?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Remove all unnecessary HAVE_CONFIG_H
[gromacs.git] / src / gromacs / gmxlib / nonbonded / nb_kernel_sparc64_hpc_ace_double / nb_kernel_ElecEwSw_VdwNone_GeomW3W3_sparc64_hpc_ace_double.c
blobf22ee531c650e805a7c8938f2da322a700d9e217
1 /*
2 * This file is part of the GROMACS molecular simulation package.
3 *
4 * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
8 *
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
13 *
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
18 *
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
23 *
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
31 *
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
34 */
35 /*
36 * Note: this file was generated by the GROMACS sparc64_hpc_ace_double kernel generator.
37 */
38 #include "config.h"
39
40 #include <math.h>
41
42 #include "../nb_kernel.h"
43 #include "types/simple.h"
44 #include "gromacs/math/vec.h"
45 #include "nrnb.h"
46
47 #include "kernelutil_sparc64_hpc_ace_double.h"
48
49 /*
50 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW3W3_VF_sparc64_hpc_ace_double
51 * Electrostatics interaction: Ewald
52 * VdW interaction: None
53 * Geometry: Water3-Water3
54 * Calculate force/pot: PotentialAndForce
55 */
56 void
57 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_VF_sparc64_hpc_ace_double
58 (t_nblist * gmx_restrict nlist,
59 rvec * gmx_restrict xx,
60 rvec * gmx_restrict ff,
61 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)
65 {
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 double precision SIMD, e.g. for the two different
69 * jnr indices corresponding to data put in the four positions in the SIMD register.
70 */
71 int i_shift_offset,i_coord_offset,outeriter,inneriter;
72 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
73 int jnrA,jnrB;
74 int j_coord_offsetA,j_coord_offsetB;
75 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
76 real rcutoff_scalar;
77 real *shiftvec,*fshift,*x,*f;
78 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
79 int vdwioffset0;
80 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
81 int vdwioffset1;
82 _fjsp_v2r8 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
83 int vdwioffset2;
84 _fjsp_v2r8 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
85 int vdwjidx0A,vdwjidx0B;
86 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
87 int vdwjidx1A,vdwjidx1B;
88 _fjsp_v2r8 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
89 int vdwjidx2A,vdwjidx2B;
90 _fjsp_v2r8 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
91 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
92 _fjsp_v2r8 dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
93 _fjsp_v2r8 dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
94 _fjsp_v2r8 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
95 _fjsp_v2r8 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
96 _fjsp_v2r8 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
97 _fjsp_v2r8 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
98 _fjsp_v2r8 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
99 _fjsp_v2r8 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
100 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
101 real *charge;
102 _fjsp_v2r8 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
103 real *ewtab;
104 _fjsp_v2r8 rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
105 real rswitch_scalar,d_scalar;
106 _fjsp_v2r8 itab_tmp;
107 _fjsp_v2r8 dummy_mask,cutoff_mask;
108 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
109 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
110 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
111
112 x = xx[0];
113 f = ff[0];
114
115 nri = nlist->nri;
116 iinr = nlist->iinr;
117 jindex = nlist->jindex;
118 jjnr = nlist->jjnr;
119 shiftidx = nlist->shift;
120 gid = nlist->gid;
121 shiftvec = fr->shift_vec[0];
122 fshift = fr->fshift[0];
123 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
124 charge = mdatoms->chargeA;
125
126 sh_ewald = gmx_fjsp_set1_v2r8(fr->ic->sh_ewald);
127 ewtab = fr->ic->tabq_coul_FDV0;
128 ewtabscale = gmx_fjsp_set1_v2r8(fr->ic->tabq_scale);
129 ewtabhalfspace = gmx_fjsp_set1_v2r8(0.5/fr->ic->tabq_scale);
130
131 /* Setup water-specific parameters */
132 inr = nlist->iinr[0];
133 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+0]));
134 iq1 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+1]));
135 iq2 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+2]));
136
137 jq0 = gmx_fjsp_set1_v2r8(charge[inr+0]);
138 jq1 = gmx_fjsp_set1_v2r8(charge[inr+1]);
139 jq2 = gmx_fjsp_set1_v2r8(charge[inr+2]);
140 qq00 = _fjsp_mul_v2r8(iq0,jq0);
141 qq01 = _fjsp_mul_v2r8(iq0,jq1);
142 qq02 = _fjsp_mul_v2r8(iq0,jq2);
143 qq10 = _fjsp_mul_v2r8(iq1,jq0);
144 qq11 = _fjsp_mul_v2r8(iq1,jq1);
145 qq12 = _fjsp_mul_v2r8(iq1,jq2);
146 qq20 = _fjsp_mul_v2r8(iq2,jq0);
147 qq21 = _fjsp_mul_v2r8(iq2,jq1);
148 qq22 = _fjsp_mul_v2r8(iq2,jq2);
149
150 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
151 rcutoff_scalar = fr->rcoulomb;
152 rcutoff = gmx_fjsp_set1_v2r8(rcutoff_scalar);
153 rcutoff2 = _fjsp_mul_v2r8(rcutoff,rcutoff);
154
155 rswitch_scalar = fr->rcoulomb_switch;
156 rswitch = gmx_fjsp_set1_v2r8(rswitch_scalar);
157 /* Setup switch parameters */
158 d_scalar = rcutoff_scalar-rswitch_scalar;
159 d = gmx_fjsp_set1_v2r8(d_scalar);
160 swV3 = gmx_fjsp_set1_v2r8(-10.0/(d_scalar*d_scalar*d_scalar));
161 swV4 = gmx_fjsp_set1_v2r8( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
162 swV5 = gmx_fjsp_set1_v2r8( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
163 swF2 = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar));
164 swF3 = gmx_fjsp_set1_v2r8( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
165 swF4 = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
166
167 /* Avoid stupid compiler warnings */
168 jnrA = jnrB = 0;
169 j_coord_offsetA = 0;
170 j_coord_offsetB = 0;
171
172 outeriter = 0;
173 inneriter = 0;
174
175 /* Start outer loop over neighborlists */
176 for(iidx=0; iidx<nri; iidx++)
177 {
178 /* Load shift vector for this list */
179 i_shift_offset = DIM*shiftidx[iidx];
180
181 /* Load limits for loop over neighbors */
182 j_index_start = jindex[iidx];
183 j_index_end = jindex[iidx+1];
184
185 /* Get outer coordinate index */
186 inr = iinr[iidx];
187 i_coord_offset = DIM*inr;
188
189 /* Load i particle coords and add shift vector */
190 gmx_fjsp_load_shift_and_3rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
191 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
192
193 fix0 = _fjsp_setzero_v2r8();
194 fiy0 = _fjsp_setzero_v2r8();
195 fiz0 = _fjsp_setzero_v2r8();
196 fix1 = _fjsp_setzero_v2r8();
197 fiy1 = _fjsp_setzero_v2r8();
198 fiz1 = _fjsp_setzero_v2r8();
199 fix2 = _fjsp_setzero_v2r8();
200 fiy2 = _fjsp_setzero_v2r8();
201 fiz2 = _fjsp_setzero_v2r8();
202
203 /* Reset potential sums */
204 velecsum = _fjsp_setzero_v2r8();
205
206 /* Start inner kernel loop */
207 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
208 {
209
210 /* Get j neighbor index, and coordinate index */
211 jnrA = jjnr[jidx];
212 jnrB = jjnr[jidx+1];
213 j_coord_offsetA = DIM*jnrA;
214 j_coord_offsetB = DIM*jnrB;
215
216 /* load j atom coordinates */
217 gmx_fjsp_load_3rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
218 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
219
220 /* Calculate displacement vector */
221 dx00 = _fjsp_sub_v2r8(ix0,jx0);
222 dy00 = _fjsp_sub_v2r8(iy0,jy0);
223 dz00 = _fjsp_sub_v2r8(iz0,jz0);
224 dx01 = _fjsp_sub_v2r8(ix0,jx1);
225 dy01 = _fjsp_sub_v2r8(iy0,jy1);
226 dz01 = _fjsp_sub_v2r8(iz0,jz1);
227 dx02 = _fjsp_sub_v2r8(ix0,jx2);
228 dy02 = _fjsp_sub_v2r8(iy0,jy2);
229 dz02 = _fjsp_sub_v2r8(iz0,jz2);
230 dx10 = _fjsp_sub_v2r8(ix1,jx0);
231 dy10 = _fjsp_sub_v2r8(iy1,jy0);
232 dz10 = _fjsp_sub_v2r8(iz1,jz0);
233 dx11 = _fjsp_sub_v2r8(ix1,jx1);
234 dy11 = _fjsp_sub_v2r8(iy1,jy1);
235 dz11 = _fjsp_sub_v2r8(iz1,jz1);
236 dx12 = _fjsp_sub_v2r8(ix1,jx2);
237 dy12 = _fjsp_sub_v2r8(iy1,jy2);
238 dz12 = _fjsp_sub_v2r8(iz1,jz2);
239 dx20 = _fjsp_sub_v2r8(ix2,jx0);
240 dy20 = _fjsp_sub_v2r8(iy2,jy0);
241 dz20 = _fjsp_sub_v2r8(iz2,jz0);
242 dx21 = _fjsp_sub_v2r8(ix2,jx1);
243 dy21 = _fjsp_sub_v2r8(iy2,jy1);
244 dz21 = _fjsp_sub_v2r8(iz2,jz1);
245 dx22 = _fjsp_sub_v2r8(ix2,jx2);
246 dy22 = _fjsp_sub_v2r8(iy2,jy2);
247 dz22 = _fjsp_sub_v2r8(iz2,jz2);
248
249 /* Calculate squared distance and things based on it */
250 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
251 rsq01 = gmx_fjsp_calc_rsq_v2r8(dx01,dy01,dz01);
252 rsq02 = gmx_fjsp_calc_rsq_v2r8(dx02,dy02,dz02);
253 rsq10 = gmx_fjsp_calc_rsq_v2r8(dx10,dy10,dz10);
254 rsq11 = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
255 rsq12 = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
256 rsq20 = gmx_fjsp_calc_rsq_v2r8(dx20,dy20,dz20);
257 rsq21 = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
258 rsq22 = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
259
260 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
261 rinv01 = gmx_fjsp_invsqrt_v2r8(rsq01);
262 rinv02 = gmx_fjsp_invsqrt_v2r8(rsq02);
263 rinv10 = gmx_fjsp_invsqrt_v2r8(rsq10);
264 rinv11 = gmx_fjsp_invsqrt_v2r8(rsq11);
265 rinv12 = gmx_fjsp_invsqrt_v2r8(rsq12);
266 rinv20 = gmx_fjsp_invsqrt_v2r8(rsq20);
267 rinv21 = gmx_fjsp_invsqrt_v2r8(rsq21);
268 rinv22 = gmx_fjsp_invsqrt_v2r8(rsq22);
269
270 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
271 rinvsq01 = _fjsp_mul_v2r8(rinv01,rinv01);
272 rinvsq02 = _fjsp_mul_v2r8(rinv02,rinv02);
273 rinvsq10 = _fjsp_mul_v2r8(rinv10,rinv10);
274 rinvsq11 = _fjsp_mul_v2r8(rinv11,rinv11);
275 rinvsq12 = _fjsp_mul_v2r8(rinv12,rinv12);
276 rinvsq20 = _fjsp_mul_v2r8(rinv20,rinv20);
277 rinvsq21 = _fjsp_mul_v2r8(rinv21,rinv21);
278 rinvsq22 = _fjsp_mul_v2r8(rinv22,rinv22);
279
280 fjx0 = _fjsp_setzero_v2r8();
281 fjy0 = _fjsp_setzero_v2r8();
282 fjz0 = _fjsp_setzero_v2r8();
283 fjx1 = _fjsp_setzero_v2r8();
284 fjy1 = _fjsp_setzero_v2r8();
285 fjz1 = _fjsp_setzero_v2r8();
286 fjx2 = _fjsp_setzero_v2r8();
287 fjy2 = _fjsp_setzero_v2r8();
288 fjz2 = _fjsp_setzero_v2r8();
289
290 /**************************
291 * CALCULATE INTERACTIONS *
292 **************************/
293
294 if (gmx_fjsp_any_lt_v2r8(rsq00,rcutoff2))
295 {
296
297 r00 = _fjsp_mul_v2r8(rsq00,rinv00);
298
299 /* EWALD ELECTROSTATICS */
300
301 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
302 ewrt = _fjsp_mul_v2r8(r00,ewtabscale);
303 itab_tmp = _fjsp_dtox_v2r8(ewrt);
304 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
305 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
306
307 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
308 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
309 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
310 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
311 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
312 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
313 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
314 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
315 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(rinv00,velec));
316 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq00,rinv00),_fjsp_sub_v2r8(rinvsq00,felec));
317
318 d = _fjsp_sub_v2r8(r00,rswitch);
319 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
320 d2 = _fjsp_mul_v2r8(d,d);
321 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
322
323 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
324
325 /* Evaluate switch function */
326 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
327 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv00,_fjsp_mul_v2r8(velec,dsw)) );
328 velec = _fjsp_mul_v2r8(velec,sw);
329 cutoff_mask = _fjsp_cmplt_v2r8(rsq00,rcutoff2);
330
331 /* Update potential sum for this i atom from the interaction with this j atom. */
332 velec = _fjsp_and_v2r8(velec,cutoff_mask);
333 velecsum = _fjsp_add_v2r8(velecsum,velec);
334
335 fscal = felec;
336
337 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
338
339 /* Update vectorial force */
340 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
341 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
342 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
343
344 fjx0 = _fjsp_madd_v2r8(dx00,fscal,fjx0);
345 fjy0 = _fjsp_madd_v2r8(dy00,fscal,fjy0);
346 fjz0 = _fjsp_madd_v2r8(dz00,fscal,fjz0);
347
348 }
349
350 /**************************
351 * CALCULATE INTERACTIONS *
352 **************************/
353
354 if (gmx_fjsp_any_lt_v2r8(rsq01,rcutoff2))
355 {
356
357 r01 = _fjsp_mul_v2r8(rsq01,rinv01);
358
359 /* EWALD ELECTROSTATICS */
360
361 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
362 ewrt = _fjsp_mul_v2r8(r01,ewtabscale);
363 itab_tmp = _fjsp_dtox_v2r8(ewrt);
364 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
365 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
366
367 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
368 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
369 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
370 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
371 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
372 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
373 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
374 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
375 velec = _fjsp_mul_v2r8(qq01,_fjsp_sub_v2r8(rinv01,velec));
376 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq01,rinv01),_fjsp_sub_v2r8(rinvsq01,felec));
377
378 d = _fjsp_sub_v2r8(r01,rswitch);
379 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
380 d2 = _fjsp_mul_v2r8(d,d);
381 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
382
383 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
384
385 /* Evaluate switch function */
386 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
387 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv01,_fjsp_mul_v2r8(velec,dsw)) );
388 velec = _fjsp_mul_v2r8(velec,sw);
389 cutoff_mask = _fjsp_cmplt_v2r8(rsq01,rcutoff2);
390
391 /* Update potential sum for this i atom from the interaction with this j atom. */
392 velec = _fjsp_and_v2r8(velec,cutoff_mask);
393 velecsum = _fjsp_add_v2r8(velecsum,velec);
394
395 fscal = felec;
396
397 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
398
399 /* Update vectorial force */
400 fix0 = _fjsp_madd_v2r8(dx01,fscal,fix0);
401 fiy0 = _fjsp_madd_v2r8(dy01,fscal,fiy0);
402 fiz0 = _fjsp_madd_v2r8(dz01,fscal,fiz0);
403
404 fjx1 = _fjsp_madd_v2r8(dx01,fscal,fjx1);
405 fjy1 = _fjsp_madd_v2r8(dy01,fscal,fjy1);
406 fjz1 = _fjsp_madd_v2r8(dz01,fscal,fjz1);
407
408 }
409
410 /**************************
411 * CALCULATE INTERACTIONS *
412 **************************/
413
414 if (gmx_fjsp_any_lt_v2r8(rsq02,rcutoff2))
415 {
416
417 r02 = _fjsp_mul_v2r8(rsq02,rinv02);
418
419 /* EWALD ELECTROSTATICS */
420
421 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
422 ewrt = _fjsp_mul_v2r8(r02,ewtabscale);
423 itab_tmp = _fjsp_dtox_v2r8(ewrt);
424 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
425 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
426
427 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
428 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
429 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
430 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
431 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
432 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
433 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
434 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
435 velec = _fjsp_mul_v2r8(qq02,_fjsp_sub_v2r8(rinv02,velec));
436 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq02,rinv02),_fjsp_sub_v2r8(rinvsq02,felec));
437
438 d = _fjsp_sub_v2r8(r02,rswitch);
439 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
440 d2 = _fjsp_mul_v2r8(d,d);
441 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
442
443 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
444
445 /* Evaluate switch function */
446 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
447 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv02,_fjsp_mul_v2r8(velec,dsw)) );
448 velec = _fjsp_mul_v2r8(velec,sw);
449 cutoff_mask = _fjsp_cmplt_v2r8(rsq02,rcutoff2);
450
451 /* Update potential sum for this i atom from the interaction with this j atom. */
452 velec = _fjsp_and_v2r8(velec,cutoff_mask);
453 velecsum = _fjsp_add_v2r8(velecsum,velec);
454
455 fscal = felec;
456
457 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
458
459 /* Update vectorial force */
460 fix0 = _fjsp_madd_v2r8(dx02,fscal,fix0);
461 fiy0 = _fjsp_madd_v2r8(dy02,fscal,fiy0);
462 fiz0 = _fjsp_madd_v2r8(dz02,fscal,fiz0);
463
464 fjx2 = _fjsp_madd_v2r8(dx02,fscal,fjx2);
465 fjy2 = _fjsp_madd_v2r8(dy02,fscal,fjy2);
466 fjz2 = _fjsp_madd_v2r8(dz02,fscal,fjz2);
467
468 }
469
470 /**************************
471 * CALCULATE INTERACTIONS *
472 **************************/
473
474 if (gmx_fjsp_any_lt_v2r8(rsq10,rcutoff2))
475 {
476
477 r10 = _fjsp_mul_v2r8(rsq10,rinv10);
478
479 /* EWALD ELECTROSTATICS */
480
481 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
482 ewrt = _fjsp_mul_v2r8(r10,ewtabscale);
483 itab_tmp = _fjsp_dtox_v2r8(ewrt);
484 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
485 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
486
487 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
488 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
489 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
490 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
491 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
492 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
493 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
494 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
495 velec = _fjsp_mul_v2r8(qq10,_fjsp_sub_v2r8(rinv10,velec));
496 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq10,rinv10),_fjsp_sub_v2r8(rinvsq10,felec));
497
498 d = _fjsp_sub_v2r8(r10,rswitch);
499 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
500 d2 = _fjsp_mul_v2r8(d,d);
501 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
502
503 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
504
505 /* Evaluate switch function */
506 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
507 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv10,_fjsp_mul_v2r8(velec,dsw)) );
508 velec = _fjsp_mul_v2r8(velec,sw);
509 cutoff_mask = _fjsp_cmplt_v2r8(rsq10,rcutoff2);
510
511 /* Update potential sum for this i atom from the interaction with this j atom. */
512 velec = _fjsp_and_v2r8(velec,cutoff_mask);
513 velecsum = _fjsp_add_v2r8(velecsum,velec);
514
515 fscal = felec;
516
517 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
518
519 /* Update vectorial force */
520 fix1 = _fjsp_madd_v2r8(dx10,fscal,fix1);
521 fiy1 = _fjsp_madd_v2r8(dy10,fscal,fiy1);
522 fiz1 = _fjsp_madd_v2r8(dz10,fscal,fiz1);
523
524 fjx0 = _fjsp_madd_v2r8(dx10,fscal,fjx0);
525 fjy0 = _fjsp_madd_v2r8(dy10,fscal,fjy0);
526 fjz0 = _fjsp_madd_v2r8(dz10,fscal,fjz0);
527
528 }
529
530 /**************************
531 * CALCULATE INTERACTIONS *
532 **************************/
533
534 if (gmx_fjsp_any_lt_v2r8(rsq11,rcutoff2))
535 {
536
537 r11 = _fjsp_mul_v2r8(rsq11,rinv11);
538
539 /* EWALD ELECTROSTATICS */
540
541 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
542 ewrt = _fjsp_mul_v2r8(r11,ewtabscale);
543 itab_tmp = _fjsp_dtox_v2r8(ewrt);
544 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
545 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
546
547 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
548 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
549 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
550 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
551 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
552 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
553 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
554 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
555 velec = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(rinv11,velec));
556 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq11,rinv11),_fjsp_sub_v2r8(rinvsq11,felec));
557
558 d = _fjsp_sub_v2r8(r11,rswitch);
559 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
560 d2 = _fjsp_mul_v2r8(d,d);
561 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
562
563 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
564
565 /* Evaluate switch function */
566 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
567 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv11,_fjsp_mul_v2r8(velec,dsw)) );
568 velec = _fjsp_mul_v2r8(velec,sw);
569 cutoff_mask = _fjsp_cmplt_v2r8(rsq11,rcutoff2);
570
571 /* Update potential sum for this i atom from the interaction with this j atom. */
572 velec = _fjsp_and_v2r8(velec,cutoff_mask);
573 velecsum = _fjsp_add_v2r8(velecsum,velec);
574
575 fscal = felec;
576
577 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
578
579 /* Update vectorial force */
580 fix1 = _fjsp_madd_v2r8(dx11,fscal,fix1);
581 fiy1 = _fjsp_madd_v2r8(dy11,fscal,fiy1);
582 fiz1 = _fjsp_madd_v2r8(dz11,fscal,fiz1);
583
584 fjx1 = _fjsp_madd_v2r8(dx11,fscal,fjx1);
585 fjy1 = _fjsp_madd_v2r8(dy11,fscal,fjy1);
586 fjz1 = _fjsp_madd_v2r8(dz11,fscal,fjz1);
587
588 }
589
590 /**************************
591 * CALCULATE INTERACTIONS *
592 **************************/
593
594 if (gmx_fjsp_any_lt_v2r8(rsq12,rcutoff2))
595 {
596
597 r12 = _fjsp_mul_v2r8(rsq12,rinv12);
598
599 /* EWALD ELECTROSTATICS */
600
601 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
602 ewrt = _fjsp_mul_v2r8(r12,ewtabscale);
603 itab_tmp = _fjsp_dtox_v2r8(ewrt);
604 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
605 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
606
607 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
608 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
609 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
610 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
611 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
612 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
613 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
614 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
615 velec = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(rinv12,velec));
616 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq12,rinv12),_fjsp_sub_v2r8(rinvsq12,felec));
617
618 d = _fjsp_sub_v2r8(r12,rswitch);
619 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
620 d2 = _fjsp_mul_v2r8(d,d);
621 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
622
623 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
624
625 /* Evaluate switch function */
626 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
627 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv12,_fjsp_mul_v2r8(velec,dsw)) );
628 velec = _fjsp_mul_v2r8(velec,sw);
629 cutoff_mask = _fjsp_cmplt_v2r8(rsq12,rcutoff2);
630
631 /* Update potential sum for this i atom from the interaction with this j atom. */
632 velec = _fjsp_and_v2r8(velec,cutoff_mask);
633 velecsum = _fjsp_add_v2r8(velecsum,velec);
634
635 fscal = felec;
636
637 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
638
639 /* Update vectorial force */
640 fix1 = _fjsp_madd_v2r8(dx12,fscal,fix1);
641 fiy1 = _fjsp_madd_v2r8(dy12,fscal,fiy1);
642 fiz1 = _fjsp_madd_v2r8(dz12,fscal,fiz1);
643
644 fjx2 = _fjsp_madd_v2r8(dx12,fscal,fjx2);
645 fjy2 = _fjsp_madd_v2r8(dy12,fscal,fjy2);
646 fjz2 = _fjsp_madd_v2r8(dz12,fscal,fjz2);
647
648 }
649
650 /**************************
651 * CALCULATE INTERACTIONS *
652 **************************/
653
654 if (gmx_fjsp_any_lt_v2r8(rsq20,rcutoff2))
655 {
656
657 r20 = _fjsp_mul_v2r8(rsq20,rinv20);
658
659 /* EWALD ELECTROSTATICS */
660
661 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
662 ewrt = _fjsp_mul_v2r8(r20,ewtabscale);
663 itab_tmp = _fjsp_dtox_v2r8(ewrt);
664 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
665 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
666
667 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
668 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
669 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
670 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
671 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
672 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
673 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
674 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
675 velec = _fjsp_mul_v2r8(qq20,_fjsp_sub_v2r8(rinv20,velec));
676 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq20,rinv20),_fjsp_sub_v2r8(rinvsq20,felec));
677
678 d = _fjsp_sub_v2r8(r20,rswitch);
679 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
680 d2 = _fjsp_mul_v2r8(d,d);
681 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
682
683 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
684
685 /* Evaluate switch function */
686 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
687 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv20,_fjsp_mul_v2r8(velec,dsw)) );
688 velec = _fjsp_mul_v2r8(velec,sw);
689 cutoff_mask = _fjsp_cmplt_v2r8(rsq20,rcutoff2);
690
691 /* Update potential sum for this i atom from the interaction with this j atom. */
692 velec = _fjsp_and_v2r8(velec,cutoff_mask);
693 velecsum = _fjsp_add_v2r8(velecsum,velec);
694
695 fscal = felec;
696
697 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
698
699 /* Update vectorial force */
700 fix2 = _fjsp_madd_v2r8(dx20,fscal,fix2);
701 fiy2 = _fjsp_madd_v2r8(dy20,fscal,fiy2);
702 fiz2 = _fjsp_madd_v2r8(dz20,fscal,fiz2);
703
704 fjx0 = _fjsp_madd_v2r8(dx20,fscal,fjx0);
705 fjy0 = _fjsp_madd_v2r8(dy20,fscal,fjy0);
706 fjz0 = _fjsp_madd_v2r8(dz20,fscal,fjz0);
707
708 }
709
710 /**************************
711 * CALCULATE INTERACTIONS *
712 **************************/
713
714 if (gmx_fjsp_any_lt_v2r8(rsq21,rcutoff2))
715 {
716
717 r21 = _fjsp_mul_v2r8(rsq21,rinv21);
718
719 /* EWALD ELECTROSTATICS */
720
721 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
722 ewrt = _fjsp_mul_v2r8(r21,ewtabscale);
723 itab_tmp = _fjsp_dtox_v2r8(ewrt);
724 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
725 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
726
727 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
728 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
729 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
730 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
731 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
732 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
733 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
734 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
735 velec = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(rinv21,velec));
736 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq21,rinv21),_fjsp_sub_v2r8(rinvsq21,felec));
737
738 d = _fjsp_sub_v2r8(r21,rswitch);
739 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
740 d2 = _fjsp_mul_v2r8(d,d);
741 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
742
743 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
744
745 /* Evaluate switch function */
746 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
747 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv21,_fjsp_mul_v2r8(velec,dsw)) );
748 velec = _fjsp_mul_v2r8(velec,sw);
749 cutoff_mask = _fjsp_cmplt_v2r8(rsq21,rcutoff2);
750
751 /* Update potential sum for this i atom from the interaction with this j atom. */
752 velec = _fjsp_and_v2r8(velec,cutoff_mask);
753 velecsum = _fjsp_add_v2r8(velecsum,velec);
754
755 fscal = felec;
756
757 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
758
759 /* Update vectorial force */
760 fix2 = _fjsp_madd_v2r8(dx21,fscal,fix2);
761 fiy2 = _fjsp_madd_v2r8(dy21,fscal,fiy2);
762 fiz2 = _fjsp_madd_v2r8(dz21,fscal,fiz2);
763
764 fjx1 = _fjsp_madd_v2r8(dx21,fscal,fjx1);
765 fjy1 = _fjsp_madd_v2r8(dy21,fscal,fjy1);
766 fjz1 = _fjsp_madd_v2r8(dz21,fscal,fjz1);
767
768 }
769
770 /**************************
771 * CALCULATE INTERACTIONS *
772 **************************/
773
774 if (gmx_fjsp_any_lt_v2r8(rsq22,rcutoff2))
775 {
776
777 r22 = _fjsp_mul_v2r8(rsq22,rinv22);
778
779 /* EWALD ELECTROSTATICS */
780
781 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
782 ewrt = _fjsp_mul_v2r8(r22,ewtabscale);
783 itab_tmp = _fjsp_dtox_v2r8(ewrt);
784 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
785 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
786
787 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
788 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
789 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
790 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
791 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
792 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
793 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
794 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
795 velec = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(rinv22,velec));
796 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq22,rinv22),_fjsp_sub_v2r8(rinvsq22,felec));
797
798 d = _fjsp_sub_v2r8(r22,rswitch);
799 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
800 d2 = _fjsp_mul_v2r8(d,d);
801 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
802
803 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
804
805 /* Evaluate switch function */
806 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
807 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv22,_fjsp_mul_v2r8(velec,dsw)) );
808 velec = _fjsp_mul_v2r8(velec,sw);
809 cutoff_mask = _fjsp_cmplt_v2r8(rsq22,rcutoff2);
810
811 /* Update potential sum for this i atom from the interaction with this j atom. */
812 velec = _fjsp_and_v2r8(velec,cutoff_mask);
813 velecsum = _fjsp_add_v2r8(velecsum,velec);
814
815 fscal = felec;
816
817 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
818
819 /* Update vectorial force */
820 fix2 = _fjsp_madd_v2r8(dx22,fscal,fix2);
821 fiy2 = _fjsp_madd_v2r8(dy22,fscal,fiy2);
822 fiz2 = _fjsp_madd_v2r8(dz22,fscal,fiz2);
823
824 fjx2 = _fjsp_madd_v2r8(dx22,fscal,fjx2);
825 fjy2 = _fjsp_madd_v2r8(dy22,fscal,fjy2);
826 fjz2 = _fjsp_madd_v2r8(dz22,fscal,fjz2);
827
828 }
829
830 gmx_fjsp_decrement_3rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
831
832 /* Inner loop uses 612 flops */
833 }
834
835 if(jidx<j_index_end)
836 {
837
838 jnrA = jjnr[jidx];
839 j_coord_offsetA = DIM*jnrA;
840
841 /* load j atom coordinates */
842 gmx_fjsp_load_3rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
843 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
844
845 /* Calculate displacement vector */
846 dx00 = _fjsp_sub_v2r8(ix0,jx0);
847 dy00 = _fjsp_sub_v2r8(iy0,jy0);
848 dz00 = _fjsp_sub_v2r8(iz0,jz0);
849 dx01 = _fjsp_sub_v2r8(ix0,jx1);
850 dy01 = _fjsp_sub_v2r8(iy0,jy1);
851 dz01 = _fjsp_sub_v2r8(iz0,jz1);
852 dx02 = _fjsp_sub_v2r8(ix0,jx2);
853 dy02 = _fjsp_sub_v2r8(iy0,jy2);
854 dz02 = _fjsp_sub_v2r8(iz0,jz2);
855 dx10 = _fjsp_sub_v2r8(ix1,jx0);
856 dy10 = _fjsp_sub_v2r8(iy1,jy0);
857 dz10 = _fjsp_sub_v2r8(iz1,jz0);
858 dx11 = _fjsp_sub_v2r8(ix1,jx1);
859 dy11 = _fjsp_sub_v2r8(iy1,jy1);
860 dz11 = _fjsp_sub_v2r8(iz1,jz1);
861 dx12 = _fjsp_sub_v2r8(ix1,jx2);
862 dy12 = _fjsp_sub_v2r8(iy1,jy2);
863 dz12 = _fjsp_sub_v2r8(iz1,jz2);
864 dx20 = _fjsp_sub_v2r8(ix2,jx0);
865 dy20 = _fjsp_sub_v2r8(iy2,jy0);
866 dz20 = _fjsp_sub_v2r8(iz2,jz0);
867 dx21 = _fjsp_sub_v2r8(ix2,jx1);
868 dy21 = _fjsp_sub_v2r8(iy2,jy1);
869 dz21 = _fjsp_sub_v2r8(iz2,jz1);
870 dx22 = _fjsp_sub_v2r8(ix2,jx2);
871 dy22 = _fjsp_sub_v2r8(iy2,jy2);
872 dz22 = _fjsp_sub_v2r8(iz2,jz2);
873
874 /* Calculate squared distance and things based on it */
875 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
876 rsq01 = gmx_fjsp_calc_rsq_v2r8(dx01,dy01,dz01);
877 rsq02 = gmx_fjsp_calc_rsq_v2r8(dx02,dy02,dz02);
878 rsq10 = gmx_fjsp_calc_rsq_v2r8(dx10,dy10,dz10);
879 rsq11 = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
880 rsq12 = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
881 rsq20 = gmx_fjsp_calc_rsq_v2r8(dx20,dy20,dz20);
882 rsq21 = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
883 rsq22 = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
884
885 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
886 rinv01 = gmx_fjsp_invsqrt_v2r8(rsq01);
887 rinv02 = gmx_fjsp_invsqrt_v2r8(rsq02);
888 rinv10 = gmx_fjsp_invsqrt_v2r8(rsq10);
889 rinv11 = gmx_fjsp_invsqrt_v2r8(rsq11);
890 rinv12 = gmx_fjsp_invsqrt_v2r8(rsq12);
891 rinv20 = gmx_fjsp_invsqrt_v2r8(rsq20);
892 rinv21 = gmx_fjsp_invsqrt_v2r8(rsq21);
893 rinv22 = gmx_fjsp_invsqrt_v2r8(rsq22);
894
895 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
896 rinvsq01 = _fjsp_mul_v2r8(rinv01,rinv01);
897 rinvsq02 = _fjsp_mul_v2r8(rinv02,rinv02);
898 rinvsq10 = _fjsp_mul_v2r8(rinv10,rinv10);
899 rinvsq11 = _fjsp_mul_v2r8(rinv11,rinv11);
900 rinvsq12 = _fjsp_mul_v2r8(rinv12,rinv12);
901 rinvsq20 = _fjsp_mul_v2r8(rinv20,rinv20);
902 rinvsq21 = _fjsp_mul_v2r8(rinv21,rinv21);
903 rinvsq22 = _fjsp_mul_v2r8(rinv22,rinv22);
904
905 fjx0 = _fjsp_setzero_v2r8();
906 fjy0 = _fjsp_setzero_v2r8();
907 fjz0 = _fjsp_setzero_v2r8();
908 fjx1 = _fjsp_setzero_v2r8();
909 fjy1 = _fjsp_setzero_v2r8();
910 fjz1 = _fjsp_setzero_v2r8();
911 fjx2 = _fjsp_setzero_v2r8();
912 fjy2 = _fjsp_setzero_v2r8();
913 fjz2 = _fjsp_setzero_v2r8();
914
915 /**************************
916 * CALCULATE INTERACTIONS *
917 **************************/
918
919 if (gmx_fjsp_any_lt_v2r8(rsq00,rcutoff2))
920 {
921
922 r00 = _fjsp_mul_v2r8(rsq00,rinv00);
923
924 /* EWALD ELECTROSTATICS */
925
926 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
927 ewrt = _fjsp_mul_v2r8(r00,ewtabscale);
928 itab_tmp = _fjsp_dtox_v2r8(ewrt);
929 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
930 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
931
932 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
933 ewtabD = _fjsp_setzero_v2r8();
934 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
935 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
936 ewtabFn = _fjsp_setzero_v2r8();
937 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
938 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
939 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
940 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(rinv00,velec));
941 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq00,rinv00),_fjsp_sub_v2r8(rinvsq00,felec));
942
943 d = _fjsp_sub_v2r8(r00,rswitch);
944 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
945 d2 = _fjsp_mul_v2r8(d,d);
946 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
947
948 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
949
950 /* Evaluate switch function */
951 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
952 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv00,_fjsp_mul_v2r8(velec,dsw)) );
953 velec = _fjsp_mul_v2r8(velec,sw);
954 cutoff_mask = _fjsp_cmplt_v2r8(rsq00,rcutoff2);
955
956 /* Update potential sum for this i atom from the interaction with this j atom. */
957 velec = _fjsp_and_v2r8(velec,cutoff_mask);
958 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
959 velecsum = _fjsp_add_v2r8(velecsum,velec);
960
961 fscal = felec;
962
963 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
964
965 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
966
967 /* Update vectorial force */
968 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
969 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
970 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
971
972 fjx0 = _fjsp_madd_v2r8(dx00,fscal,fjx0);
973 fjy0 = _fjsp_madd_v2r8(dy00,fscal,fjy0);
974 fjz0 = _fjsp_madd_v2r8(dz00,fscal,fjz0);
975
976 }
977
978 /**************************
979 * CALCULATE INTERACTIONS *
980 **************************/
981
982 if (gmx_fjsp_any_lt_v2r8(rsq01,rcutoff2))
983 {
984
985 r01 = _fjsp_mul_v2r8(rsq01,rinv01);
986
987 /* EWALD ELECTROSTATICS */
988
989 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
990 ewrt = _fjsp_mul_v2r8(r01,ewtabscale);
991 itab_tmp = _fjsp_dtox_v2r8(ewrt);
992 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
993 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
994
995 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
996 ewtabD = _fjsp_setzero_v2r8();
997 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
998 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
999 ewtabFn = _fjsp_setzero_v2r8();
1000 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1001 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1002 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1003 velec = _fjsp_mul_v2r8(qq01,_fjsp_sub_v2r8(rinv01,velec));
1004 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq01,rinv01),_fjsp_sub_v2r8(rinvsq01,felec));
1005
1006 d = _fjsp_sub_v2r8(r01,rswitch);
1007 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1008 d2 = _fjsp_mul_v2r8(d,d);
1009 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1010
1011 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1012
1013 /* Evaluate switch function */
1014 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1015 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv01,_fjsp_mul_v2r8(velec,dsw)) );
1016 velec = _fjsp_mul_v2r8(velec,sw);
1017 cutoff_mask = _fjsp_cmplt_v2r8(rsq01,rcutoff2);
1018
1019 /* Update potential sum for this i atom from the interaction with this j atom. */
1020 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1021 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1022 velecsum = _fjsp_add_v2r8(velecsum,velec);
1023
1024 fscal = felec;
1025
1026 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1027
1028 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1029
1030 /* Update vectorial force */
1031 fix0 = _fjsp_madd_v2r8(dx01,fscal,fix0);
1032 fiy0 = _fjsp_madd_v2r8(dy01,fscal,fiy0);
1033 fiz0 = _fjsp_madd_v2r8(dz01,fscal,fiz0);
1034
1035 fjx1 = _fjsp_madd_v2r8(dx01,fscal,fjx1);
1036 fjy1 = _fjsp_madd_v2r8(dy01,fscal,fjy1);
1037 fjz1 = _fjsp_madd_v2r8(dz01,fscal,fjz1);
1038
1039 }
1040
1041 /**************************
1042 * CALCULATE INTERACTIONS *
1043 **************************/
1044
1045 if (gmx_fjsp_any_lt_v2r8(rsq02,rcutoff2))
1046 {
1047
1048 r02 = _fjsp_mul_v2r8(rsq02,rinv02);
1049
1050 /* EWALD ELECTROSTATICS */
1051
1052 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1053 ewrt = _fjsp_mul_v2r8(r02,ewtabscale);
1054 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1055 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1056 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1057
1058 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1059 ewtabD = _fjsp_setzero_v2r8();
1060 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1061 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1062 ewtabFn = _fjsp_setzero_v2r8();
1063 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1064 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1065 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1066 velec = _fjsp_mul_v2r8(qq02,_fjsp_sub_v2r8(rinv02,velec));
1067 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq02,rinv02),_fjsp_sub_v2r8(rinvsq02,felec));
1068
1069 d = _fjsp_sub_v2r8(r02,rswitch);
1070 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1071 d2 = _fjsp_mul_v2r8(d,d);
1072 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1073
1074 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1075
1076 /* Evaluate switch function */
1077 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1078 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv02,_fjsp_mul_v2r8(velec,dsw)) );
1079 velec = _fjsp_mul_v2r8(velec,sw);
1080 cutoff_mask = _fjsp_cmplt_v2r8(rsq02,rcutoff2);
1081
1082 /* Update potential sum for this i atom from the interaction with this j atom. */
1083 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1084 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1085 velecsum = _fjsp_add_v2r8(velecsum,velec);
1086
1087 fscal = felec;
1088
1089 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1090
1091 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1092
1093 /* Update vectorial force */
1094 fix0 = _fjsp_madd_v2r8(dx02,fscal,fix0);
1095 fiy0 = _fjsp_madd_v2r8(dy02,fscal,fiy0);
1096 fiz0 = _fjsp_madd_v2r8(dz02,fscal,fiz0);
1097
1098 fjx2 = _fjsp_madd_v2r8(dx02,fscal,fjx2);
1099 fjy2 = _fjsp_madd_v2r8(dy02,fscal,fjy2);
1100 fjz2 = _fjsp_madd_v2r8(dz02,fscal,fjz2);
1101
1102 }
1103
1104 /**************************
1105 * CALCULATE INTERACTIONS *
1106 **************************/
1107
1108 if (gmx_fjsp_any_lt_v2r8(rsq10,rcutoff2))
1109 {
1110
1111 r10 = _fjsp_mul_v2r8(rsq10,rinv10);
1112
1113 /* EWALD ELECTROSTATICS */
1114
1115 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1116 ewrt = _fjsp_mul_v2r8(r10,ewtabscale);
1117 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1118 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1119 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1120
1121 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1122 ewtabD = _fjsp_setzero_v2r8();
1123 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1124 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1125 ewtabFn = _fjsp_setzero_v2r8();
1126 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1127 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1128 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1129 velec = _fjsp_mul_v2r8(qq10,_fjsp_sub_v2r8(rinv10,velec));
1130 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq10,rinv10),_fjsp_sub_v2r8(rinvsq10,felec));
1131
1132 d = _fjsp_sub_v2r8(r10,rswitch);
1133 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1134 d2 = _fjsp_mul_v2r8(d,d);
1135 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1136
1137 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1138
1139 /* Evaluate switch function */
1140 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1141 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv10,_fjsp_mul_v2r8(velec,dsw)) );
1142 velec = _fjsp_mul_v2r8(velec,sw);
1143 cutoff_mask = _fjsp_cmplt_v2r8(rsq10,rcutoff2);
1144
1145 /* Update potential sum for this i atom from the interaction with this j atom. */
1146 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1147 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1148 velecsum = _fjsp_add_v2r8(velecsum,velec);
1149
1150 fscal = felec;
1151
1152 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1153
1154 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1155
1156 /* Update vectorial force */
1157 fix1 = _fjsp_madd_v2r8(dx10,fscal,fix1);
1158 fiy1 = _fjsp_madd_v2r8(dy10,fscal,fiy1);
1159 fiz1 = _fjsp_madd_v2r8(dz10,fscal,fiz1);
1160
1161 fjx0 = _fjsp_madd_v2r8(dx10,fscal,fjx0);
1162 fjy0 = _fjsp_madd_v2r8(dy10,fscal,fjy0);
1163 fjz0 = _fjsp_madd_v2r8(dz10,fscal,fjz0);
1164
1165 }
1166
1167 /**************************
1168 * CALCULATE INTERACTIONS *
1169 **************************/
1170
1171 if (gmx_fjsp_any_lt_v2r8(rsq11,rcutoff2))
1172 {
1173
1174 r11 = _fjsp_mul_v2r8(rsq11,rinv11);
1175
1176 /* EWALD ELECTROSTATICS */
1177
1178 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1179 ewrt = _fjsp_mul_v2r8(r11,ewtabscale);
1180 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1181 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1182 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1183
1184 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1185 ewtabD = _fjsp_setzero_v2r8();
1186 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1187 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1188 ewtabFn = _fjsp_setzero_v2r8();
1189 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1190 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1191 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1192 velec = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(rinv11,velec));
1193 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq11,rinv11),_fjsp_sub_v2r8(rinvsq11,felec));
1194
1195 d = _fjsp_sub_v2r8(r11,rswitch);
1196 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1197 d2 = _fjsp_mul_v2r8(d,d);
1198 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1199
1200 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1201
1202 /* Evaluate switch function */
1203 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1204 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv11,_fjsp_mul_v2r8(velec,dsw)) );
1205 velec = _fjsp_mul_v2r8(velec,sw);
1206 cutoff_mask = _fjsp_cmplt_v2r8(rsq11,rcutoff2);
1207
1208 /* Update potential sum for this i atom from the interaction with this j atom. */
1209 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1210 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1211 velecsum = _fjsp_add_v2r8(velecsum,velec);
1212
1213 fscal = felec;
1214
1215 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1216
1217 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1218
1219 /* Update vectorial force */
1220 fix1 = _fjsp_madd_v2r8(dx11,fscal,fix1);
1221 fiy1 = _fjsp_madd_v2r8(dy11,fscal,fiy1);
1222 fiz1 = _fjsp_madd_v2r8(dz11,fscal,fiz1);
1223
1224 fjx1 = _fjsp_madd_v2r8(dx11,fscal,fjx1);
1225 fjy1 = _fjsp_madd_v2r8(dy11,fscal,fjy1);
1226 fjz1 = _fjsp_madd_v2r8(dz11,fscal,fjz1);
1227
1228 }
1229
1230 /**************************
1231 * CALCULATE INTERACTIONS *
1232 **************************/
1233
1234 if (gmx_fjsp_any_lt_v2r8(rsq12,rcutoff2))
1235 {
1236
1237 r12 = _fjsp_mul_v2r8(rsq12,rinv12);
1238
1239 /* EWALD ELECTROSTATICS */
1240
1241 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1242 ewrt = _fjsp_mul_v2r8(r12,ewtabscale);
1243 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1244 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1245 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1246
1247 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1248 ewtabD = _fjsp_setzero_v2r8();
1249 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1250 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1251 ewtabFn = _fjsp_setzero_v2r8();
1252 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1253 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1254 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1255 velec = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(rinv12,velec));
1256 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq12,rinv12),_fjsp_sub_v2r8(rinvsq12,felec));
1257
1258 d = _fjsp_sub_v2r8(r12,rswitch);
1259 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1260 d2 = _fjsp_mul_v2r8(d,d);
1261 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1262
1263 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1264
1265 /* Evaluate switch function */
1266 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1267 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv12,_fjsp_mul_v2r8(velec,dsw)) );
1268 velec = _fjsp_mul_v2r8(velec,sw);
1269 cutoff_mask = _fjsp_cmplt_v2r8(rsq12,rcutoff2);
1270
1271 /* Update potential sum for this i atom from the interaction with this j atom. */
1272 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1273 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1274 velecsum = _fjsp_add_v2r8(velecsum,velec);
1275
1276 fscal = felec;
1277
1278 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1279
1280 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1281
1282 /* Update vectorial force */
1283 fix1 = _fjsp_madd_v2r8(dx12,fscal,fix1);
1284 fiy1 = _fjsp_madd_v2r8(dy12,fscal,fiy1);
1285 fiz1 = _fjsp_madd_v2r8(dz12,fscal,fiz1);
1286
1287 fjx2 = _fjsp_madd_v2r8(dx12,fscal,fjx2);
1288 fjy2 = _fjsp_madd_v2r8(dy12,fscal,fjy2);
1289 fjz2 = _fjsp_madd_v2r8(dz12,fscal,fjz2);
1290
1291 }
1292
1293 /**************************
1294 * CALCULATE INTERACTIONS *
1295 **************************/
1296
1297 if (gmx_fjsp_any_lt_v2r8(rsq20,rcutoff2))
1298 {
1299
1300 r20 = _fjsp_mul_v2r8(rsq20,rinv20);
1301
1302 /* EWALD ELECTROSTATICS */
1303
1304 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1305 ewrt = _fjsp_mul_v2r8(r20,ewtabscale);
1306 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1307 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1308 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1309
1310 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1311 ewtabD = _fjsp_setzero_v2r8();
1312 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1313 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1314 ewtabFn = _fjsp_setzero_v2r8();
1315 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1316 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1317 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1318 velec = _fjsp_mul_v2r8(qq20,_fjsp_sub_v2r8(rinv20,velec));
1319 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq20,rinv20),_fjsp_sub_v2r8(rinvsq20,felec));
1320
1321 d = _fjsp_sub_v2r8(r20,rswitch);
1322 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1323 d2 = _fjsp_mul_v2r8(d,d);
1324 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1325
1326 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1327
1328 /* Evaluate switch function */
1329 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1330 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv20,_fjsp_mul_v2r8(velec,dsw)) );
1331 velec = _fjsp_mul_v2r8(velec,sw);
1332 cutoff_mask = _fjsp_cmplt_v2r8(rsq20,rcutoff2);
1333
1334 /* Update potential sum for this i atom from the interaction with this j atom. */
1335 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1336 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1337 velecsum = _fjsp_add_v2r8(velecsum,velec);
1338
1339 fscal = felec;
1340
1341 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1342
1343 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1344
1345 /* Update vectorial force */
1346 fix2 = _fjsp_madd_v2r8(dx20,fscal,fix2);
1347 fiy2 = _fjsp_madd_v2r8(dy20,fscal,fiy2);
1348 fiz2 = _fjsp_madd_v2r8(dz20,fscal,fiz2);
1349
1350 fjx0 = _fjsp_madd_v2r8(dx20,fscal,fjx0);
1351 fjy0 = _fjsp_madd_v2r8(dy20,fscal,fjy0);
1352 fjz0 = _fjsp_madd_v2r8(dz20,fscal,fjz0);
1353
1354 }
1355
1356 /**************************
1357 * CALCULATE INTERACTIONS *
1358 **************************/
1359
1360 if (gmx_fjsp_any_lt_v2r8(rsq21,rcutoff2))
1361 {
1362
1363 r21 = _fjsp_mul_v2r8(rsq21,rinv21);
1364
1365 /* EWALD ELECTROSTATICS */
1366
1367 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1368 ewrt = _fjsp_mul_v2r8(r21,ewtabscale);
1369 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1370 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1371 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1372
1373 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1374 ewtabD = _fjsp_setzero_v2r8();
1375 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1376 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1377 ewtabFn = _fjsp_setzero_v2r8();
1378 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1379 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1380 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1381 velec = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(rinv21,velec));
1382 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq21,rinv21),_fjsp_sub_v2r8(rinvsq21,felec));
1383
1384 d = _fjsp_sub_v2r8(r21,rswitch);
1385 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1386 d2 = _fjsp_mul_v2r8(d,d);
1387 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1388
1389 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1390
1391 /* Evaluate switch function */
1392 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1393 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv21,_fjsp_mul_v2r8(velec,dsw)) );
1394 velec = _fjsp_mul_v2r8(velec,sw);
1395 cutoff_mask = _fjsp_cmplt_v2r8(rsq21,rcutoff2);
1396
1397 /* Update potential sum for this i atom from the interaction with this j atom. */
1398 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1399 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1400 velecsum = _fjsp_add_v2r8(velecsum,velec);
1401
1402 fscal = felec;
1403
1404 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1405
1406 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1407
1408 /* Update vectorial force */
1409 fix2 = _fjsp_madd_v2r8(dx21,fscal,fix2);
1410 fiy2 = _fjsp_madd_v2r8(dy21,fscal,fiy2);
1411 fiz2 = _fjsp_madd_v2r8(dz21,fscal,fiz2);
1412
1413 fjx1 = _fjsp_madd_v2r8(dx21,fscal,fjx1);
1414 fjy1 = _fjsp_madd_v2r8(dy21,fscal,fjy1);
1415 fjz1 = _fjsp_madd_v2r8(dz21,fscal,fjz1);
1416
1417 }
1418
1419 /**************************
1420 * CALCULATE INTERACTIONS *
1421 **************************/
1422
1423 if (gmx_fjsp_any_lt_v2r8(rsq22,rcutoff2))
1424 {
1425
1426 r22 = _fjsp_mul_v2r8(rsq22,rinv22);
1427
1428 /* EWALD ELECTROSTATICS */
1429
1430 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1431 ewrt = _fjsp_mul_v2r8(r22,ewtabscale);
1432 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1433 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1434 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1435
1436 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1437 ewtabD = _fjsp_setzero_v2r8();
1438 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1439 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1440 ewtabFn = _fjsp_setzero_v2r8();
1441 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1442 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1443 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1444 velec = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(rinv22,velec));
1445 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq22,rinv22),_fjsp_sub_v2r8(rinvsq22,felec));
1446
1447 d = _fjsp_sub_v2r8(r22,rswitch);
1448 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1449 d2 = _fjsp_mul_v2r8(d,d);
1450 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1451
1452 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1453
1454 /* Evaluate switch function */
1455 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1456 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv22,_fjsp_mul_v2r8(velec,dsw)) );
1457 velec = _fjsp_mul_v2r8(velec,sw);
1458 cutoff_mask = _fjsp_cmplt_v2r8(rsq22,rcutoff2);
1459
1460 /* Update potential sum for this i atom from the interaction with this j atom. */
1461 velec = _fjsp_and_v2r8(velec,cutoff_mask);
1462 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
1463 velecsum = _fjsp_add_v2r8(velecsum,velec);
1464
1465 fscal = felec;
1466
1467 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1468
1469 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
1470
1471 /* Update vectorial force */
1472 fix2 = _fjsp_madd_v2r8(dx22,fscal,fix2);
1473 fiy2 = _fjsp_madd_v2r8(dy22,fscal,fiy2);
1474 fiz2 = _fjsp_madd_v2r8(dz22,fscal,fiz2);
1475
1476 fjx2 = _fjsp_madd_v2r8(dx22,fscal,fjx2);
1477 fjy2 = _fjsp_madd_v2r8(dy22,fscal,fjy2);
1478 fjz2 = _fjsp_madd_v2r8(dz22,fscal,fjz2);
1479
1480 }
1481
1482 gmx_fjsp_decrement_3rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1483
1484 /* Inner loop uses 612 flops */
1485 }
1486
1487 /* End of innermost loop */
1488
1489 gmx_fjsp_update_iforce_3atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1490 f+i_coord_offset,fshift+i_shift_offset);
1491
1492 ggid = gid[iidx];
1493 /* Update potential energies */
1494 gmx_fjsp_update_1pot_v2r8(velecsum,kernel_data->energygrp_elec+ggid);
1495
1496 /* Increment number of inner iterations */
1497 inneriter += j_index_end - j_index_start;
1498
1499 /* Outer loop uses 19 flops */
1500 }
1501
1502 /* Increment number of outer iterations */
1503 outeriter += nri;
1504
1505 /* Update outer/inner flops */
1506
1507 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_VF,outeriter*19 + inneriter*612);
1508 }
1509 /*
1510 * Gromacs nonbonded kernel: nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_sparc64_hpc_ace_double
1511 * Electrostatics interaction: Ewald
1512 * VdW interaction: None
1513 * Geometry: Water3-Water3
1514 * Calculate force/pot: Force
1515 */
1516 void
1517 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_sparc64_hpc_ace_double
1518 (t_nblist * gmx_restrict nlist,
1519 rvec * gmx_restrict xx,
1520 rvec * gmx_restrict ff,
1521 t_forcerec * gmx_restrict fr,
1522 t_mdatoms * gmx_restrict mdatoms,
1523 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1524 t_nrnb * gmx_restrict nrnb)
1525 {
1526 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1527 * just 0 for non-waters.
1528 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
1529 * jnr indices corresponding to data put in the four positions in the SIMD register.
1530 */
1531 int i_shift_offset,i_coord_offset,outeriter,inneriter;
1532 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1533 int jnrA,jnrB;
1534 int j_coord_offsetA,j_coord_offsetB;
1535 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
1536 real rcutoff_scalar;
1537 real *shiftvec,*fshift,*x,*f;
1538 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1539 int vdwioffset0;
1540 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1541 int vdwioffset1;
1542 _fjsp_v2r8 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1543 int vdwioffset2;
1544 _fjsp_v2r8 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1545 int vdwjidx0A,vdwjidx0B;
1546 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1547 int vdwjidx1A,vdwjidx1B;
1548 _fjsp_v2r8 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1549 int vdwjidx2A,vdwjidx2B;
1550 _fjsp_v2r8 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1551 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1552 _fjsp_v2r8 dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1553 _fjsp_v2r8 dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1554 _fjsp_v2r8 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1555 _fjsp_v2r8 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1556 _fjsp_v2r8 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1557 _fjsp_v2r8 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1558 _fjsp_v2r8 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1559 _fjsp_v2r8 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1560 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
1561 real *charge;
1562 _fjsp_v2r8 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1563 real *ewtab;
1564 _fjsp_v2r8 rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1565 real rswitch_scalar,d_scalar;
1566 _fjsp_v2r8 itab_tmp;
1567 _fjsp_v2r8 dummy_mask,cutoff_mask;
1568 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
1569 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
1570 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
1571
1572 x = xx[0];
1573 f = ff[0];
1574
1575 nri = nlist->nri;
1576 iinr = nlist->iinr;
1577 jindex = nlist->jindex;
1578 jjnr = nlist->jjnr;
1579 shiftidx = nlist->shift;
1580 gid = nlist->gid;
1581 shiftvec = fr->shift_vec[0];
1582 fshift = fr->fshift[0];
1583 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
1584 charge = mdatoms->chargeA;
1585
1586 sh_ewald = gmx_fjsp_set1_v2r8(fr->ic->sh_ewald);
1587 ewtab = fr->ic->tabq_coul_FDV0;
1588 ewtabscale = gmx_fjsp_set1_v2r8(fr->ic->tabq_scale);
1589 ewtabhalfspace = gmx_fjsp_set1_v2r8(0.5/fr->ic->tabq_scale);
1590
1591 /* Setup water-specific parameters */
1592 inr = nlist->iinr[0];
1593 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+0]));
1594 iq1 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+1]));
1595 iq2 = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+2]));
1596
1597 jq0 = gmx_fjsp_set1_v2r8(charge[inr+0]);
1598 jq1 = gmx_fjsp_set1_v2r8(charge[inr+1]);
1599 jq2 = gmx_fjsp_set1_v2r8(charge[inr+2]);
1600 qq00 = _fjsp_mul_v2r8(iq0,jq0);
1601 qq01 = _fjsp_mul_v2r8(iq0,jq1);
1602 qq02 = _fjsp_mul_v2r8(iq0,jq2);
1603 qq10 = _fjsp_mul_v2r8(iq1,jq0);
1604 qq11 = _fjsp_mul_v2r8(iq1,jq1);
1605 qq12 = _fjsp_mul_v2r8(iq1,jq2);
1606 qq20 = _fjsp_mul_v2r8(iq2,jq0);
1607 qq21 = _fjsp_mul_v2r8(iq2,jq1);
1608 qq22 = _fjsp_mul_v2r8(iq2,jq2);
1609
1610 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1611 rcutoff_scalar = fr->rcoulomb;
1612 rcutoff = gmx_fjsp_set1_v2r8(rcutoff_scalar);
1613 rcutoff2 = _fjsp_mul_v2r8(rcutoff,rcutoff);
1614
1615 rswitch_scalar = fr->rcoulomb_switch;
1616 rswitch = gmx_fjsp_set1_v2r8(rswitch_scalar);
1617 /* Setup switch parameters */
1618 d_scalar = rcutoff_scalar-rswitch_scalar;
1619 d = gmx_fjsp_set1_v2r8(d_scalar);
1620 swV3 = gmx_fjsp_set1_v2r8(-10.0/(d_scalar*d_scalar*d_scalar));
1621 swV4 = gmx_fjsp_set1_v2r8( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1622 swV5 = gmx_fjsp_set1_v2r8( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1623 swF2 = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar));
1624 swF3 = gmx_fjsp_set1_v2r8( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1625 swF4 = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1626
1627 /* Avoid stupid compiler warnings */
1628 jnrA = jnrB = 0;
1629 j_coord_offsetA = 0;
1630 j_coord_offsetB = 0;
1631
1632 outeriter = 0;
1633 inneriter = 0;
1634
1635 /* Start outer loop over neighborlists */
1636 for(iidx=0; iidx<nri; iidx++)
1637 {
1638 /* Load shift vector for this list */
1639 i_shift_offset = DIM*shiftidx[iidx];
1640
1641 /* Load limits for loop over neighbors */
1642 j_index_start = jindex[iidx];
1643 j_index_end = jindex[iidx+1];
1644
1645 /* Get outer coordinate index */
1646 inr = iinr[iidx];
1647 i_coord_offset = DIM*inr;
1648
1649 /* Load i particle coords and add shift vector */
1650 gmx_fjsp_load_shift_and_3rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
1651 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1652
1653 fix0 = _fjsp_setzero_v2r8();
1654 fiy0 = _fjsp_setzero_v2r8();
1655 fiz0 = _fjsp_setzero_v2r8();
1656 fix1 = _fjsp_setzero_v2r8();
1657 fiy1 = _fjsp_setzero_v2r8();
1658 fiz1 = _fjsp_setzero_v2r8();
1659 fix2 = _fjsp_setzero_v2r8();
1660 fiy2 = _fjsp_setzero_v2r8();
1661 fiz2 = _fjsp_setzero_v2r8();
1662
1663 /* Start inner kernel loop */
1664 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
1665 {
1666
1667 /* Get j neighbor index, and coordinate index */
1668 jnrA = jjnr[jidx];
1669 jnrB = jjnr[jidx+1];
1670 j_coord_offsetA = DIM*jnrA;
1671 j_coord_offsetB = DIM*jnrB;
1672
1673 /* load j atom coordinates */
1674 gmx_fjsp_load_3rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
1675 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1676
1677 /* Calculate displacement vector */
1678 dx00 = _fjsp_sub_v2r8(ix0,jx0);
1679 dy00 = _fjsp_sub_v2r8(iy0,jy0);
1680 dz00 = _fjsp_sub_v2r8(iz0,jz0);
1681 dx01 = _fjsp_sub_v2r8(ix0,jx1);
1682 dy01 = _fjsp_sub_v2r8(iy0,jy1);
1683 dz01 = _fjsp_sub_v2r8(iz0,jz1);
1684 dx02 = _fjsp_sub_v2r8(ix0,jx2);
1685 dy02 = _fjsp_sub_v2r8(iy0,jy2);
1686 dz02 = _fjsp_sub_v2r8(iz0,jz2);
1687 dx10 = _fjsp_sub_v2r8(ix1,jx0);
1688 dy10 = _fjsp_sub_v2r8(iy1,jy0);
1689 dz10 = _fjsp_sub_v2r8(iz1,jz0);
1690 dx11 = _fjsp_sub_v2r8(ix1,jx1);
1691 dy11 = _fjsp_sub_v2r8(iy1,jy1);
1692 dz11 = _fjsp_sub_v2r8(iz1,jz1);
1693 dx12 = _fjsp_sub_v2r8(ix1,jx2);
1694 dy12 = _fjsp_sub_v2r8(iy1,jy2);
1695 dz12 = _fjsp_sub_v2r8(iz1,jz2);
1696 dx20 = _fjsp_sub_v2r8(ix2,jx0);
1697 dy20 = _fjsp_sub_v2r8(iy2,jy0);
1698 dz20 = _fjsp_sub_v2r8(iz2,jz0);
1699 dx21 = _fjsp_sub_v2r8(ix2,jx1);
1700 dy21 = _fjsp_sub_v2r8(iy2,jy1);
1701 dz21 = _fjsp_sub_v2r8(iz2,jz1);
1702 dx22 = _fjsp_sub_v2r8(ix2,jx2);
1703 dy22 = _fjsp_sub_v2r8(iy2,jy2);
1704 dz22 = _fjsp_sub_v2r8(iz2,jz2);
1705
1706 /* Calculate squared distance and things based on it */
1707 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
1708 rsq01 = gmx_fjsp_calc_rsq_v2r8(dx01,dy01,dz01);
1709 rsq02 = gmx_fjsp_calc_rsq_v2r8(dx02,dy02,dz02);
1710 rsq10 = gmx_fjsp_calc_rsq_v2r8(dx10,dy10,dz10);
1711 rsq11 = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
1712 rsq12 = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
1713 rsq20 = gmx_fjsp_calc_rsq_v2r8(dx20,dy20,dz20);
1714 rsq21 = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
1715 rsq22 = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
1716
1717 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
1718 rinv01 = gmx_fjsp_invsqrt_v2r8(rsq01);
1719 rinv02 = gmx_fjsp_invsqrt_v2r8(rsq02);
1720 rinv10 = gmx_fjsp_invsqrt_v2r8(rsq10);
1721 rinv11 = gmx_fjsp_invsqrt_v2r8(rsq11);
1722 rinv12 = gmx_fjsp_invsqrt_v2r8(rsq12);
1723 rinv20 = gmx_fjsp_invsqrt_v2r8(rsq20);
1724 rinv21 = gmx_fjsp_invsqrt_v2r8(rsq21);
1725 rinv22 = gmx_fjsp_invsqrt_v2r8(rsq22);
1726
1727 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
1728 rinvsq01 = _fjsp_mul_v2r8(rinv01,rinv01);
1729 rinvsq02 = _fjsp_mul_v2r8(rinv02,rinv02);
1730 rinvsq10 = _fjsp_mul_v2r8(rinv10,rinv10);
1731 rinvsq11 = _fjsp_mul_v2r8(rinv11,rinv11);
1732 rinvsq12 = _fjsp_mul_v2r8(rinv12,rinv12);
1733 rinvsq20 = _fjsp_mul_v2r8(rinv20,rinv20);
1734 rinvsq21 = _fjsp_mul_v2r8(rinv21,rinv21);
1735 rinvsq22 = _fjsp_mul_v2r8(rinv22,rinv22);
1736
1737 fjx0 = _fjsp_setzero_v2r8();
1738 fjy0 = _fjsp_setzero_v2r8();
1739 fjz0 = _fjsp_setzero_v2r8();
1740 fjx1 = _fjsp_setzero_v2r8();
1741 fjy1 = _fjsp_setzero_v2r8();
1742 fjz1 = _fjsp_setzero_v2r8();
1743 fjx2 = _fjsp_setzero_v2r8();
1744 fjy2 = _fjsp_setzero_v2r8();
1745 fjz2 = _fjsp_setzero_v2r8();
1746
1747 /**************************
1748 * CALCULATE INTERACTIONS *
1749 **************************/
1750
1751 if (gmx_fjsp_any_lt_v2r8(rsq00,rcutoff2))
1752 {
1753
1754 r00 = _fjsp_mul_v2r8(rsq00,rinv00);
1755
1756 /* EWALD ELECTROSTATICS */
1757
1758 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1759 ewrt = _fjsp_mul_v2r8(r00,ewtabscale);
1760 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1761 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1762 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1763
1764 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1765 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
1766 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1767 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1768 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
1769 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1770 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1771 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1772 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(rinv00,velec));
1773 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq00,rinv00),_fjsp_sub_v2r8(rinvsq00,felec));
1774
1775 d = _fjsp_sub_v2r8(r00,rswitch);
1776 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1777 d2 = _fjsp_mul_v2r8(d,d);
1778 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1779
1780 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1781
1782 /* Evaluate switch function */
1783 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1784 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv00,_fjsp_mul_v2r8(velec,dsw)) );
1785 cutoff_mask = _fjsp_cmplt_v2r8(rsq00,rcutoff2);
1786
1787 fscal = felec;
1788
1789 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1790
1791 /* Update vectorial force */
1792 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
1793 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
1794 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
1795
1796 fjx0 = _fjsp_madd_v2r8(dx00,fscal,fjx0);
1797 fjy0 = _fjsp_madd_v2r8(dy00,fscal,fjy0);
1798 fjz0 = _fjsp_madd_v2r8(dz00,fscal,fjz0);
1799
1800 }
1801
1802 /**************************
1803 * CALCULATE INTERACTIONS *
1804 **************************/
1805
1806 if (gmx_fjsp_any_lt_v2r8(rsq01,rcutoff2))
1807 {
1808
1809 r01 = _fjsp_mul_v2r8(rsq01,rinv01);
1810
1811 /* EWALD ELECTROSTATICS */
1812
1813 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1814 ewrt = _fjsp_mul_v2r8(r01,ewtabscale);
1815 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1816 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1817 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1818
1819 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1820 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
1821 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1822 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1823 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
1824 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1825 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1826 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1827 velec = _fjsp_mul_v2r8(qq01,_fjsp_sub_v2r8(rinv01,velec));
1828 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq01,rinv01),_fjsp_sub_v2r8(rinvsq01,felec));
1829
1830 d = _fjsp_sub_v2r8(r01,rswitch);
1831 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1832 d2 = _fjsp_mul_v2r8(d,d);
1833 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1834
1835 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1836
1837 /* Evaluate switch function */
1838 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1839 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv01,_fjsp_mul_v2r8(velec,dsw)) );
1840 cutoff_mask = _fjsp_cmplt_v2r8(rsq01,rcutoff2);
1841
1842 fscal = felec;
1843
1844 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1845
1846 /* Update vectorial force */
1847 fix0 = _fjsp_madd_v2r8(dx01,fscal,fix0);
1848 fiy0 = _fjsp_madd_v2r8(dy01,fscal,fiy0);
1849 fiz0 = _fjsp_madd_v2r8(dz01,fscal,fiz0);
1850
1851 fjx1 = _fjsp_madd_v2r8(dx01,fscal,fjx1);
1852 fjy1 = _fjsp_madd_v2r8(dy01,fscal,fjy1);
1853 fjz1 = _fjsp_madd_v2r8(dz01,fscal,fjz1);
1854
1855 }
1856
1857 /**************************
1858 * CALCULATE INTERACTIONS *
1859 **************************/
1860
1861 if (gmx_fjsp_any_lt_v2r8(rsq02,rcutoff2))
1862 {
1863
1864 r02 = _fjsp_mul_v2r8(rsq02,rinv02);
1865
1866 /* EWALD ELECTROSTATICS */
1867
1868 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1869 ewrt = _fjsp_mul_v2r8(r02,ewtabscale);
1870 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1871 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1872 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1873
1874 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1875 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
1876 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1877 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1878 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
1879 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1880 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1881 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1882 velec = _fjsp_mul_v2r8(qq02,_fjsp_sub_v2r8(rinv02,velec));
1883 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq02,rinv02),_fjsp_sub_v2r8(rinvsq02,felec));
1884
1885 d = _fjsp_sub_v2r8(r02,rswitch);
1886 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1887 d2 = _fjsp_mul_v2r8(d,d);
1888 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1889
1890 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1891
1892 /* Evaluate switch function */
1893 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1894 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv02,_fjsp_mul_v2r8(velec,dsw)) );
1895 cutoff_mask = _fjsp_cmplt_v2r8(rsq02,rcutoff2);
1896
1897 fscal = felec;
1898
1899 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1900
1901 /* Update vectorial force */
1902 fix0 = _fjsp_madd_v2r8(dx02,fscal,fix0);
1903 fiy0 = _fjsp_madd_v2r8(dy02,fscal,fiy0);
1904 fiz0 = _fjsp_madd_v2r8(dz02,fscal,fiz0);
1905
1906 fjx2 = _fjsp_madd_v2r8(dx02,fscal,fjx2);
1907 fjy2 = _fjsp_madd_v2r8(dy02,fscal,fjy2);
1908 fjz2 = _fjsp_madd_v2r8(dz02,fscal,fjz2);
1909
1910 }
1911
1912 /**************************
1913 * CALCULATE INTERACTIONS *
1914 **************************/
1915
1916 if (gmx_fjsp_any_lt_v2r8(rsq10,rcutoff2))
1917 {
1918
1919 r10 = _fjsp_mul_v2r8(rsq10,rinv10);
1920
1921 /* EWALD ELECTROSTATICS */
1922
1923 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1924 ewrt = _fjsp_mul_v2r8(r10,ewtabscale);
1925 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1926 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1927 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1928
1929 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1930 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
1931 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1932 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1933 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
1934 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1935 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1936 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1937 velec = _fjsp_mul_v2r8(qq10,_fjsp_sub_v2r8(rinv10,velec));
1938 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq10,rinv10),_fjsp_sub_v2r8(rinvsq10,felec));
1939
1940 d = _fjsp_sub_v2r8(r10,rswitch);
1941 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1942 d2 = _fjsp_mul_v2r8(d,d);
1943 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1944
1945 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
1946
1947 /* Evaluate switch function */
1948 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1949 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv10,_fjsp_mul_v2r8(velec,dsw)) );
1950 cutoff_mask = _fjsp_cmplt_v2r8(rsq10,rcutoff2);
1951
1952 fscal = felec;
1953
1954 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
1955
1956 /* Update vectorial force */
1957 fix1 = _fjsp_madd_v2r8(dx10,fscal,fix1);
1958 fiy1 = _fjsp_madd_v2r8(dy10,fscal,fiy1);
1959 fiz1 = _fjsp_madd_v2r8(dz10,fscal,fiz1);
1960
1961 fjx0 = _fjsp_madd_v2r8(dx10,fscal,fjx0);
1962 fjy0 = _fjsp_madd_v2r8(dy10,fscal,fjy0);
1963 fjz0 = _fjsp_madd_v2r8(dz10,fscal,fjz0);
1964
1965 }
1966
1967 /**************************
1968 * CALCULATE INTERACTIONS *
1969 **************************/
1970
1971 if (gmx_fjsp_any_lt_v2r8(rsq11,rcutoff2))
1972 {
1973
1974 r11 = _fjsp_mul_v2r8(rsq11,rinv11);
1975
1976 /* EWALD ELECTROSTATICS */
1977
1978 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1979 ewrt = _fjsp_mul_v2r8(r11,ewtabscale);
1980 itab_tmp = _fjsp_dtox_v2r8(ewrt);
1981 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
1982 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
1983
1984 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
1985 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
1986 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
1987 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
1988 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
1989 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
1990 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
1991 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
1992 velec = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(rinv11,velec));
1993 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq11,rinv11),_fjsp_sub_v2r8(rinvsq11,felec));
1994
1995 d = _fjsp_sub_v2r8(r11,rswitch);
1996 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
1997 d2 = _fjsp_mul_v2r8(d,d);
1998 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
1999
2000 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2001
2002 /* Evaluate switch function */
2003 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2004 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv11,_fjsp_mul_v2r8(velec,dsw)) );
2005 cutoff_mask = _fjsp_cmplt_v2r8(rsq11,rcutoff2);
2006
2007 fscal = felec;
2008
2009 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2010
2011 /* Update vectorial force */
2012 fix1 = _fjsp_madd_v2r8(dx11,fscal,fix1);
2013 fiy1 = _fjsp_madd_v2r8(dy11,fscal,fiy1);
2014 fiz1 = _fjsp_madd_v2r8(dz11,fscal,fiz1);
2015
2016 fjx1 = _fjsp_madd_v2r8(dx11,fscal,fjx1);
2017 fjy1 = _fjsp_madd_v2r8(dy11,fscal,fjy1);
2018 fjz1 = _fjsp_madd_v2r8(dz11,fscal,fjz1);
2019
2020 }
2021
2022 /**************************
2023 * CALCULATE INTERACTIONS *
2024 **************************/
2025
2026 if (gmx_fjsp_any_lt_v2r8(rsq12,rcutoff2))
2027 {
2028
2029 r12 = _fjsp_mul_v2r8(rsq12,rinv12);
2030
2031 /* EWALD ELECTROSTATICS */
2032
2033 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2034 ewrt = _fjsp_mul_v2r8(r12,ewtabscale);
2035 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2036 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2037 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2038
2039 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2040 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
2041 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2042 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2043 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
2044 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2045 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2046 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2047 velec = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(rinv12,velec));
2048 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq12,rinv12),_fjsp_sub_v2r8(rinvsq12,felec));
2049
2050 d = _fjsp_sub_v2r8(r12,rswitch);
2051 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2052 d2 = _fjsp_mul_v2r8(d,d);
2053 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2054
2055 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2056
2057 /* Evaluate switch function */
2058 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2059 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv12,_fjsp_mul_v2r8(velec,dsw)) );
2060 cutoff_mask = _fjsp_cmplt_v2r8(rsq12,rcutoff2);
2061
2062 fscal = felec;
2063
2064 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2065
2066 /* Update vectorial force */
2067 fix1 = _fjsp_madd_v2r8(dx12,fscal,fix1);
2068 fiy1 = _fjsp_madd_v2r8(dy12,fscal,fiy1);
2069 fiz1 = _fjsp_madd_v2r8(dz12,fscal,fiz1);
2070
2071 fjx2 = _fjsp_madd_v2r8(dx12,fscal,fjx2);
2072 fjy2 = _fjsp_madd_v2r8(dy12,fscal,fjy2);
2073 fjz2 = _fjsp_madd_v2r8(dz12,fscal,fjz2);
2074
2075 }
2076
2077 /**************************
2078 * CALCULATE INTERACTIONS *
2079 **************************/
2080
2081 if (gmx_fjsp_any_lt_v2r8(rsq20,rcutoff2))
2082 {
2083
2084 r20 = _fjsp_mul_v2r8(rsq20,rinv20);
2085
2086 /* EWALD ELECTROSTATICS */
2087
2088 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2089 ewrt = _fjsp_mul_v2r8(r20,ewtabscale);
2090 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2091 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2092 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2093
2094 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2095 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
2096 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2097 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2098 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
2099 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2100 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2101 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2102 velec = _fjsp_mul_v2r8(qq20,_fjsp_sub_v2r8(rinv20,velec));
2103 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq20,rinv20),_fjsp_sub_v2r8(rinvsq20,felec));
2104
2105 d = _fjsp_sub_v2r8(r20,rswitch);
2106 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2107 d2 = _fjsp_mul_v2r8(d,d);
2108 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2109
2110 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2111
2112 /* Evaluate switch function */
2113 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2114 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv20,_fjsp_mul_v2r8(velec,dsw)) );
2115 cutoff_mask = _fjsp_cmplt_v2r8(rsq20,rcutoff2);
2116
2117 fscal = felec;
2118
2119 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2120
2121 /* Update vectorial force */
2122 fix2 = _fjsp_madd_v2r8(dx20,fscal,fix2);
2123 fiy2 = _fjsp_madd_v2r8(dy20,fscal,fiy2);
2124 fiz2 = _fjsp_madd_v2r8(dz20,fscal,fiz2);
2125
2126 fjx0 = _fjsp_madd_v2r8(dx20,fscal,fjx0);
2127 fjy0 = _fjsp_madd_v2r8(dy20,fscal,fjy0);
2128 fjz0 = _fjsp_madd_v2r8(dz20,fscal,fjz0);
2129
2130 }
2131
2132 /**************************
2133 * CALCULATE INTERACTIONS *
2134 **************************/
2135
2136 if (gmx_fjsp_any_lt_v2r8(rsq21,rcutoff2))
2137 {
2138
2139 r21 = _fjsp_mul_v2r8(rsq21,rinv21);
2140
2141 /* EWALD ELECTROSTATICS */
2142
2143 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2144 ewrt = _fjsp_mul_v2r8(r21,ewtabscale);
2145 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2146 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2147 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2148
2149 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2150 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
2151 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2152 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2153 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
2154 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2155 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2156 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2157 velec = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(rinv21,velec));
2158 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq21,rinv21),_fjsp_sub_v2r8(rinvsq21,felec));
2159
2160 d = _fjsp_sub_v2r8(r21,rswitch);
2161 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2162 d2 = _fjsp_mul_v2r8(d,d);
2163 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2164
2165 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2166
2167 /* Evaluate switch function */
2168 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2169 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv21,_fjsp_mul_v2r8(velec,dsw)) );
2170 cutoff_mask = _fjsp_cmplt_v2r8(rsq21,rcutoff2);
2171
2172 fscal = felec;
2173
2174 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2175
2176 /* Update vectorial force */
2177 fix2 = _fjsp_madd_v2r8(dx21,fscal,fix2);
2178 fiy2 = _fjsp_madd_v2r8(dy21,fscal,fiy2);
2179 fiz2 = _fjsp_madd_v2r8(dz21,fscal,fiz2);
2180
2181 fjx1 = _fjsp_madd_v2r8(dx21,fscal,fjx1);
2182 fjy1 = _fjsp_madd_v2r8(dy21,fscal,fjy1);
2183 fjz1 = _fjsp_madd_v2r8(dz21,fscal,fjz1);
2184
2185 }
2186
2187 /**************************
2188 * CALCULATE INTERACTIONS *
2189 **************************/
2190
2191 if (gmx_fjsp_any_lt_v2r8(rsq22,rcutoff2))
2192 {
2193
2194 r22 = _fjsp_mul_v2r8(rsq22,rinv22);
2195
2196 /* EWALD ELECTROSTATICS */
2197
2198 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2199 ewrt = _fjsp_mul_v2r8(r22,ewtabscale);
2200 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2201 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2202 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2203
2204 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2205 ewtabD = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
2206 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2207 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2208 ewtabFn = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
2209 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2210 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2211 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2212 velec = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(rinv22,velec));
2213 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq22,rinv22),_fjsp_sub_v2r8(rinvsq22,felec));
2214
2215 d = _fjsp_sub_v2r8(r22,rswitch);
2216 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2217 d2 = _fjsp_mul_v2r8(d,d);
2218 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2219
2220 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2221
2222 /* Evaluate switch function */
2223 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2224 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv22,_fjsp_mul_v2r8(velec,dsw)) );
2225 cutoff_mask = _fjsp_cmplt_v2r8(rsq22,rcutoff2);
2226
2227 fscal = felec;
2228
2229 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2230
2231 /* Update vectorial force */
2232 fix2 = _fjsp_madd_v2r8(dx22,fscal,fix2);
2233 fiy2 = _fjsp_madd_v2r8(dy22,fscal,fiy2);
2234 fiz2 = _fjsp_madd_v2r8(dz22,fscal,fiz2);
2235
2236 fjx2 = _fjsp_madd_v2r8(dx22,fscal,fjx2);
2237 fjy2 = _fjsp_madd_v2r8(dy22,fscal,fjy2);
2238 fjz2 = _fjsp_madd_v2r8(dz22,fscal,fjz2);
2239
2240 }
2241
2242 gmx_fjsp_decrement_3rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2243
2244 /* Inner loop uses 585 flops */
2245 }
2246
2247 if(jidx<j_index_end)
2248 {
2249
2250 jnrA = jjnr[jidx];
2251 j_coord_offsetA = DIM*jnrA;
2252
2253 /* load j atom coordinates */
2254 gmx_fjsp_load_3rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
2255 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
2256
2257 /* Calculate displacement vector */
2258 dx00 = _fjsp_sub_v2r8(ix0,jx0);
2259 dy00 = _fjsp_sub_v2r8(iy0,jy0);
2260 dz00 = _fjsp_sub_v2r8(iz0,jz0);
2261 dx01 = _fjsp_sub_v2r8(ix0,jx1);
2262 dy01 = _fjsp_sub_v2r8(iy0,jy1);
2263 dz01 = _fjsp_sub_v2r8(iz0,jz1);
2264 dx02 = _fjsp_sub_v2r8(ix0,jx2);
2265 dy02 = _fjsp_sub_v2r8(iy0,jy2);
2266 dz02 = _fjsp_sub_v2r8(iz0,jz2);
2267 dx10 = _fjsp_sub_v2r8(ix1,jx0);
2268 dy10 = _fjsp_sub_v2r8(iy1,jy0);
2269 dz10 = _fjsp_sub_v2r8(iz1,jz0);
2270 dx11 = _fjsp_sub_v2r8(ix1,jx1);
2271 dy11 = _fjsp_sub_v2r8(iy1,jy1);
2272 dz11 = _fjsp_sub_v2r8(iz1,jz1);
2273 dx12 = _fjsp_sub_v2r8(ix1,jx2);
2274 dy12 = _fjsp_sub_v2r8(iy1,jy2);
2275 dz12 = _fjsp_sub_v2r8(iz1,jz2);
2276 dx20 = _fjsp_sub_v2r8(ix2,jx0);
2277 dy20 = _fjsp_sub_v2r8(iy2,jy0);
2278 dz20 = _fjsp_sub_v2r8(iz2,jz0);
2279 dx21 = _fjsp_sub_v2r8(ix2,jx1);
2280 dy21 = _fjsp_sub_v2r8(iy2,jy1);
2281 dz21 = _fjsp_sub_v2r8(iz2,jz1);
2282 dx22 = _fjsp_sub_v2r8(ix2,jx2);
2283 dy22 = _fjsp_sub_v2r8(iy2,jy2);
2284 dz22 = _fjsp_sub_v2r8(iz2,jz2);
2285
2286 /* Calculate squared distance and things based on it */
2287 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
2288 rsq01 = gmx_fjsp_calc_rsq_v2r8(dx01,dy01,dz01);
2289 rsq02 = gmx_fjsp_calc_rsq_v2r8(dx02,dy02,dz02);
2290 rsq10 = gmx_fjsp_calc_rsq_v2r8(dx10,dy10,dz10);
2291 rsq11 = gmx_fjsp_calc_rsq_v2r8(dx11,dy11,dz11);
2292 rsq12 = gmx_fjsp_calc_rsq_v2r8(dx12,dy12,dz12);
2293 rsq20 = gmx_fjsp_calc_rsq_v2r8(dx20,dy20,dz20);
2294 rsq21 = gmx_fjsp_calc_rsq_v2r8(dx21,dy21,dz21);
2295 rsq22 = gmx_fjsp_calc_rsq_v2r8(dx22,dy22,dz22);
2296
2297 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
2298 rinv01 = gmx_fjsp_invsqrt_v2r8(rsq01);
2299 rinv02 = gmx_fjsp_invsqrt_v2r8(rsq02);
2300 rinv10 = gmx_fjsp_invsqrt_v2r8(rsq10);
2301 rinv11 = gmx_fjsp_invsqrt_v2r8(rsq11);
2302 rinv12 = gmx_fjsp_invsqrt_v2r8(rsq12);
2303 rinv20 = gmx_fjsp_invsqrt_v2r8(rsq20);
2304 rinv21 = gmx_fjsp_invsqrt_v2r8(rsq21);
2305 rinv22 = gmx_fjsp_invsqrt_v2r8(rsq22);
2306
2307 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
2308 rinvsq01 = _fjsp_mul_v2r8(rinv01,rinv01);
2309 rinvsq02 = _fjsp_mul_v2r8(rinv02,rinv02);
2310 rinvsq10 = _fjsp_mul_v2r8(rinv10,rinv10);
2311 rinvsq11 = _fjsp_mul_v2r8(rinv11,rinv11);
2312 rinvsq12 = _fjsp_mul_v2r8(rinv12,rinv12);
2313 rinvsq20 = _fjsp_mul_v2r8(rinv20,rinv20);
2314 rinvsq21 = _fjsp_mul_v2r8(rinv21,rinv21);
2315 rinvsq22 = _fjsp_mul_v2r8(rinv22,rinv22);
2316
2317 fjx0 = _fjsp_setzero_v2r8();
2318 fjy0 = _fjsp_setzero_v2r8();
2319 fjz0 = _fjsp_setzero_v2r8();
2320 fjx1 = _fjsp_setzero_v2r8();
2321 fjy1 = _fjsp_setzero_v2r8();
2322 fjz1 = _fjsp_setzero_v2r8();
2323 fjx2 = _fjsp_setzero_v2r8();
2324 fjy2 = _fjsp_setzero_v2r8();
2325 fjz2 = _fjsp_setzero_v2r8();
2326
2327 /**************************
2328 * CALCULATE INTERACTIONS *
2329 **************************/
2330
2331 if (gmx_fjsp_any_lt_v2r8(rsq00,rcutoff2))
2332 {
2333
2334 r00 = _fjsp_mul_v2r8(rsq00,rinv00);
2335
2336 /* EWALD ELECTROSTATICS */
2337
2338 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2339 ewrt = _fjsp_mul_v2r8(r00,ewtabscale);
2340 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2341 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2342 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2343
2344 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2345 ewtabD = _fjsp_setzero_v2r8();
2346 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2347 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2348 ewtabFn = _fjsp_setzero_v2r8();
2349 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2350 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2351 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2352 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(rinv00,velec));
2353 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq00,rinv00),_fjsp_sub_v2r8(rinvsq00,felec));
2354
2355 d = _fjsp_sub_v2r8(r00,rswitch);
2356 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2357 d2 = _fjsp_mul_v2r8(d,d);
2358 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2359
2360 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2361
2362 /* Evaluate switch function */
2363 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2364 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv00,_fjsp_mul_v2r8(velec,dsw)) );
2365 cutoff_mask = _fjsp_cmplt_v2r8(rsq00,rcutoff2);
2366
2367 fscal = felec;
2368
2369 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2370
2371 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2372
2373 /* Update vectorial force */
2374 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
2375 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
2376 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
2377
2378 fjx0 = _fjsp_madd_v2r8(dx00,fscal,fjx0);
2379 fjy0 = _fjsp_madd_v2r8(dy00,fscal,fjy0);
2380 fjz0 = _fjsp_madd_v2r8(dz00,fscal,fjz0);
2381
2382 }
2383
2384 /**************************
2385 * CALCULATE INTERACTIONS *
2386 **************************/
2387
2388 if (gmx_fjsp_any_lt_v2r8(rsq01,rcutoff2))
2389 {
2390
2391 r01 = _fjsp_mul_v2r8(rsq01,rinv01);
2392
2393 /* EWALD ELECTROSTATICS */
2394
2395 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2396 ewrt = _fjsp_mul_v2r8(r01,ewtabscale);
2397 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2398 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2399 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2400
2401 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2402 ewtabD = _fjsp_setzero_v2r8();
2403 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2404 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2405 ewtabFn = _fjsp_setzero_v2r8();
2406 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2407 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2408 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2409 velec = _fjsp_mul_v2r8(qq01,_fjsp_sub_v2r8(rinv01,velec));
2410 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq01,rinv01),_fjsp_sub_v2r8(rinvsq01,felec));
2411
2412 d = _fjsp_sub_v2r8(r01,rswitch);
2413 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2414 d2 = _fjsp_mul_v2r8(d,d);
2415 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2416
2417 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2418
2419 /* Evaluate switch function */
2420 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2421 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv01,_fjsp_mul_v2r8(velec,dsw)) );
2422 cutoff_mask = _fjsp_cmplt_v2r8(rsq01,rcutoff2);
2423
2424 fscal = felec;
2425
2426 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2427
2428 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2429
2430 /* Update vectorial force */
2431 fix0 = _fjsp_madd_v2r8(dx01,fscal,fix0);
2432 fiy0 = _fjsp_madd_v2r8(dy01,fscal,fiy0);
2433 fiz0 = _fjsp_madd_v2r8(dz01,fscal,fiz0);
2434
2435 fjx1 = _fjsp_madd_v2r8(dx01,fscal,fjx1);
2436 fjy1 = _fjsp_madd_v2r8(dy01,fscal,fjy1);
2437 fjz1 = _fjsp_madd_v2r8(dz01,fscal,fjz1);
2438
2439 }
2440
2441 /**************************
2442 * CALCULATE INTERACTIONS *
2443 **************************/
2444
2445 if (gmx_fjsp_any_lt_v2r8(rsq02,rcutoff2))
2446 {
2447
2448 r02 = _fjsp_mul_v2r8(rsq02,rinv02);
2449
2450 /* EWALD ELECTROSTATICS */
2451
2452 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2453 ewrt = _fjsp_mul_v2r8(r02,ewtabscale);
2454 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2455 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2456 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2457
2458 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2459 ewtabD = _fjsp_setzero_v2r8();
2460 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2461 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2462 ewtabFn = _fjsp_setzero_v2r8();
2463 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2464 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2465 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2466 velec = _fjsp_mul_v2r8(qq02,_fjsp_sub_v2r8(rinv02,velec));
2467 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq02,rinv02),_fjsp_sub_v2r8(rinvsq02,felec));
2468
2469 d = _fjsp_sub_v2r8(r02,rswitch);
2470 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2471 d2 = _fjsp_mul_v2r8(d,d);
2472 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2473
2474 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2475
2476 /* Evaluate switch function */
2477 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2478 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv02,_fjsp_mul_v2r8(velec,dsw)) );
2479 cutoff_mask = _fjsp_cmplt_v2r8(rsq02,rcutoff2);
2480
2481 fscal = felec;
2482
2483 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2484
2485 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2486
2487 /* Update vectorial force */
2488 fix0 = _fjsp_madd_v2r8(dx02,fscal,fix0);
2489 fiy0 = _fjsp_madd_v2r8(dy02,fscal,fiy0);
2490 fiz0 = _fjsp_madd_v2r8(dz02,fscal,fiz0);
2491
2492 fjx2 = _fjsp_madd_v2r8(dx02,fscal,fjx2);
2493 fjy2 = _fjsp_madd_v2r8(dy02,fscal,fjy2);
2494 fjz2 = _fjsp_madd_v2r8(dz02,fscal,fjz2);
2495
2496 }
2497
2498 /**************************
2499 * CALCULATE INTERACTIONS *
2500 **************************/
2501
2502 if (gmx_fjsp_any_lt_v2r8(rsq10,rcutoff2))
2503 {
2504
2505 r10 = _fjsp_mul_v2r8(rsq10,rinv10);
2506
2507 /* EWALD ELECTROSTATICS */
2508
2509 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2510 ewrt = _fjsp_mul_v2r8(r10,ewtabscale);
2511 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2512 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2513 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2514
2515 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2516 ewtabD = _fjsp_setzero_v2r8();
2517 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2518 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2519 ewtabFn = _fjsp_setzero_v2r8();
2520 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2521 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2522 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2523 velec = _fjsp_mul_v2r8(qq10,_fjsp_sub_v2r8(rinv10,velec));
2524 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq10,rinv10),_fjsp_sub_v2r8(rinvsq10,felec));
2525
2526 d = _fjsp_sub_v2r8(r10,rswitch);
2527 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2528 d2 = _fjsp_mul_v2r8(d,d);
2529 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2530
2531 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2532
2533 /* Evaluate switch function */
2534 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2535 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv10,_fjsp_mul_v2r8(velec,dsw)) );
2536 cutoff_mask = _fjsp_cmplt_v2r8(rsq10,rcutoff2);
2537
2538 fscal = felec;
2539
2540 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2541
2542 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2543
2544 /* Update vectorial force */
2545 fix1 = _fjsp_madd_v2r8(dx10,fscal,fix1);
2546 fiy1 = _fjsp_madd_v2r8(dy10,fscal,fiy1);
2547 fiz1 = _fjsp_madd_v2r8(dz10,fscal,fiz1);
2548
2549 fjx0 = _fjsp_madd_v2r8(dx10,fscal,fjx0);
2550 fjy0 = _fjsp_madd_v2r8(dy10,fscal,fjy0);
2551 fjz0 = _fjsp_madd_v2r8(dz10,fscal,fjz0);
2552
2553 }
2554
2555 /**************************
2556 * CALCULATE INTERACTIONS *
2557 **************************/
2558
2559 if (gmx_fjsp_any_lt_v2r8(rsq11,rcutoff2))
2560 {
2561
2562 r11 = _fjsp_mul_v2r8(rsq11,rinv11);
2563
2564 /* EWALD ELECTROSTATICS */
2565
2566 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2567 ewrt = _fjsp_mul_v2r8(r11,ewtabscale);
2568 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2569 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2570 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2571
2572 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2573 ewtabD = _fjsp_setzero_v2r8();
2574 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2575 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2576 ewtabFn = _fjsp_setzero_v2r8();
2577 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2578 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2579 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2580 velec = _fjsp_mul_v2r8(qq11,_fjsp_sub_v2r8(rinv11,velec));
2581 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq11,rinv11),_fjsp_sub_v2r8(rinvsq11,felec));
2582
2583 d = _fjsp_sub_v2r8(r11,rswitch);
2584 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2585 d2 = _fjsp_mul_v2r8(d,d);
2586 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2587
2588 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2589
2590 /* Evaluate switch function */
2591 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2592 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv11,_fjsp_mul_v2r8(velec,dsw)) );
2593 cutoff_mask = _fjsp_cmplt_v2r8(rsq11,rcutoff2);
2594
2595 fscal = felec;
2596
2597 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2598
2599 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2600
2601 /* Update vectorial force */
2602 fix1 = _fjsp_madd_v2r8(dx11,fscal,fix1);
2603 fiy1 = _fjsp_madd_v2r8(dy11,fscal,fiy1);
2604 fiz1 = _fjsp_madd_v2r8(dz11,fscal,fiz1);
2605
2606 fjx1 = _fjsp_madd_v2r8(dx11,fscal,fjx1);
2607 fjy1 = _fjsp_madd_v2r8(dy11,fscal,fjy1);
2608 fjz1 = _fjsp_madd_v2r8(dz11,fscal,fjz1);
2609
2610 }
2611
2612 /**************************
2613 * CALCULATE INTERACTIONS *
2614 **************************/
2615
2616 if (gmx_fjsp_any_lt_v2r8(rsq12,rcutoff2))
2617 {
2618
2619 r12 = _fjsp_mul_v2r8(rsq12,rinv12);
2620
2621 /* EWALD ELECTROSTATICS */
2622
2623 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2624 ewrt = _fjsp_mul_v2r8(r12,ewtabscale);
2625 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2626 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2627 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2628
2629 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2630 ewtabD = _fjsp_setzero_v2r8();
2631 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2632 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2633 ewtabFn = _fjsp_setzero_v2r8();
2634 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2635 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2636 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2637 velec = _fjsp_mul_v2r8(qq12,_fjsp_sub_v2r8(rinv12,velec));
2638 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq12,rinv12),_fjsp_sub_v2r8(rinvsq12,felec));
2639
2640 d = _fjsp_sub_v2r8(r12,rswitch);
2641 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2642 d2 = _fjsp_mul_v2r8(d,d);
2643 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2644
2645 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2646
2647 /* Evaluate switch function */
2648 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2649 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv12,_fjsp_mul_v2r8(velec,dsw)) );
2650 cutoff_mask = _fjsp_cmplt_v2r8(rsq12,rcutoff2);
2651
2652 fscal = felec;
2653
2654 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2655
2656 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2657
2658 /* Update vectorial force */
2659 fix1 = _fjsp_madd_v2r8(dx12,fscal,fix1);
2660 fiy1 = _fjsp_madd_v2r8(dy12,fscal,fiy1);
2661 fiz1 = _fjsp_madd_v2r8(dz12,fscal,fiz1);
2662
2663 fjx2 = _fjsp_madd_v2r8(dx12,fscal,fjx2);
2664 fjy2 = _fjsp_madd_v2r8(dy12,fscal,fjy2);
2665 fjz2 = _fjsp_madd_v2r8(dz12,fscal,fjz2);
2666
2667 }
2668
2669 /**************************
2670 * CALCULATE INTERACTIONS *
2671 **************************/
2672
2673 if (gmx_fjsp_any_lt_v2r8(rsq20,rcutoff2))
2674 {
2675
2676 r20 = _fjsp_mul_v2r8(rsq20,rinv20);
2677
2678 /* EWALD ELECTROSTATICS */
2679
2680 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2681 ewrt = _fjsp_mul_v2r8(r20,ewtabscale);
2682 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2683 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2684 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2685
2686 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2687 ewtabD = _fjsp_setzero_v2r8();
2688 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2689 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2690 ewtabFn = _fjsp_setzero_v2r8();
2691 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2692 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2693 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2694 velec = _fjsp_mul_v2r8(qq20,_fjsp_sub_v2r8(rinv20,velec));
2695 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq20,rinv20),_fjsp_sub_v2r8(rinvsq20,felec));
2696
2697 d = _fjsp_sub_v2r8(r20,rswitch);
2698 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2699 d2 = _fjsp_mul_v2r8(d,d);
2700 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2701
2702 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2703
2704 /* Evaluate switch function */
2705 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2706 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv20,_fjsp_mul_v2r8(velec,dsw)) );
2707 cutoff_mask = _fjsp_cmplt_v2r8(rsq20,rcutoff2);
2708
2709 fscal = felec;
2710
2711 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2712
2713 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2714
2715 /* Update vectorial force */
2716 fix2 = _fjsp_madd_v2r8(dx20,fscal,fix2);
2717 fiy2 = _fjsp_madd_v2r8(dy20,fscal,fiy2);
2718 fiz2 = _fjsp_madd_v2r8(dz20,fscal,fiz2);
2719
2720 fjx0 = _fjsp_madd_v2r8(dx20,fscal,fjx0);
2721 fjy0 = _fjsp_madd_v2r8(dy20,fscal,fjy0);
2722 fjz0 = _fjsp_madd_v2r8(dz20,fscal,fjz0);
2723
2724 }
2725
2726 /**************************
2727 * CALCULATE INTERACTIONS *
2728 **************************/
2729
2730 if (gmx_fjsp_any_lt_v2r8(rsq21,rcutoff2))
2731 {
2732
2733 r21 = _fjsp_mul_v2r8(rsq21,rinv21);
2734
2735 /* EWALD ELECTROSTATICS */
2736
2737 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2738 ewrt = _fjsp_mul_v2r8(r21,ewtabscale);
2739 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2740 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2741 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2742
2743 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2744 ewtabD = _fjsp_setzero_v2r8();
2745 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2746 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2747 ewtabFn = _fjsp_setzero_v2r8();
2748 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2749 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2750 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2751 velec = _fjsp_mul_v2r8(qq21,_fjsp_sub_v2r8(rinv21,velec));
2752 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq21,rinv21),_fjsp_sub_v2r8(rinvsq21,felec));
2753
2754 d = _fjsp_sub_v2r8(r21,rswitch);
2755 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2756 d2 = _fjsp_mul_v2r8(d,d);
2757 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2758
2759 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2760
2761 /* Evaluate switch function */
2762 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2763 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv21,_fjsp_mul_v2r8(velec,dsw)) );
2764 cutoff_mask = _fjsp_cmplt_v2r8(rsq21,rcutoff2);
2765
2766 fscal = felec;
2767
2768 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2769
2770 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2771
2772 /* Update vectorial force */
2773 fix2 = _fjsp_madd_v2r8(dx21,fscal,fix2);
2774 fiy2 = _fjsp_madd_v2r8(dy21,fscal,fiy2);
2775 fiz2 = _fjsp_madd_v2r8(dz21,fscal,fiz2);
2776
2777 fjx1 = _fjsp_madd_v2r8(dx21,fscal,fjx1);
2778 fjy1 = _fjsp_madd_v2r8(dy21,fscal,fjy1);
2779 fjz1 = _fjsp_madd_v2r8(dz21,fscal,fjz1);
2780
2781 }
2782
2783 /**************************
2784 * CALCULATE INTERACTIONS *
2785 **************************/
2786
2787 if (gmx_fjsp_any_lt_v2r8(rsq22,rcutoff2))
2788 {
2789
2790 r22 = _fjsp_mul_v2r8(rsq22,rinv22);
2791
2792 /* EWALD ELECTROSTATICS */
2793
2794 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2795 ewrt = _fjsp_mul_v2r8(r22,ewtabscale);
2796 itab_tmp = _fjsp_dtox_v2r8(ewrt);
2797 eweps = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
2798 _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
2799
2800 ewtabF = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
2801 ewtabD = _fjsp_setzero_v2r8();
2802 GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
2803 ewtabV = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
2804 ewtabFn = _fjsp_setzero_v2r8();
2805 GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
2806 felec = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
2807 velec = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
2808 velec = _fjsp_mul_v2r8(qq22,_fjsp_sub_v2r8(rinv22,velec));
2809 felec = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq22,rinv22),_fjsp_sub_v2r8(rinvsq22,felec));
2810
2811 d = _fjsp_sub_v2r8(r22,rswitch);
2812 d = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
2813 d2 = _fjsp_mul_v2r8(d,d);
2814 sw = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
2815
2816 dsw = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
2817
2818 /* Evaluate switch function */
2819 /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2820 felec = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv22,_fjsp_mul_v2r8(velec,dsw)) );
2821 cutoff_mask = _fjsp_cmplt_v2r8(rsq22,rcutoff2);
2822
2823 fscal = felec;
2824
2825 fscal = _fjsp_and_v2r8(fscal,cutoff_mask);
2826
2827 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
2828
2829 /* Update vectorial force */
2830 fix2 = _fjsp_madd_v2r8(dx22,fscal,fix2);
2831 fiy2 = _fjsp_madd_v2r8(dy22,fscal,fiy2);
2832 fiz2 = _fjsp_madd_v2r8(dz22,fscal,fiz2);
2833
2834 fjx2 = _fjsp_madd_v2r8(dx22,fscal,fjx2);
2835 fjy2 = _fjsp_madd_v2r8(dy22,fscal,fjy2);
2836 fjz2 = _fjsp_madd_v2r8(dz22,fscal,fjz2);
2837
2838 }
2839
2840 gmx_fjsp_decrement_3rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2841
2842 /* Inner loop uses 585 flops */
2843 }
2844
2845 /* End of innermost loop */
2846
2847 gmx_fjsp_update_iforce_3atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
2848 f+i_coord_offset,fshift+i_shift_offset);
2849
2850 /* Increment number of inner iterations */
2851 inneriter += j_index_end - j_index_start;
2852
2853 /* Outer loop uses 18 flops */
2854 }
2855
2856 /* Increment number of outer iterations */
2857 outeriter += nri;
2858
2859 /* Update outer/inner flops */
2860
2861 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_F,outeriter*18 + inneriter*585);
2862 }
The ultimate molecular dynamics simulation package