src/gromacs/gmxlib/nonbonded/nb_kernel_sse2_double/kernelutil_x86_sse2_double.h

   1 /*
   2  *                This source code is part of
   3  *
   4  *                 G   R   O   M   A   C   S
   5  *
   6  * Copyright (c) 2011-2012, The GROMACS Development Team
   7  *
   8  * Gromacs is a library for molecular simulation and trajectory analysis,
   9  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  10  * a full list of developers and information, check out http://www.gromacs.org
  11  *
  12  * This program is free software; you can redistribute it and/or modify it under
  13  * the terms of the GNU Lesser General Public License as published by the Free
  14  * Software Foundation; either version 2 of the License, or (at your option) any
  15  * later version.
  16  * As a special exception, you may use this file as part of a free software
  17  * library without restriction.  Specifically, if other files instantiate
  18  * templates or use macros or inline functions from this file, or you compile
  19  * this file and link it with other files to produce an executable, this
  20  * file does not by itself cause the resulting executable to be covered by
  21  * the GNU Lesser General Public License.
  22  *
  23  * In plain-speak: do not worry about classes/macros/templates either - only
  24  * changes to the library have to be LGPL, not an application linking with it.
  25  *
  26  * To help fund GROMACS development, we humbly ask that you cite
  27  * the papers people have written on it - you can find them on the website!
  28  */
  29 #ifndef _kernelutil_x86_sse2_double_h_
  30 #define _kernelutil_x86_sse2_double_h_
  31
  32 #include <math.h>
  33
  34 #include "gmx_x86_sse2.h"
  35
  36 #include <stdio.h>
  37
  38
  39 /* Normal sum of four ymm registers */
  40 #define gmx_mm_sum4_pd(t0,t1,t2,t3)  _mm_add_pd(_mm_add_pd(t0,t1),_mm_add_pd(t2,t3))
  41
  42 static int
  43 gmx_mm_any_lt(__m128d a, __m128d b)
  44 {
  45     return _mm_movemask_pd(_mm_cmplt_pd(a,b));
  46 }
  47
  48 static gmx_inline __m128d
  49 gmx_mm_calc_rsq_pd(__m128d dx, __m128d dy, __m128d dz)
  50 {
  51     return _mm_add_pd( _mm_add_pd( _mm_mul_pd(dx,dx), _mm_mul_pd(dy,dy) ), _mm_mul_pd(dz,dz) );
  52 }
  53
  54
  55 /* Load a double value from 1-2 places, merge into xmm register */
  56 static gmx_inline __m128d
  57 gmx_mm_load_2real_swizzle_pd(const double * gmx_restrict ptrA,
  58                              const double * gmx_restrict ptrB)
  59 {
  60     return _mm_unpacklo_pd(_mm_load_sd(ptrA),_mm_load_sd(ptrB));
  61 }
  62
  63 static gmx_inline __m128d
  64 gmx_mm_load_1real_pd(const double * gmx_restrict ptrA)
  65 {
  66     return _mm_load_sd(ptrA);
  67 }
  68
  69
  70 static gmx_inline void
  71 gmx_mm_store_2real_swizzle_pd(double * gmx_restrict ptrA,
  72                               double * gmx_restrict ptrB,
  73                               __m128d xmm1)
  74 {
  75     __m128d t2;
  76
  77     t2       = _mm_unpackhi_pd(xmm1,xmm1);
  78     _mm_store_sd(ptrA,xmm1);
  79     _mm_store_sd(ptrB,t2);
  80 }
  81
  82 static gmx_inline void
  83 gmx_mm_store_1real_pd(double * gmx_restrict ptrA, __m128d xmm1)
  84 {
  85     _mm_store_sd(ptrA,xmm1);
  86 }
  87
  88
  89 /* Similar to store, but increments value in memory */
  90 static gmx_inline void
  91 gmx_mm_increment_2real_swizzle_pd(double * gmx_restrict ptrA,
  92                                   double * gmx_restrict ptrB, __m128d xmm1)
  93 {
  94     __m128d t1;
  95
  96     t1   = _mm_unpackhi_pd(xmm1,xmm1);
  97     xmm1 = _mm_add_sd(xmm1,_mm_load_sd(ptrA));
  98     t1   = _mm_add_sd(t1,_mm_load_sd(ptrB));
  99     _mm_store_sd(ptrA,xmm1);
 100     _mm_store_sd(ptrB,t1);
 101 }
 102
 103 static gmx_inline void
 104 gmx_mm_increment_1real_pd(double * gmx_restrict ptrA, __m128d xmm1)
 105 {
 106     __m128d tmp;
 107
 108     tmp = gmx_mm_load_1real_pd(ptrA);
 109     tmp = _mm_add_sd(tmp,xmm1);
 110     gmx_mm_store_1real_pd(ptrA,tmp);
 111 }
 112
 113
 114 static gmx_inline void
 115 gmx_mm_load_2pair_swizzle_pd(const double * gmx_restrict p1,
 116                              const double * gmx_restrict p2,
 117                              __m128d * gmx_restrict c6,
 118                              __m128d * gmx_restrict c12)
 119 {
 120     __m128d t1,t2,t3;
 121
 122     t1   = _mm_loadu_pd(p1);
 123     t2   = _mm_loadu_pd(p2);
 124     *c6  = _mm_unpacklo_pd(t1,t2);
 125     *c12 = _mm_unpackhi_pd(t1,t2);
 126 }
 127
 128 static gmx_inline void
 129 gmx_mm_load_1pair_swizzle_pd(const double * gmx_restrict p1,
 130                              __m128d * gmx_restrict c6,
 131                              __m128d * gmx_restrict c12)
 132 {
 133     *c6     = _mm_load_sd(p1);
 134     *c12    = _mm_load_sd(p1+1);
 135 }
 136
 137
 138
 139 static gmx_inline void
 140 gmx_mm_load_shift_and_1rvec_broadcast_pd(const double * gmx_restrict xyz_shift,
 141                                          const double * gmx_restrict xyz,
 142                                          __m128d * gmx_restrict x1,
 143                                          __m128d * gmx_restrict y1,
 144                                          __m128d * gmx_restrict z1)
 145 {
 146     __m128d mem_xy,mem_z,mem_sxy,mem_sz;
 147
 148     mem_xy  = _mm_loadu_pd(xyz);
 149     mem_z   = _mm_load_sd(xyz+2);
 150     mem_sxy = _mm_loadu_pd(xyz_shift);
 151     mem_sz  = _mm_load_sd(xyz_shift+2);
 152
 153     mem_xy  = _mm_add_pd(mem_xy,mem_sxy);
 154     mem_z   = _mm_add_pd(mem_z,mem_sz);
 155
 156     *x1  = _mm_shuffle_pd(mem_xy,mem_xy,_MM_SHUFFLE2(0,0));
 157     *y1  = _mm_shuffle_pd(mem_xy,mem_xy,_MM_SHUFFLE2(1,1));
 158     *z1  = _mm_shuffle_pd(mem_z,mem_z,_MM_SHUFFLE2(0,0));
 159 }
 160
 161
 162 static gmx_inline void
 163 gmx_mm_load_shift_and_3rvec_broadcast_pd(const double * gmx_restrict xyz_shift,
 164                                          const double * gmx_restrict xyz,
 165                                          __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 166                                          __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 167                                          __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3)
 168 {
 169     __m128d t1,t2,t3,t4,t5,sxy,sz,szx,syz;
 170
 171     t1  = _mm_loadu_pd(xyz);
 172     t2  = _mm_loadu_pd(xyz+2);
 173     t3  = _mm_loadu_pd(xyz+4);
 174     t4  = _mm_loadu_pd(xyz+6);
 175     t5  = _mm_load_sd(xyz+8);
 176
 177     sxy = _mm_loadu_pd(xyz_shift);
 178     sz  = _mm_load_sd(xyz_shift+2);
 179     szx = _mm_shuffle_pd(sz,sxy,_MM_SHUFFLE2(0,0));
 180     syz = _mm_shuffle_pd(sxy,sz,_MM_SHUFFLE2(0,1));
 181
 182     t1  = _mm_add_pd(t1,sxy);
 183     t2  = _mm_add_pd(t2,szx);
 184     t3  = _mm_add_pd(t3,syz);
 185     t4  = _mm_add_pd(t4,sxy);
 186     t5  = _mm_add_sd(t5,sz);
 187
 188     *x1  = _mm_shuffle_pd(t1,t1,_MM_SHUFFLE2(0,0));
 189     *y1  = _mm_shuffle_pd(t1,t1,_MM_SHUFFLE2(1,1));
 190     *z1  = _mm_shuffle_pd(t2,t2,_MM_SHUFFLE2(0,0));
 191     *x2  = _mm_shuffle_pd(t2,t2,_MM_SHUFFLE2(1,1));
 192     *y2  = _mm_shuffle_pd(t3,t3,_MM_SHUFFLE2(0,0));
 193     *z2  = _mm_shuffle_pd(t3,t3,_MM_SHUFFLE2(1,1));
 194     *x3  = _mm_shuffle_pd(t4,t4,_MM_SHUFFLE2(0,0));
 195     *y3  = _mm_shuffle_pd(t4,t4,_MM_SHUFFLE2(1,1));
 196     *z3  = _mm_shuffle_pd(t5,t5,_MM_SHUFFLE2(0,0));
 197 }
 198
 199
 200 static gmx_inline void
 201 gmx_mm_load_shift_and_4rvec_broadcast_pd(const double * gmx_restrict xyz_shift,
 202                                          const double * gmx_restrict xyz,
 203                                          __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 204                                          __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 205                                          __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3,
 206                                          __m128d * gmx_restrict x4, __m128d * gmx_restrict y4, __m128d * gmx_restrict z4)
 207 {
 208     __m128d t1,t2,t3,t4,t5,t6,sxy,sz,szx,syz;
 209
 210     t1  = _mm_loadu_pd(xyz);
 211     t2  = _mm_loadu_pd(xyz+2);
 212     t3  = _mm_loadu_pd(xyz+4);
 213     t4  = _mm_loadu_pd(xyz+6);
 214     t5  = _mm_loadu_pd(xyz+8);
 215     t6  = _mm_loadu_pd(xyz+10);
 216
 217     sxy = _mm_loadu_pd(xyz_shift);
 218     sz  = _mm_load_sd(xyz_shift+2);
 219     szx = _mm_shuffle_pd(sz,sxy,_MM_SHUFFLE2(0,0));
 220     syz = _mm_shuffle_pd(sxy,sz,_MM_SHUFFLE2(0,1));
 221
 222     t1  = _mm_add_pd(t1,sxy);
 223     t2  = _mm_add_pd(t2,szx);
 224     t3  = _mm_add_pd(t3,syz);
 225     t4  = _mm_add_pd(t4,sxy);
 226     t5  = _mm_add_pd(t5,szx);
 227     t6  = _mm_add_pd(t6,syz);
 228
 229     *x1  = _mm_shuffle_pd(t1,t1,_MM_SHUFFLE2(0,0));
 230     *y1  = _mm_shuffle_pd(t1,t1,_MM_SHUFFLE2(1,1));
 231     *z1  = _mm_shuffle_pd(t2,t2,_MM_SHUFFLE2(0,0));
 232     *x2  = _mm_shuffle_pd(t2,t2,_MM_SHUFFLE2(1,1));
 233     *y2  = _mm_shuffle_pd(t3,t3,_MM_SHUFFLE2(0,0));
 234     *z2  = _mm_shuffle_pd(t3,t3,_MM_SHUFFLE2(1,1));
 235     *x3  = _mm_shuffle_pd(t4,t4,_MM_SHUFFLE2(0,0));
 236     *y3  = _mm_shuffle_pd(t4,t4,_MM_SHUFFLE2(1,1));
 237     *z3  = _mm_shuffle_pd(t5,t5,_MM_SHUFFLE2(0,0));
 238     *x4  = _mm_shuffle_pd(t5,t5,_MM_SHUFFLE2(1,1));
 239     *y4  = _mm_shuffle_pd(t6,t6,_MM_SHUFFLE2(0,0));
 240     *z4  = _mm_shuffle_pd(t6,t6,_MM_SHUFFLE2(1,1));
 241 }
 242
 243
 244
 245
 246 static gmx_inline void
 247 gmx_mm_load_1rvec_1ptr_swizzle_pd(const double * gmx_restrict p1,
 248                                   __m128d * gmx_restrict x, __m128d * gmx_restrict y, __m128d * gmx_restrict z)
 249 {
 250          *x            = _mm_load_sd(p1);
 251      *y            = _mm_load_sd(p1+1);
 252      *z            = _mm_load_sd(p1+2);
 253 }
 254
 255 static gmx_inline void
 256 gmx_mm_load_3rvec_1ptr_swizzle_pd(const double * gmx_restrict p1,
 257                                   __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 258                                   __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 259                                   __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3)
 260 {
 261          *x1            = _mm_load_sd(p1);
 262      *y1            = _mm_load_sd(p1+1);
 263      *z1            = _mm_load_sd(p1+2);
 264          *x2            = _mm_load_sd(p1+3);
 265      *y2            = _mm_load_sd(p1+4);
 266      *z2            = _mm_load_sd(p1+5);
 267          *x3            = _mm_load_sd(p1+6);
 268      *y3            = _mm_load_sd(p1+7);
 269      *z3            = _mm_load_sd(p1+8);
 270 }
 271
 272 static gmx_inline void
 273 gmx_mm_load_4rvec_1ptr_swizzle_pd(const double * gmx_restrict p1,
 274                                   __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 275                                   __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 276                                   __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3,
 277                                   __m128d * gmx_restrict x4, __m128d * gmx_restrict y4, __m128d * gmx_restrict z4)
 278 {
 279     *x1            = _mm_load_sd(p1);
 280     *y1            = _mm_load_sd(p1+1);
 281     *z1            = _mm_load_sd(p1+2);
 282     *x2            = _mm_load_sd(p1+3);
 283     *y2            = _mm_load_sd(p1+4);
 284     *z2            = _mm_load_sd(p1+5);
 285     *x3            = _mm_load_sd(p1+6);
 286     *y3            = _mm_load_sd(p1+7);
 287     *z3            = _mm_load_sd(p1+8);
 288     *x4            = _mm_load_sd(p1+9);
 289     *y4            = _mm_load_sd(p1+10);
 290     *z4            = _mm_load_sd(p1+11);
 291 }
 292
 293
 294 static gmx_inline void
 295 gmx_mm_load_1rvec_2ptr_swizzle_pd(const double * gmx_restrict ptrA,
 296                                   const double * gmx_restrict ptrB,
 297                                   __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1)
 298 {
 299     __m128d t1,t2,t3,t4;
 300     t1           = _mm_loadu_pd(ptrA);
 301     t2           = _mm_loadu_pd(ptrB);
 302     t3           = _mm_load_sd(ptrA+2);
 303     t4           = _mm_load_sd(ptrB+2);
 304     GMX_MM_TRANSPOSE2_PD(t1,t2);
 305     *x1          = t1;
 306     *y1          = t2;
 307     *z1          = _mm_unpacklo_pd(t3,t4);
 308 }
 309
 310
 311 static gmx_inline void
 312 gmx_mm_load_3rvec_2ptr_swizzle_pd(const double * gmx_restrict ptrA, const double * gmx_restrict ptrB,
 313                                   __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 314                                   __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 315                                   __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3)
 316 {
 317     __m128d t1,t2,t3,t4,t5,t6,t7,t8,t9,t10;
 318     t1           = _mm_loadu_pd(ptrA);
 319     t2           = _mm_loadu_pd(ptrB);
 320     t3           = _mm_loadu_pd(ptrA+2);
 321     t4           = _mm_loadu_pd(ptrB+2);
 322     t5           = _mm_loadu_pd(ptrA+4);
 323     t6           = _mm_loadu_pd(ptrB+4);
 324     t7           = _mm_loadu_pd(ptrA+6);
 325     t8           = _mm_loadu_pd(ptrB+6);
 326     t9           = _mm_load_sd(ptrA+8);
 327     t10          = _mm_load_sd(ptrB+8);
 328     GMX_MM_TRANSPOSE2_PD(t1,t2);
 329     GMX_MM_TRANSPOSE2_PD(t3,t4);
 330     GMX_MM_TRANSPOSE2_PD(t5,t6);
 331     GMX_MM_TRANSPOSE2_PD(t7,t8);
 332     *x1          = t1;
 333     *y1          = t2;
 334     *z1          = t3;
 335     *x2          = t4;
 336     *y2          = t5;
 337     *z2          = t6;
 338     *x3          = t7;
 339     *y3          = t8;
 340     *z3          = _mm_unpacklo_pd(t9,t10);
 341 }
 342
 343
 344 static gmx_inline void
 345 gmx_mm_load_4rvec_2ptr_swizzle_pd(const double * gmx_restrict ptrA, const double * gmx_restrict ptrB,
 346                                   __m128d * gmx_restrict x1, __m128d * gmx_restrict y1, __m128d * gmx_restrict z1,
 347                                   __m128d * gmx_restrict x2, __m128d * gmx_restrict y2, __m128d * gmx_restrict z2,
 348                                   __m128d * gmx_restrict x3, __m128d * gmx_restrict y3, __m128d * gmx_restrict z3,
 349                                   __m128d * gmx_restrict x4, __m128d * gmx_restrict y4, __m128d * gmx_restrict z4)
 350 {
 351     __m128d t1,t2,t3,t4,t5,t6;
 352     t1           = _mm_loadu_pd(ptrA);
 353     t2           = _mm_loadu_pd(ptrB);
 354     t3           = _mm_loadu_pd(ptrA+2);
 355     t4           = _mm_loadu_pd(ptrB+2);
 356     t5           = _mm_loadu_pd(ptrA+4);
 357     t6           = _mm_loadu_pd(ptrB+4);
 358     GMX_MM_TRANSPOSE2_PD(t1,t2);
 359     GMX_MM_TRANSPOSE2_PD(t3,t4);
 360     GMX_MM_TRANSPOSE2_PD(t5,t6);
 361     *x1          = t1;
 362     *y1          = t2;
 363     *z1          = t3;
 364     *x2          = t4;
 365     *y2          = t5;
 366     *z2          = t6;
 367     t1           = _mm_loadu_pd(ptrA+6);
 368     t2           = _mm_loadu_pd(ptrB+6);
 369     t3           = _mm_loadu_pd(ptrA+8);
 370     t4           = _mm_loadu_pd(ptrB+8);
 371     t5           = _mm_loadu_pd(ptrA+10);
 372     t6           = _mm_loadu_pd(ptrB+10);
 373     GMX_MM_TRANSPOSE2_PD(t1,t2);
 374     GMX_MM_TRANSPOSE2_PD(t3,t4);
 375     GMX_MM_TRANSPOSE2_PD(t5,t6);
 376     *x3          = t1;
 377     *y3          = t2;
 378     *z3          = t3;
 379     *x4          = t4;
 380     *y4          = t5;
 381     *z4          = t6;
 382 }
 383
 384
 385 /* Routines to decrement rvec in memory, typically use for j particle force updates */
 386 static gmx_inline void
 387 gmx_mm_decrement_1rvec_1ptr_noswizzle_pd(double * gmx_restrict ptrA,
 388                                          __m128d xy, __m128d z)
 389 {
 390     __m128d t1,t2;
 391
 392     t1 = _mm_loadu_pd(ptrA);
 393     t2 = _mm_load_sd(ptrA+2);
 394
 395     t1 = _mm_sub_pd(t1,xy);
 396     t2 = _mm_sub_sd(t2,z);
 397
 398     _mm_storeu_pd(ptrA,t1);
 399     _mm_store_sd(ptrA+2,t2);
 400 }
 401
 402 static gmx_inline void
 403 gmx_mm_decrement_3rvec_1ptr_noswizzle_pd(double * gmx_restrict ptrA,
 404                                          __m128d xy1, __m128d z1,
 405                                          __m128d xy2, __m128d z2,
 406                                          __m128d xy3, __m128d z3)
 407 {
 408     __m128d t1,t2;
 409     __m128d tA,tB,tC,tD,tE;
 410
 411     tA   = _mm_loadu_pd(ptrA);
 412     tB   = _mm_loadu_pd(ptrA+2);
 413     tC   = _mm_loadu_pd(ptrA+4);
 414     tD   = _mm_loadu_pd(ptrA+6);
 415     tE   = _mm_load_sd(ptrA+8);
 416
 417     /* xy1: y1 x1 */
 418     t1   = _mm_shuffle_pd(z1,xy2,_MM_SHUFFLE2(0,1)); /* x2 z1 */
 419     t2   = _mm_shuffle_pd(xy2,z2,_MM_SHUFFLE2(0,1)); /* z2 y2 */
 420     /* xy3: y3 x3 */
 421
 422     tA   = _mm_sub_pd(tA,xy1);
 423     tB   = _mm_sub_pd(tB,t1);
 424     tC   = _mm_sub_pd(tC,t2);
 425     tD   = _mm_sub_pd(tD,xy3);
 426     tE   = _mm_sub_sd(tE,z3);
 427
 428     _mm_storeu_pd(ptrA,tA);
 429     _mm_storeu_pd(ptrA+2,tB);
 430     _mm_storeu_pd(ptrA+4,tC);
 431     _mm_storeu_pd(ptrA+6,tD);
 432     _mm_store_sd(ptrA+8,tE);
 433 }
 434
 435 static gmx_inline void
 436 gmx_mm_decrement_4rvec_1ptr_noswizzle_pd(double * gmx_restrict ptrA,
 437                                          __m128d xy1, __m128d z1,
 438                                          __m128d xy2, __m128d z2,
 439                                          __m128d xy3, __m128d z3,
 440                                          __m128d xy4, __m128d z4)
 441 {
 442     __m128d t1,t2,t3,t4;
 443     __m128d tA,tB,tC,tD,tE,tF;
 444
 445     tA   = _mm_loadu_pd(ptrA);
 446     tB   = _mm_loadu_pd(ptrA+2);
 447     tC   = _mm_loadu_pd(ptrA+4);
 448     tD   = _mm_loadu_pd(ptrA+6);
 449     tE   = _mm_loadu_pd(ptrA+8);
 450     tF   = _mm_loadu_pd(ptrA+10);
 451
 452     /* xy1: y1 x1 */
 453     t1   = _mm_shuffle_pd(z1,xy2,_MM_SHUFFLE2(0,0)); /* x2 z1 */
 454     t2   = _mm_shuffle_pd(xy2,z2,_MM_SHUFFLE2(0,1)); /* z2 y2 */
 455     /* xy3: y3 x3 */
 456     t3   = _mm_shuffle_pd(z3,xy4,_MM_SHUFFLE2(0,0)); /* x4 z3 */
 457     t4   = _mm_shuffle_pd(xy4,z4,_MM_SHUFFLE2(0,1)); /* z4 y4 */
 458
 459     tA   = _mm_sub_pd(tA,xy1);
 460     tB   = _mm_sub_pd(tB,t1);
 461     tC   = _mm_sub_pd(tC,t2);
 462     tD   = _mm_sub_pd(tD,xy3);
 463     tE   = _mm_sub_pd(tE,t3);
 464     tF   = _mm_sub_pd(tF,t4);
 465
 466     _mm_storeu_pd(ptrA,tA);
 467     _mm_storeu_pd(ptrA+2,tB);
 468     _mm_storeu_pd(ptrA+4,tC);
 469     _mm_storeu_pd(ptrA+6,tD);
 470     _mm_storeu_pd(ptrA+8,tE);
 471     _mm_storeu_pd(ptrA+10,tF);
 472 }
 473
 474 static gmx_inline void
 475 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(double * gmx_restrict ptrA,
 476                                        __m128d x1, __m128d y1, __m128d z1)
 477 {
 478     __m128d t1,t2,t3;
 479
 480     t1           = _mm_load_sd(ptrA);
 481     t2           = _mm_load_sd(ptrA+1);
 482     t3           = _mm_load_sd(ptrA+2);
 483
 484     t1           = _mm_sub_sd(t1,x1);
 485     t2           = _mm_sub_sd(t2,y1);
 486     t3           = _mm_sub_sd(t3,z1);
 487     _mm_store_sd(ptrA,t1);
 488     _mm_store_sd(ptrA+1,t2);
 489     _mm_store_sd(ptrA+2,t3);
 490 }
 491
 492
 493 static gmx_inline void
 494 gmx_mm_decrement_3rvec_1ptr_swizzle_pd(double * gmx_restrict ptrA,
 495                                        __m128d x1, __m128d y1, __m128d z1,
 496                                        __m128d x2, __m128d y2, __m128d z2,
 497                                        __m128d x3, __m128d y3, __m128d z3)
 498 {
 499     __m128d t1,t2,t3,t4,t5;
 500
 501     t1          = _mm_loadu_pd(ptrA);
 502     t2          = _mm_loadu_pd(ptrA+2);
 503     t3          = _mm_loadu_pd(ptrA+4);
 504     t4          = _mm_loadu_pd(ptrA+6);
 505     t5          = _mm_load_sd(ptrA+8);
 506
 507     x1          = _mm_unpacklo_pd(x1,y1);
 508     z1          = _mm_unpacklo_pd(z1,x2);
 509     y2          = _mm_unpacklo_pd(y2,z2);
 510     x3          = _mm_unpacklo_pd(x3,y3);
 511     /* nothing to be done for z3 */
 512
 513     t1          = _mm_sub_pd(t1,x1);
 514     t2          = _mm_sub_pd(t2,z1);
 515     t3          = _mm_sub_pd(t3,y2);
 516     t4          = _mm_sub_pd(t4,x3);
 517     t5          = _mm_sub_sd(t5,z3);
 518     _mm_storeu_pd(ptrA,t1);
 519     _mm_storeu_pd(ptrA+2,t2);
 520     _mm_storeu_pd(ptrA+4,t3);
 521     _mm_storeu_pd(ptrA+6,t4);
 522     _mm_store_sd(ptrA+8,t5);
 523 }
 524
 525
 526 static gmx_inline void
 527 gmx_mm_decrement_4rvec_1ptr_swizzle_pd(double * gmx_restrict ptrA,
 528                                        __m128d x1, __m128d y1, __m128d z1,
 529                                        __m128d x2, __m128d y2, __m128d z2,
 530                                        __m128d x3, __m128d y3, __m128d z3,
 531                                        __m128d x4, __m128d y4, __m128d z4)
 532 {
 533     __m128d t1,t2,t3,t4,t5,t6;
 534
 535     t1          = _mm_loadu_pd(ptrA);
 536     t2          = _mm_loadu_pd(ptrA+2);
 537     t3          = _mm_loadu_pd(ptrA+4);
 538     t4          = _mm_loadu_pd(ptrA+6);
 539     t5          = _mm_loadu_pd(ptrA+8);
 540     t6          = _mm_loadu_pd(ptrA+10);
 541
 542     x1          = _mm_unpacklo_pd(x1,y1);
 543     z1          = _mm_unpacklo_pd(z1,x2);
 544     y2          = _mm_unpacklo_pd(y2,z2);
 545     x3          = _mm_unpacklo_pd(x3,y3);
 546     z3          = _mm_unpacklo_pd(z3,x4);
 547     y4          = _mm_unpacklo_pd(y4,z4);
 548
 549     _mm_storeu_pd(ptrA,    _mm_sub_pd( t1,x1 ));
 550     _mm_storeu_pd(ptrA+2,  _mm_sub_pd( t2,z1 ));
 551     _mm_storeu_pd(ptrA+4,  _mm_sub_pd( t3,y2 ));
 552     _mm_storeu_pd(ptrA+6,  _mm_sub_pd( t4,x3 ));
 553     _mm_storeu_pd(ptrA+8,  _mm_sub_pd( t5,z3 ));
 554     _mm_storeu_pd(ptrA+10, _mm_sub_pd( t6,y4 ));
 555 }
 556
 557 static gmx_inline void
 558 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(double * gmx_restrict ptrA, double * gmx_restrict ptrB,
 559                                        __m128d x1, __m128d y1, __m128d z1)
 560 {
 561     __m128d t1,t2,t3,t4,t5,t6,t7;
 562
 563     t1          = _mm_loadu_pd(ptrA);
 564     t2          = _mm_load_sd(ptrA+2);
 565     t3          = _mm_loadu_pd(ptrB);
 566     t4          = _mm_load_sd(ptrB+2);
 567
 568     t5          = _mm_unpacklo_pd(x1,y1);
 569     t6          = _mm_unpackhi_pd(x1,y1);
 570     t7          = _mm_unpackhi_pd(z1,z1);
 571
 572     t1          = _mm_sub_pd(t1,t5);
 573     t2          = _mm_sub_sd(t2,z1);
 574
 575     t3          = _mm_sub_pd(t3,t6);
 576     t4          = _mm_sub_sd(t4,t7);
 577
 578     _mm_storeu_pd(ptrA,t1);
 579     _mm_store_sd(ptrA+2,t2);
 580     _mm_storeu_pd(ptrB,t3);
 581     _mm_store_sd(ptrB+2,t4);
 582 }
 583
 584 static gmx_inline void
 585 gmx_mm_decrement_3rvec_2ptr_swizzle_pd(double * gmx_restrict ptrA, double * gmx_restrict ptrB,
 586                                        __m128d x1, __m128d y1, __m128d z1,
 587                                        __m128d x2, __m128d y2, __m128d z2,
 588                                        __m128d x3, __m128d y3, __m128d z3)
 589 {
 590     __m128d t1,t2,t3,t4,t5,t6,t7,t8,t9,t10;
 591     __m128d tA,tB,tC,tD,tE,tF,tG,tH,tI;
 592
 593     t1          = _mm_loadu_pd(ptrA);
 594     t2          = _mm_loadu_pd(ptrA+2);
 595     t3          = _mm_loadu_pd(ptrA+4);
 596     t4          = _mm_loadu_pd(ptrA+6);
 597     t5          = _mm_load_sd(ptrA+8);
 598     t6          = _mm_loadu_pd(ptrB);
 599     t7          = _mm_loadu_pd(ptrB+2);
 600     t8          = _mm_loadu_pd(ptrB+4);
 601     t9          = _mm_loadu_pd(ptrB+6);
 602     t10         = _mm_load_sd(ptrB+8);
 603
 604     tA          = _mm_unpacklo_pd(x1,y1);
 605     tB          = _mm_unpackhi_pd(x1,y1);
 606     tC          = _mm_unpacklo_pd(z1,x2);
 607     tD          = _mm_unpackhi_pd(z1,x2);
 608     tE          = _mm_unpacklo_pd(y2,z2);
 609     tF          = _mm_unpackhi_pd(y2,z2);
 610     tG          = _mm_unpacklo_pd(x3,y3);
 611     tH          = _mm_unpackhi_pd(x3,y3);
 612     tI          = _mm_unpackhi_pd(z3,z3);
 613
 614     t1          = _mm_sub_pd(t1,tA);
 615     t2          = _mm_sub_pd(t2,tC);
 616     t3          = _mm_sub_pd(t3,tE);
 617     t4          = _mm_sub_pd(t4,tG);
 618     t5          = _mm_sub_sd(t5,z3);
 619
 620     t6          = _mm_sub_pd(t6,tB);
 621     t7          = _mm_sub_pd(t7,tD);
 622     t8          = _mm_sub_pd(t8,tF);
 623     t9          = _mm_sub_pd(t9,tH);
 624     t10         = _mm_sub_sd(t10,tI);
 625
 626     _mm_storeu_pd(ptrA,t1);
 627     _mm_storeu_pd(ptrA+2,t2);
 628     _mm_storeu_pd(ptrA+4,t3);
 629     _mm_storeu_pd(ptrA+6,t4);
 630     _mm_store_sd(ptrA+8,t5);
 631     _mm_storeu_pd(ptrB,t6);
 632     _mm_storeu_pd(ptrB+2,t7);
 633     _mm_storeu_pd(ptrB+4,t8);
 634     _mm_storeu_pd(ptrB+6,t9);
 635     _mm_store_sd(ptrB+8,t10);
 636 }
 637
 638
 639 static gmx_inline void
 640 gmx_mm_decrement_4rvec_2ptr_swizzle_pd(double * gmx_restrict ptrA, double * gmx_restrict ptrB,
 641                                        __m128d x1, __m128d y1, __m128d z1,
 642                                        __m128d x2, __m128d y2, __m128d z2,
 643                                        __m128d x3, __m128d y3, __m128d z3,
 644                                        __m128d x4, __m128d y4, __m128d z4)
 645 {
 646     __m128d t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12;
 647     __m128d tA,tB,tC,tD,tE,tF,tG,tH,tI,tJ,tK,tL;
 648
 649     t1          = _mm_loadu_pd(ptrA);
 650     t2          = _mm_loadu_pd(ptrA+2);
 651     t3          = _mm_loadu_pd(ptrA+4);
 652     t4          = _mm_loadu_pd(ptrA+6);
 653     t5          = _mm_loadu_pd(ptrA+8);
 654     t6          = _mm_loadu_pd(ptrA+10);
 655     t7          = _mm_loadu_pd(ptrB);
 656     t8          = _mm_loadu_pd(ptrB+2);
 657     t9          = _mm_loadu_pd(ptrB+4);
 658     t10         = _mm_loadu_pd(ptrB+6);
 659     t11         = _mm_loadu_pd(ptrB+8);
 660     t12         = _mm_loadu_pd(ptrB+10);
 661
 662     tA          = _mm_unpacklo_pd(x1,y1);
 663     tB          = _mm_unpackhi_pd(x1,y1);
 664     tC          = _mm_unpacklo_pd(z1,x2);
 665     tD          = _mm_unpackhi_pd(z1,x2);
 666     tE          = _mm_unpacklo_pd(y2,z2);
 667     tF          = _mm_unpackhi_pd(y2,z2);
 668     tG          = _mm_unpacklo_pd(x3,y3);
 669     tH          = _mm_unpackhi_pd(x3,y3);
 670     tI          = _mm_unpacklo_pd(z3,x4);
 671     tJ          = _mm_unpackhi_pd(z3,x4);
 672     tK          = _mm_unpacklo_pd(y4,z4);
 673     tL          = _mm_unpackhi_pd(y4,z4);
 674
 675     t1          = _mm_sub_pd(t1,tA);
 676     t2          = _mm_sub_pd(t2,tC);
 677     t3          = _mm_sub_pd(t3,tE);
 678     t4          = _mm_sub_pd(t4,tG);
 679     t5          = _mm_sub_pd(t5,tI);
 680     t6          = _mm_sub_pd(t6,tK);
 681
 682     t7          = _mm_sub_pd(t7,tB);
 683     t8          = _mm_sub_pd(t8,tD);
 684     t9          = _mm_sub_pd(t9,tF);
 685     t10         = _mm_sub_pd(t10,tH);
 686     t11         = _mm_sub_pd(t11,tJ);
 687     t12         = _mm_sub_pd(t12,tL);
 688
 689     _mm_storeu_pd(ptrA,  t1);
 690     _mm_storeu_pd(ptrA+2,t2);
 691     _mm_storeu_pd(ptrA+4,t3);
 692     _mm_storeu_pd(ptrA+6,t4);
 693     _mm_storeu_pd(ptrA+8,t5);
 694     _mm_storeu_pd(ptrA+10,t6);
 695     _mm_storeu_pd(ptrB,  t7);
 696     _mm_storeu_pd(ptrB+2,t8);
 697     _mm_storeu_pd(ptrB+4,t9);
 698     _mm_storeu_pd(ptrB+6,t10);
 699     _mm_storeu_pd(ptrB+8,t11);
 700     _mm_storeu_pd(ptrB+10,t12);
 701 }
 702
 703
 704
 705
 706
 707 static gmx_inline void
 708 gmx_mm_update_iforce_1atom_swizzle_pd(__m128d fix1, __m128d fiy1, __m128d fiz1,
 709                                       double * gmx_restrict fptr,
 710                                       double * gmx_restrict fshiftptr)
 711 {
 712     __m128d t1,t2,t3;
 713
 714     /* transpose data */
 715     t1 = fix1;
 716     fix1 = _mm_unpacklo_pd(fix1,fiy1); /* y0 x0 */
 717     fiy1 = _mm_unpackhi_pd(t1,fiy1);   /* y1 x1 */
 718
 719     fix1 = _mm_add_pd(fix1,fiy1);
 720     fiz1 = _mm_add_sd( fiz1, _mm_unpackhi_pd(fiz1,fiz1 ));
 721
 722     _mm_storeu_pd( fptr, _mm_add_pd( _mm_loadu_pd(fptr), fix1 ));
 723     _mm_store_sd( fptr+2, _mm_add_sd( _mm_load_sd(fptr+2), fiz1 ));
 724
 725     _mm_storeu_pd( fshiftptr, _mm_add_pd( _mm_loadu_pd(fshiftptr), fix1 ));
 726     _mm_store_sd( fshiftptr+2, _mm_add_sd( _mm_load_sd(fshiftptr+2), fiz1 ));
 727 }
 728
 729 static gmx_inline void
 730 gmx_mm_update_iforce_3atom_swizzle_pd(__m128d fix1, __m128d fiy1, __m128d fiz1,
 731                                       __m128d fix2, __m128d fiy2, __m128d fiz2,
 732                                       __m128d fix3, __m128d fiy3, __m128d fiz3,
 733                                       double * gmx_restrict fptr,
 734                                       double * gmx_restrict fshiftptr)
 735 {
 736     __m128d t1,t2;
 737
 738     /* transpose data */
 739     GMX_MM_TRANSPOSE2_PD(fix1,fiy1);
 740     GMX_MM_TRANSPOSE2_PD(fiz1,fix2);
 741     GMX_MM_TRANSPOSE2_PD(fiy2,fiz2);
 742     t1 = fix3;
 743     fix3 = _mm_unpacklo_pd(fix3,fiy3); /* y0 x0 */
 744     fiy3 = _mm_unpackhi_pd(t1,fiy3);   /* y1 x1 */
 745
 746     fix1 = _mm_add_pd(fix1,fiy1);
 747     fiz1 = _mm_add_pd(fiz1,fix2);
 748     fiy2 = _mm_add_pd(fiy2,fiz2);
 749
 750     fix3 = _mm_add_pd(fix3,fiy3);
 751     fiz3 = _mm_add_sd( fiz3, _mm_unpackhi_pd(fiz3,fiz3));
 752
 753     _mm_storeu_pd( fptr, _mm_add_pd( _mm_loadu_pd(fptr), fix1 ));
 754     _mm_storeu_pd( fptr+2, _mm_add_pd( _mm_loadu_pd(fptr+2), fiz1 ));
 755     _mm_storeu_pd( fptr+4, _mm_add_pd( _mm_loadu_pd(fptr+4), fiy2 ));
 756     _mm_storeu_pd( fptr+6, _mm_add_pd( _mm_loadu_pd(fptr+6), fix3 ));
 757     _mm_store_sd( fptr+8, _mm_add_sd( _mm_load_sd(fptr+8), fiz3 ));
 758
 759     fix1 = _mm_add_pd(fix1,fix3);
 760     t1   = _mm_shuffle_pd(fiz1,fiy2,_MM_SHUFFLE2(0,1));
 761     fix1 = _mm_add_pd(fix1,t1); /* x and y sums */
 762
 763     t2   = _mm_shuffle_pd(fiy2,fiy2,_MM_SHUFFLE2(1,1));
 764     fiz1 = _mm_add_sd(fiz1,fiz3);
 765     fiz1 = _mm_add_sd(fiz1,t2); /* z sum */
 766
 767     _mm_storeu_pd( fshiftptr, _mm_add_pd( _mm_loadu_pd(fshiftptr), fix1 ));
 768     _mm_store_sd( fshiftptr+2, _mm_add_sd( _mm_load_sd(fshiftptr+2), fiz1 ));
 769 }
 770
 771
 772 static gmx_inline void
 773 gmx_mm_update_iforce_4atom_swizzle_pd(__m128d fix1, __m128d fiy1, __m128d fiz1,
 774                                       __m128d fix2, __m128d fiy2, __m128d fiz2,
 775                                       __m128d fix3, __m128d fiy3, __m128d fiz3,
 776                                       __m128d fix4, __m128d fiy4, __m128d fiz4,
 777                                       double * gmx_restrict fptr,
 778                                       double * gmx_restrict fshiftptr)
 779 {
 780     __m128d t1,t2;
 781
 782     /* transpose data */
 783     GMX_MM_TRANSPOSE2_PD(fix1,fiy1);
 784     GMX_MM_TRANSPOSE2_PD(fiz1,fix2);
 785     GMX_MM_TRANSPOSE2_PD(fiy2,fiz2);
 786     GMX_MM_TRANSPOSE2_PD(fix3,fiy3);
 787     GMX_MM_TRANSPOSE2_PD(fiz3,fix4);
 788     GMX_MM_TRANSPOSE2_PD(fiy4,fiz4);
 789
 790     fix1 = _mm_add_pd(fix1,fiy1);
 791     fiz1 = _mm_add_pd(fiz1,fix2);
 792     fiy2 = _mm_add_pd(fiy2,fiz2);
 793     fix3 = _mm_add_pd(fix3,fiy3);
 794     fiz3 = _mm_add_pd(fiz3,fix4);
 795     fiy4 = _mm_add_pd(fiy4,fiz4);
 796
 797     _mm_storeu_pd( fptr, _mm_add_pd( _mm_loadu_pd(fptr),       fix1 ));
 798     _mm_storeu_pd( fptr+2, _mm_add_pd( _mm_loadu_pd(fptr+2),   fiz1 ));
 799     _mm_storeu_pd( fptr+4, _mm_add_pd( _mm_loadu_pd(fptr+4),   fiy2 ));
 800     _mm_storeu_pd( fptr+6, _mm_add_pd( _mm_loadu_pd(fptr+6),   fix3 ));
 801     _mm_storeu_pd( fptr+8, _mm_add_pd( _mm_loadu_pd(fptr+8),   fiz3 ));
 802     _mm_storeu_pd( fptr+10, _mm_add_pd( _mm_loadu_pd(fptr+10), fiy4 ));
 803
 804     t1 = _mm_shuffle_pd(fiz1,fiy2,_MM_SHUFFLE2(0,1));
 805     fix1 = _mm_add_pd(fix1,t1);
 806     t2 = _mm_shuffle_pd(fiz3,fiy4,_MM_SHUFFLE2(0,1));
 807     fix3 = _mm_add_pd(fix3,t2);
 808     fix1 = _mm_add_pd(fix1,fix3); /* x and y sums */
 809
 810     fiz1 = _mm_add_sd(fiz1, _mm_unpackhi_pd(fiy2,fiy2));
 811     fiz3 = _mm_add_sd(fiz3, _mm_unpackhi_pd(fiy4,fiy4));
 812     fiz1 = _mm_add_sd(fiz1,fiz3); /* z sum */
 813
 814     _mm_storeu_pd( fshiftptr, _mm_add_pd( _mm_loadu_pd(fshiftptr), fix1 ));
 815     _mm_store_sd( fshiftptr+2, _mm_add_sd( _mm_load_sd(fshiftptr+2), fiz1 ));
 816 }
 817
 818
 819
 820 static gmx_inline void
 821 gmx_mm_update_1pot_pd(__m128d pot1, double * gmx_restrict ptrA)
 822 {
 823     pot1 = _mm_add_pd(pot1, _mm_unpackhi_pd(pot1,pot1));
 824     _mm_store_sd(ptrA,_mm_add_sd(pot1,_mm_load_sd(ptrA)));
 825 }
 826
 827 static gmx_inline void
 828 gmx_mm_update_2pot_pd(__m128d pot1, double * gmx_restrict ptrA,
 829                       __m128d pot2, double * gmx_restrict ptrB)
 830 {
 831     GMX_MM_TRANSPOSE2_PD(pot1,pot2);
 832     pot1 = _mm_add_pd(pot1,pot2);
 833     pot2 = _mm_unpackhi_pd(pot1,pot1);
 834
 835     _mm_store_sd(ptrA,_mm_add_sd(pot1,_mm_load_sd(ptrA)));
 836     _mm_store_sd(ptrB,_mm_add_sd(pot2,_mm_load_sd(ptrB)));
 837 }
 838
 839
 840 #endif /* _kernelutil_x86_sse2_double_h_ */