2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014,2015,2017,2018, 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 sse4_1_double kernel generator.
44 #include "../nb_kernel.h"
45 #include "gromacs/gmxlib/nrnb.h"
47 #include "kernelutil_x86_sse4_1_double.h"
50 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW4P1_VF_sse4_1_double
51 * Electrostatics interaction: Ewald
52 * VdW interaction: None
53 * Geometry: Water4-Particle
54 * Calculate force/pot: PotentialAndForce
57 nb_kernel_ElecEwSw_VdwNone_GeomW4P1_VF_sse4_1_double
58 (t_nblist * gmx_restrict nlist,
59 rvec * gmx_restrict xx,
60 rvec * gmx_restrict ff,
61 struct t_forcerec * gmx_restrict fr,
62 t_mdatoms * gmx_restrict mdatoms,
63 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
64 t_nrnb * gmx_restrict nrnb)
66 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
67 * just 0 for non-waters.
68 * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
69 * jnr indices corresponding to data put in the four positions in the SIMD register.
71 int i_shift_offset,i_coord_offset,outeriter,inneriter;
72 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
74 int j_coord_offsetA,j_coord_offsetB;
75 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
77 real *shiftvec,*fshift,*x,*f;
78 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
80 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
82 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
84 __m128d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
85 int vdwjidx0A,vdwjidx0B;
86 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
87 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
88 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
89 __m128d dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
90 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
93 __m128d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
95 __m128d rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
96 real rswitch_scalar,d_scalar;
97 __m128d dummy_mask,cutoff_mask;
98 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
99 __m128d one = _mm_set1_pd(1.0);
100 __m128d two = _mm_set1_pd(2.0);
106 jindex = nlist->jindex;
108 shiftidx = nlist->shift;
110 shiftvec = fr->shift_vec[0];
111 fshift = fr->fshift[0];
112 facel = _mm_set1_pd(fr->ic->epsfac);
113 charge = mdatoms->chargeA;
115 sh_ewald = _mm_set1_pd(fr->ic->sh_ewald);
116 ewtab = fr->ic->tabq_coul_FDV0;
117 ewtabscale = _mm_set1_pd(fr->ic->tabq_scale);
118 ewtabhalfspace = _mm_set1_pd(0.5/fr->ic->tabq_scale);
120 /* Setup water-specific parameters */
121 inr = nlist->iinr[0];
122 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
123 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
124 iq3 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
126 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
127 rcutoff_scalar = fr->ic->rcoulomb;
128 rcutoff = _mm_set1_pd(rcutoff_scalar);
129 rcutoff2 = _mm_mul_pd(rcutoff,rcutoff);
131 rswitch_scalar = fr->ic->rcoulomb_switch;
132 rswitch = _mm_set1_pd(rswitch_scalar);
133 /* Setup switch parameters */
134 d_scalar = rcutoff_scalar-rswitch_scalar;
135 d = _mm_set1_pd(d_scalar);
136 swV3 = _mm_set1_pd(-10.0/(d_scalar*d_scalar*d_scalar));
137 swV4 = _mm_set1_pd( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
138 swV5 = _mm_set1_pd( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
139 swF2 = _mm_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar));
140 swF3 = _mm_set1_pd( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
141 swF4 = _mm_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
143 /* Avoid stupid compiler warnings */
151 /* Start outer loop over neighborlists */
152 for(iidx=0; iidx<nri; iidx++)
154 /* Load shift vector for this list */
155 i_shift_offset = DIM*shiftidx[iidx];
157 /* Load limits for loop over neighbors */
158 j_index_start = jindex[iidx];
159 j_index_end = jindex[iidx+1];
161 /* Get outer coordinate index */
163 i_coord_offset = DIM*inr;
165 /* Load i particle coords and add shift vector */
166 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
167 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
169 fix1 = _mm_setzero_pd();
170 fiy1 = _mm_setzero_pd();
171 fiz1 = _mm_setzero_pd();
172 fix2 = _mm_setzero_pd();
173 fiy2 = _mm_setzero_pd();
174 fiz2 = _mm_setzero_pd();
175 fix3 = _mm_setzero_pd();
176 fiy3 = _mm_setzero_pd();
177 fiz3 = _mm_setzero_pd();
179 /* Reset potential sums */
180 velecsum = _mm_setzero_pd();
182 /* Start inner kernel loop */
183 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
186 /* Get j neighbor index, and coordinate index */
189 j_coord_offsetA = DIM*jnrA;
190 j_coord_offsetB = DIM*jnrB;
192 /* load j atom coordinates */
193 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
196 /* Calculate displacement vector */
197 dx10 = _mm_sub_pd(ix1,jx0);
198 dy10 = _mm_sub_pd(iy1,jy0);
199 dz10 = _mm_sub_pd(iz1,jz0);
200 dx20 = _mm_sub_pd(ix2,jx0);
201 dy20 = _mm_sub_pd(iy2,jy0);
202 dz20 = _mm_sub_pd(iz2,jz0);
203 dx30 = _mm_sub_pd(ix3,jx0);
204 dy30 = _mm_sub_pd(iy3,jy0);
205 dz30 = _mm_sub_pd(iz3,jz0);
207 /* Calculate squared distance and things based on it */
208 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
209 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
210 rsq30 = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
212 rinv10 = sse41_invsqrt_d(rsq10);
213 rinv20 = sse41_invsqrt_d(rsq20);
214 rinv30 = sse41_invsqrt_d(rsq30);
216 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
217 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
218 rinvsq30 = _mm_mul_pd(rinv30,rinv30);
220 /* Load parameters for j particles */
221 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
223 fjx0 = _mm_setzero_pd();
224 fjy0 = _mm_setzero_pd();
225 fjz0 = _mm_setzero_pd();
227 /**************************
228 * CALCULATE INTERACTIONS *
229 **************************/
231 if (gmx_mm_any_lt(rsq10,rcutoff2))
234 r10 = _mm_mul_pd(rsq10,rinv10);
236 /* Compute parameters for interactions between i and j atoms */
237 qq10 = _mm_mul_pd(iq1,jq0);
239 /* EWALD ELECTROSTATICS */
241 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
242 ewrt = _mm_mul_pd(r10,ewtabscale);
243 ewitab = _mm_cvttpd_epi32(ewrt);
244 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
245 ewitab = _mm_slli_epi32(ewitab,2);
246 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
247 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
248 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
249 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
250 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
251 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
252 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
253 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
254 velec = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
255 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
257 d = _mm_sub_pd(r10,rswitch);
258 d = _mm_max_pd(d,_mm_setzero_pd());
259 d2 = _mm_mul_pd(d,d);
260 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
262 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
264 /* Evaluate switch function */
265 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
266 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv10,_mm_mul_pd(velec,dsw)) );
267 velec = _mm_mul_pd(velec,sw);
268 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
270 /* Update potential sum for this i atom from the interaction with this j atom. */
271 velec = _mm_and_pd(velec,cutoff_mask);
272 velecsum = _mm_add_pd(velecsum,velec);
276 fscal = _mm_and_pd(fscal,cutoff_mask);
278 /* Calculate temporary vectorial force */
279 tx = _mm_mul_pd(fscal,dx10);
280 ty = _mm_mul_pd(fscal,dy10);
281 tz = _mm_mul_pd(fscal,dz10);
283 /* Update vectorial force */
284 fix1 = _mm_add_pd(fix1,tx);
285 fiy1 = _mm_add_pd(fiy1,ty);
286 fiz1 = _mm_add_pd(fiz1,tz);
288 fjx0 = _mm_add_pd(fjx0,tx);
289 fjy0 = _mm_add_pd(fjy0,ty);
290 fjz0 = _mm_add_pd(fjz0,tz);
294 /**************************
295 * CALCULATE INTERACTIONS *
296 **************************/
298 if (gmx_mm_any_lt(rsq20,rcutoff2))
301 r20 = _mm_mul_pd(rsq20,rinv20);
303 /* Compute parameters for interactions between i and j atoms */
304 qq20 = _mm_mul_pd(iq2,jq0);
306 /* EWALD ELECTROSTATICS */
308 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
309 ewrt = _mm_mul_pd(r20,ewtabscale);
310 ewitab = _mm_cvttpd_epi32(ewrt);
311 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
312 ewitab = _mm_slli_epi32(ewitab,2);
313 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
314 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
315 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
316 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
317 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
318 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
319 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
320 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
321 velec = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
322 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
324 d = _mm_sub_pd(r20,rswitch);
325 d = _mm_max_pd(d,_mm_setzero_pd());
326 d2 = _mm_mul_pd(d,d);
327 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
329 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
331 /* Evaluate switch function */
332 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
333 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv20,_mm_mul_pd(velec,dsw)) );
334 velec = _mm_mul_pd(velec,sw);
335 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
337 /* Update potential sum for this i atom from the interaction with this j atom. */
338 velec = _mm_and_pd(velec,cutoff_mask);
339 velecsum = _mm_add_pd(velecsum,velec);
343 fscal = _mm_and_pd(fscal,cutoff_mask);
345 /* Calculate temporary vectorial force */
346 tx = _mm_mul_pd(fscal,dx20);
347 ty = _mm_mul_pd(fscal,dy20);
348 tz = _mm_mul_pd(fscal,dz20);
350 /* Update vectorial force */
351 fix2 = _mm_add_pd(fix2,tx);
352 fiy2 = _mm_add_pd(fiy2,ty);
353 fiz2 = _mm_add_pd(fiz2,tz);
355 fjx0 = _mm_add_pd(fjx0,tx);
356 fjy0 = _mm_add_pd(fjy0,ty);
357 fjz0 = _mm_add_pd(fjz0,tz);
361 /**************************
362 * CALCULATE INTERACTIONS *
363 **************************/
365 if (gmx_mm_any_lt(rsq30,rcutoff2))
368 r30 = _mm_mul_pd(rsq30,rinv30);
370 /* Compute parameters for interactions between i and j atoms */
371 qq30 = _mm_mul_pd(iq3,jq0);
373 /* EWALD ELECTROSTATICS */
375 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
376 ewrt = _mm_mul_pd(r30,ewtabscale);
377 ewitab = _mm_cvttpd_epi32(ewrt);
378 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
379 ewitab = _mm_slli_epi32(ewitab,2);
380 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
381 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
382 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
383 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
384 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
385 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
386 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
387 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
388 velec = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
389 felec = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
391 d = _mm_sub_pd(r30,rswitch);
392 d = _mm_max_pd(d,_mm_setzero_pd());
393 d2 = _mm_mul_pd(d,d);
394 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
396 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
398 /* Evaluate switch function */
399 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
400 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv30,_mm_mul_pd(velec,dsw)) );
401 velec = _mm_mul_pd(velec,sw);
402 cutoff_mask = _mm_cmplt_pd(rsq30,rcutoff2);
404 /* Update potential sum for this i atom from the interaction with this j atom. */
405 velec = _mm_and_pd(velec,cutoff_mask);
406 velecsum = _mm_add_pd(velecsum,velec);
410 fscal = _mm_and_pd(fscal,cutoff_mask);
412 /* Calculate temporary vectorial force */
413 tx = _mm_mul_pd(fscal,dx30);
414 ty = _mm_mul_pd(fscal,dy30);
415 tz = _mm_mul_pd(fscal,dz30);
417 /* Update vectorial force */
418 fix3 = _mm_add_pd(fix3,tx);
419 fiy3 = _mm_add_pd(fiy3,ty);
420 fiz3 = _mm_add_pd(fiz3,tz);
422 fjx0 = _mm_add_pd(fjx0,tx);
423 fjy0 = _mm_add_pd(fjy0,ty);
424 fjz0 = _mm_add_pd(fjz0,tz);
428 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
430 /* Inner loop uses 198 flops */
437 j_coord_offsetA = DIM*jnrA;
439 /* load j atom coordinates */
440 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
443 /* Calculate displacement vector */
444 dx10 = _mm_sub_pd(ix1,jx0);
445 dy10 = _mm_sub_pd(iy1,jy0);
446 dz10 = _mm_sub_pd(iz1,jz0);
447 dx20 = _mm_sub_pd(ix2,jx0);
448 dy20 = _mm_sub_pd(iy2,jy0);
449 dz20 = _mm_sub_pd(iz2,jz0);
450 dx30 = _mm_sub_pd(ix3,jx0);
451 dy30 = _mm_sub_pd(iy3,jy0);
452 dz30 = _mm_sub_pd(iz3,jz0);
454 /* Calculate squared distance and things based on it */
455 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
456 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
457 rsq30 = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
459 rinv10 = sse41_invsqrt_d(rsq10);
460 rinv20 = sse41_invsqrt_d(rsq20);
461 rinv30 = sse41_invsqrt_d(rsq30);
463 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
464 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
465 rinvsq30 = _mm_mul_pd(rinv30,rinv30);
467 /* Load parameters for j particles */
468 jq0 = _mm_load_sd(charge+jnrA+0);
470 fjx0 = _mm_setzero_pd();
471 fjy0 = _mm_setzero_pd();
472 fjz0 = _mm_setzero_pd();
474 /**************************
475 * CALCULATE INTERACTIONS *
476 **************************/
478 if (gmx_mm_any_lt(rsq10,rcutoff2))
481 r10 = _mm_mul_pd(rsq10,rinv10);
483 /* Compute parameters for interactions between i and j atoms */
484 qq10 = _mm_mul_pd(iq1,jq0);
486 /* EWALD ELECTROSTATICS */
488 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
489 ewrt = _mm_mul_pd(r10,ewtabscale);
490 ewitab = _mm_cvttpd_epi32(ewrt);
491 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
492 ewitab = _mm_slli_epi32(ewitab,2);
493 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
494 ewtabD = _mm_setzero_pd();
495 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
496 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
497 ewtabFn = _mm_setzero_pd();
498 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
499 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
500 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
501 velec = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
502 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
504 d = _mm_sub_pd(r10,rswitch);
505 d = _mm_max_pd(d,_mm_setzero_pd());
506 d2 = _mm_mul_pd(d,d);
507 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
509 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
511 /* Evaluate switch function */
512 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
513 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv10,_mm_mul_pd(velec,dsw)) );
514 velec = _mm_mul_pd(velec,sw);
515 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
517 /* Update potential sum for this i atom from the interaction with this j atom. */
518 velec = _mm_and_pd(velec,cutoff_mask);
519 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
520 velecsum = _mm_add_pd(velecsum,velec);
524 fscal = _mm_and_pd(fscal,cutoff_mask);
526 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
528 /* Calculate temporary vectorial force */
529 tx = _mm_mul_pd(fscal,dx10);
530 ty = _mm_mul_pd(fscal,dy10);
531 tz = _mm_mul_pd(fscal,dz10);
533 /* Update vectorial force */
534 fix1 = _mm_add_pd(fix1,tx);
535 fiy1 = _mm_add_pd(fiy1,ty);
536 fiz1 = _mm_add_pd(fiz1,tz);
538 fjx0 = _mm_add_pd(fjx0,tx);
539 fjy0 = _mm_add_pd(fjy0,ty);
540 fjz0 = _mm_add_pd(fjz0,tz);
544 /**************************
545 * CALCULATE INTERACTIONS *
546 **************************/
548 if (gmx_mm_any_lt(rsq20,rcutoff2))
551 r20 = _mm_mul_pd(rsq20,rinv20);
553 /* Compute parameters for interactions between i and j atoms */
554 qq20 = _mm_mul_pd(iq2,jq0);
556 /* EWALD ELECTROSTATICS */
558 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
559 ewrt = _mm_mul_pd(r20,ewtabscale);
560 ewitab = _mm_cvttpd_epi32(ewrt);
561 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
562 ewitab = _mm_slli_epi32(ewitab,2);
563 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
564 ewtabD = _mm_setzero_pd();
565 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
566 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
567 ewtabFn = _mm_setzero_pd();
568 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
569 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
570 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
571 velec = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
572 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
574 d = _mm_sub_pd(r20,rswitch);
575 d = _mm_max_pd(d,_mm_setzero_pd());
576 d2 = _mm_mul_pd(d,d);
577 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
579 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
581 /* Evaluate switch function */
582 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
583 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv20,_mm_mul_pd(velec,dsw)) );
584 velec = _mm_mul_pd(velec,sw);
585 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
587 /* Update potential sum for this i atom from the interaction with this j atom. */
588 velec = _mm_and_pd(velec,cutoff_mask);
589 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
590 velecsum = _mm_add_pd(velecsum,velec);
594 fscal = _mm_and_pd(fscal,cutoff_mask);
596 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
598 /* Calculate temporary vectorial force */
599 tx = _mm_mul_pd(fscal,dx20);
600 ty = _mm_mul_pd(fscal,dy20);
601 tz = _mm_mul_pd(fscal,dz20);
603 /* Update vectorial force */
604 fix2 = _mm_add_pd(fix2,tx);
605 fiy2 = _mm_add_pd(fiy2,ty);
606 fiz2 = _mm_add_pd(fiz2,tz);
608 fjx0 = _mm_add_pd(fjx0,tx);
609 fjy0 = _mm_add_pd(fjy0,ty);
610 fjz0 = _mm_add_pd(fjz0,tz);
614 /**************************
615 * CALCULATE INTERACTIONS *
616 **************************/
618 if (gmx_mm_any_lt(rsq30,rcutoff2))
621 r30 = _mm_mul_pd(rsq30,rinv30);
623 /* Compute parameters for interactions between i and j atoms */
624 qq30 = _mm_mul_pd(iq3,jq0);
626 /* EWALD ELECTROSTATICS */
628 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
629 ewrt = _mm_mul_pd(r30,ewtabscale);
630 ewitab = _mm_cvttpd_epi32(ewrt);
631 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
632 ewitab = _mm_slli_epi32(ewitab,2);
633 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
634 ewtabD = _mm_setzero_pd();
635 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
636 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
637 ewtabFn = _mm_setzero_pd();
638 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
639 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
640 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
641 velec = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
642 felec = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
644 d = _mm_sub_pd(r30,rswitch);
645 d = _mm_max_pd(d,_mm_setzero_pd());
646 d2 = _mm_mul_pd(d,d);
647 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
649 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
651 /* Evaluate switch function */
652 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
653 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv30,_mm_mul_pd(velec,dsw)) );
654 velec = _mm_mul_pd(velec,sw);
655 cutoff_mask = _mm_cmplt_pd(rsq30,rcutoff2);
657 /* Update potential sum for this i atom from the interaction with this j atom. */
658 velec = _mm_and_pd(velec,cutoff_mask);
659 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
660 velecsum = _mm_add_pd(velecsum,velec);
664 fscal = _mm_and_pd(fscal,cutoff_mask);
666 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
668 /* Calculate temporary vectorial force */
669 tx = _mm_mul_pd(fscal,dx30);
670 ty = _mm_mul_pd(fscal,dy30);
671 tz = _mm_mul_pd(fscal,dz30);
673 /* Update vectorial force */
674 fix3 = _mm_add_pd(fix3,tx);
675 fiy3 = _mm_add_pd(fiy3,ty);
676 fiz3 = _mm_add_pd(fiz3,tz);
678 fjx0 = _mm_add_pd(fjx0,tx);
679 fjy0 = _mm_add_pd(fjy0,ty);
680 fjz0 = _mm_add_pd(fjz0,tz);
684 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
686 /* Inner loop uses 198 flops */
689 /* End of innermost loop */
691 gmx_mm_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
692 f+i_coord_offset+DIM,fshift+i_shift_offset);
695 /* Update potential energies */
696 gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
698 /* Increment number of inner iterations */
699 inneriter += j_index_end - j_index_start;
701 /* Outer loop uses 19 flops */
704 /* Increment number of outer iterations */
707 /* Update outer/inner flops */
709 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_VF,outeriter*19 + inneriter*198);
712 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW4P1_F_sse4_1_double
713 * Electrostatics interaction: Ewald
714 * VdW interaction: None
715 * Geometry: Water4-Particle
716 * Calculate force/pot: Force
719 nb_kernel_ElecEwSw_VdwNone_GeomW4P1_F_sse4_1_double
720 (t_nblist * gmx_restrict nlist,
721 rvec * gmx_restrict xx,
722 rvec * gmx_restrict ff,
723 struct t_forcerec * gmx_restrict fr,
724 t_mdatoms * gmx_restrict mdatoms,
725 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
726 t_nrnb * gmx_restrict nrnb)
728 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
729 * just 0 for non-waters.
730 * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
731 * jnr indices corresponding to data put in the four positions in the SIMD register.
733 int i_shift_offset,i_coord_offset,outeriter,inneriter;
734 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
736 int j_coord_offsetA,j_coord_offsetB;
737 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
739 real *shiftvec,*fshift,*x,*f;
740 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
742 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
744 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
746 __m128d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
747 int vdwjidx0A,vdwjidx0B;
748 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
749 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
750 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
751 __m128d dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
752 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
755 __m128d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
757 __m128d rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
758 real rswitch_scalar,d_scalar;
759 __m128d dummy_mask,cutoff_mask;
760 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
761 __m128d one = _mm_set1_pd(1.0);
762 __m128d two = _mm_set1_pd(2.0);
768 jindex = nlist->jindex;
770 shiftidx = nlist->shift;
772 shiftvec = fr->shift_vec[0];
773 fshift = fr->fshift[0];
774 facel = _mm_set1_pd(fr->ic->epsfac);
775 charge = mdatoms->chargeA;
777 sh_ewald = _mm_set1_pd(fr->ic->sh_ewald);
778 ewtab = fr->ic->tabq_coul_FDV0;
779 ewtabscale = _mm_set1_pd(fr->ic->tabq_scale);
780 ewtabhalfspace = _mm_set1_pd(0.5/fr->ic->tabq_scale);
782 /* Setup water-specific parameters */
783 inr = nlist->iinr[0];
784 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
785 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
786 iq3 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
788 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
789 rcutoff_scalar = fr->ic->rcoulomb;
790 rcutoff = _mm_set1_pd(rcutoff_scalar);
791 rcutoff2 = _mm_mul_pd(rcutoff,rcutoff);
793 rswitch_scalar = fr->ic->rcoulomb_switch;
794 rswitch = _mm_set1_pd(rswitch_scalar);
795 /* Setup switch parameters */
796 d_scalar = rcutoff_scalar-rswitch_scalar;
797 d = _mm_set1_pd(d_scalar);
798 swV3 = _mm_set1_pd(-10.0/(d_scalar*d_scalar*d_scalar));
799 swV4 = _mm_set1_pd( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
800 swV5 = _mm_set1_pd( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
801 swF2 = _mm_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar));
802 swF3 = _mm_set1_pd( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
803 swF4 = _mm_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
805 /* Avoid stupid compiler warnings */
813 /* Start outer loop over neighborlists */
814 for(iidx=0; iidx<nri; iidx++)
816 /* Load shift vector for this list */
817 i_shift_offset = DIM*shiftidx[iidx];
819 /* Load limits for loop over neighbors */
820 j_index_start = jindex[iidx];
821 j_index_end = jindex[iidx+1];
823 /* Get outer coordinate index */
825 i_coord_offset = DIM*inr;
827 /* Load i particle coords and add shift vector */
828 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
829 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
831 fix1 = _mm_setzero_pd();
832 fiy1 = _mm_setzero_pd();
833 fiz1 = _mm_setzero_pd();
834 fix2 = _mm_setzero_pd();
835 fiy2 = _mm_setzero_pd();
836 fiz2 = _mm_setzero_pd();
837 fix3 = _mm_setzero_pd();
838 fiy3 = _mm_setzero_pd();
839 fiz3 = _mm_setzero_pd();
841 /* Start inner kernel loop */
842 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
845 /* Get j neighbor index, and coordinate index */
848 j_coord_offsetA = DIM*jnrA;
849 j_coord_offsetB = DIM*jnrB;
851 /* load j atom coordinates */
852 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
855 /* Calculate displacement vector */
856 dx10 = _mm_sub_pd(ix1,jx0);
857 dy10 = _mm_sub_pd(iy1,jy0);
858 dz10 = _mm_sub_pd(iz1,jz0);
859 dx20 = _mm_sub_pd(ix2,jx0);
860 dy20 = _mm_sub_pd(iy2,jy0);
861 dz20 = _mm_sub_pd(iz2,jz0);
862 dx30 = _mm_sub_pd(ix3,jx0);
863 dy30 = _mm_sub_pd(iy3,jy0);
864 dz30 = _mm_sub_pd(iz3,jz0);
866 /* Calculate squared distance and things based on it */
867 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
868 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
869 rsq30 = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
871 rinv10 = sse41_invsqrt_d(rsq10);
872 rinv20 = sse41_invsqrt_d(rsq20);
873 rinv30 = sse41_invsqrt_d(rsq30);
875 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
876 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
877 rinvsq30 = _mm_mul_pd(rinv30,rinv30);
879 /* Load parameters for j particles */
880 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
882 fjx0 = _mm_setzero_pd();
883 fjy0 = _mm_setzero_pd();
884 fjz0 = _mm_setzero_pd();
886 /**************************
887 * CALCULATE INTERACTIONS *
888 **************************/
890 if (gmx_mm_any_lt(rsq10,rcutoff2))
893 r10 = _mm_mul_pd(rsq10,rinv10);
895 /* Compute parameters for interactions between i and j atoms */
896 qq10 = _mm_mul_pd(iq1,jq0);
898 /* EWALD ELECTROSTATICS */
900 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
901 ewrt = _mm_mul_pd(r10,ewtabscale);
902 ewitab = _mm_cvttpd_epi32(ewrt);
903 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
904 ewitab = _mm_slli_epi32(ewitab,2);
905 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
906 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
907 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
908 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
909 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
910 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
911 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
912 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
913 velec = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
914 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
916 d = _mm_sub_pd(r10,rswitch);
917 d = _mm_max_pd(d,_mm_setzero_pd());
918 d2 = _mm_mul_pd(d,d);
919 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
921 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
923 /* Evaluate switch function */
924 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
925 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv10,_mm_mul_pd(velec,dsw)) );
926 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
930 fscal = _mm_and_pd(fscal,cutoff_mask);
932 /* Calculate temporary vectorial force */
933 tx = _mm_mul_pd(fscal,dx10);
934 ty = _mm_mul_pd(fscal,dy10);
935 tz = _mm_mul_pd(fscal,dz10);
937 /* Update vectorial force */
938 fix1 = _mm_add_pd(fix1,tx);
939 fiy1 = _mm_add_pd(fiy1,ty);
940 fiz1 = _mm_add_pd(fiz1,tz);
942 fjx0 = _mm_add_pd(fjx0,tx);
943 fjy0 = _mm_add_pd(fjy0,ty);
944 fjz0 = _mm_add_pd(fjz0,tz);
948 /**************************
949 * CALCULATE INTERACTIONS *
950 **************************/
952 if (gmx_mm_any_lt(rsq20,rcutoff2))
955 r20 = _mm_mul_pd(rsq20,rinv20);
957 /* Compute parameters for interactions between i and j atoms */
958 qq20 = _mm_mul_pd(iq2,jq0);
960 /* EWALD ELECTROSTATICS */
962 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
963 ewrt = _mm_mul_pd(r20,ewtabscale);
964 ewitab = _mm_cvttpd_epi32(ewrt);
965 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
966 ewitab = _mm_slli_epi32(ewitab,2);
967 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
968 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
969 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
970 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
971 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
972 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
973 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
974 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
975 velec = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
976 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
978 d = _mm_sub_pd(r20,rswitch);
979 d = _mm_max_pd(d,_mm_setzero_pd());
980 d2 = _mm_mul_pd(d,d);
981 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
983 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
985 /* Evaluate switch function */
986 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
987 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv20,_mm_mul_pd(velec,dsw)) );
988 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
992 fscal = _mm_and_pd(fscal,cutoff_mask);
994 /* Calculate temporary vectorial force */
995 tx = _mm_mul_pd(fscal,dx20);
996 ty = _mm_mul_pd(fscal,dy20);
997 tz = _mm_mul_pd(fscal,dz20);
999 /* Update vectorial force */
1000 fix2 = _mm_add_pd(fix2,tx);
1001 fiy2 = _mm_add_pd(fiy2,ty);
1002 fiz2 = _mm_add_pd(fiz2,tz);
1004 fjx0 = _mm_add_pd(fjx0,tx);
1005 fjy0 = _mm_add_pd(fjy0,ty);
1006 fjz0 = _mm_add_pd(fjz0,tz);
1010 /**************************
1011 * CALCULATE INTERACTIONS *
1012 **************************/
1014 if (gmx_mm_any_lt(rsq30,rcutoff2))
1017 r30 = _mm_mul_pd(rsq30,rinv30);
1019 /* Compute parameters for interactions between i and j atoms */
1020 qq30 = _mm_mul_pd(iq3,jq0);
1022 /* EWALD ELECTROSTATICS */
1024 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1025 ewrt = _mm_mul_pd(r30,ewtabscale);
1026 ewitab = _mm_cvttpd_epi32(ewrt);
1027 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1028 ewitab = _mm_slli_epi32(ewitab,2);
1029 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1030 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
1031 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
1032 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
1033 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
1034 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
1035 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
1036 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
1037 velec = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
1038 felec = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
1040 d = _mm_sub_pd(r30,rswitch);
1041 d = _mm_max_pd(d,_mm_setzero_pd());
1042 d2 = _mm_mul_pd(d,d);
1043 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
1045 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
1047 /* Evaluate switch function */
1048 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1049 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv30,_mm_mul_pd(velec,dsw)) );
1050 cutoff_mask = _mm_cmplt_pd(rsq30,rcutoff2);
1054 fscal = _mm_and_pd(fscal,cutoff_mask);
1056 /* Calculate temporary vectorial force */
1057 tx = _mm_mul_pd(fscal,dx30);
1058 ty = _mm_mul_pd(fscal,dy30);
1059 tz = _mm_mul_pd(fscal,dz30);
1061 /* Update vectorial force */
1062 fix3 = _mm_add_pd(fix3,tx);
1063 fiy3 = _mm_add_pd(fiy3,ty);
1064 fiz3 = _mm_add_pd(fiz3,tz);
1066 fjx0 = _mm_add_pd(fjx0,tx);
1067 fjy0 = _mm_add_pd(fjy0,ty);
1068 fjz0 = _mm_add_pd(fjz0,tz);
1072 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
1074 /* Inner loop uses 189 flops */
1077 if(jidx<j_index_end)
1081 j_coord_offsetA = DIM*jnrA;
1083 /* load j atom coordinates */
1084 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
1087 /* Calculate displacement vector */
1088 dx10 = _mm_sub_pd(ix1,jx0);
1089 dy10 = _mm_sub_pd(iy1,jy0);
1090 dz10 = _mm_sub_pd(iz1,jz0);
1091 dx20 = _mm_sub_pd(ix2,jx0);
1092 dy20 = _mm_sub_pd(iy2,jy0);
1093 dz20 = _mm_sub_pd(iz2,jz0);
1094 dx30 = _mm_sub_pd(ix3,jx0);
1095 dy30 = _mm_sub_pd(iy3,jy0);
1096 dz30 = _mm_sub_pd(iz3,jz0);
1098 /* Calculate squared distance and things based on it */
1099 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1100 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1101 rsq30 = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
1103 rinv10 = sse41_invsqrt_d(rsq10);
1104 rinv20 = sse41_invsqrt_d(rsq20);
1105 rinv30 = sse41_invsqrt_d(rsq30);
1107 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
1108 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
1109 rinvsq30 = _mm_mul_pd(rinv30,rinv30);
1111 /* Load parameters for j particles */
1112 jq0 = _mm_load_sd(charge+jnrA+0);
1114 fjx0 = _mm_setzero_pd();
1115 fjy0 = _mm_setzero_pd();
1116 fjz0 = _mm_setzero_pd();
1118 /**************************
1119 * CALCULATE INTERACTIONS *
1120 **************************/
1122 if (gmx_mm_any_lt(rsq10,rcutoff2))
1125 r10 = _mm_mul_pd(rsq10,rinv10);
1127 /* Compute parameters for interactions between i and j atoms */
1128 qq10 = _mm_mul_pd(iq1,jq0);
1130 /* EWALD ELECTROSTATICS */
1132 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1133 ewrt = _mm_mul_pd(r10,ewtabscale);
1134 ewitab = _mm_cvttpd_epi32(ewrt);
1135 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1136 ewitab = _mm_slli_epi32(ewitab,2);
1137 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1138 ewtabD = _mm_setzero_pd();
1139 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
1140 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
1141 ewtabFn = _mm_setzero_pd();
1142 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
1143 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
1144 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
1145 velec = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
1146 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
1148 d = _mm_sub_pd(r10,rswitch);
1149 d = _mm_max_pd(d,_mm_setzero_pd());
1150 d2 = _mm_mul_pd(d,d);
1151 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
1153 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
1155 /* Evaluate switch function */
1156 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1157 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv10,_mm_mul_pd(velec,dsw)) );
1158 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
1162 fscal = _mm_and_pd(fscal,cutoff_mask);
1164 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1166 /* Calculate temporary vectorial force */
1167 tx = _mm_mul_pd(fscal,dx10);
1168 ty = _mm_mul_pd(fscal,dy10);
1169 tz = _mm_mul_pd(fscal,dz10);
1171 /* Update vectorial force */
1172 fix1 = _mm_add_pd(fix1,tx);
1173 fiy1 = _mm_add_pd(fiy1,ty);
1174 fiz1 = _mm_add_pd(fiz1,tz);
1176 fjx0 = _mm_add_pd(fjx0,tx);
1177 fjy0 = _mm_add_pd(fjy0,ty);
1178 fjz0 = _mm_add_pd(fjz0,tz);
1182 /**************************
1183 * CALCULATE INTERACTIONS *
1184 **************************/
1186 if (gmx_mm_any_lt(rsq20,rcutoff2))
1189 r20 = _mm_mul_pd(rsq20,rinv20);
1191 /* Compute parameters for interactions between i and j atoms */
1192 qq20 = _mm_mul_pd(iq2,jq0);
1194 /* EWALD ELECTROSTATICS */
1196 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1197 ewrt = _mm_mul_pd(r20,ewtabscale);
1198 ewitab = _mm_cvttpd_epi32(ewrt);
1199 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1200 ewitab = _mm_slli_epi32(ewitab,2);
1201 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1202 ewtabD = _mm_setzero_pd();
1203 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
1204 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
1205 ewtabFn = _mm_setzero_pd();
1206 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
1207 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
1208 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
1209 velec = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
1210 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
1212 d = _mm_sub_pd(r20,rswitch);
1213 d = _mm_max_pd(d,_mm_setzero_pd());
1214 d2 = _mm_mul_pd(d,d);
1215 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
1217 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
1219 /* Evaluate switch function */
1220 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1221 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv20,_mm_mul_pd(velec,dsw)) );
1222 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
1226 fscal = _mm_and_pd(fscal,cutoff_mask);
1228 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1230 /* Calculate temporary vectorial force */
1231 tx = _mm_mul_pd(fscal,dx20);
1232 ty = _mm_mul_pd(fscal,dy20);
1233 tz = _mm_mul_pd(fscal,dz20);
1235 /* Update vectorial force */
1236 fix2 = _mm_add_pd(fix2,tx);
1237 fiy2 = _mm_add_pd(fiy2,ty);
1238 fiz2 = _mm_add_pd(fiz2,tz);
1240 fjx0 = _mm_add_pd(fjx0,tx);
1241 fjy0 = _mm_add_pd(fjy0,ty);
1242 fjz0 = _mm_add_pd(fjz0,tz);
1246 /**************************
1247 * CALCULATE INTERACTIONS *
1248 **************************/
1250 if (gmx_mm_any_lt(rsq30,rcutoff2))
1253 r30 = _mm_mul_pd(rsq30,rinv30);
1255 /* Compute parameters for interactions between i and j atoms */
1256 qq30 = _mm_mul_pd(iq3,jq0);
1258 /* EWALD ELECTROSTATICS */
1260 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1261 ewrt = _mm_mul_pd(r30,ewtabscale);
1262 ewitab = _mm_cvttpd_epi32(ewrt);
1263 eweps = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1264 ewitab = _mm_slli_epi32(ewitab,2);
1265 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1266 ewtabD = _mm_setzero_pd();
1267 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
1268 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
1269 ewtabFn = _mm_setzero_pd();
1270 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
1271 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
1272 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
1273 velec = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
1274 felec = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
1276 d = _mm_sub_pd(r30,rswitch);
1277 d = _mm_max_pd(d,_mm_setzero_pd());
1278 d2 = _mm_mul_pd(d,d);
1279 sw = _mm_add_pd(one,_mm_mul_pd(d2,_mm_mul_pd(d,_mm_add_pd(swV3,_mm_mul_pd(d,_mm_add_pd(swV4,_mm_mul_pd(d,swV5)))))));
1281 dsw = _mm_mul_pd(d2,_mm_add_pd(swF2,_mm_mul_pd(d,_mm_add_pd(swF3,_mm_mul_pd(d,swF4)))));
1283 /* Evaluate switch function */
1284 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1285 felec = _mm_sub_pd( _mm_mul_pd(felec,sw) , _mm_mul_pd(rinv30,_mm_mul_pd(velec,dsw)) );
1286 cutoff_mask = _mm_cmplt_pd(rsq30,rcutoff2);
1290 fscal = _mm_and_pd(fscal,cutoff_mask);
1292 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1294 /* Calculate temporary vectorial force */
1295 tx = _mm_mul_pd(fscal,dx30);
1296 ty = _mm_mul_pd(fscal,dy30);
1297 tz = _mm_mul_pd(fscal,dz30);
1299 /* Update vectorial force */
1300 fix3 = _mm_add_pd(fix3,tx);
1301 fiy3 = _mm_add_pd(fiy3,ty);
1302 fiz3 = _mm_add_pd(fiz3,tz);
1304 fjx0 = _mm_add_pd(fjx0,tx);
1305 fjy0 = _mm_add_pd(fjy0,ty);
1306 fjz0 = _mm_add_pd(fjz0,tz);
1310 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1312 /* Inner loop uses 189 flops */
1315 /* End of innermost loop */
1317 gmx_mm_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1318 f+i_coord_offset+DIM,fshift+i_shift_offset);
1320 /* Increment number of inner iterations */
1321 inneriter += j_index_end - j_index_start;
1323 /* Outer loop uses 18 flops */
1326 /* Increment number of outer iterations */
1329 /* Update outer/inner flops */
1331 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_F,outeriter*18 + inneriter*189);