2 * This file is part of the GROMACS molecular simulation package.
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.
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.
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.
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.
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.
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.
36 * Note: this file was generated by the GROMACS sparc64_hpc_ace_double kernel generator.
44 #include "../nb_kernel.h"
45 #include "gromacs/legacyheaders/types/simple.h"
46 #include "gromacs/math/vec.h"
47 #include "gromacs/legacyheaders/nrnb.h"
49 #include "kernelutil_sparc64_hpc_ace_double.h"
52 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwLJ_GeomP1P1_VF_sparc64_hpc_ace_double
53 * Electrostatics interaction: ReactionField
54 * VdW interaction: LennardJones
55 * Geometry: Particle-Particle
56 * Calculate force/pot: PotentialAndForce
59 nb_kernel_ElecRF_VdwLJ_GeomP1P1_VF_sparc64_hpc_ace_double
60 (t_nblist * gmx_restrict nlist,
61 rvec * gmx_restrict xx,
62 rvec * gmx_restrict ff,
63 t_forcerec * gmx_restrict fr,
64 t_mdatoms * gmx_restrict mdatoms,
65 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
66 t_nrnb * gmx_restrict nrnb)
68 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
69 * just 0 for non-waters.
70 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
71 * jnr indices corresponding to data put in the four positions in the SIMD register.
73 int i_shift_offset,i_coord_offset,outeriter,inneriter;
74 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
76 int j_coord_offsetA,j_coord_offsetB;
77 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
79 real *shiftvec,*fshift,*x,*f;
80 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
82 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
83 int vdwjidx0A,vdwjidx0B;
84 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
85 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
86 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
89 _fjsp_v2r8 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
92 _fjsp_v2r8 one_sixth = gmx_fjsp_set1_v2r8(1.0/6.0);
93 _fjsp_v2r8 one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
95 _fjsp_v2r8 dummy_mask,cutoff_mask;
96 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
97 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
98 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
105 jindex = nlist->jindex;
107 shiftidx = nlist->shift;
109 shiftvec = fr->shift_vec[0];
110 fshift = fr->fshift[0];
111 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
112 charge = mdatoms->chargeA;
113 krf = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
114 krf2 = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
115 crf = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
116 nvdwtype = fr->ntype;
118 vdwtype = mdatoms->typeA;
120 /* Avoid stupid compiler warnings */
128 /* Start outer loop over neighborlists */
129 for(iidx=0; iidx<nri; iidx++)
131 /* Load shift vector for this list */
132 i_shift_offset = DIM*shiftidx[iidx];
134 /* Load limits for loop over neighbors */
135 j_index_start = jindex[iidx];
136 j_index_end = jindex[iidx+1];
138 /* Get outer coordinate index */
140 i_coord_offset = DIM*inr;
142 /* Load i particle coords and add shift vector */
143 gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
145 fix0 = _fjsp_setzero_v2r8();
146 fiy0 = _fjsp_setzero_v2r8();
147 fiz0 = _fjsp_setzero_v2r8();
149 /* Load parameters for i particles */
150 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_load1_v2r8(charge+inr+0));
151 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
153 /* Reset potential sums */
154 velecsum = _fjsp_setzero_v2r8();
155 vvdwsum = _fjsp_setzero_v2r8();
157 /* Start inner kernel loop */
158 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
161 /* Get j neighbor index, and coordinate index */
164 j_coord_offsetA = DIM*jnrA;
165 j_coord_offsetB = DIM*jnrB;
167 /* load j atom coordinates */
168 gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
171 /* Calculate displacement vector */
172 dx00 = _fjsp_sub_v2r8(ix0,jx0);
173 dy00 = _fjsp_sub_v2r8(iy0,jy0);
174 dz00 = _fjsp_sub_v2r8(iz0,jz0);
176 /* Calculate squared distance and things based on it */
177 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
179 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
181 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
183 /* Load parameters for j particles */
184 jq0 = gmx_fjsp_load_2real_swizzle_v2r8(charge+jnrA+0,charge+jnrB+0);
185 vdwjidx0A = 2*vdwtype[jnrA+0];
186 vdwjidx0B = 2*vdwtype[jnrB+0];
188 /**************************
189 * CALCULATE INTERACTIONS *
190 **************************/
192 /* Compute parameters for interactions between i and j atoms */
193 qq00 = _fjsp_mul_v2r8(iq0,jq0);
194 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
195 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
197 /* REACTION-FIELD ELECTROSTATICS */
198 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq00,rinv00),crf));
199 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
201 /* LENNARD-JONES DISPERSION/REPULSION */
203 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
204 vvdw6 = _fjsp_mul_v2r8(c6_00,rinvsix);
205 vvdw12 = _fjsp_mul_v2r8(c12_00,_fjsp_mul_v2r8(rinvsix,rinvsix));
206 vvdw = _fjsp_msub_v2r8( vvdw12,one_twelfth, _fjsp_mul_v2r8(vvdw6,one_sixth) );
207 fvdw = _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw12,vvdw6),rinvsq00);
209 /* Update potential sum for this i atom from the interaction with this j atom. */
210 velecsum = _fjsp_add_v2r8(velecsum,velec);
211 vvdwsum = _fjsp_add_v2r8(vvdwsum,vvdw);
213 fscal = _fjsp_add_v2r8(felec,fvdw);
215 /* Update vectorial force */
216 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
217 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
218 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
220 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
222 /* Inner loop uses 47 flops */
229 j_coord_offsetA = DIM*jnrA;
231 /* load j atom coordinates */
232 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
235 /* Calculate displacement vector */
236 dx00 = _fjsp_sub_v2r8(ix0,jx0);
237 dy00 = _fjsp_sub_v2r8(iy0,jy0);
238 dz00 = _fjsp_sub_v2r8(iz0,jz0);
240 /* Calculate squared distance and things based on it */
241 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
243 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
245 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
247 /* Load parameters for j particles */
248 jq0 = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(),charge+jnrA+0);
249 vdwjidx0A = 2*vdwtype[jnrA+0];
251 /**************************
252 * CALCULATE INTERACTIONS *
253 **************************/
255 /* Compute parameters for interactions between i and j atoms */
256 qq00 = _fjsp_mul_v2r8(iq0,jq0);
257 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
258 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
260 /* REACTION-FIELD ELECTROSTATICS */
261 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq00,rinv00),crf));
262 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
264 /* LENNARD-JONES DISPERSION/REPULSION */
266 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
267 vvdw6 = _fjsp_mul_v2r8(c6_00,rinvsix);
268 vvdw12 = _fjsp_mul_v2r8(c12_00,_fjsp_mul_v2r8(rinvsix,rinvsix));
269 vvdw = _fjsp_msub_v2r8( vvdw12,one_twelfth, _fjsp_mul_v2r8(vvdw6,one_sixth) );
270 fvdw = _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw12,vvdw6),rinvsq00);
272 /* Update potential sum for this i atom from the interaction with this j atom. */
273 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
274 velecsum = _fjsp_add_v2r8(velecsum,velec);
275 vvdw = _fjsp_unpacklo_v2r8(vvdw,_fjsp_setzero_v2r8());
276 vvdwsum = _fjsp_add_v2r8(vvdwsum,vvdw);
278 fscal = _fjsp_add_v2r8(felec,fvdw);
280 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
282 /* Update vectorial force */
283 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
284 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
285 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
287 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
289 /* Inner loop uses 47 flops */
292 /* End of innermost loop */
294 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
295 f+i_coord_offset,fshift+i_shift_offset);
298 /* Update potential energies */
299 gmx_fjsp_update_1pot_v2r8(velecsum,kernel_data->energygrp_elec+ggid);
300 gmx_fjsp_update_1pot_v2r8(vvdwsum,kernel_data->energygrp_vdw+ggid);
302 /* Increment number of inner iterations */
303 inneriter += j_index_end - j_index_start;
305 /* Outer loop uses 9 flops */
308 /* Increment number of outer iterations */
311 /* Update outer/inner flops */
313 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*9 + inneriter*47);
316 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwLJ_GeomP1P1_F_sparc64_hpc_ace_double
317 * Electrostatics interaction: ReactionField
318 * VdW interaction: LennardJones
319 * Geometry: Particle-Particle
320 * Calculate force/pot: Force
323 nb_kernel_ElecRF_VdwLJ_GeomP1P1_F_sparc64_hpc_ace_double
324 (t_nblist * gmx_restrict nlist,
325 rvec * gmx_restrict xx,
326 rvec * gmx_restrict ff,
327 t_forcerec * gmx_restrict fr,
328 t_mdatoms * gmx_restrict mdatoms,
329 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
330 t_nrnb * gmx_restrict nrnb)
332 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
333 * just 0 for non-waters.
334 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
335 * jnr indices corresponding to data put in the four positions in the SIMD register.
337 int i_shift_offset,i_coord_offset,outeriter,inneriter;
338 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
340 int j_coord_offsetA,j_coord_offsetB;
341 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
343 real *shiftvec,*fshift,*x,*f;
344 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
346 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
347 int vdwjidx0A,vdwjidx0B;
348 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
349 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
350 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
353 _fjsp_v2r8 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
356 _fjsp_v2r8 one_sixth = gmx_fjsp_set1_v2r8(1.0/6.0);
357 _fjsp_v2r8 one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
359 _fjsp_v2r8 dummy_mask,cutoff_mask;
360 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
361 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
362 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
369 jindex = nlist->jindex;
371 shiftidx = nlist->shift;
373 shiftvec = fr->shift_vec[0];
374 fshift = fr->fshift[0];
375 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
376 charge = mdatoms->chargeA;
377 krf = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
378 krf2 = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
379 crf = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
380 nvdwtype = fr->ntype;
382 vdwtype = mdatoms->typeA;
384 /* Avoid stupid compiler warnings */
392 /* Start outer loop over neighborlists */
393 for(iidx=0; iidx<nri; iidx++)
395 /* Load shift vector for this list */
396 i_shift_offset = DIM*shiftidx[iidx];
398 /* Load limits for loop over neighbors */
399 j_index_start = jindex[iidx];
400 j_index_end = jindex[iidx+1];
402 /* Get outer coordinate index */
404 i_coord_offset = DIM*inr;
406 /* Load i particle coords and add shift vector */
407 gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
409 fix0 = _fjsp_setzero_v2r8();
410 fiy0 = _fjsp_setzero_v2r8();
411 fiz0 = _fjsp_setzero_v2r8();
413 /* Load parameters for i particles */
414 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_load1_v2r8(charge+inr+0));
415 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
417 /* Start inner kernel loop */
418 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
421 /* Get j neighbor index, and coordinate index */
424 j_coord_offsetA = DIM*jnrA;
425 j_coord_offsetB = DIM*jnrB;
427 /* load j atom coordinates */
428 gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
431 /* Calculate displacement vector */
432 dx00 = _fjsp_sub_v2r8(ix0,jx0);
433 dy00 = _fjsp_sub_v2r8(iy0,jy0);
434 dz00 = _fjsp_sub_v2r8(iz0,jz0);
436 /* Calculate squared distance and things based on it */
437 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
439 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
441 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
443 /* Load parameters for j particles */
444 jq0 = gmx_fjsp_load_2real_swizzle_v2r8(charge+jnrA+0,charge+jnrB+0);
445 vdwjidx0A = 2*vdwtype[jnrA+0];
446 vdwjidx0B = 2*vdwtype[jnrB+0];
448 /**************************
449 * CALCULATE INTERACTIONS *
450 **************************/
452 /* Compute parameters for interactions between i and j atoms */
453 qq00 = _fjsp_mul_v2r8(iq0,jq0);
454 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
455 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
457 /* REACTION-FIELD ELECTROSTATICS */
458 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
460 /* LENNARD-JONES DISPERSION/REPULSION */
462 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
463 fvdw = _fjsp_mul_v2r8(_fjsp_msub_v2r8(c12_00,rinvsix,c6_00),_fjsp_mul_v2r8(rinvsix,rinvsq00));
465 fscal = _fjsp_add_v2r8(felec,fvdw);
467 /* Update vectorial force */
468 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
469 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
470 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
472 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
474 /* Inner loop uses 37 flops */
481 j_coord_offsetA = DIM*jnrA;
483 /* load j atom coordinates */
484 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
487 /* Calculate displacement vector */
488 dx00 = _fjsp_sub_v2r8(ix0,jx0);
489 dy00 = _fjsp_sub_v2r8(iy0,jy0);
490 dz00 = _fjsp_sub_v2r8(iz0,jz0);
492 /* Calculate squared distance and things based on it */
493 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
495 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
497 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
499 /* Load parameters for j particles */
500 jq0 = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(),charge+jnrA+0);
501 vdwjidx0A = 2*vdwtype[jnrA+0];
503 /**************************
504 * CALCULATE INTERACTIONS *
505 **************************/
507 /* Compute parameters for interactions between i and j atoms */
508 qq00 = _fjsp_mul_v2r8(iq0,jq0);
509 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
510 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
512 /* REACTION-FIELD ELECTROSTATICS */
513 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
515 /* LENNARD-JONES DISPERSION/REPULSION */
517 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
518 fvdw = _fjsp_mul_v2r8(_fjsp_msub_v2r8(c12_00,rinvsix,c6_00),_fjsp_mul_v2r8(rinvsix,rinvsq00));
520 fscal = _fjsp_add_v2r8(felec,fvdw);
522 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
524 /* Update vectorial force */
525 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
526 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
527 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
529 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
531 /* Inner loop uses 37 flops */
534 /* End of innermost loop */
536 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
537 f+i_coord_offset,fshift+i_shift_offset);
539 /* Increment number of inner iterations */
540 inneriter += j_index_end - j_index_start;
542 /* Outer loop uses 7 flops */
545 /* Increment number of outer iterations */
548 /* Update outer/inner flops */
550 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*7 + inneriter*37);