Szilard Pall [Fri, 1 Oct 2010 16:42:38 +0000 (18:42 +0200)]
make install-mdrun now installes shared libs (fix for bugzilla 549) + mnir fixes
"make install-mdrun" correctly installing the shared gromacs libraries
when the case, including the pkgconfig data and symbolic links.
Also made the openmm_wrapper and gmgx_gpu_utils libraries always build
and link staticlally (no need for these externally).
Christoph Junghans [Thu, 30 Sep 2010 10:01:20 +0000 (12:01 +0200)]
configure: added '--disable-la-files' option
Berk Hess [Thu, 30 Sep 2010 14:07:01 +0000 (16:07 +0200)]
changed the grompp charge group radii to always check against rlist iso againt rlistlong
Sander Pronk [Thu, 30 Sep 2010 08:56:22 +0000 (10:56 +0200)]
Error message on enabling FORTRAN and threads.
Roland Schulz [Thu, 30 Sep 2010 02:49:55 +0000 (22:49 -0400)]
made vmdio work with older api version (e.g. vmd 1.8.6)
improved error message
Roland Schulz [Mon, 27 Sep 2010 22:27:57 +0000 (18:27 -0400)]
include "error.h" was not neccessary and caused problems with xlc (bug 575)
Roland Schulz [Sun, 26 Sep 2010 05:41:40 +0000 (01:41 -0400)]
1) gmx_fio_seek was not returning correct return value, thus trjacat thought the seek failed and trjcat -overwrite wasn't working
2) checkpoint didn't check the return value of gmx_fio_seek
3) log file wasn't truncated if md5 file was available in checkpoint
4) made comments more clear
Rossen Apostolov [Sat, 25 Sep 2010 19:59:20 +0000 (21:59 +0200)]
Reverted changes to md_openmm.c, get rid of 'nan' in averages.
David van der Spoel [Sat, 25 Sep 2010 13:13:36 +0000 (15:13 +0200)]
Added software-related quote
Rossen Apostolov [Sat, 25 Sep 2010 12:21:52 +0000 (14:21 +0200)]
mdrun-gpu now doesn't output individual energy term, only the totals.
Closes bug 568.
Berk Hess [Sat, 25 Sep 2010 07:37:42 +0000 (09:37 +0200)]
fixed case of opls atom types in share/top/ions.itp to make old opls topologies work with the new ff setup
David van der Spoel [Fri, 24 Sep 2010 06:45:52 +0000 (08:45 +0200)]
Updated to reflect bug 557.
Mark Abraham [Thu, 23 Sep 2010 04:41:41 +0000 (14:41 +1000)]
Fixes for g_chi for bugzilla 569, so that it now uses the GROMACS 4.5 residuetypes.dat database, not the old aminoacids.dat. These fixes use the new gmx_residuetypes API functions contained in the previous commit (
bbf195d).
Mark Abraham [Thu, 23 Sep 2010 04:40:30 +0000 (14:40 +1000)]
Extensions to the gmx_residuetypes_t API to meet the needs of g_chi
Roland Schulz [Thu, 23 Sep 2010 04:05:01 +0000 (00:05 -0400)]
vmdio read the box but didn't mark it as beeing read (bBox)
Roland Shulz [Wed, 22 Sep 2010 21:33:39 +0000 (17:33 -0400)]
MCH3S was missing
Sander Pronk [Wed, 22 Sep 2010 14:51:29 +0000 (16:51 +0200)]
Added missing includes to gmx_tune_pme.c
Berk Hess [Wed, 22 Sep 2010 14:22:29 +0000 (16:22 +0200)]
added a warning/note for bond oscillational periods of a few times the time step
Berk Hess [Tue, 21 Sep 2010 15:43:33 +0000 (17:43 +0200)]
added check for tau_t<=0 with integrator sd
Berk Hess [Tue, 21 Sep 2010 15:19:32 +0000 (17:19 +0200)]
fixed incorrect gmxcheck comparison of ir->separate_dhdl_file
Carsten Kutzner [Tue, 21 Sep 2010 14:40:45 +0000 (16:40 +0200)]
Added missing #ifdef around MPI directive
Rossen Apostolov [Tue, 21 Sep 2010 14:06:04 +0000 (16:06 +0200)]
Fixed missing installation of the openmm_api_wrapper; added INSTALL.*make* for distribution.
Berk Hess [Tue, 21 Sep 2010 13:28:32 +0000 (15:28 +0200)]
improved the initial guess and fitting weights for autocorrelation fitting with -beginfit>0
Berk Hess [Tue, 21 Sep 2010 13:27:53 +0000 (15:27 +0200)]
removed g_analyze confusing fitting messages about fitting the function when fitting only the acf
Szilard Pall [Mon, 20 Sep 2010 21:26:23 +0000 (23:26 +0200)]
FindFFT?[F] scripts can use a path provided in FFTWx[F]_ROOT_DIR environment variable when looking for libs/headers
Berk Hess [Mon, 20 Sep 2010 14:50:01 +0000 (16:50 +0200)]
fixed too small fixed array in grompp top reading (only used with incorrect input
Rossen Apostolov [Mon, 20 Sep 2010 12:27:15 +0000 (14:27 +0200)]
Added do_multiprot in src/contrib, thanks to Ran Friedman
David van der Spoel [Mon, 20 Sep 2010 11:54:17 +0000 (13:54 +0200)]
Merge branch 'release-4-5-patches' of git@git.gromacs.org:gromacs into release-4-5-patches
David van der Spoel [Mon, 20 Sep 2010 11:50:05 +0000 (13:50 +0200)]
Fixed bug bugzilla.gromacs.org/show_bug.cgi?id=564 by adding a call to a routine freeing datastructures for lsq fitting. The program still uses memory
linearly with the number of timesteps in the mulsq variable. This should be fixed later on.
Carsten Kutzner [Mon, 20 Sep 2010 11:03:26 +0000 (13:03 +0200)]
gmx_pme_error now uses at most all charges for the self error estimate (even for -self values > 1)
Per Larsson [Mon, 20 Sep 2010 09:00:20 +0000 (11:00 +0200)]
Changes to GB non-polar solvation calculations
When using GB, users now explicitly have to ask for the non-polar
solvation to be calculated. Previously this was always done despite
the sa_algorithm=no in the mdp-file.
Grompp now also honors the value of sa_surface_algorithm in the mdp-file.
The non-polar energy gets printed, so that all partial
energy terms add upp to the total potetial energy.
All of these changes are documented in the online manual.
Teemu Murtola [Mon, 20 Sep 2010 07:22:16 +0000 (09:22 +0200)]
Fixes for selection modifier initialization.
Selection keywords merge, plus, and permute did not work properly in all
cases, and resulted in double frees if there was a syntax error. There
are still some pathological cases like "name CA plus x < 5" that don't
work correctly, but those should rarely be needed, and will take more
time to fix.
Roland Schulz [Sun, 19 Sep 2010 22:09:20 +0000 (15:09 -0700)]
fixed box vector read in vmdio
Carsten Kutzner [Fri, 17 Sep 2010 13:20:48 +0000 (15:20 +0200)]
Fixed reproducibility of g_pme_error results in parallel
Sander Pronk [Fri, 17 Sep 2010 12:19:24 +0000 (14:19 +0200)]
Fixed more checkpointing issues with .edr dh data.
Sander Pronk [Fri, 17 Sep 2010 11:35:50 +0000 (13:35 +0200)]
Checkpoints backward-compatible + more consistency checks in g_bar.
Sander Pronk [Fri, 17 Sep 2010 10:57:07 +0000 (12:57 +0200)]
Fixed checkpointing bug with g_bar energies
Berk Hess [Fri, 17 Sep 2010 10:19:58 +0000 (12:19 +0200)]
Revert "added GB non-polar solvation energy term to the log and energy file output"
This reverts commit
8c51a787e9cbe208361c8428b949213962bde589.
Christoph Junghans [Fri, 17 Sep 2010 09:51:46 +0000 (11:51 +0200)]
configure: power6 and fortran can now be enabled at the same time.
This is not extremely useful for the kernels, because power6 kernels
will overwrite fortran kernels, but it allows other fortran stuff,
like the libxdrf fortran interface, on power6.
Christoph Junghans [Fri, 17 Sep 2010 09:45:22 +0000 (11:45 +0200)]
configure: fortran kernels can also be compiled with gfortran (g77 is not included in >=gcc-4.0 anymore)
Berk Hess [Fri, 17 Sep 2010 08:11:32 +0000 (10:11 +0200)]
added GB non-polar solvation energy term to the log and energy file output
Berk Hess [Fri, 17 Sep 2010 08:10:51 +0000 (10:10 +0200)]
fixed free-energy giving a fatal-error when GB12, GB13 or GB14 terms are present
Berk Hess [Fri, 17 Sep 2010 07:00:08 +0000 (09:00 +0200)]
added hint to g_traj warning
kyleb [Thu, 16 Sep 2010 21:26:22 +0000 (14:26 -0700)]
Changed gmx_log_open to fix FAH Checkpointing; see Bugzilla 538.
David van der Spoel [Thu, 16 Sep 2010 17:08:52 +0000 (19:08 +0200)]
Corrected name of one of the developers listed in header.
Berk Hess [Thu, 16 Sep 2010 15:47:52 +0000 (17:47 +0200)]
fixed the maximum print for g_traj -cv and -cf
Sander Pronk [Thu, 16 Sep 2010 15:48:55 +0000 (17:48 +0200)]
eneconv now warns when throwing away blocks with dH data.
Pär Bjelkmar [Thu, 16 Sep 2010 14:28:51 +0000 (16:28 +0200)]
Merge branch 'release-4-5-patches' of git://git.gromacs.org/gromacs into release-4-5-patches
Sander Pronk [Thu, 16 Sep 2010 14:29:09 +0000 (16:29 +0200)]
Improved performance of g_bar while reading .edr files with small nstenergy.
Pär Bjelkmar [Thu, 16 Sep 2010 14:15:31 +0000 (16:15 +0200)]
Atom-based charge groups in CHARMM.
Berk Hess [Thu, 16 Sep 2010 13:41:29 +0000 (15:41 +0200)]
fixed some minor issues with g_traj -cv and -cf
Berk Hess [Thu, 16 Sep 2010 12:45:40 +0000 (14:45 +0200)]
tau_t=-1 now works in grompp and with SD
Berk Hess [Wed, 15 Sep 2010 15:44:48 +0000 (17:44 +0200)]
fixed g_traj segv when ran without pbc
Berk Hess [Wed, 15 Sep 2010 15:28:41 +0000 (17:28 +0200)]
fixed charge group radius LJ check, which could produce too low numbers in the grompp notice/warning print
Berk Hess [Wed, 15 Sep 2010 15:22:25 +0000 (17:22 +0200)]
added missing F_PIDIHS case to topsort free energy perturbation check
Berk Hess [Wed, 15 Sep 2010 13:32:45 +0000 (15:32 +0200)]
added GMX_MAX_THREADS env.var. to set the maximum number of threads for thread-MPI
Christoph Junghans [Wed, 15 Sep 2010 09:02:27 +0000 (11:02 +0200)]
configure: added missing quotes
Sander Pronk [Wed, 15 Sep 2010 08:49:54 +0000 (10:49 +0200)]
More scheduler-friendly locks for gmxfio
Christoph Junghans [Wed, 15 Sep 2010 08:44:30 +0000 (10:44 +0200)]
Fixed synchronization counter update for pwr6 kernels
Teemu Murtola [Tue, 14 Sep 2010 13:51:53 +0000 (15:51 +0200)]
Fixed version generation script to work on Macs.
There was a non-standard \| construct in a sed script that did work with
some sed implementations, but not with others.
Berk Hess [Tue, 14 Sep 2010 13:28:23 +0000 (15:28 +0200)]
fixed memory allocation and access issues when reading xpm files
Berk Hess [Tue, 14 Sep 2010 11:11:27 +0000 (13:11 +0200)]
grompp option -pp now always writes to cwd
Berk Hess [Tue, 14 Sep 2010 10:50:43 +0000 (12:50 +0200)]
g_density now normalizes every frame separately
Carsten Kutzner [Tue, 14 Sep 2010 10:52:10 +0000 (12:52 +0200)]
Removed unused code from g_pme_error
Berk Hess [Tue, 14 Sep 2010 10:10:15 +0000 (12:10 +0200)]
g_density now normalizes with the average volume instead of the volume of the last frame
Berk Hess [Tue, 14 Sep 2010 08:57:15 +0000 (10:57 +0200)]
fixed incorrect reference position of position restraints with free energy calculations with posA!=posB and lambda!=0
David van der Spoel [Tue, 14 Sep 2010 07:04:25 +0000 (09:04 +0200)]
Fixed bug 553 by checking whether an atom name is longer than 2 characters
and has a digit in the third character. In that case the name is expected
to be protein. However, in this case e.g. Si12 will be designated as Sulphur.
The alternative would be to check whether the present residue is an amino acid
but that may be problematic as well.
Mark Abraham [Tue, 14 Sep 2010 02:50:16 +0000 (12:50 +1000)]
When grompp failed to find the forcefield, the error messages had some string parameters in the wrong order
Mark Abraham [Tue, 14 Sep 2010 02:49:08 +0000 (12:49 +1000)]
Formatting fixes to documentation
Per Larsson [Mon, 13 Sep 2010 12:14:29 +0000 (14:14 +0200)]
When using GBSA, the note encouraging PME over plain cut-offs is no longer printed
Rossen Apostolov [Mon, 13 Sep 2010 11:26:38 +0000 (13:26 +0200)]
Added dependencies for openmm_api_wrapper
Rossen Apostolov [Mon, 13 Sep 2010 10:45:49 +0000 (12:45 +0200)]
Added library dependency for gmx_gpu_utils
Berk Hess [Mon, 13 Sep 2010 08:39:50 +0000 (10:39 +0200)]
fixed grompp -t frame choice for trajectories with frames with and without v
Berk Hess [Mon, 13 Sep 2010 07:58:55 +0000 (09:58 +0200)]
state switches between the cpt and tpr file in mdrun are now only allowed when an env.var. is set
Berk Hess [Mon, 13 Sep 2010 07:52:24 +0000 (09:52 +0200)]
added check and warning for no velocities in grompp -t file
David van der Spoel [Sun, 12 Sep 2010 14:49:31 +0000 (16:49 +0200)]
Fixed demo to work with new pdb2gmx that requires two input values.
Berk Hess [Fri, 10 Sep 2010 20:01:00 +0000 (22:01 +0200)]
turned on checkpointing with nstlist=0
Berk Hess [Fri, 10 Sep 2010 13:33:09 +0000 (15:33 +0200)]
removed the lower bound check on test particle insertion energies (this was due to table issues which are now resolved)
Berk Hess [Fri, 10 Sep 2010 09:35:13 +0000 (11:35 +0200)]
removed tip5p from charmm27.ff/watermodels.dat, since several other files would need to be updated as well and we don't want to encourage the use of tip5p
Berk Hess [Fri, 10 Sep 2010 09:28:26 +0000 (11:28 +0200)]
made sure the energy history from a checkpoint file is ignored when not appending mdrun output
Teemu Murtola [Fri, 10 Sep 2010 07:08:50 +0000 (09:08 +0200)]
CMake documentation improvements (bug #539).
Added a bit of information to CMake cache documentation strings as
discussed in bug #539.
Mark Abraham [Thu, 9 Sep 2010 23:17:20 +0000 (09:17 +1000)]
Added TIP5p to CHARMM forcefield by analogy with AMBER03 TIP5p and existing CHARMM TIP[34]p
Mark Abraham [Wed, 8 Sep 2010 04:59:04 +0000 (14:59 +1000)]
Spelling fix
Sander Pronk [Thu, 9 Sep 2010 15:18:50 +0000 (17:18 +0200)]
Fixed a deadlock while closing files.
Berk Hess [Thu, 9 Sep 2010 14:49:46 +0000 (16:49 +0200)]
added SSE overflow check for TPI
Berk Hess [Thu, 9 Sep 2010 14:44:15 +0000 (16:44 +0200)]
the force tables are now continued linearly from nearly 0 to 0 to treat completely overlapping atoms in energy minimization or TPI
Sander Pronk [Thu, 9 Sep 2010 13:56:16 +0000 (15:56 +0200)]
Fixed segv in trajana due to wrong cell index calculation.
Gerrit Groenhof [Thu, 9 Sep 2010 07:13:31 +0000 (09:13 +0200)]
added environment variable to run orca in parallel
Sander Pronk [Wed, 8 Sep 2010 19:43:41 +0000 (21:43 +0200)]
Fixed segv while reading overly long lines in xvgr.c
hess [Wed, 8 Sep 2010 14:48:02 +0000 (16:48 +0200)]
mdrun now also generates a fatal error for nnodes with large prime divisors (when -dd is not set)
Szilard Pall [Wed, 8 Sep 2010 14:10:23 +0000 (16:10 +0200)]
Fixed binary/lib suffixing inconsistencies (bug 539)
Berk Hess [Wed, 8 Sep 2010 11:51:28 +0000 (13:51 +0200)]
improved the automatic choice of the number of PME nodes
Per Larsson [Wed, 8 Sep 2010 11:41:43 +0000 (13:41 +0200)]
Fixed atomtype H0 for GBSA in amber03
Berk Hess [Wed, 8 Sep 2010 09:39:18 +0000 (11:39 +0200)]
added MCH3S atom type for Charmm for virtual sites in Met
Mark Abraham [Wed, 8 Sep 2010 01:56:52 +0000 (11:56 +1000)]
Added note to g_covar about memory costs and managing segfaults
Berk Hess [Tue, 7 Sep 2010 14:32:40 +0000 (16:32 +0200)]
changed the order of the outer loops in the fft5d transpose routines
Berk Hess [Tue, 7 Sep 2010 11:16:20 +0000 (13:16 +0200)]
removed empty edr and xtc files with nm
Berk Hess [Tue, 7 Sep 2010 11:15:50 +0000 (13:15 +0200)]
removed (nan) average printing with normal modes
Berk Hess [Tue, 7 Sep 2010 11:15:10 +0000 (13:15 +0200)]
fixed crash with parallel nm with cut-off's
Szilard Pall [Tue, 7 Sep 2010 08:31:48 +0000 (10:31 +0200)]
fixed warnings caused by some uninitialized variables