2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
5 * Copyright (c) 2001-2004, The GROMACS development team,
6 * check out http://www.gromacs.org for more information.
7 * Copyright (c) 2012, by the GROMACS development team, led by
8 * David van der Spoel, Berk Hess, Erik Lindahl, and including many
9 * others, as listed in the AUTHORS file in the top-level source
10 * directory and at http://www.gromacs.org.
12 * GROMACS is free software; you can redistribute it and/or
13 * modify it under the terms of the GNU Lesser General Public License
14 * as published by the Free Software Foundation; either version 2.1
15 * of the License, or (at your option) any later version.
17 * GROMACS is distributed in the hope that it will be useful,
18 * but WITHOUT ANY WARRANTY; without even the implied warranty of
19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
20 * Lesser General Public License for more details.
22 * You should have received a copy of the GNU Lesser General Public
23 * License along with GROMACS; if not, see
24 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
25 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
27 * If you want to redistribute modifications to GROMACS, please
28 * consider that scientific software is very special. Version
29 * control is crucial - bugs must be traceable. We will be happy to
30 * consider code for inclusion in the official distribution, but
31 * derived work must not be called official GROMACS. Details are found
32 * in the README & COPYING files - if they are missing, get the
33 * official version at http://www.gromacs.org.
35 * To help us fund GROMACS development, we humbly ask that you cite
36 * the research papers on the package. Check out http://www.gromacs.org.
60 #include "eigensolver.h"
63 #include "mtop_util.h"
64 #include "sparsematrix.h"
69 static double cv_corr(double nu,double T)
71 double x = PLANCK*nu/(BOLTZ*T);
77 return BOLTZ*KILO*(ex*sqr(x)/sqr(ex-1) - 1);
80 static double u_corr(double nu,double T)
82 double x = PLANCK*nu/(BOLTZ*T);
88 return BOLTZ*T*(0.5*x - 1 + x/(ex-1));
91 static int get_nharm_mt(gmx_moltype_t *mt)
93 static int harm_func[] = { F_BONDS };
97 for(i=0; (i<asize(harm_func)); i++)
100 nh += mt->ilist[ft].nr/(interaction_function[ft].nratoms+1);
105 static int get_nvsite_mt(gmx_moltype_t *mt)
107 static int vs_func[] = { F_VSITE2, F_VSITE3, F_VSITE3FD, F_VSITE3FAD,
108 F_VSITE3OUT, F_VSITE4FD, F_VSITE4FDN, F_VSITEN };
112 for(i=0; (i<asize(vs_func)); i++)
115 nh += mt->ilist[ft].nr/(interaction_function[ft].nratoms+1);
120 static int get_nharm(gmx_mtop_t *mtop,int *nvsites)
126 for(j=0; (j<mtop->nmolblock); j++)
128 mt = mtop->molblock[j].type;
129 nh += mtop->molblock[j].nmol * get_nharm_mt(&(mtop->moltype[mt]));
130 nv += mtop->molblock[j].nmol * get_nvsite_mt(&(mtop->moltype[mt]));
137 nma_full_hessian(real * hess,
150 natoms = top->atoms.nr;
152 /* divide elements hess[i][j] by sqrt(mas[i])*sqrt(mas[j]) when required */
156 for (i=0; (i<natoms); i++)
158 for (j=0; (j<DIM); j++)
160 for (k=0; (k<natoms); k++)
162 mass_fac=gmx_invsqrt(top->atoms.atom[i].m*top->atoms.atom[k].m);
163 for (l=0; (l<DIM); l++)
164 hess[(i*DIM+j)*ndim+k*DIM+l]*=mass_fac;
170 /* call diagonalization routine. */
172 fprintf(stderr,"\nDiagonalizing to find vectors %d through %d...\n",begin,end);
175 eigensolver(hess,ndim,begin-1,end-1,eigenvalues,eigenvectors);
177 /* And scale the output eigenvectors */
178 if (bM && eigenvectors!=NULL)
180 for(i=0;i<(end-begin+1);i++)
182 for(j=0;j<natoms;j++)
184 mass_fac = gmx_invsqrt(top->atoms.atom[j].m);
185 for (k=0; (k<DIM); k++)
187 eigenvectors[i*ndim+j*DIM+k] *= mass_fac;
197 nma_sparse_hessian(gmx_sparsematrix_t * sparse_hessian,
211 natoms = top->atoms.nr;
214 /* Cannot check symmetry since we only store half matrix */
215 /* divide elements hess[i][j] by sqrt(mas[i])*sqrt(mas[j]) when required */
219 for (iatom=0; (iatom<natoms); iatom++)
221 for (j=0; (j<DIM); j++)
224 for(k=0;k<sparse_hessian->ndata[row];k++)
226 col = sparse_hessian->data[row][k].col;
228 mass_fac=gmx_invsqrt(top->atoms.atom[iatom].m*top->atoms.atom[katom].m);
229 sparse_hessian->data[row][k].value *=mass_fac;
234 fprintf(stderr,"\nDiagonalizing to find eigenvectors 1 through %d...\n",neig);
237 sparse_eigensolver(sparse_hessian,neig,eigenvalues,eigenvectors,10000000);
239 /* Scale output eigenvectors */
240 if (bM && eigenvectors!=NULL)
244 for(j=0;j<natoms;j++)
246 mass_fac = gmx_invsqrt(top->atoms.atom[j].m);
247 for (k=0; (k<DIM); k++)
249 eigenvectors[i*ndim+j*DIM+k] *= mass_fac;
258 int gmx_nmeig(int argc,char *argv[])
260 const char *desc[] = {
261 "[TT]g_nmeig[tt] calculates the eigenvectors/values of a (Hessian) matrix,",
262 "which can be calculated with [TT]mdrun[tt].",
263 "The eigenvectors are written to a trajectory file ([TT]-v[tt]).",
264 "The structure is written first with t=0. The eigenvectors",
265 "are written as frames with the eigenvector number as timestamp.",
266 "The eigenvectors can be analyzed with [TT]g_anaeig[tt].",
267 "An ensemble of structures can be generated from the eigenvectors with",
268 "[TT]g_nmens[tt]. When mass weighting is used, the generated eigenvectors",
269 "will be scaled back to plain Cartesian coordinates before generating the",
270 "output. In this case, they will no longer be exactly orthogonal in the",
271 "standard Cartesian norm, but in the mass-weighted norm they would be.[PAR]",
272 "This program can be optionally used to compute quantum corrections to heat capacity",
273 "and enthalpy by providing an extra file argument [TT]-qcorr[tt]. See the GROMACS",
274 "manual, Chapter 1, for details. The result includes subtracting a harmonic",
275 "degree of freedom at the given temperature.",
276 "The total correction is printed on the terminal screen.",
277 "The recommended way of getting the corrections out is:[PAR]",
278 "[TT]g_nmeig -s topol.tpr -f nm.mtx -first 7 -last 10000 -T 300 -qc [-constr][tt][PAR]",
279 "The [TT]-constr[tt] option should be used when bond constraints were used during the",
280 "simulation [BB]for all the covalent bonds[bb]. If this is not the case, ",
281 "you need to analyze the [TT]quant_corr.xvg[tt] file yourself.[PAR]",
282 "To make things more flexible, the program can also take virtual sites into account",
283 "when computing quantum corrections. When selecting [TT]-constr[tt] and",
284 "[TT]-qc[tt], the [TT]-begin[tt] and [TT]-end[tt] options will be set automatically as well.",
285 "Again, if you think you know it better, please check the [TT]eigenfreq.xvg[tt]",
289 static gmx_bool bM=TRUE,bCons=FALSE;
290 static int begin=1,end=50,maxspec=4000;
291 static real T=298.15,width=1;
294 { "-m", FALSE, etBOOL, {&bM},
295 "Divide elements of Hessian by product of sqrt(mass) of involved "
296 "atoms prior to diagonalization. This should be used for 'Normal Modes' "
298 { "-first", FALSE, etINT, {&begin},
299 "First eigenvector to write away" },
300 { "-last", FALSE, etINT, {&end},
301 "Last eigenvector to write away" },
302 { "-maxspec", FALSE, etINT, {&maxspec},
303 "Highest frequency (1/cm) to consider in the spectrum" },
304 { "-T", FALSE, etREAL, {&T},
305 "Temperature for computing quantum heat capacity and enthalpy when using normal mode calculations to correct classical simulations" },
306 { "-constr", FALSE, etBOOL, {&bCons},
307 "If constraints were used in the simulation but not in the normal mode analysis (this is the recommended way of doing it) you will need to set this for computing the quantum corrections." },
308 { "-width", FALSE, etREAL, {&width},
309 "Width (sigma) of the gaussian peaks (1/cm) when generating a spectrum" }
320 real rdum,mass_fac,qcvtot,qutot,qcv,qu;
321 int natoms,ndim,nrow,ncol,count,nharm,nvsite;
327 int version,generation;
329 real factor_gmx_to_omega2;
330 real factor_omega_to_wavenumber;
335 const char *qcleg[] = { "Heat Capacity cV (J/mol K)",
336 "Enthalpy H (kJ/mol)" };
337 real * full_hessian = NULL;
338 gmx_sparsematrix_t * sparse_hessian = NULL;
341 { efMTX, "-f", "hessian", ffREAD },
342 { efTPX, NULL, NULL, ffREAD },
343 { efXVG, "-of", "eigenfreq", ffWRITE },
344 { efXVG, "-ol", "eigenval", ffWRITE },
345 { efXVG, "-os", "spectrum", ffOPTWR },
346 { efXVG, "-qc", "quant_corr", ffOPTWR },
347 { efTRN, "-v", "eigenvec", ffWRITE }
349 #define NFILE asize(fnm)
351 cr = init_par(&argc,&argv);
354 CopyRight(stderr,argv[0]);
356 parse_common_args(&argc,argv,PCA_BE_NICE | (MASTER(cr) ? 0 : PCA_QUIET),
357 NFILE,fnm,asize(pa),pa,asize(desc),desc,0,NULL,&oenv);
359 /* Read tpr file for volume and number of harmonic terms */
360 read_tpxheader(ftp2fn(efTPX,NFILE,fnm),&tpx,TRUE,&version,&generation);
361 snew(top_x,tpx.natoms);
363 read_tpx(ftp2fn(efTPX,NFILE,fnm),NULL,box,&natoms,
364 top_x,NULL,NULL,&mtop);
367 nharm = get_nharm(&mtop,&nvsite);
374 top = gmx_mtop_t_to_t_topology(&mtop);
379 if (opt2bSet("-qc",NFILE,fnm))
381 begin = 7+DIM*nvsite;
388 printf("Using begin = %d and end = %d\n",begin,end);
390 /*open Hessian matrix */
391 gmx_mtxio_read(ftp2fn(efMTX,NFILE,fnm),&nrow,&ncol,&full_hessian,&sparse_hessian);
393 /* Memory for eigenvalues and eigenvectors (begin..end) */
394 snew(eigenvalues,nrow);
395 snew(eigenvectors,nrow*(end-begin+1));
397 /* If the Hessian is in sparse format we can calculate max (ndim-1) eigenvectors,
398 * and they must start at the first one. If this is not valid we convert to full matrix
399 * storage, but warn the user that we might run out of memory...
401 if((sparse_hessian != NULL) && (begin!=1 || end==ndim))
405 fprintf(stderr,"Cannot use sparse Hessian with first eigenvector != 1.\n");
409 fprintf(stderr,"Cannot use sparse Hessian to calculate all eigenvectors.\n");
412 fprintf(stderr,"Will try to allocate memory and convert to full matrix representation...\n");
414 snew(full_hessian,nrow*ncol);
415 for(i=0;i<nrow*ncol;i++)
418 for(i=0;i<sparse_hessian->nrow;i++)
420 for(j=0;j<sparse_hessian->ndata[i];j++)
422 k = sparse_hessian->data[i][j].col;
423 value = sparse_hessian->data[i][j].value;
424 full_hessian[i*ndim+k] = value;
425 full_hessian[k*ndim+i] = value;
428 gmx_sparsematrix_destroy(sparse_hessian);
429 sparse_hessian = NULL;
430 fprintf(stderr,"Converted sparse to full matrix storage.\n");
433 if (full_hessian != NULL)
435 /* Using full matrix storage */
436 nma_full_hessian(full_hessian,nrow,bM,&top,begin,end,
437 eigenvalues,eigenvectors);
441 /* Sparse memory storage, allocate memory for eigenvectors */
442 snew(eigenvectors,ncol*end);
443 nma_sparse_hessian(sparse_hessian,bM,&top,end,eigenvalues,eigenvectors);
446 /* check the output, first 6 eigenvalues should be reasonably small */
448 for (i=begin-1; (i<6); i++)
450 if (fabs(eigenvalues[i]) > 1.0e-3)
455 fprintf(stderr,"\nOne of the lowest 6 eigenvalues has a non-zero value.\n");
456 fprintf(stderr,"This could mean that the reference structure was not\n");
457 fprintf(stderr,"properly energy minimized.\n");
460 /* now write the output */
461 fprintf (stderr,"Writing eigenvalues...\n");
462 out=xvgropen(opt2fn("-ol",NFILE,fnm),
463 "Eigenvalues","Eigenvalue index","Eigenvalue [Gromacs units]",
465 if (output_env_get_print_xvgr_codes(oenv)) {
467 fprintf(out,"@ subtitle \"mass weighted\"\n");
469 fprintf(out,"@ subtitle \"not mass weighted\"\n");
472 for (i=0; i<=(end-begin); i++)
473 fprintf (out,"%6d %15g\n",begin+i,eigenvalues[i]);
477 if (opt2bSet("-qc",NFILE,fnm)) {
478 qc = xvgropen(opt2fn("-qc",NFILE,fnm),"Quantum Corrections","Eigenvector index","",oenv);
479 xvgr_legend(qc,asize(qcleg),qcleg,oenv);
484 printf("Writing eigenfrequencies - negative eigenvalues will be set to zero.\n");
486 out=xvgropen(opt2fn("-of",NFILE,fnm),
487 "Eigenfrequencies","Eigenvector index","Wavenumber [cm\\S-1\\N]",
489 if (output_env_get_print_xvgr_codes(oenv)) {
491 fprintf(out,"@ subtitle \"mass weighted\"\n");
493 fprintf(out,"@ subtitle \"not mass weighted\"\n");
497 if (opt2bSet("-os",NFILE,fnm) && (maxspec > 0))
499 snew(spectrum,maxspec);
500 spec=xvgropen(opt2fn("-os",NFILE,fnm),
501 "Vibrational spectrum based on harmonic approximation",
502 "\\f{12}w\\f{4} (cm\\S-1\\N)",
503 "Intensity [Gromacs units]",
505 for(i=0; (i<maxspec); i++)
511 /* Gromacs units are kJ/(mol*nm*nm*amu),
512 * where amu is the atomic mass unit.
514 * For the eigenfrequencies we want to convert this to spectroscopic absorption
515 * wavenumbers given in cm^(-1), which is the frequency divided by the speed of
516 * light. Do this by first converting to omega^2 (units 1/s), take the square
517 * root, and finally divide by the speed of light (nm/ps in gromacs).
519 factor_gmx_to_omega2 = 1.0E21/(AVOGADRO*AMU);
520 factor_omega_to_wavenumber = 1.0E-5/(2.0*M_PI*SPEED_OF_LIGHT);
522 for (i=begin; (i<=end); i++)
524 value = eigenvalues[i-begin];
527 omega = sqrt(value*factor_gmx_to_omega2);
528 nu = 1e-12*omega/(2*M_PI);
529 value = omega*factor_omega_to_wavenumber;
530 fprintf (out,"%6d %15g\n",i,value);
533 wfac = eigenvalues[i-begin]/(width*sqrt(2*M_PI));
534 for(j=0; (j<maxspec); j++)
536 spectrum[j] += wfac*exp(-sqr(j-value)/(2*sqr(width)));
547 fprintf (qc,"%6d %15g %15g\n",i,qcv,qu);
555 for(j=0; (j<maxspec); j++)
557 fprintf(spec,"%10g %10g\n",1.0*j,spectrum[j]);
562 printf("Quantum corrections for harmonic degrees of freedom\n");
563 printf("Use appropriate -first and -last options to get reliable results.\n");
564 printf("There were %d constraints and %d vsites in the simulation\n",
566 printf("Total correction to cV = %g J/mol K\n",qcvtot);
567 printf("Total correction to H = %g kJ/mol\n",qutot);
569 please_cite(stdout,"Caleman2011b");
571 /* Writing eigenvectors. Note that if mass scaling was used, the eigenvectors
572 * were scaled back from mass weighted cartesian to plain cartesian in the
573 * nma_full_hessian() or nma_sparse_hessian() routines. Mass scaled vectors
574 * will not be strictly orthogonal in plain cartesian scalar products.
576 write_eigenvectors(opt2fn("-v",NFILE,fnm),natoms,eigenvectors,FALSE,begin,end,
577 eWXR_NO,NULL,FALSE,top_x,bM,eigenvalues);