps(qq32,VV);
 588             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 589             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq32,FF),_mm_mul_ps(vftabscale,rinv32)));
 590
 591             /* Update potential sum for this i atom from the interaction with this j atom. */
 592             velecsum         = _mm_add_ps(velecsum,velec);
 593
 594             fscal            = felec;
 595
 596              /* Update vectorial force */
 597             fix3             = _mm_macc_ps(dx32,fscal,fix3);
 598             fiy3             = _mm_macc_ps(dy32,fscal,fiy3);
 599             fiz3             = _mm_macc_ps(dz32,fscal,fiz3);
 600
 601             fjx2             = _mm_macc_ps(dx32,fscal,fjx2);
 602             fjy2             = _mm_macc_ps(dy32,fscal,fjy2);
 603             fjz2             = _mm_macc_ps(dz32,fscal,fjz2);
 604
 605             /**************************
 606              * CALCULATE INTERACTIONS *
 607              **************************/
 608
 609             r33              = _mm_mul_ps(rsq33,rinv33);
 610
 611             /* Calculate table index by multiplying r with table scale and truncate to integer */
 612             rt               = _mm_mul_ps(r33,vftabscale);
 613             vfitab           = _mm_cvttps_epi32(rt);
 614 #ifdef __XOP__
 615             vfeps            = _mm_frcz_ps(rt);
 616 #else
 617             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 618 #endif
 619             twovfeps         = _mm_add_ps(vfeps,vfeps);
 620             vfitab           = _mm_slli_epi32(vfitab,2);
 621
 622             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 623             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 624             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 625             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 626             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 627             _MM_TRANSPOSE4_PS(Y,F,G,H);
 628             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 629             VV               = _mm_macc_ps(vfeps,Fp,Y);
 630             velec            = _mm_mul_ps(qq33,VV);
 631             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 632             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq33,FF),_mm_mul_ps(vftabscale,rinv33)));
 633
 634             /* Update potential sum for this i atom from the interaction with this j atom. */
 635             velecsum         = _mm_add_ps(velecsum,velec);
 636
 637             fscal            = felec;
 638
 639              /* Update vectorial force */
 640             fix3             = _mm_macc_ps(dx33,fscal,fix3);
 641             fiy3             = _mm_macc_ps(dy33,fscal,fiy3);
 642             fiz3             = _mm_macc_ps(dz33,fscal,fiz3);
 643
 644             fjx3             = _mm_macc_ps(dx33,fscal,fjx3);
 645             fjy3             = _mm_macc_ps(dy33,fscal,fjy3);
 646             fjz3             = _mm_macc_ps(dz33,fscal,fjz3);
 647
 648             fjptrA             = f+j_coord_offsetA;
 649             fjptrB             = f+j_coord_offsetB;
 650             fjptrC             = f+j_coord_offsetC;
 651             fjptrD             = f+j_coord_offsetD;
 652
 653             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
 654                                                    fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 655
 656             /* Inner loop uses 414 flops */
 657         }
 658
 659         if(jidx<j_index_end)
 660         {
 661
 662             /* Get j neighbor index, and coordinate index */
 663             jnrlistA         = jjnr[jidx];
 664             jnrlistB         = jjnr[jidx+1];
 665             jnrlistC         = jjnr[jidx+2];
 666             jnrlistD         = jjnr[jidx+3];
 667             /* Sign of each element will be negative for non-real atoms.
 668              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 669              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 670              */
 671             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 672             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 673             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 674             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 675             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 676             j_coord_offsetA  = DIM*jnrA;
 677             j_coord_offsetB  = DIM*jnrB;
 678             j_coord_offsetC  = DIM*jnrC;
 679             j_coord_offsetD  = DIM*jnrD;
 680
 681             /* load j atom coordinates */
 682             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 683                                               x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 684                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 685
 686             /* Calculate displacement vector */
 687             dx11             = _mm_sub_ps(ix1,jx1);
 688             dy11             = _mm_sub_ps(iy1,jy1);
 689             dz11             = _mm_sub_ps(iz1,jz1);
 690             dx12             = _mm_sub_ps(ix1,jx2);
 691             dy12             = _mm_sub_ps(iy1,jy2);
 692             dz12             = _mm_sub_ps(iz1,jz2);
 693             dx13             = _mm_sub_ps(ix1,jx3);
 694             dy13             = _mm_sub_ps(iy1,jy3);
 695             dz13             = _mm_sub_ps(iz1,jz3);
 696             dx21             = _mm_sub_ps(ix2,jx1);
 697             dy21             = _mm_sub_ps(iy2,jy1);
 698             dz21             = _mm_sub_ps(iz2,jz1);
 699             dx22             = _mm_sub_ps(ix2,jx2);
 700             dy22             = _mm_sub_ps(iy2,jy2);
 701             dz22             = _mm_sub_ps(iz2,jz2);
 702             dx23             = _mm_sub_ps(ix2,jx3);
 703             dy23             = _mm_sub_ps(iy2,jy3);
 704             dz23             = _mm_sub_ps(iz2,jz3);
 705             dx31             = _mm_sub_ps(ix3,jx1);
 706             dy31             = _mm_sub_ps(iy3,jy1);
 707             dz31             = _mm_sub_ps(iz3,jz1);
 708             dx32             = _mm_sub_ps(ix3,jx2);
 709             dy32             = _mm_sub_ps(iy3,jy2);
 710             dz32             = _mm_sub_ps(iz3,jz2);
 711             dx33             = _mm_sub_ps(ix3,jx3);
 712             dy33             = _mm_sub_ps(iy3,jy3);
 713             dz33             = _mm_sub_ps(iz3,jz3);
 714
 715             /* Calculate squared distance and things based on it */
 716             rsq11            = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
 717             rsq12            = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
 718             rsq13            = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
 719             rsq21            = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
 720             rsq22            = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
 721             rsq23            = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
 722             rsq31            = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
 723             rsq32            = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
 724             rsq33            = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
 725
 726             rinv11           = gmx_mm_invsqrt_ps(rsq11);
 727             rinv12           = gmx_mm_invsqrt_ps(rsq12);
 728             rinv13           = gmx_mm_invsqrt_ps(rsq13);
 729             rinv21           = gmx_mm_invsqrt_ps(rsq21);
 730             rinv22           = gmx_mm_invsqrt_ps(rsq22);
 731             rinv23           = gmx_mm_invsqrt_ps(rsq23);
 732             rinv31           = gmx_mm_invsqrt_ps(rsq31);
 733             rinv32           = gmx_mm_invsqrt_ps(rsq32);
 734             rinv33           = gmx_mm_invsqrt_ps(rsq33);
 735
 736             fjx1             = _mm_setzero_ps();
 737             fjy1             = _mm_setzero_ps();
 738             fjz1             = _mm_setzero_ps();
 739             fjx2             = _mm_setzero_ps();
 740             fjy2             = _mm_setzero_ps();
 741             fjz2             = _mm_setzero_ps();
 742             fjx3             = _mm_setzero_ps();
 743             fjy3             = _mm_setzero_ps();
 744             fjz3             = _mm_setzero_ps();
 745
 746             /**************************
 747              * CALCULATE INTERACTIONS *
 748              **************************/
 749
 750             r11              = _mm_mul_ps(rsq11,rinv11);
 751             r11              = _mm_andnot_ps(dummy_mask,r11);
 752
 753             /* Calculate table index by multiplying r with table scale and truncate to integer */
 754             rt               = _mm_mul_ps(r11,vftabscale);
 755             vfitab           = _mm_cvttps_epi32(rt);
 756 #ifdef __XOP__
 757             vfeps            = _mm_frcz_ps(rt);
 758 #else
 759             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 760 #endif
 761             twovfeps         = _mm_add_ps(vfeps,vfeps);
 762             vfitab           = _mm_slli_epi32(vfitab,2);
 763
 764             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 765             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 766             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 767             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 768             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 769             _MM_TRANSPOSE4_PS(Y,F,G,H);
 770             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 771             VV               = _mm_macc_ps(vfeps,Fp,Y);
 772             velec            = _mm_mul_ps(qq11,VV);
 773             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 774             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq11,FF),_mm_mul_ps(vftabscale,rinv11)));
 775
 776             /* Update potential sum for this i atom from the interaction with this j atom. */
 777             velec            = _mm_andnot_ps(dummy_mask,velec);
 778             velecsum         = _mm_add_ps(velecsum,velec);
 779
 780             fscal            = felec;
 781
 782             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 783
 784              /* Update vectorial force */
 785             fix1             = _mm_macc_ps(dx11,fscal,fix1);
 786             fiy1             = _mm_macc_ps(dy11,fscal,fiy1);
 787             fiz1             = _mm_macc_ps(dz11,fscal,fiz1);
 788
 789             fjx1             = _mm_macc_ps(dx11,fscal,fjx1);
 790             fjy1             = _mm_macc_ps(dy11,fscal,fjy1);
 791             fjz1             = _mm_macc_ps(dz11,fscal,fjz1);
 792
 793             /**************************
 794              * CALCULATE INTERACTIONS *
 795              **************************/
 796
 797             r12              = _mm_mul_ps(rsq12,rinv12);
 798             r12              = _mm_andnot_ps(dummy_mask,r12);
 799
 800             /* Calculate table index by multiplying r with table scale and truncate to integer */
 801             rt               = _mm_mul_ps(r12,vftabscale);
 802             vfitab           = _mm_cvttps_epi32(rt);
 803 #ifdef __XOP__
 804             vfeps            = _mm_frcz_ps(rt);
 805 #else
 806             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 807 #endif
 808             twovfeps         = _mm_add_ps(vfeps,vfeps);
 809             vfitab           = _mm_slli_epi32(vfitab,2);
 810
 811             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 812             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 813             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 814             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 815             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 816             _MM_TRANSPOSE4_PS(Y,F,G,H);
 817             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 818             VV               = _mm_macc_ps(vfeps,Fp,Y);
 819             velec            = _mm_mul_ps(qq12,VV);
 820             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 821             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq12,FF),_mm_mul_ps(vftabscale,rinv12)));
 822
 823             /* Update potential sum for this i atom from the interaction with this j atom. */
 824             velec            = _mm_andnot_ps(dummy_mask,velec);
 825             velecsum         = _mm_add_ps(velecsum,velec);
 826
 827             fscal            = felec;
 828
 829             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 830
 831              /* Update vectorial force */
 832             fix1             = _mm_macc_ps(dx12,fscal,fix1);
 833             fiy1             = _mm_macc_ps(dy12,fscal,fiy1);
 834             fiz1             = _mm_macc_ps(dz12,fscal,fiz1);
 835
 836             fjx2             = _mm_macc_ps(dx12,fscal,fjx2);
 837             fjy2             = _mm_macc_ps(dy12,fscal,fjy2);
 838             fjz2             = _mm_macc_ps(dz12,fscal,fjz2);
 839
 840             /**************************
 841              * CALCULATE INTERACTIONS *
 842              **************************/
 843
 844             r13              = _mm_mul_ps(rsq13,rinv13);
 845             r13              = _mm_andnot_ps(dummy_mask,r13);
 846
 847             /* Calculate table index by multiplying r with table scale and truncate to integer */
 848             rt               = _mm_mul_ps(r13,vftabscale);
 849             vfitab           = _mm_cvttps_epi32(rt);
 850 #ifdef __XOP__
 851             vfeps            = _mm_frcz_ps(rt);
 852 #else
 853             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 854 #endif
 855             twovfeps         = _mm_add_ps(vfeps,vfeps);
 856             vfitab           = _mm_slli_epi32(vfitab,2);
 857
 858             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 859             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 860             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 861             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 862             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 863             _MM_TRANSPOSE4_PS(Y,F,G,H);
 864             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 865             VV               = _mm_macc_ps(vfeps,Fp,Y);
 866             velec            = _mm_mul_ps(qq13,VV);
 867             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 868             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq13,FF),_mm_mul_ps(vftabscale,rinv13)));
 869
 870             /* Update potential sum for this i atom from the interaction with this j atom. */
 871             velec            = _mm_andnot_ps(dummy_mask,velec);
 872             velecsum         = _mm_add_ps(velecsum,velec);
 873
 874             fscal            = felec;
 875
 876             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 877
 878              /* Update vectorial force */
 879             fix1             = _mm_macc_ps(dx13,fscal,fix1);
 880             fiy1             = _mm_macc_ps(dy13,fscal,fiy1);
 881             fiz1             = _mm_macc_ps(dz13,fscal,fiz1);
 882
 883             fjx3             = _mm_macc_ps(dx13,fscal,fjx3);
 884             fjy3             = _mm_macc_ps(dy13,fscal,fjy3);
 885             fjz3             = _mm_macc_ps(dz13,fscal,fjz3);
 886
 887             /**************************
 888              * CALCULATE INTERACTIONS *
 889              **************************/
 890
 891             r21              = _mm_mul_ps(rsq21,rinv21);
 892             r21              = _mm_andnot_ps(dummy_mask,r21);
 893
 894             /* Calculate table index by multiplying r with table scale and truncate to integer */
 895             rt               = _mm_mul_ps(r21,vftabscale);
 896             vfitab           = _mm_cvttps_epi32(rt);
 897 #ifdef __XOP__
 898             vfeps            = _mm_frcz_ps(rt);
 899 #else
 900             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 901 #endif
 902             twovfeps         = _mm_add_ps(vfeps,vfeps);
 903             vfitab           = _mm_slli_epi32(vfitab,2);
 904
 905             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 906             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 907             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 908             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 909             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 910             _MM_TRANSPOSE4_PS(Y,F,G,H);
 911             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 912             VV               = _mm_macc_ps(vfeps,Fp,Y);
 913             velec            = _mm_mul_ps(qq21,VV);
 914             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 915             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq21,FF),_mm_mul_ps(vftabscale,rinv21)));
 916
 917             /* Update potential sum for this i atom from the interaction with this j atom. */
 918             velec            = _mm_andnot_ps(dummy_mask,velec);
 919             velecsum         = _mm_add_ps(velecsum,velec);
 920
 921             fscal            = felec;
 922
 923             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 924
 925              /* Update vectorial force */
 926             fix2             = _mm_macc_ps(dx21,fscal,fix2);
 927             fiy2             = _mm_macc_ps(dy21,fscal,fiy2);
 928             fiz2             = _mm_macc_ps(dz21,fscal,fiz2);
 929
 930             fjx1             = _mm_macc_ps(dx21,fscal,fjx1);
 931             fjy1             = _mm_macc_ps(dy21,fscal,fjy1);
 932             fjz1             = _mm_macc_ps(dz21,fscal,fjz1);
 933
 934             /**************************
 935              * CALCULATE INTERACTIONS *
 936              **************************/
 937
 938             r22              = _mm_mul_ps(rsq22,rinv22);
 939             r22              = _mm_andnot_ps(dummy_mask,r22);
 940
 941             /* Calculate table index by multiplying r with table scale and truncate to integer */
 942             rt               = _mm_mul_ps(r22,vftabscale);
 943             vfitab           = _mm_cvttps_epi32(rt);
 944 #ifdef __XOP__
 945             vfeps            = _mm_frcz_ps(rt);
 946 #else
 947             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 948 #endif
 949             twovfeps         = _mm_add_ps(vfeps,vfeps);
 950             vfitab           = _mm_slli_epi32(vfitab,2);
 951
 952             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 953             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
 954             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
 955             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
 956             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
 957             _MM_TRANSPOSE4_PS(Y,F,G,H);
 958             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
 959             VV               = _mm_macc_ps(vfeps,Fp,Y);
 960             velec            = _mm_mul_ps(qq22,VV);
 961             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
 962             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq22,FF),_mm_mul_ps(vftabscale,rinv22)));
 963
 964             /* Update potential sum for this i atom from the interaction with this j atom. */
 965             velec            = _mm_andnot_ps(dummy_mask,velec);
 966             velecsum         = _mm_add_ps(velecsum,velec);
 967
 968             fscal            = felec;
 969
 970             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 971
 972              /* Update vectorial force */
 973             fix2             = _mm_macc_ps(dx22,fscal,fix2);
 974             fiy2             = _mm_macc_ps(dy22,fscal,fiy2);
 975             fiz2             = _mm_macc_ps(dz22,fscal,fiz2);
 976
 977             fjx2             = _mm_macc_ps(dx22,fscal,fjx2);
 978             fjy2             = _mm_macc_ps(dy22,fscal,fjy2);
 979             fjz2             = _mm_macc_ps(dz22,fscal,fjz2);
 980
 981             /**************************
 982              * CALCULATE INTERACTIONS *
 983              **************************/
 984
 985             r23              = _mm_mul_ps(rsq23,rinv23);
 986             r23              = _mm_andnot_ps(dummy_mask,r23);
 987
 988             /* Calculate table index by multiplying r with table scale and truncate to integer */
 989             rt               = _mm_mul_ps(r23,vftabscale);
 990             vfitab           = _mm_cvttps_epi32(rt);
 991 #ifdef __XOP__
 992             vfeps            = _mm_frcz_ps(rt);
 993 #else
 994             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 995 #endif
 996             twovfeps         = _mm_add_ps(vfeps,vfeps);
 997             vfitab           = _mm_slli_epi32(vfitab,2);
 998
 999             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1000             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1001             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1002             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1003             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1004             _MM_TRANSPOSE4_PS(Y,F,G,H);
1005             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1006             VV               = _mm_macc_ps(vfeps,Fp,Y);
1007             velec            = _mm_mul_ps(qq23,VV);
1008             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1009             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq23,FF),_mm_mul_ps(vftabscale,rinv23)));
1010
1011             /* Update potential sum for this i atom from the interaction with this j atom. */
1012             velec            = _mm_andnot_ps(dummy_mask,velec);
1013             velecsum         = _mm_add_ps(velecsum,velec);
1014
1015             fscal            = felec;
1016
1017             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1018
1019              /* Update vectorial force */
1020             fix2             = _mm_macc_ps(dx23,fscal,fix2);
1021             fiy2             = _mm_macc_ps(dy23,fscal,fiy2);
1022             fiz2             = _mm_macc_ps(dz23,fscal,fiz2);
1023
1024             fjx3             = _mm_macc_ps(dx23,fscal,fjx3);
1025             fjy3             = _mm_macc_ps(dy23,fscal,fjy3);
1026             fjz3             = _mm_macc_ps(dz23,fscal,fjz3);
1027
1028             /**************************
1029              * CALCULATE INTERACTIONS *
1030              **************************/
1031
1032             r31              = _mm_mul_ps(rsq31,rinv31);
1033             r31              = _mm_andnot_ps(dummy_mask,r31);
1034
1035             /* Calculate table index by multiplying r with table scale and truncate to integer */
1036             rt               = _mm_mul_ps(r31,vftabscale);
1037             vfitab           = _mm_cvttps_epi32(rt);
1038 #ifdef __XOP__
1039             vfeps            = _mm_frcz_ps(rt);
1040 #else
1041             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1042 #endif
1043             twovfeps         = _mm_add_ps(vfeps,vfeps);
1044             vfitab           = _mm_slli_epi32(vfitab,2);
1045
1046             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1047             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1048             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1049             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1050             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1051             _MM_TRANSPOSE4_PS(Y,F,G,H);
1052             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1053             VV               = _mm_macc_ps(vfeps,Fp,Y);
1054             velec            = _mm_mul_ps(qq31,VV);
1055             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1056             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq31,FF),_mm_mul_ps(vftabscale,rinv31)));
1057
1058             /* Update potential sum for this i atom from the interaction with this j atom. */
1059             velec            = _mm_andnot_ps(dummy_mask,velec);
1060             velecsum         = _mm_add_ps(velecsum,velec);
1061
1062             fscal            = felec;
1063
1064             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1065
1066              /* Update vectorial force */
1067             fix3             = _mm_macc_ps(dx31,fscal,fix3);
1068             fiy3             = _mm_macc_ps(dy31,fscal,fiy3);
1069             fiz3             = _mm_macc_ps(dz31,fscal,fiz3);
1070
1071             fjx1             = _mm_macc_ps(dx31,fscal,fjx1);
1072             fjy1             = _mm_macc_ps(dy31,fscal,fjy1);
1073             fjz1             = _mm_macc_ps(dz31,fscal,fjz1);
1074
1075             /**************************
1076              * CALCULATE INTERACTIONS *
1077              **************************/
1078
1079             r32              = _mm_mul_ps(rsq32,rinv32);
1080             r32              = _mm_andnot_ps(dummy_mask,r32);
1081
1082             /* Calculate table index by multiplying r with table scale and truncate to integer */
1083             rt               = _mm_mul_ps(r32,vftabscale);
1084             vfitab           = _mm_cvttps_epi32(rt);
1085 #ifdef __XOP__
1086             vfeps            = _mm_frcz_ps(rt);
1087 #else
1088             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1089 #endif
1090             twovfeps         = _mm_add_ps(vfeps,vfeps);
1091             vfitab           = _mm_slli_epi32(vfitab,2);
1092
1093             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1094             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1095             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1096             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1097             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1098             _MM_TRANSPOSE4_PS(Y,F,G,H);
1099             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1100             VV               = _mm_macc_ps(vfeps,Fp,Y);
1101             velec            = _mm_mul_ps(qq32,VV);
1102             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1103             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq32,FF),_mm_mul_ps(vftabscale,rinv32)));
1104
1105             /* Update potential sum for this i atom from the interaction with this j atom. */
1106             velec            = _mm_andnot_ps(dummy_mask,velec);
1107             velecsum         = _mm_add_ps(velecsum,velec);
1108
1109             fscal            = felec;
1110
1111             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1112
1113              /* Update vectorial force */
1114             fix3             = _mm_macc_ps(dx32,fscal,fix3);
1115             fiy3             = _mm_macc_ps(dy32,fscal,fiy3);
1116             fiz3             = _mm_macc_ps(dz32,fscal,fiz3);
1117
1118             fjx2             = _mm_macc_ps(dx32,fscal,fjx2);
1119             fjy2             = _mm_macc_ps(dy32,fscal,fjy2);
1120             fjz2             = _mm_macc_ps(dz32,fscal,fjz2);
1121
1122             /**************************
1123              * CALCULATE INTERACTIONS *
1124              **************************/
1125
1126             r33              = _mm_mul_ps(rsq33,rinv33);
1127             r33              = _mm_andnot_ps(dummy_mask,r33);
1128
1129             /* Calculate table index by multiplying r with table scale and truncate to integer */
1130             rt               = _mm_mul_ps(r33,vftabscale);
1131             vfitab           = _mm_cvttps_epi32(rt);
1132 #ifdef __XOP__
1133             vfeps            = _mm_frcz_ps(rt);
1134 #else
1135             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1136 #endif
1137             twovfeps         = _mm_add_ps(vfeps,vfeps);
1138             vfitab           = _mm_slli_epi32(vfitab,2);
1139
1140             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1141             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1142             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1143             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1144             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1145             _MM_TRANSPOSE4_PS(Y,F,G,H);
1146             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1147             VV               = _mm_macc_ps(vfeps,Fp,Y);
1148             velec            = _mm_mul_ps(qq33,VV);
1149             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1150             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq33,FF),_mm_mul_ps(vftabscale,rinv33)));
1151
1152             /* Update potential sum for this i atom from the interaction with this j atom. */
1153             velec            = _mm_andnot_ps(dummy_mask,velec);
1154             velecsum         = _mm_add_ps(velecsum,velec);
1155
1156             fscal            = felec;
1157
1158             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1159
1160              /* Update vectorial force */
1161             fix3             = _mm_macc_ps(dx33,fscal,fix3);
1162             fiy3             = _mm_macc_ps(dy33,fscal,fiy3);
1163             fiz3             = _mm_macc_ps(dz33,fscal,fiz3);
1164
1165             fjx3             = _mm_macc_ps(dx33,fscal,fjx3);
1166             fjy3             = _mm_macc_ps(dy33,fscal,fjy3);
1167             fjz3             = _mm_macc_ps(dz33,fscal,fjz3);
1168
1169             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1170             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1171             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1172             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1173
1174             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1175                                                    fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1176
1177             /* Inner loop uses 423 flops */
1178         }
1179
1180         /* End of innermost loop */
1181
1182         gmx_mm_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1183                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
1184
1185         ggid                        = gid[iidx];
1186         /* Update potential energies */
1187         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1188
1189         /* Increment number of inner iterations */
1190         inneriter                  += j_index_end - j_index_start;
1191
1192         /* Outer loop uses 19 flops */
1193     }
1194
1195     /* Increment number of outer iterations */
1196     outeriter        += nri;
1197
1198     /* Update outer/inner flops */
1199
1200     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_VF,outeriter*19 + inneriter*423);
1201 }
1202 /*
1203  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwNone_GeomW4W4_F_avx_128_fma_single
1204  * Electrostatics interaction: CubicSplineTable
1205  * VdW interaction:            None
1206  * Geometry:                   Water4-Water4
1207  * Calculate force/pot:        Force
1208  */
1209 void
1210 nb_kernel_ElecCSTab_VdwNone_GeomW4W4_F_avx_128_fma_single
1211                     (t_nblist * gmx_restrict                nlist,
1212                      rvec * gmx_restrict                    xx,
1213                      rvec * gmx_restrict                    ff,
1214                      t_forcerec * gmx_restrict              fr,
1215                      t_mdatoms * gmx_restrict               mdatoms,
1216                      nb_kernel_data_t * gmx_restrict        kernel_data,
1217                      t_nrnb * gmx_restrict                  nrnb)
1218 {
1219     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1220      * just 0 for non-waters.
1221      * Suffixes A,B,C,D refer to j loop unrolling done with AVX_128, e.g. for the four different
1222      * jnr indices corresponding to data put in the four positions in the SIMD register.
1223      */
1224     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1225     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1226     int              jnrA,jnrB,jnrC,jnrD;
1227     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1228     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1229     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1230     real             rcutoff_scalar;
1231     real             *shiftvec,*fshift,*x,*f;
1232     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1233     real             scratch[4*DIM];
1234     __m128           fscal,rcutoff,rcutoff2,jidxall;
1235     int              vdwioffset1;
1236     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1237     int              vdwioffset2;
1238     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1239     int              vdwioffset3;
1240     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1241     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1242     __m128           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1243     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1244     __m128           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1245     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
1246     __m128           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1247     __m128           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1248     __m128           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1249     __m128           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1250     __m128           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1251     __m128           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1252     __m128           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1253     __m128           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1254     __m128           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1255     __m128           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1256     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
1257     real             *charge;
1258     __m128i          vfitab;
1259     __m128i          ifour       = _mm_set1_epi32(4);
1260     __m128           rt,vfeps,twovfeps,vftabscale,Y,F,G,H,Fp,VV,FF;
1261     real             *vftab;
1262     __m128           dummy_mask,cutoff_mask;
1263     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
1264     __m128           one     = _mm_set1_ps(1.0);
1265     __m128           two     = _mm_set1_ps(2.0);
1266     x                = xx[0];
1267     f                = ff[0];
1268
1269     nri              = nlist->nri;
1270     iinr             = nlist->iinr;
1271     jindex           = nlist->jindex;
1272     jjnr             = nlist->jjnr;
1273     shiftidx         = nlist->shift;
1274     gid              = nlist->gid;
1275     shiftvec         = fr->shift_vec[0];
1276     fshift           = fr->fshift[0];
1277     facel            = _mm_set1_ps(fr->epsfac);
1278     charge           = mdatoms->chargeA;
1279
1280     vftab            = kernel_data->table_elec->data;
1281     vftabscale       = _mm_set1_ps(kernel_data->table_elec->scale);
1282
1283     /* Setup water-specific parameters */
1284     inr              = nlist->iinr[0];
1285     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
1286     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
1287     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
1288
1289     jq1              = _mm_set1_ps(charge[inr+1]);
1290     jq2              = _mm_set1_ps(charge[inr+2]);
1291     jq3              = _mm_set1_ps(charge[inr+3]);
1292     qq11             = _mm_mul_ps(iq1,jq1);
1293     qq12             = _mm_mul_ps(iq1,jq2);
1294     qq13             = _mm_mul_ps(iq1,jq3);
1295     qq21             = _mm_mul_ps(iq2,jq1);
1296     qq22             = _mm_mul_ps(iq2,jq2);
1297     qq23             = _mm_mul_ps(iq2,jq3);
1298     qq31             = _mm_mul_ps(iq3,jq1);
1299     qq32             = _mm_mul_ps(iq3,jq2);
1300     qq33             = _mm_mul_ps(iq3,jq3);
1301
1302     /* Avoid stupid compiler warnings */
1303     jnrA = jnrB = jnrC = jnrD = 0;
1304     j_coord_offsetA = 0;
1305     j_coord_offsetB = 0;
1306     j_coord_offsetC = 0;
1307     j_coord_offsetD = 0;
1308
1309     outeriter        = 0;
1310     inneriter        = 0;
1311
1312     for(iidx=0;iidx<4*DIM;iidx++)
1313     {
1314         scratch[iidx] = 0.0;
1315     }
1316
1317     /* Start outer loop over neighborlists */
1318     for(iidx=0; iidx<nri; iidx++)
1319     {
1320         /* Load shift vector for this list */
1321         i_shift_offset   = DIM*shiftidx[iidx];
1322
1323         /* Load limits for loop over neighbors */
1324         j_index_start    = jindex[iidx];
1325         j_index_end      = jindex[iidx+1];
1326
1327         /* Get outer coordinate index */
1328         inr              = iinr[iidx];
1329         i_coord_offset   = DIM*inr;
1330
1331         /* Load i particle coords and add shift vector */
1332         gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
1333                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1334
1335         fix1             = _mm_setzero_ps();
1336         fiy1             = _mm_setzero_ps();
1337         fiz1             = _mm_setzero_ps();
1338         fix2             = _mm_setzero_ps();
1339         fiy2             = _mm_setzero_ps();
1340         fiz2             = _mm_setzero_ps();
1341         fix3             = _mm_setzero_ps();
1342         fiy3             = _mm_setzero_ps();
1343         fiz3             = _mm_setzero_ps();
1344
1345         /* Start inner kernel loop */
1346         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1347         {
1348
1349             /* Get j neighbor index, and coordinate index */
1350             jnrA             = jjnr[jidx];
1351             jnrB             = jjnr[jidx+1];
1352             jnrC             = jjnr[jidx+2];
1353             jnrD             = jjnr[jidx+3];
1354             j_coord_offsetA  = DIM*jnrA;
1355             j_coord_offsetB  = DIM*jnrB;
1356             j_coord_offsetC  = DIM*jnrC;
1357             j_coord_offsetD  = DIM*jnrD;
1358
1359             /* load j atom coordinates */
1360             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1361                                               x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1362                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1363
1364             /* Calculate displacement vector */
1365             dx11             = _mm_sub_ps(ix1,jx1);
1366             dy11             = _mm_sub_ps(iy1,jy1);
1367             dz11             = _mm_sub_ps(iz1,jz1);
1368             dx12             = _mm_sub_ps(ix1,jx2);
1369             dy12             = _mm_sub_ps(iy1,jy2);
1370             dz12             = _mm_sub_ps(iz1,jz2);
1371             dx13             = _mm_sub_ps(ix1,jx3);
1372             dy13             = _mm_sub_ps(iy1,jy3);
1373             dz13             = _mm_sub_ps(iz1,jz3);
1374             dx21             = _mm_sub_ps(ix2,jx1);
1375             dy21             = _mm_sub_ps(iy2,jy1);
1376             dz21             = _mm_sub_ps(iz2,jz1);
1377             dx22             = _mm_sub_ps(ix2,jx2);
1378             dy22             = _mm_sub_ps(iy2,jy2);
1379             dz22             = _mm_sub_ps(iz2,jz2);
1380             dx23             = _mm_sub_ps(ix2,jx3);
1381             dy23             = _mm_sub_ps(iy2,jy3);
1382             dz23             = _mm_sub_ps(iz2,jz3);
1383             dx31             = _mm_sub_ps(ix3,jx1);
1384             dy31             = _mm_sub_ps(iy3,jy1);
1385             dz31             = _mm_sub_ps(iz3,jz1);
1386             dx32             = _mm_sub_ps(ix3,jx2);
1387             dy32             = _mm_sub_ps(iy3,jy2);
1388             dz32             = _mm_sub_ps(iz3,jz2);
1389             dx33             = _mm_sub_ps(ix3,jx3);
1390             dy33             = _mm_sub_ps(iy3,jy3);
1391             dz33             = _mm_sub_ps(iz3,jz3);
1392
1393             /* Calculate squared distance and things based on it */
1394             rsq11            = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
1395             rsq12            = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
1396             rsq13            = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
1397             rsq21            = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
1398             rsq22            = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
1399             rsq23            = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
1400             rsq31            = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
1401             rsq32            = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
1402             rsq33            = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
1403
1404             rinv11           = gmx_mm_invsqrt_ps(rsq11);
1405             rinv12           = gmx_mm_invsqrt_ps(rsq12);
1406             rinv13           = gmx_mm_invsqrt_ps(rsq13);
1407             rinv21           = gmx_mm_invsqrt_ps(rsq21);
1408             rinv22           = gmx_mm_invsqrt_ps(rsq22);
1409             rinv23           = gmx_mm_invsqrt_ps(rsq23);
1410             rinv31           = gmx_mm_invsqrt_ps(rsq31);
1411             rinv32           = gmx_mm_invsqrt_ps(rsq32);
1412             rinv33           = gmx_mm_invsqrt_ps(rsq33);
1413
1414             fjx1             = _mm_setzero_ps();
1415             fjy1             = _mm_setzero_ps();
1416             fjz1             = _mm_setzero_ps();
1417             fjx2             = _mm_setzero_ps();
1418             fjy2             = _mm_setzero_ps();
1419             fjz2             = _mm_setzero_ps();
1420             fjx3             = _mm_setzero_ps();
1421             fjy3             = _mm_setzero_ps();
1422             fjz3             = _mm_setzero_ps();
1423
1424             /**************************
1425              * CALCULATE INTERACTIONS *
1426              **************************/
1427
1428             r11              = _mm_mul_ps(rsq11,rinv11);
1429
1430             /* Calculate table index by multiplying r with table scale and truncate to integer */
1431             rt               = _mm_mul_ps(r11,vftabscale);
1432             vfitab           = _mm_cvttps_epi32(rt);
1433 #ifdef __XOP__
1434             vfeps            = _mm_frcz_ps(rt);
1435 #else
1436             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1437 #endif
1438             twovfeps         = _mm_add_ps(vfeps,vfeps);
1439             vfitab           = _mm_slli_epi32(vfitab,2);
1440
1441             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1442             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1443             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1444             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1445             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1446             _MM_TRANSPOSE4_PS(Y,F,G,H);
1447             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1448             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1449             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq11,FF),_mm_mul_ps(vftabscale,rinv11)));
1450
1451             fscal            = felec;
1452
1453              /* Update vectorial force */
1454             fix1             = _mm_macc_ps(dx11,fscal,fix1);
1455             fiy1             = _mm_macc_ps(dy11,fscal,fiy1);
1456             fiz1             = _mm_macc_ps(dz11,fscal,fiz1);
1457
1458             fjx1             = _mm_macc_ps(dx11,fscal,fjx1);
1459             fjy1             = _mm_macc_ps(dy11,fscal,fjy1);
1460             fjz1             = _mm_macc_ps(dz11,fscal,fjz1);
1461
1462             /**************************
1463              * CALCULATE INTERACTIONS *
1464              **************************/
1465
1466             r12              = _mm_mul_ps(rsq12,rinv12);
1467
1468             /* Calculate table index by multiplying r with table scale and truncate to integer */
1469             rt               = _mm_mul_ps(r12,vftabscale);
1470             vfitab           = _mm_cvttps_epi32(rt);
1471 #ifdef __XOP__
1472             vfeps            = _mm_frcz_ps(rt);
1473 #else
1474             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1475 #endif
1476             twovfeps         = _mm_add_ps(vfeps,vfeps);
1477             vfitab           = _mm_slli_epi32(vfitab,2);
1478
1479             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1480             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1481             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1482             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1483             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1484             _MM_TRANSPOSE4_PS(Y,F,G,H);
1485             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1486             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1487             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq12,FF),_mm_mul_ps(vftabscale,rinv12)));
1488
1489             fscal            = felec;
1490
1491              /* Update vectorial force */
1492             fix1             = _mm_macc_ps(dx12,fscal,fix1);
1493             fiy1             = _mm_macc_ps(dy12,fscal,fiy1);
1494             fiz1             = _mm_macc_ps(dz12,fscal,fiz1);
1495
1496             fjx2             = _mm_macc_ps(dx12,fscal,fjx2);
1497             fjy2             = _mm_macc_ps(dy12,fscal,fjy2);
1498             fjz2             = _mm_macc_ps(dz12,fscal,fjz2);
1499
1500             /**************************
1501              * CALCULATE INTERACTIONS *
1502              **************************/
1503
1504             r13              = _mm_mul_ps(rsq13,rinv13);
1505
1506             /* Calculate table index by multiplying r with table scale and truncate to integer */
1507             rt               = _mm_mul_ps(r13,vftabscale);
1508             vfitab           = _mm_cvttps_epi32(rt);
1509 #ifdef __XOP__
1510             vfeps            = _mm_frcz_ps(rt);
1511 #else
1512             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1513 #endif
1514             twovfeps         = _mm_add_ps(vfeps,vfeps);
1515             vfitab           = _mm_slli_epi32(vfitab,2);
1516
1517             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1518             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1519             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1520             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1521             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1522             _MM_TRANSPOSE4_PS(Y,F,G,H);
1523             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1524             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1525             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq13,FF),_mm_mul_ps(vftabscale,rinv13)));
1526
1527             fscal            = felec;
1528
1529              /* Update vectorial force */
1530             fix1             = _mm_macc_ps(dx13,fscal,fix1);
1531             fiy1             = _mm_macc_ps(dy13,fscal,fiy1);
1532             fiz1             = _mm_macc_ps(dz13,fscal,fiz1);
1533
1534             fjx3             = _mm_macc_ps(dx13,fscal,fjx3);
1535             fjy3             = _mm_macc_ps(dy13,fscal,fjy3);
1536             fjz3             = _mm_macc_ps(dz13,fscal,fjz3);
1537
1538             /**************************
1539              * CALCULATE INTERACTIONS *
1540              **************************/
1541
1542             r21              = _mm_mul_ps(rsq21,rinv21);
1543
1544             /* Calculate table index by multiplying r with table scale and truncate to integer */
1545             rt               = _mm_mul_ps(r21,vftabscale);
1546             vfitab           = _mm_cvttps_epi32(rt);
1547 #ifdef __XOP__
1548             vfeps            = _mm_frcz_ps(rt);
1549 #else
1550             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1551 #endif
1552             twovfeps         = _mm_add_ps(vfeps,vfeps);
1553             vfitab           = _mm_slli_epi32(vfitab,2);
1554
1555             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1556             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1557             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1558             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1559             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1560             _MM_TRANSPOSE4_PS(Y,F,G,H);
1561             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1562             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1563             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq21,FF),_mm_mul_ps(vftabscale,rinv21)));
1564
1565             fscal            = felec;
1566
1567              /* Update vectorial force */
1568             fix2             = _mm_macc_ps(dx21,fscal,fix2);
1569             fiy2             = _mm_macc_ps(dy21,fscal,fiy2);
1570             fiz2             = _mm_macc_ps(dz21,fscal,fiz2);
1571
1572             fjx1             = _mm_macc_ps(dx21,fscal,fjx1);
1573             fjy1             = _mm_macc_ps(dy21,fscal,fjy1);
1574             fjz1             = _mm_macc_ps(dz21,fscal,fjz1);
1575
1576             /**************************
1577              * CALCULATE INTERACTIONS *
1578              **************************/
1579
1580             r22              = _mm_mul_ps(rsq22,rinv22);
1581
1582             /* Calculate table index by multiplying r with table scale and truncate to integer */
1583             rt               = _mm_mul_ps(r22,vftabscale);
1584             vfitab           = _mm_cvttps_epi32(rt);
1585 #ifdef __XOP__
1586             vfeps            = _mm_frcz_ps(rt);
1587 #else
1588             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1589 #endif
1590             twovfeps         = _mm_add_ps(vfeps,vfeps);
1591             vfitab           = _mm_slli_epi32(vfitab,2);
1592
1593             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1594             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1595             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1596             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1597             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1598             _MM_TRANSPOSE4_PS(Y,F,G,H);
1599             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1600             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1601             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq22,FF),_mm_mul_ps(vftabscale,rinv22)));
1602
1603             fscal            = felec;
1604
1605              /* Update vectorial force */
1606             fix2             = _mm_macc_ps(dx22,fscal,fix2);
1607             fiy2             = _mm_macc_ps(dy22,fscal,fiy2);
1608             fiz2             = _mm_macc_ps(dz22,fscal,fiz2);
1609
1610             fjx2             = _mm_macc_ps(dx22,fscal,fjx2);
1611             fjy2             = _mm_macc_ps(dy22,fscal,fjy2);
1612             fjz2             = _mm_macc_ps(dz22,fscal,fjz2);
1613
1614             /**************************
1615              * CALCULATE INTERACTIONS *
1616              **************************/
1617
1618             r23              = _mm_mul_ps(rsq23,rinv23);
1619
1620             /* Calculate table index by multiplying r with table scale and truncate to integer */
1621             rt               = _mm_mul_ps(r23,vftabscale);
1622             vfitab           = _mm_cvttps_epi32(rt);
1623 #ifdef __XOP__
1624             vfeps            = _mm_frcz_ps(rt);
1625 #else
1626             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1627 #endif
1628             twovfeps         = _mm_add_ps(vfeps,vfeps);
1629             vfitab           = _mm_slli_epi32(vfitab,2);
1630
1631             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1632             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1633             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1634             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1635             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1636             _MM_TRANSPOSE4_PS(Y,F,G,H);
1637             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1638             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1639             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq23,FF),_mm_mul_ps(vftabscale,rinv23)));
1640
1641             fscal            = felec;
1642
1643              /* Update vectorial force */
1644             fix2             = _mm_macc_ps(dx23,fscal,fix2);
1645             fiy2             = _mm_macc_ps(dy23,fscal,fiy2);
1646             fiz2             = _mm_macc_ps(dz23,fscal,fiz2);
1647
1648             fjx3             = _mm_macc_ps(dx23,fscal,fjx3);
1649             fjy3             = _mm_macc_ps(dy23,fscal,fjy3);
1650             fjz3             = _mm_macc_ps(dz23,fscal,fjz3);
1651
1652             /**************************
1653              * CALCULATE INTERACTIONS *
1654              **************************/
1655
1656             r31              = _mm_mul_ps(rsq31,rinv31);
1657
1658             /* Calculate table index by multiplying r with table scale and truncate to integer */
1659             rt               = _mm_mul_ps(r31,vftabscale);
1660             vfitab           = _mm_cvttps_epi32(rt);
1661 #ifdef __XOP__
1662             vfeps            = _mm_frcz_ps(rt);
1663 #else
1664             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1665 #endif
1666             twovfeps         = _mm_add_ps(vfeps,vfeps);
1667             vfitab           = _mm_slli_epi32(vfitab,2);
1668
1669             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1670             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1671             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1672             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1673             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1674             _MM_TRANSPOSE4_PS(Y,F,G,H);
1675             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1676             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1677             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq31,FF),_mm_mul_ps(vftabscale,rinv31)));
1678
1679             fscal            = felec;
1680
1681              /* Update vectorial force */
1682             fix3             = _mm_macc_ps(dx31,fscal,fix3);
1683             fiy3             = _mm_macc_ps(dy31,fscal,fiy3);
1684             fiz3             = _mm_macc_ps(dz31,fscal,fiz3);
1685
1686             fjx1             = _mm_macc_ps(dx31,fscal,fjx1);
1687             fjy1             = _mm_macc_ps(dy31,fscal,fjy1);
1688             fjz1             = _mm_macc_ps(dz31,fscal,fjz1);
1689
1690             /**************************
1691              * CALCULATE INTERACTIONS *
1692              **************************/
1693
1694             r32              = _mm_mul_ps(rsq32,rinv32);
1695
1696             /* Calculate table index by multiplying r with table scale and truncate to integer */
1697             rt               = _mm_mul_ps(r32,vftabscale);
1698             vfitab           = _mm_cvttps_epi32(rt);
1699 #ifdef __XOP__
1700             vfeps            = _mm_frcz_ps(rt);
1701 #else
1702             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1703 #endif
1704             twovfeps         = _mm_add_ps(vfeps,vfeps);
1705             vfitab           = _mm_slli_epi32(vfitab,2);
1706
1707             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1708             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1709             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1710             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1711             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1712             _MM_TRANSPOSE4_PS(Y,F,G,H);
1713             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1714             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1715             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq32,FF),_mm_mul_ps(vftabscale,rinv32)));
1716
1717             fscal            = felec;
1718
1719              /* Update vectorial force */
1720             fix3             = _mm_macc_ps(dx32,fscal,fix3);
1721             fiy3             = _mm_macc_ps(dy32,fscal,fiy3);
1722             fiz3             = _mm_macc_ps(dz32,fscal,fiz3);
1723
1724             fjx2             = _mm_macc_ps(dx32,fscal,fjx2);
1725             fjy2             = _mm_macc_ps(dy32,fscal,fjy2);
1726             fjz2             = _mm_macc_ps(dz32,fscal,fjz2);
1727
1728             /**************************
1729              * CALCULATE INTERACTIONS *
1730              **************************/
1731
1732             r33              = _mm_mul_ps(rsq33,rinv33);
1733
1734             /* Calculate table index by multiplying r with table scale and truncate to integer */
1735             rt               = _mm_mul_ps(r33,vftabscale);
1736             vfitab           = _mm_cvttps_epi32(rt);
1737 #ifdef __XOP__
1738             vfeps            = _mm_frcz_ps(rt);
1739 #else
1740             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1741 #endif
1742             twovfeps         = _mm_add_ps(vfeps,vfeps);
1743             vfitab           = _mm_slli_epi32(vfitab,2);
1744
1745             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1746             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1747             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1748             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1749             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1750             _MM_TRANSPOSE4_PS(Y,F,G,H);
1751             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1752             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1753             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq33,FF),_mm_mul_ps(vftabscale,rinv33)));
1754
1755             fscal            = felec;
1756
1757              /* Update vectorial force */
1758             fix3             = _mm_macc_ps(dx33,fscal,fix3);
1759             fiy3             = _mm_macc_ps(dy33,fscal,fiy3);
1760             fiz3             = _mm_macc_ps(dz33,fscal,fiz3);
1761
1762             fjx3             = _mm_macc_ps(dx33,fscal,fjx3);
1763             fjy3             = _mm_macc_ps(dy33,fscal,fjy3);
1764             fjz3             = _mm_macc_ps(dz33,fscal,fjz3);
1765
1766             fjptrA             = f+j_coord_offsetA;
1767             fjptrB             = f+j_coord_offsetB;
1768             fjptrC             = f+j_coord_offsetC;
1769             fjptrD             = f+j_coord_offsetD;
1770
1771             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1772                                                    fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1773
1774             /* Inner loop uses 378 flops */
1775         }
1776
1777         if(jidx<j_index_end)
1778         {
1779
1780             /* Get j neighbor index, and coordinate index */
1781             jnrlistA         = jjnr[jidx];
1782             jnrlistB         = jjnr[jidx+1];
1783             jnrlistC         = jjnr[jidx+2];
1784             jnrlistD         = jjnr[jidx+3];
1785             /* Sign of each element will be negative for non-real atoms.
1786              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1787              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1788              */
1789             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1790             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1791             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1792             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1793             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1794             j_coord_offsetA  = DIM*jnrA;
1795             j_coord_offsetB  = DIM*jnrB;
1796             j_coord_offsetC  = DIM*jnrC;
1797             j_coord_offsetD  = DIM*jnrD;
1798
1799             /* load j atom coordinates */
1800             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1801                                               x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1802                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1803
1804             /* Calculate displacement vector */
1805             dx11             = _mm_sub_ps(ix1,jx1);
1806             dy11             = _mm_sub_ps(iy1,jy1);
1807             dz11             = _mm_sub_ps(iz1,jz1);
1808             dx12             = _mm_sub_ps(ix1,jx2);
1809             dy12             = _mm_sub_ps(iy1,jy2);
1810             dz12             = _mm_sub_ps(iz1,jz2);
1811             dx13             = _mm_sub_ps(ix1,jx3);
1812             dy13             = _mm_sub_ps(iy1,jy3);
1813             dz13             = _mm_sub_ps(iz1,jz3);
1814             dx21             = _mm_sub_ps(ix2,jx1);
1815             dy21             = _mm_sub_ps(iy2,jy1);
1816             dz21             = _mm_sub_ps(iz2,jz1);
1817             dx22             = _mm_sub_ps(ix2,jx2);
1818             dy22             = _mm_sub_ps(iy2,jy2);
1819             dz22             = _mm_sub_ps(iz2,jz2);
1820             dx23             = _mm_sub_ps(ix2,jx3);
1821             dy23             = _mm_sub_ps(iy2,jy3);
1822             dz23             = _mm_sub_ps(iz2,jz3);
1823             dx31             = _mm_sub_ps(ix3,jx1);
1824             dy31             = _mm_sub_ps(iy3,jy1);
1825             dz31             = _mm_sub_ps(iz3,jz1);
1826             dx32             = _mm_sub_ps(ix3,jx2);
1827             dy32             = _mm_sub_ps(iy3,jy2);
1828             dz32             = _mm_sub_ps(iz3,jz2);
1829             dx33             = _mm_sub_ps(ix3,jx3);
1830             dy33             = _mm_sub_ps(iy3,jy3);
1831             dz33             = _mm_sub_ps(iz3,jz3);
1832
1833             /* Calculate squared distance and things based on it */
1834             rsq11            = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
1835             rsq12            = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
1836             rsq13            = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
1837             rsq21            = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
1838             rsq22            = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
1839             rsq23            = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
1840             rsq31            = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
1841             rsq32            = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
1842             rsq33            = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
1843
1844             rinv11           = gmx_mm_invsqrt_ps(rsq11);
1845             rinv12           = gmx_mm_invsqrt_ps(rsq12);
1846             rinv13           = gmx_mm_invsqrt_ps(rsq13);
1847             rinv21           = gmx_mm_invsqrt_ps(rsq21);
1848             rinv22           = gmx_mm_invsqrt_ps(rsq22);
1849             rinv23           = gmx_mm_invsqrt_ps(rsq23);
1850             rinv31           = gmx_mm_invsqrt_ps(rsq31);
1851             rinv32           = gmx_mm_invsqrt_ps(rsq32);
1852             rinv33           = gmx_mm_invsqrt_ps(rsq33);
1853
1854             fjx1             = _mm_setzero_ps();
1855             fjy1             = _mm_setzero_ps();
1856             fjz1             = _mm_setzero_ps();
1857             fjx2             = _mm_setzero_ps();
1858             fjy2             = _mm_setzero_ps();
1859             fjz2             = _mm_setzero_ps();
1860             fjx3             = _mm_setzero_ps();
1861             fjy3             = _mm_setzero_ps();
1862             fjz3             = _mm_setzero_ps();
1863
1864             /**************************
1865              * CALCULATE INTERACTIONS *
1866              **************************/
1867
1868             r11              = _mm_mul_ps(rsq11,rinv11);
1869             r11              = _mm_andnot_ps(dummy_mask,r11);
1870
1871             /* Calculate table index by multiplying r with table scale and truncate to integer */
1872             rt               = _mm_mul_ps(r11,vftabscale);
1873             vfitab           = _mm_cvttps_epi32(rt);
1874 #ifdef __XOP__
1875             vfeps            = _mm_frcz_ps(rt);
1876 #else
1877             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1878 #endif
1879             twovfeps         = _mm_add_ps(vfeps,vfeps);
1880             vfitab           = _mm_slli_epi32(vfitab,2);
1881
1882             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1883             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1884             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1885             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1886             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1887             _MM_TRANSPOSE4_PS(Y,F,G,H);
1888             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1889             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1890             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq11,FF),_mm_mul_ps(vftabscale,rinv11)));
1891
1892             fscal            = felec;
1893
1894             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1895
1896              /* Update vectorial force */
1897             fix1             = _mm_macc_ps(dx11,fscal,fix1);
1898             fiy1             = _mm_macc_ps(dy11,fscal,fiy1);
1899             fiz1             = _mm_macc_ps(dz11,fscal,fiz1);
1900
1901             fjx1             = _mm_macc_ps(dx11,fscal,fjx1);
1902             fjy1             = _mm_macc_ps(dy11,fscal,fjy1);
1903             fjz1             = _mm_macc_ps(dz11,fscal,fjz1);
1904
1905             /**************************
1906              * CALCULATE INTERACTIONS *
1907              **************************/
1908
1909             r12              = _mm_mul_ps(rsq12,rinv12);
1910             r12              = _mm_andnot_ps(dummy_mask,r12);
1911
1912             /* Calculate table index by multiplying r with table scale and truncate to integer */
1913             rt               = _mm_mul_ps(r12,vftabscale);
1914             vfitab           = _mm_cvttps_epi32(rt);
1915 #ifdef __XOP__
1916             vfeps            = _mm_frcz_ps(rt);
1917 #else
1918             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1919 #endif
1920             twovfeps         = _mm_add_ps(vfeps,vfeps);
1921             vfitab           = _mm_slli_epi32(vfitab,2);
1922
1923             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1924             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1925             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1926             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1927             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1928             _MM_TRANSPOSE4_PS(Y,F,G,H);
1929             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1930             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1931             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq12,FF),_mm_mul_ps(vftabscale,rinv12)));
1932
1933             fscal            = felec;
1934
1935             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1936
1937              /* Update vectorial force */
1938             fix1             = _mm_macc_ps(dx12,fscal,fix1);
1939             fiy1             = _mm_macc_ps(dy12,fscal,fiy1);
1940             fiz1             = _mm_macc_ps(dz12,fscal,fiz1);
1941
1942             fjx2             = _mm_macc_ps(dx12,fscal,fjx2);
1943             fjy2             = _mm_macc_ps(dy12,fscal,fjy2);
1944             fjz2             = _mm_macc_ps(dz12,fscal,fjz2);
1945
1946             /**************************
1947              * CALCULATE INTERACTIONS *
1948              **************************/
1949
1950             r13              = _mm_mul_ps(rsq13,rinv13);
1951             r13              = _mm_andnot_ps(dummy_mask,r13);
1952
1953             /* Calculate table index by multiplying r with table scale and truncate to integer */
1954             rt               = _mm_mul_ps(r13,vftabscale);
1955             vfitab           = _mm_cvttps_epi32(rt);
1956 #ifdef __XOP__
1957             vfeps            = _mm_frcz_ps(rt);
1958 #else
1959             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1960 #endif
1961             twovfeps         = _mm_add_ps(vfeps,vfeps);
1962             vfitab           = _mm_slli_epi32(vfitab,2);
1963
1964             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1965             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
1966             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
1967             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
1968             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
1969             _MM_TRANSPOSE4_PS(Y,F,G,H);
1970             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
1971             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
1972             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq13,FF),_mm_mul_ps(vftabscale,rinv13)));
1973
1974             fscal            = felec;
1975
1976             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1977
1978              /* Update vectorial force */
1979             fix1             = _mm_macc_ps(dx13,fscal,fix1);
1980             fiy1             = _mm_macc_ps(dy13,fscal,fiy1);
1981             fiz1             = _mm_macc_ps(dz13,fscal,fiz1);
1982
1983             fjx3             = _mm_macc_ps(dx13,fscal,fjx3);
1984             fjy3             = _mm_macc_ps(dy13,fscal,fjy3);
1985             fjz3             = _mm_macc_ps(dz13,fscal,fjz3);
1986
1987             /**************************
1988              * CALCULATE INTERACTIONS *
1989              **************************/
1990
1991             r21              = _mm_mul_ps(rsq21,rinv21);
1992             r21              = _mm_andnot_ps(dummy_mask,r21);
1993
1994             /* Calculate table index by multiplying r with table scale and truncate to integer */
1995             rt               = _mm_mul_ps(r21,vftabscale);
1996             vfitab           = _mm_cvttps_epi32(rt);
1997 #ifdef __XOP__
1998             vfeps            = _mm_frcz_ps(rt);
1999 #else
2000             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2001 #endif
2002             twovfeps         = _mm_add_ps(vfeps,vfeps);
2003             vfitab           = _mm_slli_epi32(vfitab,2);
2004
2005             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2006             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2007             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2008             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2009             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2010             _MM_TRANSPOSE4_PS(Y,F,G,H);
2011             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2012             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2013             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq21,FF),_mm_mul_ps(vftabscale,rinv21)));
2014
2015             fscal            = felec;
2016
2017             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2018
2019              /* Update vectorial force */
2020             fix2             = _mm_macc_ps(dx21,fscal,fix2);
2021             fiy2             = _mm_macc_ps(dy21,fscal,fiy2);
2022             fiz2             = _mm_macc_ps(dz21,fscal,fiz2);
2023
2024             fjx1             = _mm_macc_ps(dx21,fscal,fjx1);
2025             fjy1             = _mm_macc_ps(dy21,fscal,fjy1);
2026             fjz1             = _mm_macc_ps(dz21,fscal,fjz1);
2027
2028             /**************************
2029              * CALCULATE INTERACTIONS *
2030              **************************/
2031
2032             r22              = _mm_mul_ps(rsq22,rinv22);
2033             r22              = _mm_andnot_ps(dummy_mask,r22);
2034
2035             /* Calculate table index by multiplying r with table scale and truncate to integer */
2036             rt               = _mm_mul_ps(r22,vftabscale);
2037             vfitab           = _mm_cvttps_epi32(rt);
2038 #ifdef __XOP__
2039             vfeps            = _mm_frcz_ps(rt);
2040 #else
2041             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2042 #endif
2043             twovfeps         = _mm_add_ps(vfeps,vfeps);
2044             vfitab           = _mm_slli_epi32(vfitab,2);
2045
2046             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2047             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2048             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2049             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2050             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2051             _MM_TRANSPOSE4_PS(Y,F,G,H);
2052             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2053             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2054             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq22,FF),_mm_mul_ps(vftabscale,rinv22)));
2055
2056             fscal            = felec;
2057
2058             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2059
2060              /* Update vectorial force */
2061             fix2             = _mm_macc_ps(dx22,fscal,fix2);
2062             fiy2             = _mm_macc_ps(dy22,fscal,fiy2);
2063             fiz2             = _mm_macc_ps(dz22,fscal,fiz2);
2064
2065             fjx2             = _mm_macc_ps(dx22,fscal,fjx2);
2066             fjy2             = _mm_macc_ps(dy22,fscal,fjy2);
2067             fjz2             = _mm_macc_ps(dz22,fscal,fjz2);
2068
2069             /**************************
2070              * CALCULATE INTERACTIONS *
2071              **************************/
2072
2073             r23              = _mm_mul_ps(rsq23,rinv23);
2074             r23              = _mm_andnot_ps(dummy_mask,r23);
2075
2076             /* Calculate table index by multiplying r with table scale and truncate to integer */
2077             rt               = _mm_mul_ps(r23,vftabscale);
2078             vfitab           = _mm_cvttps_epi32(rt);
2079 #ifdef __XOP__
2080             vfeps            = _mm_frcz_ps(rt);
2081 #else
2082             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2083 #endif
2084             twovfeps         = _mm_add_ps(vfeps,vfeps);
2085             vfitab           = _mm_slli_epi32(vfitab,2);
2086
2087             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2088             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2089             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2090             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2091             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2092             _MM_TRANSPOSE4_PS(Y,F,G,H);
2093             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2094             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2095             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq23,FF),_mm_mul_ps(vftabscale,rinv23)));
2096
2097             fscal            = felec;
2098
2099             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2100
2101              /* Update vectorial force */
2102             fix2             = _mm_macc_ps(dx23,fscal,fix2);
2103             fiy2             = _mm_macc_ps(dy23,fscal,fiy2);
2104             fiz2             = _mm_macc_ps(dz23,fscal,fiz2);
2105
2106             fjx3             = _mm_macc_ps(dx23,fscal,fjx3);
2107             fjy3             = _mm_macc_ps(dy23,fscal,fjy3);
2108             fjz3             = _mm_macc_ps(dz23,fscal,fjz3);
2109
2110             /**************************
2111              * CALCULATE INTERACTIONS *
2112              **************************/
2113
2114             r31              = _mm_mul_ps(rsq31,rinv31);
2115             r31              = _mm_andnot_ps(dummy_mask,r31);
2116
2117             /* Calculate table index by multiplying r with table scale and truncate to integer */
2118             rt               = _mm_mul_ps(r31,vftabscale);
2119             vfitab           = _mm_cvttps_epi32(rt);
2120 #ifdef __XOP__
2121             vfeps            = _mm_frcz_ps(rt);
2122 #else
2123             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2124 #endif
2125             twovfeps         = _mm_add_ps(vfeps,vfeps);
2126             vfitab           = _mm_slli_epi32(vfitab,2);
2127
2128             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2129             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2130             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2131             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2132             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2133             _MM_TRANSPOSE4_PS(Y,F,G,H);
2134             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2135             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2136             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq31,FF),_mm_mul_ps(vftabscale,rinv31)));
2137
2138             fscal            = felec;
2139
2140             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2141
2142              /* Update vectorial force */
2143             fix3             = _mm_macc_ps(dx31,fscal,fix3);
2144             fiy3             = _mm_macc_ps(dy31,fscal,fiy3);
2145             fiz3             = _mm_macc_ps(dz31,fscal,fiz3);
2146
2147             fjx1             = _mm_macc_ps(dx31,fscal,fjx1);
2148             fjy1             = _mm_macc_ps(dy31,fscal,fjy1);
2149             fjz1             = _mm_macc_ps(dz31,fscal,fjz1);
2150
2151             /**************************
2152              * CALCULATE INTERACTIONS *
2153              **************************/
2154
2155             r32              = _mm_mul_ps(rsq32,rinv32);
2156             r32              = _mm_andnot_ps(dummy_mask,r32);
2157
2158             /* Calculate table index by multiplying r with table scale and truncate to integer */
2159             rt               = _mm_mul_ps(r32,vftabscale);
2160             vfitab           = _mm_cvttps_epi32(rt);
2161 #ifdef __XOP__
2162             vfeps            = _mm_frcz_ps(rt);
2163 #else
2164             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2165 #endif
2166             twovfeps         = _mm_add_ps(vfeps,vfeps);
2167             vfitab           = _mm_slli_epi32(vfitab,2);
2168
2169             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2170             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2171             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2172             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2173             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2174             _MM_TRANSPOSE4_PS(Y,F,G,H);
2175             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2176             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2177             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq32,FF),_mm_mul_ps(vftabscale,rinv32)));
2178
2179             fscal            = felec;
2180
2181             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2182
2183              /* Update vectorial force */
2184             fix3             = _mm_macc_ps(dx32,fscal,fix3);
2185             fiy3             = _mm_macc_ps(dy32,fscal,fiy3);
2186             fiz3             = _mm_macc_ps(dz32,fscal,fiz3);
2187
2188             fjx2             = _mm_macc_ps(dx32,fscal,fjx2);
2189             fjy2             = _mm_macc_ps(dy32,fscal,fjy2);
2190             fjz2             = _mm_macc_ps(dz32,fscal,fjz2);
2191
2192             /**************************
2193              * CALCULATE INTERACTIONS *
2194              **************************/
2195
2196             r33              = _mm_mul_ps(rsq33,rinv33);
2197             r33              = _mm_andnot_ps(dummy_mask,r33);
2198
2199             /* Calculate table index by multiplying r with table scale and truncate to integer */
2200             rt               = _mm_mul_ps(r33,vftabscale);
2201             vfitab           = _mm_cvttps_epi32(rt);
2202 #ifdef __XOP__
2203             vfeps            = _mm_frcz_ps(rt);
2204 #else
2205             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
2206 #endif
2207             twovfeps         = _mm_add_ps(vfeps,vfeps);
2208             vfitab           = _mm_slli_epi32(vfitab,2);
2209
2210             /* CUBIC SPLINE TABLE ELECTROSTATICS */
2211             Y                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,0) );
2212             F                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,1) );
2213             G                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,2) );
2214             H                = _mm_load_ps( vftab + _mm_extract_epi32(vfitab,3) );
2215             _MM_TRANSPOSE4_PS(Y,F,G,H);
2216             Fp               = _mm_macc_ps(vfeps,_mm_macc_ps(H,vfeps,G),F);
2217             FF               = _mm_macc_ps(vfeps,_mm_macc_ps(twovfeps,H,G),Fp);
2218             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq33,FF),_mm_mul_ps(vftabscale,rinv33)));
2219
2220             fscal            = felec;
2221
2222             fscal            = _mm_andnot_ps(dummy_mask,fscal);
2223
2224              /* Update vectorial force */
2225             fix3             = _mm_macc_ps(dx33,fscal,fix3);
2226             fiy3             = _mm_macc_ps(dy33,fscal,fiy3);
2227             fiz3             = _mm_macc_ps(dz33,fscal,fiz3);
2228
2229             fjx3             = _mm_macc_ps(dx33,fscal,fjx3);
2230             fjy3             = _mm_macc_ps(dy33,fscal,fjy3);
2231             fjz3             = _mm_macc_ps(dz33,fscal,fjz3);
2232
2233             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2234             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2235             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2236             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2237
2238             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
2239                                                    fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2240
2241             /* Inner loop uses 387 flops */
2242         }
2243
2244         /* End of innermost loop */
2245
2246         gmx_mm_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
2247                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
2248
2249         /* Increment number of inner iterations */
2250         inneriter                  += j_index_end - j_index_start;
2251
2252         /* Outer loop uses 18 flops */
2253     }
2254
2255     /* Increment number of outer iterations */
2256     outeriter        += nri;
2257
2258     /* Update outer/inner flops */
2259
2260     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_F,outeriter*18 + inneriter*387);
2261 }