src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecCoul_VdwLJ_GeomW4P1_avx_128_fma_double.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_128_fma_double kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_128_fma_double.h"
  48 #include "kernelutil_x86_avx_128_fma_double.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_128_fma_double
  52  * Electrostatics interaction: Coulomb
  53  * VdW interaction:            LennardJones
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_128_fma_double
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB;
  75     int              j_coord_offsetA,j_coord_offsetB;
  76     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  77     real             rcutoff_scalar;
  78     real             *shiftvec,*fshift,*x,*f;
  79     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  80     int              vdwioffset0;
  81     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  82     int              vdwioffset1;
  83     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  84     int              vdwioffset2;
  85     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  86     int              vdwioffset3;
  87     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  88     int              vdwjidx0A,vdwjidx0B;
  89     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  90     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  91     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  92     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  93     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  94     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  95     real             *charge;
  96     int              nvdwtype;
  97     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  98     int              *vdwtype;
  99     real             *vdwparam;
 100     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 101     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 102     __m128d          dummy_mask,cutoff_mask;
 103     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 104     __m128d          one     = _mm_set1_pd(1.0);
 105     __m128d          two     = _mm_set1_pd(2.0);
 106     x                = xx[0];
 107     f                = ff[0];
 108
 109     nri              = nlist->nri;
 110     iinr             = nlist->iinr;
 111     jindex           = nlist->jindex;
 112     jjnr             = nlist->jjnr;
 113     shiftidx         = nlist->shift;
 114     gid              = nlist->gid;
 115     shiftvec         = fr->shift_vec[0];
 116     fshift           = fr->fshift[0];
 117     facel            = _mm_set1_pd(fr->epsfac);
 118     charge           = mdatoms->chargeA;
 119     nvdwtype         = fr->ntype;
 120     vdwparam         = fr->nbfp;
 121     vdwtype          = mdatoms->typeA;
 122
 123     /* Setup water-specific parameters */
 124     inr              = nlist->iinr[0];
 125     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 126     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 127     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 128     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 129
 130     /* Avoid stupid compiler warnings */
 131     jnrA = jnrB = 0;
 132     j_coord_offsetA = 0;
 133     j_coord_offsetB = 0;
 134
 135     outeriter        = 0;
 136     inneriter        = 0;
 137
 138     /* Start outer loop over neighborlists */
 139     for(iidx=0; iidx<nri; iidx++)
 140     {
 141         /* Load shift vector for this list */
 142         i_shift_offset   = DIM*shiftidx[iidx];
 143
 144         /* Load limits for loop over neighbors */
 145         j_index_start    = jindex[iidx];
 146         j_index_end      = jindex[iidx+1];
 147
 148         /* Get outer coordinate index */
 149         inr              = iinr[iidx];
 150         i_coord_offset   = DIM*inr;
 151
 152         /* Load i particle coords and add shift vector */
 153         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 154                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 155
 156         fix0             = _mm_setzero_pd();
 157         fiy0             = _mm_setzero_pd();
 158         fiz0             = _mm_setzero_pd();
 159         fix1             = _mm_setzero_pd();
 160         fiy1             = _mm_setzero_pd();
 161         fiz1             = _mm_setzero_pd();
 162         fix2             = _mm_setzero_pd();
 163         fiy2             = _mm_setzero_pd();
 164         fiz2             = _mm_setzero_pd();
 165         fix3             = _mm_setzero_pd();
 166         fiy3             = _mm_setzero_pd();
 167         fiz3             = _mm_setzero_pd();
 168
 169         /* Reset potential sums */
 170         velecsum         = _mm_setzero_pd();
 171         vvdwsum          = _mm_setzero_pd();
 172
 173         /* Start inner kernel loop */
 174         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 175         {
 176
 177             /* Get j neighbor index, and coordinate index */
 178             jnrA             = jjnr[jidx];
 179             jnrB             = jjnr[jidx+1];
 180             j_coord_offsetA  = DIM*jnrA;
 181             j_coord_offsetB  = DIM*jnrB;
 182
 183             /* load j atom coordinates */
 184             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 185                                               &jx0,&jy0,&jz0);
 186
 187             /* Calculate displacement vector */
 188             dx00             = _mm_sub_pd(ix0,jx0);
 189             dy00             = _mm_sub_pd(iy0,jy0);
 190             dz00             = _mm_sub_pd(iz0,jz0);
 191             dx10             = _mm_sub_pd(ix1,jx0);
 192             dy10             = _mm_sub_pd(iy1,jy0);
 193             dz10             = _mm_sub_pd(iz1,jz0);
 194             dx20             = _mm_sub_pd(ix2,jx0);
 195             dy20             = _mm_sub_pd(iy2,jy0);
 196             dz20             = _mm_sub_pd(iz2,jz0);
 197             dx30             = _mm_sub_pd(ix3,jx0);
 198             dy30             = _mm_sub_pd(iy3,jy0);
 199             dz30             = _mm_sub_pd(iz3,jz0);
 200
 201             /* Calculate squared distance and things based on it */
 202             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 203             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 204             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 205             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 206
 207             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 208             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 209             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 210
 211             rinvsq00         = gmx_mm_inv_pd(rsq00);
 212             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 213             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 214             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 215
 216             /* Load parameters for j particles */
 217             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 218             vdwjidx0A        = 2*vdwtype[jnrA+0];
 219             vdwjidx0B        = 2*vdwtype[jnrB+0];
 220
 221             fjx0             = _mm_setzero_pd();
 222             fjy0             = _mm_setzero_pd();
 223             fjz0             = _mm_setzero_pd();
 224
 225             /**************************
 226              * CALCULATE INTERACTIONS *
 227              **************************/
 228
 229             /* Compute parameters for interactions between i and j atoms */
 230             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 231                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 232
 233             /* LENNARD-JONES DISPERSION/REPULSION */
 234
 235             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 236             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 237             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 238             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 239             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 240
 241             /* Update potential sum for this i atom from the interaction with this j atom. */
 242             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 243
 244             fscal            = fvdw;
 245
 246             /* Update vectorial force */
 247             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 248             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 249             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 250
 251             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 252             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 253             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 254
 255             /**************************
 256              * CALCULATE INTERACTIONS *
 257              **************************/
 258
 259             /* Compute parameters for interactions between i and j atoms */
 260             qq10             = _mm_mul_pd(iq1,jq0);
 261
 262             /* COULOMB ELECTROSTATICS */
 263             velec            = _mm_mul_pd(qq10,rinv10);
 264             felec            = _mm_mul_pd(velec,rinvsq10);
 265
 266             /* Update potential sum for this i atom from the interaction with this j atom. */
 267             velecsum         = _mm_add_pd(velecsum,velec);
 268
 269             fscal            = felec;
 270
 271             /* Update vectorial force */
 272             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 273             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 274             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 275
 276             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 277             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 278             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 279
 280             /**************************
 281              * CALCULATE INTERACTIONS *
 282              **************************/
 283
 284             /* Compute parameters for interactions between i and j atoms */
 285             qq20             = _mm_mul_pd(iq2,jq0);
 286
 287             /* COULOMB ELECTROSTATICS */
 288             velec            = _mm_mul_pd(qq20,rinv20);
 289             felec            = _mm_mul_pd(velec,rinvsq20);
 290
 291             /* Update potential sum for this i atom from the interaction with this j atom. */
 292             velecsum         = _mm_add_pd(velecsum,velec);
 293
 294             fscal            = felec;
 295
 296             /* Update vectorial force */
 297             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 298             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 299             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 300
 301             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 302             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 303             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 304
 305             /**************************
 306              * CALCULATE INTERACTIONS *
 307              **************************/
 308
 309             /* Compute parameters for interactions between i and j atoms */
 310             qq30             = _mm_mul_pd(iq3,jq0);
 311
 312             /* COULOMB ELECTROSTATICS */
 313             velec            = _mm_mul_pd(qq30,rinv30);
 314             felec            = _mm_mul_pd(velec,rinvsq30);
 315
 316             /* Update potential sum for this i atom from the interaction with this j atom. */
 317             velecsum         = _mm_add_pd(velecsum,velec);
 318
 319             fscal            = felec;
 320
 321             /* Update vectorial force */
 322             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 323             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 324             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 325
 326             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 327             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 328             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 329
 330             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 331
 332             /* Inner loop uses 131 flops */
 333         }
 334
 335         if(jidx<j_index_end)
 336         {
 337
 338             jnrA             = jjnr[jidx];
 339             j_coord_offsetA  = DIM*jnrA;
 340
 341             /* load j atom coordinates */
 342             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 343                                               &jx0,&jy0,&jz0);
 344
 345             /* Calculate displacement vector */
 346             dx00             = _mm_sub_pd(ix0,jx0);
 347             dy00             = _mm_sub_pd(iy0,jy0);
 348             dz00             = _mm_sub_pd(iz0,jz0);
 349             dx10             = _mm_sub_pd(ix1,jx0);
 350             dy10             = _mm_sub_pd(iy1,jy0);
 351             dz10             = _mm_sub_pd(iz1,jz0);
 352             dx20             = _mm_sub_pd(ix2,jx0);
 353             dy20             = _mm_sub_pd(iy2,jy0);
 354             dz20             = _mm_sub_pd(iz2,jz0);
 355             dx30             = _mm_sub_pd(ix3,jx0);
 356             dy30             = _mm_sub_pd(iy3,jy0);
 357             dz30             = _mm_sub_pd(iz3,jz0);
 358
 359             /* Calculate squared distance and things based on it */
 360             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 361             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 362             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 363             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 364
 365             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 366             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 367             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 368
 369             rinvsq00         = gmx_mm_inv_pd(rsq00);
 370             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 371             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 372             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 373
 374             /* Load parameters for j particles */
 375             jq0              = _mm_load_sd(charge+jnrA+0);
 376             vdwjidx0A        = 2*vdwtype[jnrA+0];
 377
 378             fjx0             = _mm_setzero_pd();
 379             fjy0             = _mm_setzero_pd();
 380             fjz0             = _mm_setzero_pd();
 381
 382             /**************************
 383              * CALCULATE INTERACTIONS *
 384              **************************/
 385
 386             /* Compute parameters for interactions between i and j atoms */
 387             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
 388
 389             /* LENNARD-JONES DISPERSION/REPULSION */
 390
 391             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 392             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 393             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 394             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 395             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 396
 397             /* Update potential sum for this i atom from the interaction with this j atom. */
 398             vvdw             = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
 399             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 400
 401             fscal            = fvdw;
 402
 403             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 404
 405             /* Update vectorial force */
 406             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 407             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 408             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 409
 410             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 411             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 412             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 413
 414             /**************************
 415              * CALCULATE INTERACTIONS *
 416              **************************/
 417
 418             /* Compute parameters for interactions between i and j atoms */
 419             qq10             = _mm_mul_pd(iq1,jq0);
 420
 421             /* COULOMB ELECTROSTATICS */
 422             velec            = _mm_mul_pd(qq10,rinv10);
 423             felec            = _mm_mul_pd(velec,rinvsq10);
 424
 425             /* Update potential sum for this i atom from the interaction with this j atom. */
 426             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 427             velecsum         = _mm_add_pd(velecsum,velec);
 428
 429             fscal            = felec;
 430
 431             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 432
 433             /* Update vectorial force */
 434             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 435             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 436             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 437
 438             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 439             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 440             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 441
 442             /**************************
 443              * CALCULATE INTERACTIONS *
 444              **************************/
 445
 446             /* Compute parameters for interactions between i and j atoms */
 447             qq20             = _mm_mul_pd(iq2,jq0);
 448
 449             /* COULOMB ELECTROSTATICS */
 450             velec            = _mm_mul_pd(qq20,rinv20);
 451             felec            = _mm_mul_pd(velec,rinvsq20);
 452
 453             /* Update potential sum for this i atom from the interaction with this j atom. */
 454             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 455             velecsum         = _mm_add_pd(velecsum,velec);
 456
 457             fscal            = felec;
 458
 459             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 460
 461             /* Update vectorial force */
 462             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 463             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 464             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 465
 466             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 467             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 468             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 469
 470             /**************************
 471              * CALCULATE INTERACTIONS *
 472              **************************/
 473
 474             /* Compute parameters for interactions between i and j atoms */
 475             qq30             = _mm_mul_pd(iq3,jq0);
 476
 477             /* COULOMB ELECTROSTATICS */
 478             velec            = _mm_mul_pd(qq30,rinv30);
 479             felec            = _mm_mul_pd(velec,rinvsq30);
 480
 481             /* Update potential sum for this i atom from the interaction with this j atom. */
 482             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 483             velecsum         = _mm_add_pd(velecsum,velec);
 484
 485             fscal            = felec;
 486
 487             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 488
 489             /* Update vectorial force */
 490             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 491             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 492             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 493
 494             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 495             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 496             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 497
 498             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 499
 500             /* Inner loop uses 131 flops */
 501         }
 502
 503         /* End of innermost loop */
 504
 505         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 506                                               f+i_coord_offset,fshift+i_shift_offset);
 507
 508         ggid                        = gid[iidx];
 509         /* Update potential energies */
 510         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 511         gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 512
 513         /* Increment number of inner iterations */
 514         inneriter                  += j_index_end - j_index_start;
 515
 516         /* Outer loop uses 26 flops */
 517     }
 518
 519     /* Increment number of outer iterations */
 520     outeriter        += nri;
 521
 522     /* Update outer/inner flops */
 523
 524     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*131);
 525 }
 526 /*
 527  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_128_fma_double
 528  * Electrostatics interaction: Coulomb
 529  * VdW interaction:            LennardJones
 530  * Geometry:                   Water4-Particle
 531  * Calculate force/pot:        Force
 532  */
 533 void
 534 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_128_fma_double
 535                     (t_nblist                    * gmx_restrict       nlist,
 536                      rvec                        * gmx_restrict          xx,
 537                      rvec                        * gmx_restrict          ff,
 538                      t_forcerec                  * gmx_restrict          fr,
 539                      t_mdatoms                   * gmx_restrict     mdatoms,
 540                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 541                      t_nrnb                      * gmx_restrict        nrnb)
 542 {
 543     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 544      * just 0 for non-waters.
 545      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 546      * jnr indices corresponding to data put in the four positions in the SIMD register.
 547      */
 548     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 549     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 550     int              jnrA,jnrB;
 551     int              j_coord_offsetA,j_coord_offsetB;
 552     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 553     real             rcutoff_scalar;
 554     real             *shiftvec,*fshift,*x,*f;
 555     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 556     int              vdwioffset0;
 557     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 558     int              vdwioffset1;
 559     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 560     int              vdwioffset2;
 561     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 562     int              vdwioffset3;
 563     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 564     int              vdwjidx0A,vdwjidx0B;
 565     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 566     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 567     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 568     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 569     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 570     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 571     real             *charge;
 572     int              nvdwtype;
 573     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 574     int              *vdwtype;
 575     real             *vdwparam;
 576     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 577     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 578     __m128d          dummy_mask,cutoff_mask;
 579     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 580     __m128d          one     = _mm_set1_pd(1.0);
 581     __m128d          two     = _mm_set1_pd(2.0);
 582     x                = xx[0];
 583     f                = ff[0];
 584
 585     nri              = nlist->nri;
 586     iinr             = nlist->iinr;
 587     jindex           = nlist->jindex;
 588     jjnr             = nlist->jjnr;
 589     shiftidx         = nlist->shift;
 590     gid              = nlist->gid;
 591     shiftvec         = fr->shift_vec[0];
 592     fshift           = fr->fshift[0];
 593     facel            = _mm_set1_pd(fr->epsfac);
 594     charge           = mdatoms->chargeA;
 595     nvdwtype         = fr->ntype;
 596     vdwparam         = fr->nbfp;
 597     vdwtype          = mdatoms->typeA;
 598
 599     /* Setup water-specific parameters */
 600     inr              = nlist->iinr[0];
 601     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 602     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 603     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 604     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 605
 606     /* Avoid stupid compiler warnings */
 607     jnrA = jnrB = 0;
 608     j_coord_offsetA = 0;
 609     j_coord_offsetB = 0;
 610
 611     outeriter        = 0;
 612     inneriter        = 0;
 613
 614     /* Start outer loop over neighborlists */
 615     for(iidx=0; iidx<nri; iidx++)
 616     {
 617         /* Load shift vector for this list */
 618         i_shift_offset   = DIM*shiftidx[iidx];
 619
 620         /* Load limits for loop over neighbors */
 621         j_index_start    = jindex[iidx];
 622         j_index_end      = jindex[iidx+1];
 623
 624         /* Get outer coordinate index */
 625         inr              = iinr[iidx];
 626         i_coord_offset   = DIM*inr;
 627
 628         /* Load i particle coords and add shift vector */
 629         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 630                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 631
 632         fix0             = _mm_setzero_pd();
 633         fiy0             = _mm_setzero_pd();
 634         fiz0             = _mm_setzero_pd();
 635         fix1             = _mm_setzero_pd();
 636         fiy1             = _mm_setzero_pd();
 637         fiz1             = _mm_setzero_pd();
 638         fix2             = _mm_setzero_pd();
 639         fiy2             = _mm_setzero_pd();
 640         fiz2             = _mm_setzero_pd();
 641         fix3             = _mm_setzero_pd();
 642         fiy3             = _mm_setzero_pd();
 643         fiz3             = _mm_setzero_pd();
 644
 645         /* Start inner kernel loop */
 646         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 647         {
 648
 649             /* Get j neighbor index, and coordinate index */
 650             jnrA             = jjnr[jidx];
 651             jnrB             = jjnr[jidx+1];
 652             j_coord_offsetA  = DIM*jnrA;
 653             j_coord_offsetB  = DIM*jnrB;
 654
 655             /* load j atom coordinates */
 656             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 657                                               &jx0,&jy0,&jz0);
 658
 659             /* Calculate displacement vector */
 660             dx00             = _mm_sub_pd(ix0,jx0);
 661             dy00             = _mm_sub_pd(iy0,jy0);
 662             dz00             = _mm_sub_pd(iz0,jz0);
 663             dx10             = _mm_sub_pd(ix1,jx0);
 664             dy10             = _mm_sub_pd(iy1,jy0);
 665             dz10             = _mm_sub_pd(iz1,jz0);
 666             dx20             = _mm_sub_pd(ix2,jx0);
 667             dy20             = _mm_sub_pd(iy2,jy0);
 668             dz20             = _mm_sub_pd(iz2,jz0);
 669             dx30             = _mm_sub_pd(ix3,jx0);
 670             dy30             = _mm_sub_pd(iy3,jy0);
 671             dz30             = _mm_sub_pd(iz3,jz0);
 672
 673             /* Calculate squared distance and things based on it */
 674             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 675             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 676             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 677             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 678
 679             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 680             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 681             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 682
 683             rinvsq00         = gmx_mm_inv_pd(rsq00);
 684             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 685             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 686             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 687
 688             /* Load parameters for j particles */
 689             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 690             vdwjidx0A        = 2*vdwtype[jnrA+0];
 691             vdwjidx0B        = 2*vdwtype[jnrB+0];
 692
 693             fjx0             = _mm_setzero_pd();
 694             fjy0             = _mm_setzero_pd();
 695             fjz0             = _mm_setzero_pd();
 696
 697             /**************************
 698              * CALCULATE INTERACTIONS *
 699              **************************/
 700
 701             /* Compute parameters for interactions between i and j atoms */
 702             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 703                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 704
 705             /* LENNARD-JONES DISPERSION/REPULSION */
 706
 707             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 708             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
 709
 710             fscal            = fvdw;
 711
 712             /* Update vectorial force */
 713             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 714             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 715             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 716
 717             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 718             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 719             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 720
 721             /**************************
 722              * CALCULATE INTERACTIONS *
 723              **************************/
 724
 725             /* Compute parameters for interactions between i and j atoms */
 726             qq10             = _mm_mul_pd(iq1,jq0);
 727
 728             /* COULOMB ELECTROSTATICS */
 729             velec            = _mm_mul_pd(qq10,rinv10);
 730             felec            = _mm_mul_pd(velec,rinvsq10);
 731
 732             fscal            = felec;
 733
 734             /* Update vectorial force */
 735             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 736             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 737             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 738
 739             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 740             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 741             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 742
 743             /**************************
 744              * CALCULATE INTERACTIONS *
 745              **************************/
 746
 747             /* Compute parameters for interactions between i and j atoms */
 748             qq20             = _mm_mul_pd(iq2,jq0);
 749
 750             /* COULOMB ELECTROSTATICS */
 751             velec            = _mm_mul_pd(qq20,rinv20);
 752             felec            = _mm_mul_pd(velec,rinvsq20);
 753
 754             fscal            = felec;
 755
 756             /* Update vectorial force */
 757             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 758             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 759             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 760
 761             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 762             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 763             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 764
 765             /**************************
 766              * CALCULATE INTERACTIONS *
 767              **************************/
 768
 769             /* Compute parameters for interactions between i and j atoms */
 770             qq30             = _mm_mul_pd(iq3,jq0);
 771
 772             /* COULOMB ELECTROSTATICS */
 773             velec            = _mm_mul_pd(qq30,rinv30);
 774             felec            = _mm_mul_pd(velec,rinvsq30);
 775
 776             fscal            = felec;
 777
 778             /* Update vectorial force */
 779             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 780             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 781             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 782
 783             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 784             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 785             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 786
 787             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 788
 789             /* Inner loop uses 123 flops */
 790         }
 791
 792         if(jidx<j_index_end)
 793         {
 794
 795             jnrA             = jjnr[jidx];
 796             j_coord_offsetA  = DIM*jnrA;
 797
 798             /* load j atom coordinates */
 799             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 800                                               &jx0,&jy0,&jz0);
 801
 802             /* Calculate displacement vector */
 803             dx00             = _mm_sub_pd(ix0,jx0);
 804             dy00             = _mm_sub_pd(iy0,jy0);
 805             dz00             = _mm_sub_pd(iz0,jz0);
 806             dx10             = _mm_sub_pd(ix1,jx0);
 807             dy10             = _mm_sub_pd(iy1,jy0);
 808             dz10             = _mm_sub_pd(iz1,jz0);
 809             dx20             = _mm_sub_pd(ix2,jx0);
 810             dy20             = _mm_sub_pd(iy2,jy0);
 811             dz20             = _mm_sub_pd(iz2,jz0);
 812             dx30             = _mm_sub_pd(ix3,jx0);
 813             dy30             = _mm_sub_pd(iy3,jy0);
 814             dz30             = _mm_sub_pd(iz3,jz0);
 815
 816             /* Calculate squared distance and things based on it */
 817             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 818             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 819             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 820             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 821
 822             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 823             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 824             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 825
 826             rinvsq00         = gmx_mm_inv_pd(rsq00);
 827             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 828             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 829             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 830
 831             /* Load parameters for j particles */
 832             jq0              = _mm_load_sd(charge+jnrA+0);
 833             vdwjidx0A        = 2*vdwtype[jnrA+0];
 834
 835             fjx0             = _mm_setzero_pd();
 836             fjy0             = _mm_setzero_pd();
 837             fjz0             = _mm_setzero_pd();
 838
 839             /**************************
 840              * CALCULATE INTERACTIONS *
 841              **************************/
 842
 843             /* Compute parameters for interactions between i and j atoms */
 844             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
 845
 846             /* LENNARD-JONES DISPERSION/REPULSION */
 847
 848             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 849             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
 850
 851             fscal            = fvdw;
 852
 853             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 854
 855             /* Update vectorial force */
 856             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 857             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 858             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 859
 860             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 861             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 862             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 863
 864             /**************************
 865              * CALCULATE INTERACTIONS *
 866              **************************/
 867
 868             /* Compute parameters for interactions between i and j atoms */
 869             qq10             = _mm_mul_pd(iq1,jq0);
 870
 871             /* COULOMB ELECTROSTATICS */
 872             velec            = _mm_mul_pd(qq10,rinv10);
 873             felec            = _mm_mul_pd(velec,rinvsq10);
 874
 875             fscal            = felec;
 876
 877             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 878
 879             /* Update vectorial force */
 880             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 881             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 882             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 883
 884             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 885             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 886             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 887
 888             /**************************
 889              * CALCULATE INTERACTIONS *
 890              **************************/
 891
 892             /* Compute parameters for interactions between i and j atoms */
 893             qq20             = _mm_mul_pd(iq2,jq0);
 894
 895             /* COULOMB ELECTROSTATICS */
 896             velec            = _mm_mul_pd(qq20,rinv20);
 897             felec            = _mm_mul_pd(velec,rinvsq20);
 898
 899             fscal            = felec;
 900
 901             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 902
 903             /* Update vectorial force */
 904             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 905             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 906             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 907
 908             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 909             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 910             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 911
 912             /**************************
 913              * CALCULATE INTERACTIONS *
 914              **************************/
 915
 916             /* Compute parameters for interactions between i and j atoms */
 917             qq30             = _mm_mul_pd(iq3,jq0);
 918
 919             /* COULOMB ELECTROSTATICS */
 920             velec            = _mm_mul_pd(qq30,rinv30);
 921             felec            = _mm_mul_pd(velec,rinvsq30);
 922
 923             fscal            = felec;
 924
 925             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 926
 927             /* Update vectorial force */
 928             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 929             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 930             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 931
 932             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 933             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 934             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 935
 936             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 937
 938             /* Inner loop uses 123 flops */
 939         }
 940
 941         /* End of innermost loop */
 942
 943         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 944                                               f+i_coord_offset,fshift+i_shift_offset);
 945
 946         /* Increment number of inner iterations */
 947         inneriter                  += j_index_end - j_index_start;
 948
 949         /* Outer loop uses 24 flops */
 950     }
 951
 952     /* Increment number of outer iterations */
 953     outeriter        += nri;
 954
 955     /* Update outer/inner flops */
 956
 957     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*123);
 958 }