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 * Copyright (c) 2013,2014, by the GROMACS development team, led by
7 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
8 * and including many others, as listed in the AUTHORS file in the
9 * top-level source directory and at http://www.gromacs.org.
11 * GROMACS is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU Lesser General Public License
13 * as published by the Free Software Foundation; either version 2.1
14 * of the License, or (at your option) any later version.
16 * GROMACS is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * Lesser General Public License for more details.
21 * You should have received a copy of the GNU Lesser General Public
22 * License along with GROMACS; if not, see
23 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
24 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
26 * If you want to redistribute modifications to GROMACS, please
27 * consider that scientific software is very special. Version
28 * control is crucial - bugs must be traceable. We will be happy to
29 * consider code for inclusion in the official distribution, but
30 * derived work must not be called official GROMACS. Details are found
31 * in the README & COPYING files - if they are missing, get the
32 * official version at http://www.gromacs.org.
34 * To help us fund GROMACS development, we humbly ask that you cite
35 * the research papers on the package. Check out http://www.gromacs.org.
48 #include "gromacs/utility/futil.h"
53 #include "mtop_util.h"
54 #include "gromacs/math/units.h"
57 #include "gromacs/commandline/pargs.h"
58 #include "gromacs/fileio/mtxio.h"
59 #include "gromacs/fileio/xvgr.h"
60 #include "gromacs/linearalgebra/eigensolver.h"
61 #include "gromacs/linearalgebra/sparsematrix.h"
62 #include "gromacs/math/vec.h"
63 #include "gromacs/utility/smalloc.h"
65 static double cv_corr(double nu, double T)
67 double x = PLANCK*nu/(BOLTZ*T);
76 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);
91 return BOLTZ*T*(0.5*x - 1 + x/(ex-1));
95 static int get_nharm_mt(gmx_moltype_t *mt)
97 static int harm_func[] = { F_BONDS };
101 for (i = 0; (i < asize(harm_func)); i++)
104 nh += mt->ilist[ft].nr/(interaction_function[ft].nratoms+1);
109 static int get_nvsite_mt(gmx_moltype_t *mt)
111 static int vs_func[] = {
112 F_VSITE2, F_VSITE3, F_VSITE3FD, F_VSITE3FAD,
113 F_VSITE3OUT, F_VSITE4FD, F_VSITE4FDN, F_VSITEN
118 for (i = 0; (i < asize(vs_func)); i++)
121 nh += mt->ilist[ft].nr/(interaction_function[ft].nratoms+1);
126 static int get_nharm(gmx_mtop_t *mtop, int *nvsites)
132 for (j = 0; (j < mtop->nmolblock); j++)
134 mt = mtop->molblock[j].type;
135 nh += mtop->molblock[j].nmol * get_nharm_mt(&(mtop->moltype[mt]));
136 nv += mtop->molblock[j].nmol * get_nvsite_mt(&(mtop->moltype[mt]));
143 nma_full_hessian(real * hess,
156 natoms = top->atoms.nr;
158 /* divide elements hess[i][j] by sqrt(mas[i])*sqrt(mas[j]) when required */
162 for (i = 0; (i < natoms); i++)
164 for (j = 0; (j < DIM); j++)
166 for (k = 0; (k < natoms); k++)
168 mass_fac = gmx_invsqrt(top->atoms.atom[i].m*top->atoms.atom[k].m);
169 for (l = 0; (l < DIM); l++)
171 hess[(i*DIM+j)*ndim+k*DIM+l] *= mass_fac;
178 /* call diagonalization routine. */
180 fprintf(stderr, "\nDiagonalizing to find vectors %d through %d...\n", begin, end);
183 eigensolver(hess, ndim, begin-1, end-1, eigenvalues, eigenvectors);
185 /* And scale the output eigenvectors */
186 if (bM && eigenvectors != NULL)
188 for (i = 0; i < (end-begin+1); i++)
190 for (j = 0; j < natoms; j++)
192 mass_fac = gmx_invsqrt(top->atoms.atom[j].m);
193 for (k = 0; (k < DIM); k++)
195 eigenvectors[i*ndim+j*DIM+k] *= mass_fac;
205 nma_sparse_hessian(gmx_sparsematrix_t * sparse_hessian,
219 natoms = top->atoms.nr;
222 /* Cannot check symmetry since we only store half matrix */
223 /* divide elements hess[i][j] by sqrt(mas[i])*sqrt(mas[j]) when required */
227 for (iatom = 0; (iatom < natoms); iatom++)
229 for (j = 0; (j < DIM); j++)
232 for (k = 0; k < sparse_hessian->ndata[row]; k++)
234 col = sparse_hessian->data[row][k].col;
236 mass_fac = gmx_invsqrt(top->atoms.atom[iatom].m*top->atoms.atom[katom].m);
237 sparse_hessian->data[row][k].value *= mass_fac;
242 fprintf(stderr, "\nDiagonalizing to find eigenvectors 1 through %d...\n", neig);
245 sparse_eigensolver(sparse_hessian, neig, eigenvalues, eigenvectors, 10000000);
247 /* Scale output eigenvectors */
248 if (bM && eigenvectors != NULL)
250 for (i = 0; i < neig; i++)
252 for (j = 0; j < natoms; j++)
254 mass_fac = gmx_invsqrt(top->atoms.atom[j].m);
255 for (k = 0; (k < DIM); k++)
257 eigenvectors[i*ndim+j*DIM+k] *= mass_fac;
266 int gmx_nmeig(int argc, char *argv[])
268 const char *desc[] = {
269 "[THISMODULE] calculates the eigenvectors/values of a (Hessian) matrix,",
270 "which can be calculated with [gmx-mdrun].",
271 "The eigenvectors are written to a trajectory file ([TT]-v[tt]).",
272 "The structure is written first with t=0. The eigenvectors",
273 "are written as frames with the eigenvector number as timestamp.",
274 "The eigenvectors can be analyzed with [gmx-anaeig].",
275 "An ensemble of structures can be generated from the eigenvectors with",
276 "[gmx-nmens]. When mass weighting is used, the generated eigenvectors",
277 "will be scaled back to plain Cartesian coordinates before generating the",
278 "output. In this case, they will no longer be exactly orthogonal in the",
279 "standard Cartesian norm, but in the mass-weighted norm they would be.[PAR]",
280 "This program can be optionally used to compute quantum corrections to heat capacity",
281 "and enthalpy by providing an extra file argument [TT]-qcorr[tt]. See the GROMACS",
282 "manual, Chapter 1, for details. The result includes subtracting a harmonic",
283 "degree of freedom at the given temperature.",
284 "The total correction is printed on the terminal screen.",
285 "The recommended way of getting the corrections out is:[PAR]",
286 "[TT]gmx nmeig -s topol.tpr -f nm.mtx -first 7 -last 10000 -T 300 -qc [-constr][tt][PAR]",
287 "The [TT]-constr[tt] option should be used when bond constraints were used during the",
288 "simulation [BB]for all the covalent bonds[bb]. If this is not the case, ",
289 "you need to analyze the [TT]quant_corr.xvg[tt] file yourself.[PAR]",
290 "To make things more flexible, the program can also take virtual sites into account",
291 "when computing quantum corrections. When selecting [TT]-constr[tt] and",
292 "[TT]-qc[tt], the [TT]-begin[tt] and [TT]-end[tt] options will be set automatically as well.",
293 "Again, if you think you know it better, please check the [TT]eigenfreq.xvg[tt]",
297 static gmx_bool bM = TRUE, bCons = FALSE;
298 static int begin = 1, end = 50, maxspec = 4000;
299 static real T = 298.15, width = 1;
302 { "-m", FALSE, etBOOL, {&bM},
303 "Divide elements of Hessian by product of sqrt(mass) of involved "
304 "atoms prior to diagonalization. This should be used for 'Normal Modes' "
306 { "-first", FALSE, etINT, {&begin},
307 "First eigenvector to write away" },
308 { "-last", FALSE, etINT, {&end},
309 "Last eigenvector to write away" },
310 { "-maxspec", FALSE, etINT, {&maxspec},
311 "Highest frequency (1/cm) to consider in the spectrum" },
312 { "-T", FALSE, etREAL, {&T},
313 "Temperature for computing quantum heat capacity and enthalpy when using normal mode calculations to correct classical simulations" },
314 { "-constr", FALSE, etBOOL, {&bCons},
315 "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." },
316 { "-width", FALSE, etREAL, {&width},
317 "Width (sigma) of the gaussian peaks (1/cm) when generating a spectrum" }
319 FILE *out, *qc, *spec;
328 real rdum, mass_fac, qcvtot, qutot, qcv, qu;
329 int natoms, ndim, nrow, ncol, count, nharm, nvsite;
331 int i, j, k, l, d, gnx;
335 int version, generation;
336 real value, omega, nu;
337 real factor_gmx_to_omega2;
338 real factor_omega_to_wavenumber;
339 real *spectrum = NULL;
342 const char *qcleg[] = {
343 "Heat Capacity cV (J/mol K)",
344 "Enthalpy H (kJ/mol)"
346 real * full_hessian = NULL;
347 gmx_sparsematrix_t * sparse_hessian = NULL;
350 { efMTX, "-f", "hessian", ffREAD },
351 { efTPX, NULL, NULL, ffREAD },
352 { efXVG, "-of", "eigenfreq", ffWRITE },
353 { efXVG, "-ol", "eigenval", ffWRITE },
354 { efXVG, "-os", "spectrum", ffOPTWR },
355 { efXVG, "-qc", "quant_corr", ffOPTWR },
356 { efTRN, "-v", "eigenvec", ffWRITE }
358 #define NFILE asize(fnm)
360 if (!parse_common_args(&argc, argv, PCA_BE_NICE,
361 NFILE, fnm, asize(pa), pa, asize(desc), desc, 0, NULL, &oenv))
366 /* Read tpr file for volume and number of harmonic terms */
367 read_tpxheader(ftp2fn(efTPX, NFILE, fnm), &tpx, TRUE, &version, &generation);
368 snew(top_x, tpx.natoms);
370 read_tpx(ftp2fn(efTPX, NFILE, fnm), NULL, box, &natoms,
371 top_x, NULL, NULL, &mtop);
374 nharm = get_nharm(&mtop, &nvsite);
381 top = gmx_mtop_t_to_t_topology(&mtop);
386 if (opt2bSet("-qc", NFILE, fnm))
388 begin = 7+DIM*nvsite;
399 printf("Using begin = %d and end = %d\n", begin, end);
401 /*open Hessian matrix */
402 gmx_mtxio_read(ftp2fn(efMTX, NFILE, fnm), &nrow, &ncol, &full_hessian, &sparse_hessian);
404 /* Memory for eigenvalues and eigenvectors (begin..end) */
405 snew(eigenvalues, nrow);
406 snew(eigenvectors, nrow*(end-begin+1));
408 /* If the Hessian is in sparse format we can calculate max (ndim-1) eigenvectors,
409 * and they must start at the first one. If this is not valid we convert to full matrix
410 * storage, but warn the user that we might run out of memory...
412 if ((sparse_hessian != NULL) && (begin != 1 || end == ndim))
416 fprintf(stderr, "Cannot use sparse Hessian with first eigenvector != 1.\n");
418 else if (end == ndim)
420 fprintf(stderr, "Cannot use sparse Hessian to calculate all eigenvectors.\n");
423 fprintf(stderr, "Will try to allocate memory and convert to full matrix representation...\n");
425 snew(full_hessian, nrow*ncol);
426 for (i = 0; i < nrow*ncol; i++)
431 for (i = 0; i < sparse_hessian->nrow; i++)
433 for (j = 0; j < sparse_hessian->ndata[i]; j++)
435 k = sparse_hessian->data[i][j].col;
436 value = sparse_hessian->data[i][j].value;
437 full_hessian[i*ndim+k] = value;
438 full_hessian[k*ndim+i] = value;
441 gmx_sparsematrix_destroy(sparse_hessian);
442 sparse_hessian = NULL;
443 fprintf(stderr, "Converted sparse to full matrix storage.\n");
446 if (full_hessian != NULL)
448 /* Using full matrix storage */
449 nma_full_hessian(full_hessian, nrow, bM, &top, begin, end,
450 eigenvalues, eigenvectors);
454 /* Sparse memory storage, allocate memory for eigenvectors */
455 snew(eigenvectors, ncol*end);
456 nma_sparse_hessian(sparse_hessian, bM, &top, end, eigenvalues, eigenvectors);
459 /* check the output, first 6 eigenvalues should be reasonably small */
461 for (i = begin-1; (i < 6); i++)
463 if (fabs(eigenvalues[i]) > 1.0e-3)
470 fprintf(stderr, "\nOne of the lowest 6 eigenvalues has a non-zero value.\n");
471 fprintf(stderr, "This could mean that the reference structure was not\n");
472 fprintf(stderr, "properly energy minimized.\n");
475 /* now write the output */
476 fprintf (stderr, "Writing eigenvalues...\n");
477 out = xvgropen(opt2fn("-ol", NFILE, fnm),
478 "Eigenvalues", "Eigenvalue index", "Eigenvalue [Gromacs units]",
480 if (output_env_get_print_xvgr_codes(oenv))
484 fprintf(out, "@ subtitle \"mass weighted\"\n");
488 fprintf(out, "@ subtitle \"not mass weighted\"\n");
492 for (i = 0; i <= (end-begin); i++)
494 fprintf (out, "%6d %15g\n", begin+i, eigenvalues[i]);
499 if (opt2bSet("-qc", NFILE, fnm))
501 qc = xvgropen(opt2fn("-qc", NFILE, fnm), "Quantum Corrections", "Eigenvector index", "", oenv);
502 xvgr_legend(qc, asize(qcleg), qcleg, oenv);
509 printf("Writing eigenfrequencies - negative eigenvalues will be set to zero.\n");
511 out = xvgropen(opt2fn("-of", NFILE, fnm),
512 "Eigenfrequencies", "Eigenvector index", "Wavenumber [cm\\S-1\\N]",
514 if (output_env_get_print_xvgr_codes(oenv))
518 fprintf(out, "@ subtitle \"mass weighted\"\n");
522 fprintf(out, "@ subtitle \"not mass weighted\"\n");
527 if (opt2bSet("-os", NFILE, fnm) && (maxspec > 0))
529 snew(spectrum, maxspec);
530 spec = xvgropen(opt2fn("-os", NFILE, fnm),
531 "Vibrational spectrum based on harmonic approximation",
532 "\\f{12}w\\f{4} (cm\\S-1\\N)",
533 "Intensity [Gromacs units]",
535 for (i = 0; (i < maxspec); i++)
541 /* Gromacs units are kJ/(mol*nm*nm*amu),
542 * where amu is the atomic mass unit.
544 * For the eigenfrequencies we want to convert this to spectroscopic absorption
545 * wavenumbers given in cm^(-1), which is the frequency divided by the speed of
546 * light. Do this by first converting to omega^2 (units 1/s), take the square
547 * root, and finally divide by the speed of light (nm/ps in gromacs).
549 factor_gmx_to_omega2 = 1.0E21/(AVOGADRO*AMU);
550 factor_omega_to_wavenumber = 1.0E-5/(2.0*M_PI*SPEED_OF_LIGHT);
552 for (i = begin; (i <= end); i++)
554 value = eigenvalues[i-begin];
559 omega = sqrt(value*factor_gmx_to_omega2);
560 nu = 1e-12*omega/(2*M_PI);
561 value = omega*factor_omega_to_wavenumber;
562 fprintf (out, "%6d %15g\n", i, value);
565 wfac = eigenvalues[i-begin]/(width*sqrt(2*M_PI));
566 for (j = 0; (j < maxspec); j++)
568 spectrum[j] += wfac*exp(-sqr(j-value)/(2*sqr(width)));
573 qcv = cv_corr(nu, T);
580 fprintf (qc, "%6d %15g %15g\n", i, qcv, qu);
588 for (j = 0; (j < maxspec); j++)
590 fprintf(spec, "%10g %10g\n", 1.0*j, spectrum[j]);
596 printf("Quantum corrections for harmonic degrees of freedom\n");
597 printf("Use appropriate -first and -last options to get reliable results.\n");
598 printf("There were %d constraints and %d vsites in the simulation\n",
600 printf("Total correction to cV = %g J/mol K\n", qcvtot);
601 printf("Total correction to H = %g kJ/mol\n", qutot);
603 please_cite(stdout, "Caleman2011b");
605 /* Writing eigenvectors. Note that if mass scaling was used, the eigenvectors
606 * were scaled back from mass weighted cartesian to plain cartesian in the
607 * nma_full_hessian() or nma_sparse_hessian() routines. Mass scaled vectors
608 * will not be strictly orthogonal in plain cartesian scalar products.
610 write_eigenvectors(opt2fn("-v", NFILE, fnm), natoms, eigenvectors, FALSE, begin, end,
611 eWXR_NO, NULL, FALSE, top_x, bM, eigenvalues);