?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Double precision SSE2 kernels
[gromacs.git] / src / gmxlib / nonbonded / nb_kernel_sse2_double / nb_kernel_ElecEwSh_VdwNone_GeomW3P1_sse2_double.c
blob8bd73156d0228eb156a39e3897bf5d8936fe1dc3
1 /*
2 * Note: this file was generated by the Gromacs sse2_double kernel generator.
3 *
4 * This source code is part of
5 *
6 * G R O M A C S
7 *
8 * Copyright (c) 2001-2012, The GROMACS Development Team
9 *
10 * Gromacs is a library for molecular simulation and trajectory analysis,
11 * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
12 * a full list of developers and information, check out http://www.gromacs.org
13 *
14 * This program is free software; you can redistribute it and/or modify it under
15 * the terms of the GNU Lesser General Public License as published by the Free
16 * Software Foundation; either version 2 of the License, or (at your option) any
17 * later version.
18 *
19 * To help fund GROMACS development, we humbly ask that you cite
20 * the papers people have written on it - you can find them on the website.
21 */
22 #ifdef HAVE_CONFIG_H
23 #include <config.h>
24 #endif
25
26 #include <math.h>
27
28 #include "../nb_kernel.h"
29 #include "types/simple.h"
30 #include "vec.h"
31 #include "nrnb.h"
32
33 #include "gmx_math_x86_sse2_double.h"
34 #include "kernelutil_x86_sse2_double.h"
35
36 /*
37 * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwNone_GeomW3P1_VF_sse2_double
38 * Electrostatics interaction: Ewald
39 * VdW interaction: None
40 * Geometry: Water3-Particle
41 * Calculate force/pot: PotentialAndForce
42 */
43 void
44 nb_kernel_ElecEwSh_VdwNone_GeomW3P1_VF_sse2_double
45 (t_nblist * gmx_restrict nlist,
46 rvec * gmx_restrict xx,
47 rvec * gmx_restrict ff,
48 t_forcerec * gmx_restrict fr,
49 t_mdatoms * gmx_restrict mdatoms,
50 nb_kernel_data_t * gmx_restrict kernel_data,
51 t_nrnb * gmx_restrict nrnb)
52 {
53 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
54 * just 0 for non-waters.
55 * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
56 * jnr indices corresponding to data put in the four positions in the SIMD register.
57 */
58 int i_shift_offset,i_coord_offset,outeriter,inneriter;
59 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
60 int jnrA,jnrB;
61 int j_coord_offsetA,j_coord_offsetB;
62 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
63 real rcutoff_scalar;
64 real *shiftvec,*fshift,*x,*f;
65 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
66 int vdwioffset0;
67 __m128d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
68 int vdwioffset1;
69 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
70 int vdwioffset2;
71 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
72 int vdwjidx0A,vdwjidx0B;
73 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
74 __m128d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
75 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
76 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
77 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
78 real *charge;
79 __m128i ewitab;
80 __m128d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
81 real *ewtab;
82 __m128d dummy_mask,cutoff_mask;
83 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
84 __m128d one = _mm_set1_pd(1.0);
85 __m128d two = _mm_set1_pd(2.0);
86 x = xx[0];
87 f = ff[0];
88
89 nri = nlist->nri;
90 iinr = nlist->iinr;
91 jindex = nlist->jindex;
92 jjnr = nlist->jjnr;
93 shiftidx = nlist->shift;
94 gid = nlist->gid;
95 shiftvec = fr->shift_vec[0];
96 fshift = fr->fshift[0];
97 facel = _mm_set1_pd(fr->epsfac);
98 charge = mdatoms->chargeA;
99
100 sh_ewald = _mm_set1_pd(fr->ic->sh_ewald);
101 ewtab = fr->ic->tabq_coul_FDV0;
102 ewtabscale = _mm_set1_pd(fr->ic->tabq_scale);
103 ewtabhalfspace = _mm_set1_pd(0.5/fr->ic->tabq_scale);
104
105 /* Setup water-specific parameters */
106 inr = nlist->iinr[0];
107 iq0 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
108 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
109 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
110
111 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
112 rcutoff_scalar = fr->rcoulomb;
113 rcutoff = _mm_set1_pd(rcutoff_scalar);
114 rcutoff2 = _mm_mul_pd(rcutoff,rcutoff);
115
116 /* Avoid stupid compiler warnings */
117 jnrA = jnrB = 0;
118 j_coord_offsetA = 0;
119 j_coord_offsetB = 0;
120
121 outeriter = 0;
122 inneriter = 0;
123
124 /* Start outer loop over neighborlists */
125 for(iidx=0; iidx<nri; iidx++)
126 {
127 /* Load shift vector for this list */
128 i_shift_offset = DIM*shiftidx[iidx];
129
130 /* Load limits for loop over neighbors */
131 j_index_start = jindex[iidx];
132 j_index_end = jindex[iidx+1];
133
134 /* Get outer coordinate index */
135 inr = iinr[iidx];
136 i_coord_offset = DIM*inr;
137
138 /* Load i particle coords and add shift vector */
139 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
140 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
141
142 fix0 = _mm_setzero_pd();
143 fiy0 = _mm_setzero_pd();
144 fiz0 = _mm_setzero_pd();
145 fix1 = _mm_setzero_pd();
146 fiy1 = _mm_setzero_pd();
147 fiz1 = _mm_setzero_pd();
148 fix2 = _mm_setzero_pd();
149 fiy2 = _mm_setzero_pd();
150 fiz2 = _mm_setzero_pd();
151
152 /* Reset potential sums */
153 velecsum = _mm_setzero_pd();
154
155 /* Start inner kernel loop */
156 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
157 {
158
159 /* Get j neighbor index, and coordinate index */
160 jnrA = jjnr[jidx];
161 jnrB = jjnr[jidx+1];
162 j_coord_offsetA = DIM*jnrA;
163 j_coord_offsetB = DIM*jnrB;
164
165 /* load j atom coordinates */
166 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
167 &jx0,&jy0,&jz0);
168
169 /* Calculate displacement vector */
170 dx00 = _mm_sub_pd(ix0,jx0);
171 dy00 = _mm_sub_pd(iy0,jy0);
172 dz00 = _mm_sub_pd(iz0,jz0);
173 dx10 = _mm_sub_pd(ix1,jx0);
174 dy10 = _mm_sub_pd(iy1,jy0);
175 dz10 = _mm_sub_pd(iz1,jz0);
176 dx20 = _mm_sub_pd(ix2,jx0);
177 dy20 = _mm_sub_pd(iy2,jy0);
178 dz20 = _mm_sub_pd(iz2,jz0);
179
180 /* Calculate squared distance and things based on it */
181 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
182 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
183 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
184
185 rinv00 = gmx_mm_invsqrt_pd(rsq00);
186 rinv10 = gmx_mm_invsqrt_pd(rsq10);
187 rinv20 = gmx_mm_invsqrt_pd(rsq20);
188
189 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
190 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
191 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
192
193 /* Load parameters for j particles */
194 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
195
196 fjx0 = _mm_setzero_pd();
197 fjy0 = _mm_setzero_pd();
198 fjz0 = _mm_setzero_pd();
199
200 /**************************
201 * CALCULATE INTERACTIONS *
202 **************************/
203
204 if (gmx_mm_any_lt(rsq00,rcutoff2))
205 {
206
207 r00 = _mm_mul_pd(rsq00,rinv00);
208
209 /* Compute parameters for interactions between i and j atoms */
210 qq00 = _mm_mul_pd(iq0,jq0);
211
212 /* EWALD ELECTROSTATICS */
213
214 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
215 ewrt = _mm_mul_pd(r00,ewtabscale);
216 ewitab = _mm_cvttpd_epi32(ewrt);
217 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
218 ewitab = _mm_slli_epi32(ewitab,2);
219 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
220 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
221 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
222 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
223 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
224 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
225 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
226 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
227 velec = _mm_mul_pd(qq00,_mm_sub_pd(_mm_sub_pd(rinv00,sh_ewald),velec));
228 felec = _mm_mul_pd(_mm_mul_pd(qq00,rinv00),_mm_sub_pd(rinvsq00,felec));
229
230 cutoff_mask = _mm_cmplt_pd(rsq00,rcutoff2);
231
232 /* Update potential sum for this i atom from the interaction with this j atom. */
233 velec = _mm_and_pd(velec,cutoff_mask);
234 velecsum = _mm_add_pd(velecsum,velec);
235
236 fscal = felec;
237
238 fscal = _mm_and_pd(fscal,cutoff_mask);
239
240 /* Calculate temporary vectorial force */
241 tx = _mm_mul_pd(fscal,dx00);
242 ty = _mm_mul_pd(fscal,dy00);
243 tz = _mm_mul_pd(fscal,dz00);
244
245 /* Update vectorial force */
246 fix0 = _mm_add_pd(fix0,tx);
247 fiy0 = _mm_add_pd(fiy0,ty);
248 fiz0 = _mm_add_pd(fiz0,tz);
249
250 fjx0 = _mm_add_pd(fjx0,tx);
251 fjy0 = _mm_add_pd(fjy0,ty);
252 fjz0 = _mm_add_pd(fjz0,tz);
253
254 }
255
256 /**************************
257 * CALCULATE INTERACTIONS *
258 **************************/
259
260 if (gmx_mm_any_lt(rsq10,rcutoff2))
261 {
262
263 r10 = _mm_mul_pd(rsq10,rinv10);
264
265 /* Compute parameters for interactions between i and j atoms */
266 qq10 = _mm_mul_pd(iq1,jq0);
267
268 /* EWALD ELECTROSTATICS */
269
270 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
271 ewrt = _mm_mul_pd(r10,ewtabscale);
272 ewitab = _mm_cvttpd_epi32(ewrt);
273 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
274 ewitab = _mm_slli_epi32(ewitab,2);
275 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
276 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
277 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
278 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
279 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
280 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
281 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
282 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
283 velec = _mm_mul_pd(qq10,_mm_sub_pd(_mm_sub_pd(rinv10,sh_ewald),velec));
284 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
285
286 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
287
288 /* Update potential sum for this i atom from the interaction with this j atom. */
289 velec = _mm_and_pd(velec,cutoff_mask);
290 velecsum = _mm_add_pd(velecsum,velec);
291
292 fscal = felec;
293
294 fscal = _mm_and_pd(fscal,cutoff_mask);
295
296 /* Calculate temporary vectorial force */
297 tx = _mm_mul_pd(fscal,dx10);
298 ty = _mm_mul_pd(fscal,dy10);
299 tz = _mm_mul_pd(fscal,dz10);
300
301 /* Update vectorial force */
302 fix1 = _mm_add_pd(fix1,tx);
303 fiy1 = _mm_add_pd(fiy1,ty);
304 fiz1 = _mm_add_pd(fiz1,tz);
305
306 fjx0 = _mm_add_pd(fjx0,tx);
307 fjy0 = _mm_add_pd(fjy0,ty);
308 fjz0 = _mm_add_pd(fjz0,tz);
309
310 }
311
312 /**************************
313 * CALCULATE INTERACTIONS *
314 **************************/
315
316 if (gmx_mm_any_lt(rsq20,rcutoff2))
317 {
318
319 r20 = _mm_mul_pd(rsq20,rinv20);
320
321 /* Compute parameters for interactions between i and j atoms */
322 qq20 = _mm_mul_pd(iq2,jq0);
323
324 /* EWALD ELECTROSTATICS */
325
326 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
327 ewrt = _mm_mul_pd(r20,ewtabscale);
328 ewitab = _mm_cvttpd_epi32(ewrt);
329 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
330 ewitab = _mm_slli_epi32(ewitab,2);
331 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
332 ewtabD = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,1) );
333 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
334 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
335 ewtabFn = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,1) +2);
336 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
337 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
338 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
339 velec = _mm_mul_pd(qq20,_mm_sub_pd(_mm_sub_pd(rinv20,sh_ewald),velec));
340 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
341
342 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
343
344 /* Update potential sum for this i atom from the interaction with this j atom. */
345 velec = _mm_and_pd(velec,cutoff_mask);
346 velecsum = _mm_add_pd(velecsum,velec);
347
348 fscal = felec;
349
350 fscal = _mm_and_pd(fscal,cutoff_mask);
351
352 /* Calculate temporary vectorial force */
353 tx = _mm_mul_pd(fscal,dx20);
354 ty = _mm_mul_pd(fscal,dy20);
355 tz = _mm_mul_pd(fscal,dz20);
356
357 /* Update vectorial force */
358 fix2 = _mm_add_pd(fix2,tx);
359 fiy2 = _mm_add_pd(fiy2,ty);
360 fiz2 = _mm_add_pd(fiz2,tz);
361
362 fjx0 = _mm_add_pd(fjx0,tx);
363 fjy0 = _mm_add_pd(fjy0,ty);
364 fjz0 = _mm_add_pd(fjz0,tz);
365
366 }
367
368 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
369
370 /* Inner loop uses 141 flops */
371 }
372
373 if(jidx<j_index_end)
374 {
375
376 jnrA = jjnr[jidx];
377 j_coord_offsetA = DIM*jnrA;
378
379 /* load j atom coordinates */
380 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
381 &jx0,&jy0,&jz0);
382
383 /* Calculate displacement vector */
384 dx00 = _mm_sub_pd(ix0,jx0);
385 dy00 = _mm_sub_pd(iy0,jy0);
386 dz00 = _mm_sub_pd(iz0,jz0);
387 dx10 = _mm_sub_pd(ix1,jx0);
388 dy10 = _mm_sub_pd(iy1,jy0);
389 dz10 = _mm_sub_pd(iz1,jz0);
390 dx20 = _mm_sub_pd(ix2,jx0);
391 dy20 = _mm_sub_pd(iy2,jy0);
392 dz20 = _mm_sub_pd(iz2,jz0);
393
394 /* Calculate squared distance and things based on it */
395 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
396 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
397 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
398
399 rinv00 = gmx_mm_invsqrt_pd(rsq00);
400 rinv10 = gmx_mm_invsqrt_pd(rsq10);
401 rinv20 = gmx_mm_invsqrt_pd(rsq20);
402
403 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
404 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
405 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
406
407 /* Load parameters for j particles */
408 jq0 = _mm_load_sd(charge+jnrA+0);
409
410 fjx0 = _mm_setzero_pd();
411 fjy0 = _mm_setzero_pd();
412 fjz0 = _mm_setzero_pd();
413
414 /**************************
415 * CALCULATE INTERACTIONS *
416 **************************/
417
418 if (gmx_mm_any_lt(rsq00,rcutoff2))
419 {
420
421 r00 = _mm_mul_pd(rsq00,rinv00);
422
423 /* Compute parameters for interactions between i and j atoms */
424 qq00 = _mm_mul_pd(iq0,jq0);
425
426 /* EWALD ELECTROSTATICS */
427
428 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
429 ewrt = _mm_mul_pd(r00,ewtabscale);
430 ewitab = _mm_cvttpd_epi32(ewrt);
431 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
432 ewitab = _mm_slli_epi32(ewitab,2);
433 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
434 ewtabD = _mm_setzero_pd();
435 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
436 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
437 ewtabFn = _mm_setzero_pd();
438 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
439 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
440 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
441 velec = _mm_mul_pd(qq00,_mm_sub_pd(_mm_sub_pd(rinv00,sh_ewald),velec));
442 felec = _mm_mul_pd(_mm_mul_pd(qq00,rinv00),_mm_sub_pd(rinvsq00,felec));
443
444 cutoff_mask = _mm_cmplt_pd(rsq00,rcutoff2);
445
446 /* Update potential sum for this i atom from the interaction with this j atom. */
447 velec = _mm_and_pd(velec,cutoff_mask);
448 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
449 velecsum = _mm_add_pd(velecsum,velec);
450
451 fscal = felec;
452
453 fscal = _mm_and_pd(fscal,cutoff_mask);
454
455 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
456
457 /* Calculate temporary vectorial force */
458 tx = _mm_mul_pd(fscal,dx00);
459 ty = _mm_mul_pd(fscal,dy00);
460 tz = _mm_mul_pd(fscal,dz00);
461
462 /* Update vectorial force */
463 fix0 = _mm_add_pd(fix0,tx);
464 fiy0 = _mm_add_pd(fiy0,ty);
465 fiz0 = _mm_add_pd(fiz0,tz);
466
467 fjx0 = _mm_add_pd(fjx0,tx);
468 fjy0 = _mm_add_pd(fjy0,ty);
469 fjz0 = _mm_add_pd(fjz0,tz);
470
471 }
472
473 /**************************
474 * CALCULATE INTERACTIONS *
475 **************************/
476
477 if (gmx_mm_any_lt(rsq10,rcutoff2))
478 {
479
480 r10 = _mm_mul_pd(rsq10,rinv10);
481
482 /* Compute parameters for interactions between i and j atoms */
483 qq10 = _mm_mul_pd(iq1,jq0);
484
485 /* EWALD ELECTROSTATICS */
486
487 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
488 ewrt = _mm_mul_pd(r10,ewtabscale);
489 ewitab = _mm_cvttpd_epi32(ewrt);
490 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
491 ewitab = _mm_slli_epi32(ewitab,2);
492 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
493 ewtabD = _mm_setzero_pd();
494 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
495 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
496 ewtabFn = _mm_setzero_pd();
497 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
498 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
499 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
500 velec = _mm_mul_pd(qq10,_mm_sub_pd(_mm_sub_pd(rinv10,sh_ewald),velec));
501 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
502
503 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
504
505 /* Update potential sum for this i atom from the interaction with this j atom. */
506 velec = _mm_and_pd(velec,cutoff_mask);
507 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
508 velecsum = _mm_add_pd(velecsum,velec);
509
510 fscal = felec;
511
512 fscal = _mm_and_pd(fscal,cutoff_mask);
513
514 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
515
516 /* Calculate temporary vectorial force */
517 tx = _mm_mul_pd(fscal,dx10);
518 ty = _mm_mul_pd(fscal,dy10);
519 tz = _mm_mul_pd(fscal,dz10);
520
521 /* Update vectorial force */
522 fix1 = _mm_add_pd(fix1,tx);
523 fiy1 = _mm_add_pd(fiy1,ty);
524 fiz1 = _mm_add_pd(fiz1,tz);
525
526 fjx0 = _mm_add_pd(fjx0,tx);
527 fjy0 = _mm_add_pd(fjy0,ty);
528 fjz0 = _mm_add_pd(fjz0,tz);
529
530 }
531
532 /**************************
533 * CALCULATE INTERACTIONS *
534 **************************/
535
536 if (gmx_mm_any_lt(rsq20,rcutoff2))
537 {
538
539 r20 = _mm_mul_pd(rsq20,rinv20);
540
541 /* Compute parameters for interactions between i and j atoms */
542 qq20 = _mm_mul_pd(iq2,jq0);
543
544 /* EWALD ELECTROSTATICS */
545
546 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
547 ewrt = _mm_mul_pd(r20,ewtabscale);
548 ewitab = _mm_cvttpd_epi32(ewrt);
549 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
550 ewitab = _mm_slli_epi32(ewitab,2);
551 ewtabF = _mm_load_pd( ewtab + gmx_mm_extract_epi32(ewitab,0) );
552 ewtabD = _mm_setzero_pd();
553 GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
554 ewtabV = _mm_load_sd( ewtab + gmx_mm_extract_epi32(ewitab,0) +2);
555 ewtabFn = _mm_setzero_pd();
556 GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
557 felec = _mm_add_pd(ewtabF,_mm_mul_pd(eweps,ewtabD));
558 velec = _mm_sub_pd(ewtabV,_mm_mul_pd(_mm_mul_pd(ewtabhalfspace,eweps),_mm_add_pd(ewtabF,felec)));
559 velec = _mm_mul_pd(qq20,_mm_sub_pd(_mm_sub_pd(rinv20,sh_ewald),velec));
560 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
561
562 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
563
564 /* Update potential sum for this i atom from the interaction with this j atom. */
565 velec = _mm_and_pd(velec,cutoff_mask);
566 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
567 velecsum = _mm_add_pd(velecsum,velec);
568
569 fscal = felec;
570
571 fscal = _mm_and_pd(fscal,cutoff_mask);
572
573 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
574
575 /* Calculate temporary vectorial force */
576 tx = _mm_mul_pd(fscal,dx20);
577 ty = _mm_mul_pd(fscal,dy20);
578 tz = _mm_mul_pd(fscal,dz20);
579
580 /* Update vectorial force */
581 fix2 = _mm_add_pd(fix2,tx);
582 fiy2 = _mm_add_pd(fiy2,ty);
583 fiz2 = _mm_add_pd(fiz2,tz);
584
585 fjx0 = _mm_add_pd(fjx0,tx);
586 fjy0 = _mm_add_pd(fjy0,ty);
587 fjz0 = _mm_add_pd(fjz0,tz);
588
589 }
590
591 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
592
593 /* Inner loop uses 141 flops */
594 }
595
596 /* End of innermost loop */
597
598 gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
599 f+i_coord_offset,fshift+i_shift_offset);
600
601 ggid = gid[iidx];
602 /* Update potential energies */
603 gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
604
605 /* Increment number of inner iterations */
606 inneriter += j_index_end - j_index_start;
607
608 /* Outer loop uses 19 flops */
609 }
610
611 /* Increment number of outer iterations */
612 outeriter += nri;
613
614 /* Update outer/inner flops */
615
616 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_VF,outeriter*19 + inneriter*141);
617 }
618 /*
619 * Gromacs nonbonded kernel: nb_kernel_ElecEwSh_VdwNone_GeomW3P1_F_sse2_double
620 * Electrostatics interaction: Ewald
621 * VdW interaction: None
622 * Geometry: Water3-Particle
623 * Calculate force/pot: Force
624 */
625 void
626 nb_kernel_ElecEwSh_VdwNone_GeomW3P1_F_sse2_double
627 (t_nblist * gmx_restrict nlist,
628 rvec * gmx_restrict xx,
629 rvec * gmx_restrict ff,
630 t_forcerec * gmx_restrict fr,
631 t_mdatoms * gmx_restrict mdatoms,
632 nb_kernel_data_t * gmx_restrict kernel_data,
633 t_nrnb * gmx_restrict nrnb)
634 {
635 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
636 * just 0 for non-waters.
637 * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
638 * jnr indices corresponding to data put in the four positions in the SIMD register.
639 */
640 int i_shift_offset,i_coord_offset,outeriter,inneriter;
641 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
642 int jnrA,jnrB;
643 int j_coord_offsetA,j_coord_offsetB;
644 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
645 real rcutoff_scalar;
646 real *shiftvec,*fshift,*x,*f;
647 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
648 int vdwioffset0;
649 __m128d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
650 int vdwioffset1;
651 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
652 int vdwioffset2;
653 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
654 int vdwjidx0A,vdwjidx0B;
655 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
656 __m128d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
657 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
658 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
659 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
660 real *charge;
661 __m128i ewitab;
662 __m128d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
663 real *ewtab;
664 __m128d dummy_mask,cutoff_mask;
665 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
666 __m128d one = _mm_set1_pd(1.0);
667 __m128d two = _mm_set1_pd(2.0);
668 x = xx[0];
669 f = ff[0];
670
671 nri = nlist->nri;
672 iinr = nlist->iinr;
673 jindex = nlist->jindex;
674 jjnr = nlist->jjnr;
675 shiftidx = nlist->shift;
676 gid = nlist->gid;
677 shiftvec = fr->shift_vec[0];
678 fshift = fr->fshift[0];
679 facel = _mm_set1_pd(fr->epsfac);
680 charge = mdatoms->chargeA;
681
682 sh_ewald = _mm_set1_pd(fr->ic->sh_ewald);
683 ewtab = fr->ic->tabq_coul_F;
684 ewtabscale = _mm_set1_pd(fr->ic->tabq_scale);
685 ewtabhalfspace = _mm_set1_pd(0.5/fr->ic->tabq_scale);
686
687 /* Setup water-specific parameters */
688 inr = nlist->iinr[0];
689 iq0 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
690 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
691 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
692
693 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
694 rcutoff_scalar = fr->rcoulomb;
695 rcutoff = _mm_set1_pd(rcutoff_scalar);
696 rcutoff2 = _mm_mul_pd(rcutoff,rcutoff);
697
698 /* Avoid stupid compiler warnings */
699 jnrA = jnrB = 0;
700 j_coord_offsetA = 0;
701 j_coord_offsetB = 0;
702
703 outeriter = 0;
704 inneriter = 0;
705
706 /* Start outer loop over neighborlists */
707 for(iidx=0; iidx<nri; iidx++)
708 {
709 /* Load shift vector for this list */
710 i_shift_offset = DIM*shiftidx[iidx];
711
712 /* Load limits for loop over neighbors */
713 j_index_start = jindex[iidx];
714 j_index_end = jindex[iidx+1];
715
716 /* Get outer coordinate index */
717 inr = iinr[iidx];
718 i_coord_offset = DIM*inr;
719
720 /* Load i particle coords and add shift vector */
721 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
722 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
723
724 fix0 = _mm_setzero_pd();
725 fiy0 = _mm_setzero_pd();
726 fiz0 = _mm_setzero_pd();
727 fix1 = _mm_setzero_pd();
728 fiy1 = _mm_setzero_pd();
729 fiz1 = _mm_setzero_pd();
730 fix2 = _mm_setzero_pd();
731 fiy2 = _mm_setzero_pd();
732 fiz2 = _mm_setzero_pd();
733
734 /* Start inner kernel loop */
735 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
736 {
737
738 /* Get j neighbor index, and coordinate index */
739 jnrA = jjnr[jidx];
740 jnrB = jjnr[jidx+1];
741 j_coord_offsetA = DIM*jnrA;
742 j_coord_offsetB = DIM*jnrB;
743
744 /* load j atom coordinates */
745 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
746 &jx0,&jy0,&jz0);
747
748 /* Calculate displacement vector */
749 dx00 = _mm_sub_pd(ix0,jx0);
750 dy00 = _mm_sub_pd(iy0,jy0);
751 dz00 = _mm_sub_pd(iz0,jz0);
752 dx10 = _mm_sub_pd(ix1,jx0);
753 dy10 = _mm_sub_pd(iy1,jy0);
754 dz10 = _mm_sub_pd(iz1,jz0);
755 dx20 = _mm_sub_pd(ix2,jx0);
756 dy20 = _mm_sub_pd(iy2,jy0);
757 dz20 = _mm_sub_pd(iz2,jz0);
758
759 /* Calculate squared distance and things based on it */
760 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
761 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
762 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
763
764 rinv00 = gmx_mm_invsqrt_pd(rsq00);
765 rinv10 = gmx_mm_invsqrt_pd(rsq10);
766 rinv20 = gmx_mm_invsqrt_pd(rsq20);
767
768 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
769 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
770 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
771
772 /* Load parameters for j particles */
773 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
774
775 fjx0 = _mm_setzero_pd();
776 fjy0 = _mm_setzero_pd();
777 fjz0 = _mm_setzero_pd();
778
779 /**************************
780 * CALCULATE INTERACTIONS *
781 **************************/
782
783 if (gmx_mm_any_lt(rsq00,rcutoff2))
784 {
785
786 r00 = _mm_mul_pd(rsq00,rinv00);
787
788 /* Compute parameters for interactions between i and j atoms */
789 qq00 = _mm_mul_pd(iq0,jq0);
790
791 /* EWALD ELECTROSTATICS */
792
793 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
794 ewrt = _mm_mul_pd(r00,ewtabscale);
795 ewitab = _mm_cvttpd_epi32(ewrt);
796 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
797 gmx_mm_load_2pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
798 &ewtabF,&ewtabFn);
799 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
800 felec = _mm_mul_pd(_mm_mul_pd(qq00,rinv00),_mm_sub_pd(rinvsq00,felec));
801
802 cutoff_mask = _mm_cmplt_pd(rsq00,rcutoff2);
803
804 fscal = felec;
805
806 fscal = _mm_and_pd(fscal,cutoff_mask);
807
808 /* Calculate temporary vectorial force */
809 tx = _mm_mul_pd(fscal,dx00);
810 ty = _mm_mul_pd(fscal,dy00);
811 tz = _mm_mul_pd(fscal,dz00);
812
813 /* Update vectorial force */
814 fix0 = _mm_add_pd(fix0,tx);
815 fiy0 = _mm_add_pd(fiy0,ty);
816 fiz0 = _mm_add_pd(fiz0,tz);
817
818 fjx0 = _mm_add_pd(fjx0,tx);
819 fjy0 = _mm_add_pd(fjy0,ty);
820 fjz0 = _mm_add_pd(fjz0,tz);
821
822 }
823
824 /**************************
825 * CALCULATE INTERACTIONS *
826 **************************/
827
828 if (gmx_mm_any_lt(rsq10,rcutoff2))
829 {
830
831 r10 = _mm_mul_pd(rsq10,rinv10);
832
833 /* Compute parameters for interactions between i and j atoms */
834 qq10 = _mm_mul_pd(iq1,jq0);
835
836 /* EWALD ELECTROSTATICS */
837
838 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
839 ewrt = _mm_mul_pd(r10,ewtabscale);
840 ewitab = _mm_cvttpd_epi32(ewrt);
841 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
842 gmx_mm_load_2pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
843 &ewtabF,&ewtabFn);
844 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
845 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
846
847 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
848
849 fscal = felec;
850
851 fscal = _mm_and_pd(fscal,cutoff_mask);
852
853 /* Calculate temporary vectorial force */
854 tx = _mm_mul_pd(fscal,dx10);
855 ty = _mm_mul_pd(fscal,dy10);
856 tz = _mm_mul_pd(fscal,dz10);
857
858 /* Update vectorial force */
859 fix1 = _mm_add_pd(fix1,tx);
860 fiy1 = _mm_add_pd(fiy1,ty);
861 fiz1 = _mm_add_pd(fiz1,tz);
862
863 fjx0 = _mm_add_pd(fjx0,tx);
864 fjy0 = _mm_add_pd(fjy0,ty);
865 fjz0 = _mm_add_pd(fjz0,tz);
866
867 }
868
869 /**************************
870 * CALCULATE INTERACTIONS *
871 **************************/
872
873 if (gmx_mm_any_lt(rsq20,rcutoff2))
874 {
875
876 r20 = _mm_mul_pd(rsq20,rinv20);
877
878 /* Compute parameters for interactions between i and j atoms */
879 qq20 = _mm_mul_pd(iq2,jq0);
880
881 /* EWALD ELECTROSTATICS */
882
883 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
884 ewrt = _mm_mul_pd(r20,ewtabscale);
885 ewitab = _mm_cvttpd_epi32(ewrt);
886 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
887 gmx_mm_load_2pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
888 &ewtabF,&ewtabFn);
889 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
890 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
891
892 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
893
894 fscal = felec;
895
896 fscal = _mm_and_pd(fscal,cutoff_mask);
897
898 /* Calculate temporary vectorial force */
899 tx = _mm_mul_pd(fscal,dx20);
900 ty = _mm_mul_pd(fscal,dy20);
901 tz = _mm_mul_pd(fscal,dz20);
902
903 /* Update vectorial force */
904 fix2 = _mm_add_pd(fix2,tx);
905 fiy2 = _mm_add_pd(fiy2,ty);
906 fiz2 = _mm_add_pd(fiz2,tz);
907
908 fjx0 = _mm_add_pd(fjx0,tx);
909 fjy0 = _mm_add_pd(fjy0,ty);
910 fjz0 = _mm_add_pd(fjz0,tz);
911
912 }
913
914 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
915
916 /* Inner loop uses 120 flops */
917 }
918
919 if(jidx<j_index_end)
920 {
921
922 jnrA = jjnr[jidx];
923 j_coord_offsetA = DIM*jnrA;
924
925 /* load j atom coordinates */
926 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
927 &jx0,&jy0,&jz0);
928
929 /* Calculate displacement vector */
930 dx00 = _mm_sub_pd(ix0,jx0);
931 dy00 = _mm_sub_pd(iy0,jy0);
932 dz00 = _mm_sub_pd(iz0,jz0);
933 dx10 = _mm_sub_pd(ix1,jx0);
934 dy10 = _mm_sub_pd(iy1,jy0);
935 dz10 = _mm_sub_pd(iz1,jz0);
936 dx20 = _mm_sub_pd(ix2,jx0);
937 dy20 = _mm_sub_pd(iy2,jy0);
938 dz20 = _mm_sub_pd(iz2,jz0);
939
940 /* Calculate squared distance and things based on it */
941 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
942 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
943 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
944
945 rinv00 = gmx_mm_invsqrt_pd(rsq00);
946 rinv10 = gmx_mm_invsqrt_pd(rsq10);
947 rinv20 = gmx_mm_invsqrt_pd(rsq20);
948
949 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
950 rinvsq10 = _mm_mul_pd(rinv10,rinv10);
951 rinvsq20 = _mm_mul_pd(rinv20,rinv20);
952
953 /* Load parameters for j particles */
954 jq0 = _mm_load_sd(charge+jnrA+0);
955
956 fjx0 = _mm_setzero_pd();
957 fjy0 = _mm_setzero_pd();
958 fjz0 = _mm_setzero_pd();
959
960 /**************************
961 * CALCULATE INTERACTIONS *
962 **************************/
963
964 if (gmx_mm_any_lt(rsq00,rcutoff2))
965 {
966
967 r00 = _mm_mul_pd(rsq00,rinv00);
968
969 /* Compute parameters for interactions between i and j atoms */
970 qq00 = _mm_mul_pd(iq0,jq0);
971
972 /* EWALD ELECTROSTATICS */
973
974 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
975 ewrt = _mm_mul_pd(r00,ewtabscale);
976 ewitab = _mm_cvttpd_epi32(ewrt);
977 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
978 gmx_mm_load_1pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
979 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
980 felec = _mm_mul_pd(_mm_mul_pd(qq00,rinv00),_mm_sub_pd(rinvsq00,felec));
981
982 cutoff_mask = _mm_cmplt_pd(rsq00,rcutoff2);
983
984 fscal = felec;
985
986 fscal = _mm_and_pd(fscal,cutoff_mask);
987
988 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
989
990 /* Calculate temporary vectorial force */
991 tx = _mm_mul_pd(fscal,dx00);
992 ty = _mm_mul_pd(fscal,dy00);
993 tz = _mm_mul_pd(fscal,dz00);
994
995 /* Update vectorial force */
996 fix0 = _mm_add_pd(fix0,tx);
997 fiy0 = _mm_add_pd(fiy0,ty);
998 fiz0 = _mm_add_pd(fiz0,tz);
999
1000 fjx0 = _mm_add_pd(fjx0,tx);
1001 fjy0 = _mm_add_pd(fjy0,ty);
1002 fjz0 = _mm_add_pd(fjz0,tz);
1003
1004 }
1005
1006 /**************************
1007 * CALCULATE INTERACTIONS *
1008 **************************/
1009
1010 if (gmx_mm_any_lt(rsq10,rcutoff2))
1011 {
1012
1013 r10 = _mm_mul_pd(rsq10,rinv10);
1014
1015 /* Compute parameters for interactions between i and j atoms */
1016 qq10 = _mm_mul_pd(iq1,jq0);
1017
1018 /* EWALD ELECTROSTATICS */
1019
1020 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1021 ewrt = _mm_mul_pd(r10,ewtabscale);
1022 ewitab = _mm_cvttpd_epi32(ewrt);
1023 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
1024 gmx_mm_load_1pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
1025 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
1026 felec = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
1027
1028 cutoff_mask = _mm_cmplt_pd(rsq10,rcutoff2);
1029
1030 fscal = felec;
1031
1032 fscal = _mm_and_pd(fscal,cutoff_mask);
1033
1034 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1035
1036 /* Calculate temporary vectorial force */
1037 tx = _mm_mul_pd(fscal,dx10);
1038 ty = _mm_mul_pd(fscal,dy10);
1039 tz = _mm_mul_pd(fscal,dz10);
1040
1041 /* Update vectorial force */
1042 fix1 = _mm_add_pd(fix1,tx);
1043 fiy1 = _mm_add_pd(fiy1,ty);
1044 fiz1 = _mm_add_pd(fiz1,tz);
1045
1046 fjx0 = _mm_add_pd(fjx0,tx);
1047 fjy0 = _mm_add_pd(fjy0,ty);
1048 fjz0 = _mm_add_pd(fjz0,tz);
1049
1050 }
1051
1052 /**************************
1053 * CALCULATE INTERACTIONS *
1054 **************************/
1055
1056 if (gmx_mm_any_lt(rsq20,rcutoff2))
1057 {
1058
1059 r20 = _mm_mul_pd(rsq20,rinv20);
1060
1061 /* Compute parameters for interactions between i and j atoms */
1062 qq20 = _mm_mul_pd(iq2,jq0);
1063
1064 /* EWALD ELECTROSTATICS */
1065
1066 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1067 ewrt = _mm_mul_pd(r20,ewtabscale);
1068 ewitab = _mm_cvttpd_epi32(ewrt);
1069 eweps = _mm_sub_pd(ewrt,_mm_cvtepi32_pd(ewitab));
1070 gmx_mm_load_1pair_swizzle_pd(ewtab+gmx_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
1071 felec = _mm_add_pd(_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF),_mm_mul_pd(eweps,ewtabFn));
1072 felec = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
1073
1074 cutoff_mask = _mm_cmplt_pd(rsq20,rcutoff2);
1075
1076 fscal = felec;
1077
1078 fscal = _mm_and_pd(fscal,cutoff_mask);
1079
1080 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1081
1082 /* Calculate temporary vectorial force */
1083 tx = _mm_mul_pd(fscal,dx20);
1084 ty = _mm_mul_pd(fscal,dy20);
1085 tz = _mm_mul_pd(fscal,dz20);
1086
1087 /* Update vectorial force */
1088 fix2 = _mm_add_pd(fix2,tx);
1089 fiy2 = _mm_add_pd(fiy2,ty);
1090 fiz2 = _mm_add_pd(fiz2,tz);
1091
1092 fjx0 = _mm_add_pd(fjx0,tx);
1093 fjy0 = _mm_add_pd(fjy0,ty);
1094 fjz0 = _mm_add_pd(fjz0,tz);
1095
1096 }
1097
1098 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1099
1100 /* Inner loop uses 120 flops */
1101 }
1102
1103 /* End of innermost loop */
1104
1105 gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1106 f+i_coord_offset,fshift+i_shift_offset);
1107
1108 /* Increment number of inner iterations */
1109 inneriter += j_index_end - j_index_start;
1110
1111 /* Outer loop uses 18 flops */
1112 }
1113
1114 /* Increment number of outer iterations */
1115 outeriter += nri;
1116
1117 /* Update outer/inner flops */
1118
1119 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_F,outeriter*18 + inneriter*120);
1120 }
The ultimate molecular dynamics simulation package