src/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecCSTab_VdwNone_GeomP1P1_avx_128_fma_double.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_128_fma_double kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 #include "gmx_math_x86_avx_128_fma_double.h"
  34 #include "kernelutil_x86_avx_128_fma_double.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwNone_GeomP1P1_VF_avx_128_fma_double
  38  * Electrostatics interaction: CubicSplineTable
  39  * VdW interaction:            None
  40  * Geometry:                   Particle-Particle
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecCSTab_VdwNone_GeomP1P1_VF_avx_128_fma_double
  45                     (t_nblist * gmx_restrict                nlist,
  46                      rvec * gmx_restrict                    xx,
  47                      rvec * gmx_restrict                    ff,
  48                      t_forcerec * gmx_restrict              fr,
  49                      t_mdatoms * gmx_restrict               mdatoms,
  50                      nb_kernel_data_t * gmx_restrict        kernel_data,
  51                      t_nrnb * gmx_restrict                  nrnb)
  52 {
  53     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  54      * just 0 for non-waters.
  55      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two 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;
  61     int              j_coord_offsetA,j_coord_offsetB;
  62     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  63     real             rcutoff_scalar;
  64     real             *shiftvec,*fshift,*x,*f;
  65     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  66     int              vdwioffset0;
  67     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  68     int              vdwjidx0A,vdwjidx0B;
  69     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  70     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  71     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  72     real             *charge;
  73     __m128i          vfitab;
  74     __m128i          ifour       = _mm_set1_epi32(4);
  75     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
  76     real             *vftab;
  77     __m128d          dummy_mask,cutoff_mask;
  78     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
  79     __m128d          one     = _mm_set1_pd(1.0);
  80     __m128d          two     = _mm_set1_pd(2.0);
  81     x                = xx[0];
  82     f                = ff[0];
  83
  84     nri              = nlist->nri;
  85     iinr             = nlist->iinr;
  86     jindex           = nlist->jindex;
  87     jjnr             = nlist->jjnr;
  88     shiftidx         = nlist->shift;
  89     gid              = nlist->gid;
  90     shiftvec         = fr->shift_vec[0];
  91     fshift           = fr->fshift[0];
  92     facel            = _mm_set1_pd(fr->epsfac);
  93     charge           = mdatoms->chargeA;
  94
  95     vftab            = kernel_data->table_elec->data;
  96     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
  97
  98     /* Avoid stupid compiler warnings */
  99     jnrA = jnrB = 0;
 100     j_coord_offsetA = 0;
 101     j_coord_offsetB = 0;
 102
 103     outeriter        = 0;
 104     inneriter        = 0;
 105
 106     /* Start outer loop over neighborlists */
 107     for(iidx=0; iidx<nri; iidx++)
 108     {
 109         /* Load shift vector for this list */
 110         i_shift_offset   = DIM*shiftidx[iidx];
 111
 112         /* Load limits for loop over neighbors */
 113         j_index_start    = jindex[iidx];
 114         j_index_end      = jindex[iidx+1];
 115
 116         /* Get outer coordinate index */
 117         inr              = iinr[iidx];
 118         i_coord_offset   = DIM*inr;
 119
 120         /* Load i particle coords and add shift vector */
 121         gmx_mm_load_shift_and_1rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 122
 123         fix0             = _mm_setzero_pd();
 124         fiy0             = _mm_setzero_pd();
 125         fiz0             = _mm_setzero_pd();
 126
 127         /* Load parameters for i particles */
 128         iq0              = _mm_mul_pd(facel,_mm_load1_pd(charge+inr+0));
 129
 130         /* Reset potential sums */
 131         velecsum         = _mm_setzero_pd();
 132
 133         /* Start inner kernel loop */
 134         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 135         {
 136
 137             /* Get j neighbor index, and coordinate index */
 138             jnrA             = jjnr[jidx];
 139             jnrB             = jjnr[jidx+1];
 140             j_coord_offsetA  = DIM*jnrA;
 141             j_coord_offsetB  = DIM*jnrB;
 142
 143             /* load j atom coordinates */
 144             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 145                                               &jx0,&jy0,&jz0);
 146
 147             /* Calculate displacement vector */
 148             dx00             = _mm_sub_pd(ix0,jx0);
 149             dy00             = _mm_sub_pd(iy0,jy0);
 150             dz00             = _mm_sub_pd(iz0,jz0);
 151
 152             /* Calculate squared distance and things based on it */
 153             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 154
 155             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 156
 157             /* Load parameters for j particles */
 158             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 159
 160             /**************************
 161              * CALCULATE INTERACTIONS *
 162              **************************/
 163
 164             r00              = _mm_mul_pd(rsq00,rinv00);
 165
 166             /* Compute parameters for interactions between i and j atoms */
 167             qq00             = _mm_mul_pd(iq0,jq0);
 168
 169             /* Calculate table index by multiplying r with table scale and truncate to integer */
 170             rt               = _mm_mul_pd(r00,vftabscale);
 171             vfitab           = _mm_cvttpd_epi32(rt);
 172 #ifdef __XOP__
 173             vfeps            = _mm_frcz_pd(rt);
 174 #else
 175             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 176 #endif
 177             twovfeps         = _mm_add_pd(vfeps,vfeps);
 178             vfitab           = _mm_slli_epi32(vfitab,2);
 179
 180             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 181             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 182             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 183             GMX_MM_TRANSPOSE2_PD(Y,F);
 184             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 185             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 186             GMX_MM_TRANSPOSE2_PD(G,H);
 187             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 188             VV               = _mm_macc_pd(vfeps,Fp,Y);
 189             velec            = _mm_mul_pd(qq00,VV);
 190             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 191             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
 192
 193             /* Update potential sum for this i atom from the interaction with this j atom. */
 194             velecsum         = _mm_add_pd(velecsum,velec);
 195
 196             fscal            = felec;
 197
 198             /* Update vectorial force */
 199             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 200             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 201             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 202
 203             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,
 204                                                    _mm_mul_pd(dx00,fscal),
 205                                                    _mm_mul_pd(dy00,fscal),
 206                                                    _mm_mul_pd(dz00,fscal));
 207
 208             /* Inner loop uses 46 flops */
 209         }
 210
 211         if(jidx<j_index_end)
 212         {
 213
 214             jnrA             = jjnr[jidx];
 215             j_coord_offsetA  = DIM*jnrA;
 216
 217             /* load j atom coordinates */
 218             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 219                                               &jx0,&jy0,&jz0);
 220
 221             /* Calculate displacement vector */
 222             dx00             = _mm_sub_pd(ix0,jx0);
 223             dy00             = _mm_sub_pd(iy0,jy0);
 224             dz00             = _mm_sub_pd(iz0,jz0);
 225
 226             /* Calculate squared distance and things based on it */
 227             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 228
 229             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 230
 231             /* Load parameters for j particles */
 232             jq0              = _mm_load_sd(charge+jnrA+0);
 233
 234             /**************************
 235              * CALCULATE INTERACTIONS *
 236              **************************/
 237
 238             r00              = _mm_mul_pd(rsq00,rinv00);
 239
 240             /* Compute parameters for interactions between i and j atoms */
 241             qq00             = _mm_mul_pd(iq0,jq0);
 242
 243             /* Calculate table index by multiplying r with table scale and truncate to integer */
 244             rt               = _mm_mul_pd(r00,vftabscale);
 245             vfitab           = _mm_cvttpd_epi32(rt);
 246 #ifdef __XOP__
 247             vfeps            = _mm_frcz_pd(rt);
 248 #else
 249             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 250 #endif
 251             twovfeps         = _mm_add_pd(vfeps,vfeps);
 252             vfitab           = _mm_slli_epi32(vfitab,2);
 253
 254             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 255             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 256             F                = _mm_setzero_pd();
 257             GMX_MM_TRANSPOSE2_PD(Y,F);
 258             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 259             H                = _mm_setzero_pd();
 260             GMX_MM_TRANSPOSE2_PD(G,H);
 261             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 262             VV               = _mm_macc_pd(vfeps,Fp,Y);
 263             velec            = _mm_mul_pd(qq00,VV);
 264             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 265             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
 266
 267             /* Update potential sum for this i atom from the interaction with this j atom. */
 268             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 269             velecsum         = _mm_add_pd(velecsum,velec);
 270
 271             fscal            = felec;
 272
 273             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 274
 275             /* Update vectorial force */
 276             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 277             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 278             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 279
 280             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,
 281                                                    _mm_mul_pd(dx00,fscal),
 282                                                    _mm_mul_pd(dy00,fscal),
 283                                                    _mm_mul_pd(dz00,fscal));
 284
 285             /* Inner loop uses 46 flops */
 286         }
 287
 288         /* End of innermost loop */
 289
 290         gmx_mm_update_iforce_1atom_swizzle_pd(fix0,fiy0,fiz0,
 291                                               f+i_coord_offset,fshift+i_shift_offset);
 292
 293         ggid                        = gid[iidx];
 294         /* Update potential energies */
 295         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 296
 297         /* Increment number of inner iterations */
 298         inneriter                  += j_index_end - j_index_start;
 299
 300         /* Outer loop uses 8 flops */
 301     }
 302
 303     /* Increment number of outer iterations */
 304     outeriter        += nri;
 305
 306     /* Update outer/inner flops */
 307
 308     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*8 + inneriter*46);
 309 }
 310 /*
 311  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwNone_GeomP1P1_F_avx_128_fma_double
 312  * Electrostatics interaction: CubicSplineTable
 313  * VdW interaction:            None
 314  * Geometry:                   Particle-Particle
 315  * Calculate force/pot:        Force
 316  */
 317 void
 318 nb_kernel_ElecCSTab_VdwNone_GeomP1P1_F_avx_128_fma_double
 319                     (t_nblist * gmx_restrict                nlist,
 320                      rvec * gmx_restrict                    xx,
 321                      rvec * gmx_restrict                    ff,
 322                      t_forcerec * gmx_restrict              fr,
 323                      t_mdatoms * gmx_restrict               mdatoms,
 324                      nb_kernel_data_t * gmx_restrict        kernel_data,
 325                      t_nrnb * gmx_restrict                  nrnb)
 326 {
 327     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 328      * just 0 for non-waters.
 329      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 330      * jnr indices corresponding to data put in the four positions in the SIMD register.
 331      */
 332     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 333     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 334     int              jnrA,jnrB;
 335     int              j_coord_offsetA,j_coord_offsetB;
 336     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 337     real             rcutoff_scalar;
 338     real             *shiftvec,*fshift,*x,*f;
 339     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 340     int              vdwioffset0;
 341     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 342     int              vdwjidx0A,vdwjidx0B;
 343     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 344     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 345     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 346     real             *charge;
 347     __m128i          vfitab;
 348     __m128i          ifour       = _mm_set1_epi32(4);
 349     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 350     real             *vftab;
 351     __m128d          dummy_mask,cutoff_mask;
 352     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 353     __m128d          one     = _mm_set1_pd(1.0);
 354     __m128d          two     = _mm_set1_pd(2.0);
 355     x                = xx[0];
 356     f                = ff[0];
 357
 358     nri              = nlist->nri;
 359     iinr             = nlist->iinr;
 360     jindex           = nlist->jindex;
 361     jjnr             = nlist->jjnr;
 362     shiftidx         = nlist->shift;
 363     gid              = nlist->gid;
 364     shiftvec         = fr->shift_vec[0];
 365     fshift           = fr->fshift[0];
 366     facel            = _mm_set1_pd(fr->epsfac);
 367     charge           = mdatoms->chargeA;
 368
 369     vftab            = kernel_data->table_elec->data;
 370     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
 371
 372     /* Avoid stupid compiler warnings */
 373     jnrA = jnrB = 0;
 374     j_coord_offsetA = 0;
 375     j_coord_offsetB = 0;
 376
 377     outeriter        = 0;
 378     inneriter        = 0;
 379
 380     /* Start outer loop over neighborlists */
 381     for(iidx=0; iidx<nri; iidx++)
 382     {
 383         /* Load shift vector for this list */
 384         i_shift_offset   = DIM*shiftidx[iidx];
 385
 386         /* Load limits for loop over neighbors */
 387         j_index_start    = jindex[iidx];
 388         j_index_end      = jindex[iidx+1];
 389
 390         /* Get outer coordinate index */
 391         inr              = iinr[iidx];
 392         i_coord_offset   = DIM*inr;
 393
 394         /* Load i particle coords and add shift vector */
 395         gmx_mm_load_shift_and_1rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 396
 397         fix0             = _mm_setzero_pd();
 398         fiy0             = _mm_setzero_pd();
 399         fiz0             = _mm_setzero_pd();
 400
 401         /* Load parameters for i particles */
 402         iq0              = _mm_mul_pd(facel,_mm_load1_pd(charge+inr+0));
 403
 404         /* Start inner kernel loop */
 405         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 406         {
 407
 408             /* Get j neighbor index, and coordinate index */
 409             jnrA             = jjnr[jidx];
 410             jnrB             = jjnr[jidx+1];
 411             j_coord_offsetA  = DIM*jnrA;
 412             j_coord_offsetB  = DIM*jnrB;
 413
 414             /* load j atom coordinates */
 415             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 416                                               &jx0,&jy0,&jz0);
 417
 418             /* Calculate displacement vector */
 419             dx00             = _mm_sub_pd(ix0,jx0);
 420             dy00             = _mm_sub_pd(iy0,jy0);
 421             dz00             = _mm_sub_pd(iz0,jz0);
 422
 423             /* Calculate squared distance and things based on it */
 424             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 425
 426             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 427
 428             /* Load parameters for j particles */
 429             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 430
 431             /**************************
 432              * CALCULATE INTERACTIONS *
 433              **************************/
 434
 435             r00              = _mm_mul_pd(rsq00,rinv00);
 436
 437             /* Compute parameters for interactions between i and j atoms */
 438             qq00             = _mm_mul_pd(iq0,jq0);
 439
 440             /* Calculate table index by multiplying r with table scale and truncate to integer */
 441             rt               = _mm_mul_pd(r00,vftabscale);
 442             vfitab           = _mm_cvttpd_epi32(rt);
 443 #ifdef __XOP__
 444             vfeps            = _mm_frcz_pd(rt);
 445 #else
 446             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 447 #endif
 448             twovfeps         = _mm_add_pd(vfeps,vfeps);
 449             vfitab           = _mm_slli_epi32(vfitab,2);
 450
 451             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 452             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 453             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 454             GMX_MM_TRANSPOSE2_PD(Y,F);
 455             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 456             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 457             GMX_MM_TRANSPOSE2_PD(G,H);
 458             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 459             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 460             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
 461
 462             fscal            = felec;
 463
 464             /* Update vectorial force */
 465             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 466             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 467             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 468
 469             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,
 470                                                    _mm_mul_pd(dx00,fscal),
 471                                                    _mm_mul_pd(dy00,fscal),
 472                                                    _mm_mul_pd(dz00,fscal));
 473
 474             /* Inner loop uses 42 flops */
 475         }
 476
 477         if(jidx<j_index_end)
 478         {
 479
 480             jnrA             = jjnr[jidx];
 481             j_coord_offsetA  = DIM*jnrA;
 482
 483             /* load j atom coordinates */
 484             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 485                                               &jx0,&jy0,&jz0);
 486
 487             /* Calculate displacement vector */
 488             dx00             = _mm_sub_pd(ix0,jx0);
 489             dy00             = _mm_sub_pd(iy0,jy0);
 490             dz00             = _mm_sub_pd(iz0,jz0);
 491
 492             /* Calculate squared distance and things based on it */
 493             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 494
 495             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 496
 497             /* Load parameters for j particles */
 498             jq0              = _mm_load_sd(charge+jnrA+0);
 499
 500             /**************************
 501              * CALCULATE INTERACTIONS *
 502              **************************/
 503
 504             r00              = _mm_mul_pd(rsq00,rinv00);
 505
 506             /* Compute parameters for interactions between i and j atoms */
 507             qq00             = _mm_mul_pd(iq0,jq0);
 508
 509             /* Calculate table index by multiplying r with table scale and truncate to integer */
 510             rt               = _mm_mul_pd(r00,vftabscale);
 511             vfitab           = _mm_cvttpd_epi32(rt);
 512 #ifdef __XOP__
 513             vfeps            = _mm_frcz_pd(rt);
 514 #else
 515             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 516 #endif
 517             twovfeps         = _mm_add_pd(vfeps,vfeps);
 518             vfitab           = _mm_slli_epi32(vfitab,2);
 519
 520             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 521             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 522             F                = _mm_setzero_pd();
 523             GMX_MM_TRANSPOSE2_PD(Y,F);
 524             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 525             H                = _mm_setzero_pd();
 526             GMX_MM_TRANSPOSE2_PD(G,H);
 527             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 528             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 529             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
 530
 531             fscal            = felec;
 532
 533             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 534
 535             /* Update vectorial force */
 536             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 537             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 538             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 539
 540             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,
 541                                                    _mm_mul_pd(dx00,fscal),
 542                                                    _mm_mul_pd(dy00,fscal),
 543                                                    _mm_mul_pd(dz00,fscal));
 544
 545             /* Inner loop uses 42 flops */
 546         }
 547
 548         /* End of innermost loop */
 549
 550         gmx_mm_update_iforce_1atom_swizzle_pd(fix0,fiy0,fiz0,
 551                                               f+i_coord_offset,fshift+i_shift_offset);
 552
 553         /* Increment number of inner iterations */
 554         inneriter                  += j_index_end - j_index_start;
 555
 556         /* Outer loop uses 7 flops */
 557     }
 558
 559     /* Increment number of outer iterations */
 560     outeriter        += nri;
 561
 562     /* Update outer/inner flops */
 563
 564     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*7 + inneriter*42);
 565 }