ChangeLog.c33
| | |
|
CHARMM 33 |
| c33a1 |
Developmental |
August 15, 2005 |
| c33a2 |
Developmental |
February 15, 2006 |
| c33b1 |
Release |
August 15, 2006 |
| c33b2 |
Release |
February 15, 2007 |
|
| | |
|
About CHARMM33 Development |
|
|
|
At the July 2005 CHARMM meeting (Cornell University, July 8-9, 2005),
we decided to release CHARMM c32b1 with incorporating enhancements and fixes
to the developmental version c32a2. The modified source code and features
are described in ChangeLog.c32, [3] c32b1 Change Log.
As an on-going project, CHARMM development will be carried out
with CHARMM version 33a series. The following describes new features
and enhancements incorporated in the c33 development period.
The following are introduced and modified during the development period
of February 15, 2005 through August 15, 2005. The developments are
based on CHARMM version c32a2. Some enhancements and fixes to c32a2
are also incorporated into c32b1 as described in the c32b1 release note
(ChangeLog.c32).
1.1 University of Wisconsin Developments
Developers : Qiang Cui
Email Address : cui@chem.wisc.edu
Institution : University of Wisconsin, Madison
Date : July 1, 2005
1.1.1 General compilation of SCC-DFTB in CHARMM
"$(LIB)/sccdftbint.a \" should be added to all platform Makefiles of
build/UNX. Otherwise, SCC-DFTB will NOT be compiled even with the T keyword.
build/UNX/sccdftb.mk was updated; optimization level (-O2) was enforced.
Files Modified:
build/UNX/sccdftb.mk
..............................................................................
1.1.2 Bug-fixes/improvements for SCC-DFTB/MM options
When group-based non-bond scheme is used with QM/MM, e.g., extended
electrostatics in conjunction with GSBP, the van der Waals interactions
among QM atoms are appropriately removed. In addition, the same formula
is used when fast-off is selected for extended electrostatics
A simple bug is also fixed for GHO.
Files Modified:
source/fcm/gamess.fcm
source/nbonds/nbondg.src
source/nbonds/enbondg.src
source/nbonds/enbond.src
source/nbonds/heurist.src
source/quantum/qmene.src
source/sccdftbint/stbgho.src
..............................................................................
1.1.3 The SCC-DFTB/MM Ewald Summation
The Ewald summation with SCC-DFTB/MM can now be executed in the
parallel mode, in which the QM/MM Ewald sums are distributed over different
nodes (SCC-DFTB itself is computed only on one node)
Files Added:
source/sccdftbint/sccdftbsrc/dylcao.src
source/sccdftbint/sccdftbsrc/externalchgrad.src
source/sccdftbint/sccdftbsrc/ktable_ewscc.src
source/sccdftbint/sccdftbsrc/long_sccmm1.src
source/sccdftbint/sccdftbsrc/long_sccmm2.src
Files Removed:
source/sccdftbint/sccdftbsrc/dylcao.f
source/sccdftbint/sccdftbsrc/externalchgrad.f
source/sccdftbint/sccdftbsrc/ktable_ewscc.f
source/sccdftbint/sccdftbsrc/long_sccmm1.f
source/sccdftbint/sccdftbsrc/long_sccmm2.f
Files Modified:
source/fcm/energy.fcm
source/fcm/parallel.fcm
source/fcm/sccdftb.fcm
source/energy/energy.src
source/sccdftbint/sccdftbini.src
source/sccdftbint/sccdftbsrc/eglcao.f
source/sccdftbint/sccdftbsrc/gettab.f
source/sccdftbint/sccdftbsrc/short_range.f
..............................................................................
1.1.4 Rectangular GSBP-QM/MM
The GSBP-QM/MM procotol has been extended to the rectangular case;
several simple bugs were also corrected
Files Added:
source/misc/gsbpscc2.src
Files Modified:
source/misc/gsbp.src
source/misc/gsbpscc.src
..............................................................................
1.1.5 QM Atom Dispersion Interaction
Dispersion Interaction among QM atoms can now be calculated using
an empirical formulation (ref:Elstner et al. JCP, 114, 5149, 2001)
Files Added:
source/sccdftbint/sccdftbsrc/dispersion_egr.f
source/sccdftbint/sccdftbsrc/dispersionread.f
New Testcase:
test/c33test/h4o2_disp.inp
..............................................................................
1.1.6 SCC-DFTB Hydrogen Bonding and Proton Affinities
A damping function has been applied to the gamma for XH (use the keyword,
HBON in SCCDFTB) Charge-dependence of the Hubbard parameters have been added
(use keywords: DHUB, DHGA) The repulsive potential can also be adjusted using
an added Gaussian function modulated by a switching function in a given region.
This can be a rather flexible way to develop system specific parameters in the
SCC-DFTB framework (keyword GAUS)
Files Modified (module/file.src):
source/fcm/blockscc.fcm
source/sccdftbint/sccdftbsrc/gamma.f
source/sccdftbint/sccdftbsrc/repulsive.f
New Testcase:
test/c33test/h4o2_hbond.inp
test/c33test/h4o2_dhub.inp
..............................................................................
1.1.7 QM/MM Normal Mode Analysis
The current implementation involves only finite difference of analytical
first derivatives for QM/MM potentials, although analytic QM/MM hessian has
been reported and implemented (JCP, 2000,112,1133). It is convenient to fix
a major part of the system while doing finite difference on a small region
containing the QM part; see the included test case for an example.
Files Modified (module/file.src):
source/vibran/vibran.src
source/vibran/vibcom.src
source/vibran/redbas.src
source/vibran/raise.src
New Testcase:
test/c33test/h4o2_freq.inp
------------------------------------------------------------------------------
1.2 TSRI Developments
Developers : Mike Crowley, Charles L. Brooks, III
Email Address : crowley@sripps.edu, brooks@scripps.edu
Institution : The Scripps Research Institute
Date : June 30, 2005
1.2.1 Minor cleanup and bug fixes
Files Modified:
source/dynamc/dynamc.src some pref labels added
source/emap/emapfield.src removed extraneous "+" characters and fixed heap call
source/energy/gbmv.src bugfixes
source/fcm/gbmv.fcm bugfixes
source/gener/mkpres.src bugfix format statement
source/io/rtfio.src very minor cleanup
source/manip/shake.src cleanup and remove PARASCAL code for fast shake
source/manipfsshake.src
source/misc/mmfp.src bugfix format statement
source/misc/pbeq.src bugfix format statements
..............................................................................
1.2.2 Modifications for fortran90/95 standard
Modifications for fortran 90 and 95, mostly in removing assigned and
computed goto's and removing byte declarations. Aim is to get all code
up to date with the current standards especially for coding that will
no longer be supported. Added PREF keyword FORTRAN90 and added it to
all pref.dat generated by install.com for compilers that support fortran90/95
Added ifort and g95 to install.com
Files Modified:
install.com
build/UNX/Makefile_gnu
build/UNX/Makefile_ibmaix
build/UNX/Makefile_ibmaixmp
source/correl/correl.src
source/correl/solana.src
source/dimb/nmdimb.src
source/energy/energy.src
source/energy/epath.src
source/energy/gbmv.src
source/fcm/rtf.fcm
source/gener/genpsf.src
source/machdep/cstuff.c get pointer for g95
source/machdep/machio.src open with append syntax
source/machdep/machutil.src g95/osx get pointer stuff
source/machdep/space.src g95/osx get pointer stuff
source/minmiz/tnpack.src
source/rxncor/adiab.src
source/rxncor/travel.src
source/rxncor/travel2.src
source/util/selcta.src
source/util/string.src
source/vibran/vibran.src
source/vibran/vibutil.src
..............................................................................
1.2.3 Periodic boundary cubic list builder
Faster method for building list when simple periodic
boundary conditions are used. Does not use the image or crystal code, no image
atoms are created or updated. It removes the time taken to update image atoms.
PREF keyword: PBCUBES
incompatible with IMCUBES keyword. Only one can be specified in pref.dat
Files Added:
fcm/pbcubes.fcm
nbonds/pblist.src
Files Modified:
source/cff/enbfast_cff.src
source/cff/ewald_cff.src
source/charmm/iniall.src
source/energy/energy.src
source/energy/gbim.src
source/energy/gbmv.src
source/energy/gbsw.src
source/energy/genborn.src
source/energy/intere.src
source/fcm/deflts.fcm
source/fcm/image.fcm
source/fcn/inbnd.fcm
source/image/images.src
source/image/nbndgcm.src
source/image/upimag.src
source/image/xtlfrq.src
source/io/parmio.src
source/mbond/mbdiag.src
source/misc/hbuild.src
source/mmff/enbscalar_mm.src
source/nbonds/ace.src
source/nbonds/enbfast.src
source/nbonds/enbond.src
source/nbonds/enbonda.src
source/nbonds/ewald.src
source/nbonds/nbonds.src
source/nbonds/nbutil.src
..............................................................................
1.2.4 APBS Interface
Developers : Robert Konecny, Nathan Baker
Email Address : rok@ucsd.edu
Institution : University of California San Diego
Date : JUne 30, 2005
Added the CHARMM/APBS interface. APBS (Adaptive Poisson-Boltzmann
Solver) is a software package for evaluation of the electrostatic
potential for very large biomolecular systems (http://agave.wustl.edu/apbs/).
This interface makes most of the APBS functionality accessible from within
CHARMM. More information about the interface can be found at
http://mccammon.ucsd.edu/iapbs
Files Added:
source/misc/apbs.src
source/fcm/apbs.fcm
Files Modified:
source/energy/energy.src
source/misc/pbeq.src
build/UNX/misc.mk
build/UNX/Makefile_gnu
install.com
New Documentation:
doc/apbs.doc
New Testcase:
test/c33test/apbs_elstat.inp
test/c33test/apbs_md.inp
test/c33test/apbs_vis.inp
------------------------------------------------------------------------------
1.3 Developments at the National Institute of Health
Developers : Jhih-Wei Chu, H. Lee Woodcock, Xiongwu Wu and
Bernard R. Brooks
Email Address : chu@hec.utah.edu, hlwood@nih.gov, wuxw@nhlbi.nih.gov,
brb@nih.gov
Institution : Laboratory of Biophysical Chemistry, NHLBI, NIH
Date : July 1, 2005
1.3.1 NEB minimization problem fix
The SD code is modified so that it can be used with NEB. "If(QPNEB)" is
used to separate the NEB related code and the original part. NEB related parts
in epath.src and energy.src are modified to make sure the NEB method works
(the NEBF command).
Warning added when NEB+climbing image is used together in ABNR.
Files Modified:
source/energy/energy.src
source/energy/epath.src
source/minmiz/steepd.src
source/minmiz/abnr.src
Modified Documentation:
doc/minmiz.doc
doc/replica.doc
New Testcases:
test/c33test/neb_diala_abnr.inp
test/c33test/neb_diala_sd.inp
test/c33test/neb_diala_noener.inp
data/diala_replica_initial.crd
..............................................................................
1.3.2 IPS Bugfix
For partial homogenous systems (2D as PXY or 1D as PZ), the boundary
energies and forces need to be calculated, which was done by looping over
all atom pairs and is slow. A speed up was done by calculating boundary
forces every NIPSFRQ steps. However, a large NIPSFRQ introduces large errors
in the calculation results.
This fix changes the way in which boundary energies and forces are
calculated so that the calculation is very efficient and can be done every
dynamcs step without significantly slowing down a simulation. NIPSFRQ is set
to 1 by default and can be set to large values for simulations where
conformational change is slow.
Files involved:
source/fcm/nbips.fcm Updated common variables
source/energy/energy.src Modified parallelization for EXIPS, EXIPS2D,
and EXIPS3D calculation.
source/nbonds/enbond.src Changes in ENBIPS2D arguments.
source/nbonds/enbonda.src Common variable updates.
source/nbonds/enbondg.src Common variable updates.
source/nbonds/enbfast.src Common variable updates.
source/nbonds/enbips.src Change the way in which IPSSYS and IPSSYS2D
is calculated.
source/quantum/qmene.src Change to make consistant in ENBOND arguments.
..............................................................................
1.3.3 QM/MM and Pathway Improvements
RIMP2 support to the Q-Chem/CHARMM interface is added.
The discription of Off-path simulation work term is changed.
The endpoints are now treated in a more accurate manner.
Files Modified:
source/energy/epath.src
source/gukint/gukini.src
source/fcm/gamess.fcm
------------------------------------------------------------------------------
1.4 Developments at NIC-Slovenia
Names : Milan Hodoscek
Email Address : milan@cmm.ki.si
Institution : National Institutes of Chemistry, Ljubljana, Slovenia
Date : June 30, 2004
1.4.1 IBM Pwr4 and Linux development:
Support for parallel CHARMM on IBM power4 CPU architecture using
GNU/Linux as an operating system.
Files Added:
build/UNX/Makefile_ibmlnxmp
tool/ibmlnxmp_fixlibs
Files Modified:
source/energy/rmsd.src
source/machdep/cstuff.c - one #if
source/machdep/machio.src - order changed for ##IF
doc/parallel.doc
install.com
..............................................................................
1.4.2 GNU 64 bit support added:
Support for parallel CHARMM on AMD64 architecture was
added. Two compilers are supported: PathScale and g77. Since g77
doesn't have -i8 option, user has to manually increase HEAPDM variable
in the fcm/heap.fcm file to use CHARMM for larger systems. In the case
of PathScale malloc can increase the HEAP automatically. The keyword
to add in the usual install.com command is AMD64 for g77 and PS for
PathScale.
modified files:
build/UNX/Makefile_gnu
install.com
doc/parallel.doc
..............................................................................
1.4.3 QM/MM:
Support to compile CHARMM with the original DDI (Data
Distributed Interface) library from GAMESS was added, by introducing
new pref.dat keywords DDIMPI, and DDI. This allows to use parallel
version of MP2 method to work within the QM/MM setup and other methods
which use one sided communication scheme.
Files Modified:
source/gukint/gukini.src
source/gamint/ddi.src
source/machdep/paral1.src
source/machdep/paral2.src
source/machdep/paral3.src
source/machdep/startup.src
doc/gamess.doc
test/cquantumtest/gmstst.inp
..............................................................................
1.4.4 Distributed parallel replica development.
Initial implementation of the support for complete replication
of the molecular system: each processor handels on replica. It is
designed to support a variety of methods such as replica path, replica
exchange, and other string methods, where the whole system must be
replicated and there is usually very little communication involved
during the simulation. It also support independent calculations in the
parallel environement, such as runing several simulations with the
different initial conditions. This feature is useful at some
supercomputer centers which give discounts on jobs which use more
CPUs. Currently the following is already implemented:
- open commmand
- reading coordinate files
- writing coordinate files (also PDB format)
- dynamic trajectories ( each system writes its own trajectory file )
- dynamic restart files
- replica exchange
- replica path method (pathr only)
Files Added:
source/gener/repdstr.src
source/fcm/repdstr.fcm
doc/repdstr.doc
test/c32test/repdstrio.inp
Files Modified:
source/charmm/charmm_main.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynio.src
source/energy/energy.src
source/energy/epath.src
source/fcm/number.fcm
source/flucq/fluqdyn.src
source/io/coorio.src
source/io/mainio.src
source/machdep/machio.src
source/manip/corman.src
source/quantum/qmset.src
source/util/parse.src
source/util/title.src
..............................................................................
1.4.5 New random generator fixes (##NEWRNG)
The need for better random number generator increases with
longer simulations and larger systems. The ##NEWRNG was implemented
into CHARMM a while ago, but was neglected until recently when it was
fixed for proper usage in CHARMM. The newest changes are for its
parallel behavior, ie each CPU has the same random number sequence,
unless Langevine dynamics is used. In this case (this is the only
case!) each CPU has its own random number sequence so that parallel
distributed parts of the system don't follow the same random
forces. Please, note that the use of different RNG will produce
trajectories and energies which cannot be compared.
Files Modified:
source/charmm/iniall.src
source/correl/solana.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynio.src
source/dynamc/dynlng.src
source/dynamc/dynutil.src
source/manip/intcor.src
source/mc/mc.src
source/misc/genetic.src
source/misc/pbeq.src
source/util/clcg.src
..............................................................................
1.4.6 Spatially distributed parallel method in CHARMM
The biggest bottleneck in current parallel CHARMM computations
is the communication between the processors, because it is global. In
order to reduce the amount of communication we started to develop a
new method which divides the system by space which reduces the
communication by making it local to each region. The method under
development is based on BYCC nonbonded routine. Currently we finished
the first phase of the method, ie one can perform the dynamics
simulations and store the trajectories.
Files Added:
source/fcm/spacdec.fcm
Files Modified:
source/charmm/iniall.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynamvv2.src
source/dynamc/dynamvv.src
source/dynamc/dynlng.src
source/dynamc/mts.src
source/dynamc/prssre.src
source/energy/ecnstr.src
source/energy/ediff.src
source/energy/eintern.src
source/energy/enefscal.src
source/energy/energy.src
source/energy/epath.src
source/energy/printe.src ! GRMS problem not solved yet!!!??
source/fcm/actclus.fcm ! Bob ??
source/fcm/dimens.fcm
source/fcm/parallel.fcm
source/image/upimag.src
source/machdep/paral1.src
source/manip/holonom.src
source/manip/shake.src
source/misc/aspenermb.src
source/misc/aspener.src
source/nbonds/ewaldf.src
source/nbonds/exelec.src
source/nbonds/heurist.src
source/nbonds/nbndcc.src ! Bob
source/nbonds/nbonda.src
source/nbonds/nbondg.src
source/nbonds/nbonds.src ! Bob
source/nbonds/nbutil.src ! Bob
source/pert/epert.src
source/util/util.src
------------------------------------------------------------------------------
1.5 Restricted Perturbation - TMD Method
Name : Arjan van der Vaart
Email Address : vaart@fas.harvard.edu
Institution : Harvard University
Date : June 30, 2005
The restricted perturbation - targeted molecular dynamics (RP-TMD)
method can be used for the simulation of conformational transitions.
The method yields lower free energy pathways than the standard TMD
method, and is more efficient.
The code is seamlessly integrated with the existing TMD code,
and a new documents (tmd.doc) has been added. The method has been
parallelized; this required the introduction of a new MPI wrapper "GCOMBMAX"
in paral1.src. This wrapper will work with gnu based machines; but may fail
for other machines (SUN, CRAY, etc).
A new parameter (tenm6=0.000001) has been added to "fcm/number.fcm";
this required the deletion of the "tenm6" variable in "quantum/qmset.src"
and "flucq/fluqdyn.src". The implementation includes a small bugfix for
the TMD routines (the selection arrays were swapped; a call in "mcio.src"
had too many arguments).
See the reference J.Chem.Phys. 122, 114903 (2005)
[Note] As stated above, the new MPI wrapper "GCOMBMAX" in paral1.src
works for gnu based machines, but may fail for others (SUN, CRAY, etc.).
Files Modified:
source/dynamc/dynamc.src
source/dynamc/tmd.src
source/fcm/number.fcm
source/fcm/tmd.fcm
source/flucq/fluqdyn.src
source/machdep/paral1.src
source/mc/mcio.src
source/quantum/qmset.src
Modified Documentation:
doc/tmd.doc
------------------------------------------------------------------------------
1.6 Developments at Cornell University
Developers : Edward Harder, Hyung-June Woo, Jose D. Faraldo-Gomez,
Yuqing Deng, Benoit Roux
Email Address : edh2004@med.cornell.edu, woo@chem.unr.edu,
jdf2002@med.cornell.edu, yud2001@med.cornell.edu,
benoit.roux@med.cornell.edu
Institution : Weill Cornell Medical College
Date : June 25, 2005
1.6.1 The Drude mass dependence bugfix
This fixes bugs for the Drude mass dependance when using SCF dynamics
and another bug for restarting of VV2 dynamics in the NVE ensemble.
When simulating drude models using the SCF evaluation of the drude particle
positions (drude mass independant) the trajectories were dependent on the
drude mass. Also, VV2 dynamics using no thermostat and no barostat couldn't
be restarted. Subroutine READYN would complain about an "EOF during read".
Files involved:
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
..............................................................................
1.6.2 Dipole time series bugfix
Dipole time series produced is incorrect for use in the calculation
of the dielectric constant. If MAXA is not set in the input file the default
maximum atom selection is arbitrarily set to 100 which is often insufficient
for calculating bulk dipole moment time series. To fix this the default value
has been set to the number of atoms in the system "NATOM".
Files involved:
source/correl/correl.src
..............................................................................
1.6.3 Langevin Dynamics into VV2 dynamcis integrator
The addition of Langevin Dynamics to VV2 allows for a stochastic
temperature control of the nuclear degrees of freedom in contrast to the
already implemented Nose thermostat. In TPCONTROL the Langevin Dynamics
is run by placing the LANG flag through the Nose thermostat call.
The Nose thermostat is effectively turned off by setting a very large
TAU value. The friction coefficient is declared on the TPCONTROL line
with the variable NHGAM.
Note also that barostat scaling also scales the Drude dipoles
as the default setting (tpcontrol.src)
Files Modified:
source/dynamc/dynamvv2.src
source/dynamc/tpcontrol.src
source/fcm/psf.fcm
source/fcm/nose.fcm
..............................................................................
1.6.4 Implementation of GSBP cavity potential
Impelemented a cavity potential for GSBP and cleaned up of the GCMC/GSBP
code. Bug fix for the PBEQ calls in GSBP generalized multipole calculation.
Bug fix for the reentrant surface computation in GSBP.
Files Modified:
source/fcm/pbeq.fcm
source/gener/psfsum.src
source/mc/mvgcmc.src
source/misc/gsbp.src
source/misc/pbeq.src
Modified Documentation:
doc/pbeq.doc
New Testcase:
test/c32test/gcmc_test.inp
..............................................................................
1.6.5 WCA separation in atom based and fast energy routines
The LJ potential is spit into repulsive and attractive parts
in EVDW and ENBFS8 for the van der Waals distance switch cutoff.
Three addition arguments are added to EVDW and ENBFS8.
Fixed the WCA scalar initialization when the psf is read from a file.
Fixed the WCA and RSCLF maping when atoms are deleted from the psf.
Fixed a memory leak for SSBP free energy long range correction.
Files Modified:
source/cheq/fqener.src
source/gener/modpsf.src
source/io/psfres.src
source/nbonds/enbfast.src
source/nbonds/enbond.src
source/nbonds/enbonda.src
source/misc/hbuild.src
source/misc/genetic.src
source/pert/pert.src
Modified Documentation:
doc/pert.doc
------------------------------------------------------------------------------
1.7 Karolinska Institute Fixes and Enhancements
Developer : Lennart Nilsson
Email Address : Lennart.Nilsson@biosci.ki.se
Institution : Karolinska Institute
Date : June 30, 2005
1) Modification. Allow a subset of a timeseries to be copied in CORREL
MANTIM COPY.
Files Modified:
source/correl/mantim.src
doc/correl.doc
2) Modification. Allow PBC to be handled in COOR HBONd|CONT. Together with
G. Lamoureux. Improved (simpler) setup of non-bond exclusion list.
Files Modified:
source/manip/hbanal.src
doc/corman.doc
Files Added:
test/c32test/hba2.inp
test/gnulni/hba2.out
3) Modification. Allow accumulation of statistics (average + sdev) for
relaxation parameters computed from different trajectories (or different
parts of a trajectory) in the NMR module. Also added code to use first
specified frame of trajectory as reference coord for NMR DYNA ORIE NOCOMP
if no comparison coord are present.
Files Modified:
source/misc/nmr.src
doc/nmr.doc
4) Bugfix: Calling sequence to ALLHP wrong in emapfield.src
File modified:
source/emap/emapfield.src
5) Bugifx: Protect parllel comm call from being executed when not needed.
ECONT array only needs to be distributed when QECONT is true.
Files Modified:
source/energy/energy.src
6) Bugfix: Coding error for the new RECT crystal type.
Files Modified:
source/image/crystal.src)
7) Bugfix. Array dimensions were wrong, and part of the logic flawed for the
check that the main axis did not swap direction 180 degrees, in the code to
return all three axes for the inertial moment timeseries
(ENTER ... INER ALL). Some new code added to check for axis swap for the
other two axes as well.
Files Modified:
source/correl/anacor.src
doc/correl.doc
8) Minor modifications.
Files Modified:
install.com Updated options for SCALI; set chmsocket for cc
as well as gcc when needed
source/nbonds/ewwnb.src Output format change in informational message
doc/io.doc Clarification of WRITE/PRINT NBOND and
keyword FIRST for WRITE COOR PDB FIRST
doc/nmr.doc Clarification of the difference between total and
other rates.
test/c32test/dscfvib.inp Corrected erroneous use of ?vibran .ne. 1
test/c32test/entropy.inp (should be ?novibran .eq. 1)
------------------------------------------------------------------------------
1.8 Michigan State University Developments
Developer : Michael Feig, Seiichiro Tanizaki
Email Address : feig@msu.edu
Institution : Michigan State University
Date : June 29, 2005
1.8.1 GBMV improvements
1) Implementation of heterogeneous dielectric generalized Born model
within the GBMV module
(see paper: Tanizaki & Feig, J. Chem. Phys. (2005) 122, 124706)
2) Improved GBMV performance (with FAST option) through code
optimization and modification.
3) Options to change radial integration grid in GBMV
Files Modified:
source/energy/gbmv.src
source/fcm/gbmv.fcm
Modified Documentation:
doc/gbmv.doc
Modified Testcase:
test/c32test/hdgb_geneps.inp
test/c32test/hdgb.inp
test/data/hdgb_eps.dat
..............................................................................
1.8.2 Trajectory rewind control
Minor rewrite of some input/analysis routines to avoid file rewinds
if REWTrue is used where possible. This option is intended primarily for
feeding trajectory data through named pipes in order to facilitate the
extension of the MMTSB Tool Set to support trajectory analysis functions.
[NOTE] Preprocessor key NOREWIND is not included in install.com.
This option should be turned on only if needed.
Files Modified:
source/charmm/iniall.src
source/charmm/miscom.src
source/correl/solana.src
source/dynamc/dynio.src
source/fcm/stream.fcm
source/gener/hbonds.src
source/io/coorio.src
source/io/mainio.src
source/io/parmio.src
source/io/psfres.src
source/io/rtfio.src
source/machdep/machio.src
source/util/datstr.src
source/util/title.src
source/vibran/rbquas.src
source/vibran/redbas.src
Modified Documentation:
doc/miscom.doc
------------------------------------------------------------------------------
1.9 Relaxation Parameters Calculation for Axially Symmetric Top Molecules
Developers : Matthias Buck
Email Address : Matthias.Buck@case.edu
Institution : Case Western Reserve University Medical School
Date : June 15, 2005
Previously relaxation parameters T1, T2 and NOE were calculated
separately for the CSA and Dipolar contributions (i.e. the NMR module
needed to be run twice with different keywords and then the relax.
contributions had to be added to get the total). This is now combined,
but the individual contributions are still given.
anisotropy is now implemented for an axially symmetric molecule, i.e.
Dy ~ Dz for the principle axes of interia or diffusion tensor, so that
Dparallel and Dperpendicular and a single angle to the long axis can be used.
Files Modified:
source/misc/nmr.src
Modified Documentation:
doc/nmr.doc
New Testcase:
test/c33test/nmrtest2.inp
------------------------------------------------------------------------------
1.10 University of Chicago Developments
Developers : Aaron R. Dinner
Email Address : dinner@uchicago.edu
Institution : The University of Chicago
Date : June 30, 2004
1.10.1 Enhancements to the MC and DYNAMC modules
1) The FEWEr keyword is applied to RROT BYHEavy MC moves to enable
straightforward linking of rotations and translations when some
heavy atoms have hydrogens and others do not.
2) The NLIMit variable for CROT MC moves was previously limited to one
and now can be as high as five.
3) Dynamics can be terminated automatically when sufficient time is spent
in a user defined basin, and the basin is returned in the substitution
parameter INBAsin (A = -1, B = 1, and neither = 0). This feature is
very useful for calculating the commitment probability for a structure.
Files Modified:
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/tps.src
source/mc/mvrtrn.src
source/mc/mvcrot.src
source/mc/movead.src
Modified Documentation:
doc/mc.doc
..............................................................................
1.10.2 Bugfixes
(1) Memory allocation error fix
Scratch arrays are allocated for charge equilbration during the
DYNAMC call but not freed. In addition to memory leak, arrays are allocated
as HEAP then used as STACK.
Files involved:
source/mc/mvhmc.src
(2) FORMAT fix
Format statement seems OK to me but causes crash. Width of logical must
be specified.
Files involved:
source/nbonds/ewwnb.src
(3) RXNCOR fix
Forces but not energies from umbrella potential term are added to totals.
Umbrella potential term is not added to total energy or printed.
Added UMBR to ETERM array and made corresponding changes in energy utilities.
Files involved:
source/energy/energy.src
source/energy/eutil.src
source/energy/printe.src
source/fcm/energy.fcm
source/fcm/rxncom.fcm
source/rxncor/rxnene.src
(4) MC fix for MVTYP GCMC and GRID/CBIAS
Error-check for no blockers is done before counting them.
Files involved:
source/mc/mvgcmc.src
(5) GCMC counter fix
GCMC counters not updated when runnning with finite rather than
periodic boundary.
Files involved:
source/mc/mcener.src
------------------------------------------------------------------------------
1.11 University of Minnesota Developments
Developers : Kwangho Nam (N), Dan T. Major (M), Jingzhi Pu (P),
Shuhua Ma, Darrin M. York, and Jiali Gao
Email Address : pu@comp.chem.umn.edu
Institution : University of Minnesota
Date : June 29, 2005
1.11.1 Semiempirical QM/MM Developments
1) Particle mesh Ewald (PME) for combined QM/MM simulations
The particle-mesh Ewald (PME) method that has been implemented in CHARMM
for point atomic charges has been extended to energy minimization and molecular
dynamics simulations with combined QM/MM potentials. The method has been
describe in J. Chem. Theor. Comput. 1995, 1, 2, and it uses a scheme that
translates the QM wave functions into atomic partial charges for image
summations. Since long range interaction is mainly dominated by monopole
interactions, the separation of real-space Coulomb potential from total
electrostatic potential on QM atom sites results in a reasonable and
cost-effective approximation to treating long range effects in combined
QM/MM potentials. Currently, user must specifically use the Group based
cutoffs for non-bonded interactions in the QM/MM-Ewald method. The method
can be conveniently extended to ab initio and DFT QM/MM potentials,
which will be made available in the near future. Preliminary studies of
ion-pair association in aqueous solution have shown that the approach can
effectively incorporate long-range electrostatic effects into QM/MM
calculations. In addition to running CPT dynamics, the current MOPAC/CHARMM
interface has been parallelized with QM/MM-Ewald method. The users can
use a total number of processors up to the same number as the QM atoms.
The documentation for use of the QM/MM-PME options is included in qmmm.doc.
For additional information, contact Kwangho Nam (nam@chem.umn.edu).
2) Implementation of the MNDO97 package for QM/MM with d-orbitals
The MNDO97 package that include d-orbitals in MNDO, AM1, PM3 types
of model Hamiltonians, developed by Walter Thiel, has been incorporated into
CHARMM. This extends the semiempirical QM/MM options for inclusion of
d-orbitals, especially on second-row heavy atoms. The current implementation
is separate from the MOPAC version of the semiempirical QM/MM module.
The MNDO97/CHARMM interface was initially made by Paul Bash, and was
subsequently extensively modified by Darrin York and his group. The
generalized hybrid orbital (GHO) method was incorporated into the MNDO97/
CHARMM interface by Jingzhi Pu, while the PME-QM/MM method was implemented
by Kwangho Nam. A separated documentation file is created, mndo97.doc,
which contains a complete list of the QM/MM options as well as key words
for the MNDO97 code. Additional information may be addressed to
Darrin York (york@chem.umn.edu and nam@chem.umn.edu).
3) Enhancement of the simple valence bond (SVB) potential for QM/MM correction
The SVB module is extended to include two reaction coordinates. The two
reaction coordinates may be chosen as independent corrections to the two
diabatic processes, or allowed to couple to yield an adiabatic correction
potential of the two-dimensional potential surface of interest. The method
is useful for studying mechanisms that can be either concerted or stepwise
such as ester and peptide hydrolysis reactions. If the reaction coordinate is
composed of 3 atoms, the valence bond coupling term may be exponential and/or
gaussian. If a reaction coordinate is defined as distance between 2 atoms,
the energy term may be chosen to be exponential or gaussian. The additional
key words are fully documented in qmmm.doc (major@chem.umn.edu).
4) Enhancement of QM/MM module to allow reading external parameters
Routines to allow reading semi-empirical parameters (MNDO,AM1,PM3) into
the CHARMM MOPAC module directly from the CHARMM command level were added.
This makes test of re-parameterized semiempirical parameters convenient.
This avoids the need to recompile each a new set of parameters is used.
The user must provide a parameter file similar to that used in the MOPAC
package. A specific routine was added for parallel CHARMM, so that the new
parameters are scattered correctly to each node (major@chem.umn.edu).
5) Bug fix for the GHO method interfaced to SCC-DFTB
Several variables are explicitly declared to avoid errors during
compilation in the GHO-SCC-DFTB code under some platforms. This are minor
changes, and if you were able to compile the code, this change shall not
affect you. If you were having problems during compilation and have fixed
the problem by yourself, it does not affect you. If you have given up after
encountering this problem, this fix will be helpful (pu@comp.chem.umn.edu).
Files Added:
/source/fcm/mndo97.fcm
/source/mndint/mndgho.src
/source/mndint/mndini.src
/source/mndint/mndo97q (directory)
/build/UNX/mndint.mk
/build/UNX/mndo97.mk_altix
/build/UNX/mndo97.mk_gnu
/build/UNX/mndo97.mk_ibmaix
/build/UNX/mndo97.mk_sgi
Files Modified:
source/charmm/charmm_main.src
source/charmm/iniall.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynamcv.src
source/dynamc/dynamvv.sr
source/energy/energy.src
source/fcm/am1parm.fcm
source/fcm/contrl.fcm
source/fcm/gamess.fcm
source/fcm/leps.fcm
source/fcm/nbndqm.fcm
source/fcm/qmlink.fcm
source/fcm/quantm.fcm
source/fcm/scfblk.fcm
source/flucq/flucq.src
source/flucq/fluqqmmm.src
source/gener/update.src
source/gimage/crystal.src
source/manip/scalar.src
source/manip/rotlsq.src
source/minmiz/tnpack.src
source/nbonds/enbaexp.src
source/nbonds/enbfast.src
source/nbonds/enbfs8p.src
source/nbonds/enbonda.src
source/nbonds/enbondg.src
source/nbonds/ewald.src
source/nbonds/ewaldf.src
source/nbonds/nbonda.src
source/nbonds/nbonds.src
source/nbonds/pme.src
source/quantum/addlnat.src
source/quantum/qmene.src
source/quantum/qmjunc.src
source/quantum/qmleps.src
source/quantum/qmnbnd.src
source/quantum/qmpac.src
source/quantum/qmset.src
source/sccdftbint/stbgho.src
source/shapes/shapedyn.src
source/util/matrix.src
source/util/array.src
install.com
New/Modified Documentation:
doc/qmmm.doc
doc/mndo97.doc (new file)
New/Modified Testcase:
test/cquantum/mndo97.inp
test/cquantum/testleps2.inp
test/data/mndo.inp
test/data/mol1_mndo97.crd
..............................................................................
1.11.2 Path-Integral simulations with QM/MM potentials.
A path-integral (PI) bisection sampling method to account for the
nuclear quantum correction to a classical molecular dynamics trajectory
has been implemented for use with combined QM/MM potentials. The
general approach is based on the so-called quantum-classical path (QCP)
correction described by Hwang et al. (Hwang, J.-K.; Chu, Z.T.;
Yadav, A.; Warshel, A. J. Phys. Chem. 1991, 95, 8445). It treats the
classical position as the centroid of the quasi-particles (beads) in
the path integral representation such that the quantum correction to
the classical potential of mean force from (classical) molecular dynamics
simulations can be made. Of course, one can always run path integral
calculations to obtain the QM PMF directly. The advantage of the former
can be less time-consuming. The current code makes use of the
bisection algorithm for free-particle sampling (Pollock, E.L.;
Ceperley, D.M. Phys. Rev. B 1984, 30, 2555), and converged results
can be easily achieved. See also, D. T. Major, J. Gao in J. Mol.
Graphics Modeling, 2005, in press (major@chem.umn.edu).
The current version is implemented only for CHARMM MOPAC, although
extension to SCCDFTB and other QM/MM modules are straightforward and
will be included next.
Files Added:
source/quantum/qub.src serial PI program
source/quantum/mpiqub.src additional routines needed for PI program
parallel version
source/quantum/qubene.src energy evaluation routines - for fast PI
source/fcm/qub.fcm
Files Modified:
source/charmm/charmm_main.src
source/quantum/qmene.src
source/quantum/qmset.src
source/quantum/qmjunc.src
source/quantum/qmpac.src
source/fcm/parallel.fcm
source/fcm/quantum.mk
New Documentation:
doc/qub.doc
New Testcase:
test/cquantumtest/qmpathint1.inp
test/cquantumtest/qmpathint2.inp
..............................................................................
1.11.3 Implementation of Polarizable Intermolecular Potential Functions (PIPF)
The Polarizable Inter- and intra-molecular Potential Functions (PIPF),
based up on a point dipole interaction model, has been implemented.
In the PIPF model, each polarizable interaction site is represented by an
atomic polarizability that interacts with all other polarizable sites
through a Thole damping procedure, and the total polarization energy is
determined from the induced dipoles. The PIPF potentials have been described
previously in several preliminary studies of organic liquids representing
the functional groups of amino acids. Relevant references include:
1. J. Gao, D. Habibollahzadeh, L. Shao, J. Phys. Chem. 1995, 99, 16460-16467.
"A Polarizable Intermolecular Potential Functions for Simulations of Liquid
Alcohols."
2. J. Gao, J. J. Pavelites, D. Habibollahzadeh, J. Phys. Chem. 1996, 100,
2689-2697. "Simulations of Liquid Amides using a Polarizable Intermolecular
Potential Functions (PIPF)."
In the current version in CHARMM, two approaches have been implemented:
(a) the induced point dipoles are fully converged using an iterative, self-
consistent procedure (although an equivalent matrix inversion scheme can also
be adopted), and (b) each induced dipole is treated as an independent degree
of freedom and evolves in time in molecular dynamics simulations by using the
extended Lagrangian scheme. For the latter algorithm, the dynamical dipoles
are coupled two a low temperature heat baths to allow them to fluctuate
at their fully converged values. The current PIPF implementation can be used
for systems with periodic boundary conditions as well as non-periodic
boundary conditions. Analytical first derivatives are available for both
methods to allow energy minimization and/or molecular dynamics simulations.
The CHARMM atom based non-bonded list is used to determine the interacting
centers in polarization calculations. Two of the damping functions suggested
by Thole are available to alleviate the problem of over polarization at
short inter-atomic distances between polarizable sites. In addition,
the 1-4 intramolecular polarizations can be optionally removed.
Some utility routines are provided to analyze the average dipole properties
for pure liquid simulations.
Files Added:
build/UNX/pipf.mk makefile for compiling PIPF under CHARMM
source/fcm/pipf.fcm common block for PIPF
source/pipf/dpfimg.src derivatives for systems with images
source/pipf/dpipf.src gradient driver for self-consistent dipoles
source/pipf/epfimg.src PIPF energy driver for systems with images
source/pipf/epipf.src energy driver for self-consistent dipoles
source/pipf/pfdyn.src parser for PIPF options and code for PIPF with dynamical
dipoles using the extended Lagrangian method
Files Modified:
install.com switch to install PIPF is added
source/charmm/charmm_main.src parse pipf keyword
source/dynamc/dcntrl.src adapt to dynamical dipole dynamics
source/dynamc/dynamc.src
source/dynamc/dynamc4.src
source/dynamc/dynamvv.src
source/dynamc/dynio.src dynamical dipoles are added to trajectory I/O
source/energy/energy.src adapt to the dynamical dipole initilization and
image operations
source/energy/eutil.src add entries for the polarization energy and
dipole kinetic energy
source/energy/printe.src print polarization energy information
source/fcm/energy.fcm include PIPF energy terms
source/image/upimag.src update dipoles for image atoms
source/io/coorio.src calls to READYN are modified to accomendate
arguments for PIPF
source/mbond/mbback.src call to WRIDYN is modified to accomendate
source/mc/mcio.src arguments for PIPF
source/nbonds/enbond.src add calls to PIPF energy calculations
New/Modified Documentation:
doc/pipf.doc new doc for PIPF
doc/energy.doc energy term to skip polarization energy
in PIPF is added
New Testcase:
test/c33test/pipf_test1.inp
test/c33test/pipf_test2.inp
test/c33test/pipf_test3.inp
test/data/pol2_125.crd
Other Comments: The PIPF potential complements the fluctuating charge and
the drude oscillator schemes already implemented in CHARMM. It would be of
interest to allow these methods to be used jointly.
..............................................................................
1.11.4 Miscellaneous changes and bug fixes
1) Monte Carlo call to MOPAC/CHARMM energy
MC energy call to the QM(MOPAC)/MM energies has been updated to account
for developments in the Gao group. An additional function call that
generates a new guess density matrix was added prior to the energy call
to aid in convergence and to eliminate hysteresis that hampers long-time
energy conservation in molecular dynamics simulation. Currently, a new density
matrix is generated prior to each MC or MD step. Additional tests may
be needed to optimize the frequency that a new density matrix is generated.
2) Bug fixes on the RXNCor module
Errors introduced as a result of changes of multi-dimension reaction
coordinate definition have been corrected. These bugs affect reading
external biasing potential functions.
3) Constraints on the distance difference in PERT
Modification of the TSM module to allow constraining the difference
between two distances. This one allow user to do geometry optimization and
dynamics with fixing distance difference as a constraints, which is
useful in many ways. The implementation uses Wilson projection method
to project out forces along the atoms that defined the two bonds.
Files Modified:
source/mc/mcener.src
source/pert/icfix.src
source/pert/icfcnf.src
source/emap/emapfield.src
source/rxncor/rxndef.src
source/rxncor/rxnene.src
------------------------------------------------------------------------------
1.12 HQBM Restraints and Ensemble Codes
Developers : Robert Best
Email Address : best@helix.nih.gov
Institution : National Institutes of Health
Date : June 29, 2005
The HQBM module has been expanded to include a number of restraints
based on experimental data. These include hydrogen exchange protection
factors and NOE's, J-couplings and order parameters from NMR.
In addition, an ensemble module has been developed. This allows
(i) ensemble averaging of the above restraints.
(ii) replica exchange simulation (preliminary support).
Stopmark 3 is new features have been added to install.com. This will set
up the build/arch directory such that CHARMM can be built by running
'make' from the build/arch directory.
Files Added:
source/fcm/ensemble.fcm
source/machdep/ensemble.src
install.nickname.com
Files Modified:
source/charmm/charmm_main.src
source/charmm/iniall.src
source/charmm/miscom.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynio.src
source/energy/energy.src
source/energy/hqbm.src
source/fcm/hqbm.fcm
source/io/aceio.src
source/io/coorio.src
source/io/parmio.src
source/io/psfres.src
source/io/rtfio.src
source/io/univio.src
source/machdep/machio.src
source/machdep/machutil.src
source/machdep/startup.src
source/misc/eef1.src
source/util/parse.src
source/util/title.src
build/UNX/dynamc.mk
build/UNX/energy.mk
build/UNX/io.mk
build/UNX/machdep.mk
install.com
New/Modified Documentation:
doc/hqbm.doc
doc/ensemble.doc
New/Modified Testcase:
test/c33test/hqbm_single_test.inp
test/c33test/hqbm_rc3_ens.inp
test/c33test/hqbm_rc4_ens.inp
test/c33test/hqbm_rc8.inp
test/c33test/ensemble_rex.inp
test/data/logp_1ypa.dat
test/data/ci2_jcons.dat
test/data/ci2_noe.dat
test/data/ci2_phi.dat
test/data/ci2_s2.dat
test/data/1ypa.psf
test/data/1ypa_minim.pdb
------------------------------------------------------------------------------
1.13 University of Maryland, School of Pharmacy Developments
Developers : Alex MacKerell, Igor Vorobyov, Victor Anisimov
Pedro Lopes
Email Address : alex@outerbanks.umaryland.edu
Institution : University of Maryland, School of Pharmacy
Date : July 14, 2005
1.13.1 Parameters for classical Drude oscillator model.
Force field parameters for the developmental polarizable model
based on a classical Drude oscillator have been added. Currently,
only parameters for the SWM4-DP water models, alkanes and ions are
available. Toppar files along with example inputs are in the
toppar/drude directory.
1.13.2 Parameters for silicate surfaces.
Force field parameters for silicate and aluminosilicate surfaces
are now availble in the toppar/silicate directory. Included are code
and example inputs to create and generate a quartz 001 surface. Note
that these parameters have been designed to be compatible with the
CHARMM22 and 27 all-atom force fields.
Documentation available in parmfile.doc
------------------------------------------------------------------------------
1.14 Minor Enhancements and Fixes
1.14.1 SPCISM update
Developer : Sergio A. Hassan
Email Address : mago@helix.nih.gov
Institution : CMM/DCB/CIT, National Institutes of Health (NIH)
Date : June 16, 2005
SCPISM doc is updated
File Involved:
doc/scpism.doc
..............................................................................
1.14.2 IMM1 Testcase Updated
Developer : Themis Lazaridis
Email Address : tlazaridis@ccny.cuny.edu
Institution : City College of New York/CUNY
Date : July 12, 2005
The testcase include checking the energy against the expected value.
File Involved:
test/c31test/testimm1.inp
..............................................................................
1.14.3 SHELL Bugfix
Developers : Tibor Rudas, Stefan Boresch
Email Address : tibi@mdy.univie.ac.at, stefan@mdy.univie.ac.at
Institution : Molecular Dynamics and Biomolecular Simulation group
at the Department of Biomolecular Structural Chemistry,
University of Vienna
Date : June 28, 2005
If solvent is decomposed into shells around a solute with shells
extending to less than 8A or to more than 20A the cube size for
the cubing based algorithm is reset. In the former case this leads
(only) to a less efficient - but correct - decomposition while the
generated shells can be incomplete in the latter case.
Files involved:
source/correl/shell.src
..............................................................................
1.14.4 SASA Segmentation Fault Error Fixed
Developers : Riccardo Pellarin and Amedeo Caflisch
Email Address : pellarin@bioc.unizh.ch, caflisch@bioc.unizh.ch
Institution : University of Zuerich, Switzerland
Date : June 21, 2005
When one tries to simulate a sequence that does not contain any TRP or
PHE or TYR, (or in any case where the nonbond explicit exclusion list is empty)
one would get a segmentation fault immediately after the command call.
As an output the following instance error message is normally obtained:
FREHP> sasini.src/SASINI releases 66 words of PXBL at **********
BAD BASE, ********* OR LEN2, 66 PASSED TO FREEHP
However we experienced, in some machines and in some particular compilations,
that program prints only "segmentation fault" or even stops without any alert.
Files involved:
source/misc/sasini.src
..............................................................................
1.14.5 PERT MD Simulation Fix
Developer : Stefan Boresch
Email Address : stefan@mdy.univie.ac.at
Institution : University of Vienna
Date : July 18, 2005
PERT MD runs die with error "GENERATING TOO MANY IM CONTACTS".
Use of PERT automatically disables the BYCB method. Thus, no
BYCB only code should be executed. However, the error
message "GENERATING TOO MANY IM CONTACTS" is triggered by
a sanity check of the BYCB method, which shouldn't be active
in PERT. The code in questions is protected by an ##IF IMCUBES
block, but not protected against being executed if BYCB isn't
used (such as is always the case in PERT runs).
Files involved:
source/nbonds/nbonds.src
The following are introduced and modified during the development period
of August 15, 2005 through February 15, 2006. The developments are
based on CHARMM version c33a1. Some enhancements and fixes to c32b1
are also incorporated into c32b2 as described in the c32b2 release note
(ChangeLog.c32).
------------------------------------------------------------------------------
2.1 TSRI Developments
Developers : Mike Crowley, Charles L. Brooks, III,
Sandeep Patel, Jianhan Chen,
Jana Khandogin, Olgun Guvench
Email Address : crowley@sripps.edu, brooks@scripps.edu
sandeep@scripps.edu, jianhanc@scripps.edu,
janakhan@scripps.edu, guvench@scripps.edu
Institution : The Scripps Research Institute
Date : December 31, 2005
2.1.1 CHARMM parallel routines
Current charmm parallel routines have extra overhead that can be avoided
by calling mpi routines directly. Many data copies can be avoided and use of
specific mpi features can be used. This is still in testing mode to come up
with either the best method or a choice between methods depending on
architecture. The reduces and gathers are the key mpi routines targeted in
dynamc.src and energy.src.
Currently only supports 64 bit architectures (ibmaixmp64, itanium, altix)
ALLMPI - turns on calling mpi routines directly instead of through charmm
wrappers
MPIINPLACE - uses the data-in-place feature of gather/reduce operations
MPIBUF - uses mpi buffers
Files Modified:
source/dynamc/dynamc.src
source/energy/energy.src
..............................................................................
2.1.2 CHEQ additions for mixed polarizable/nonpolarizable systems
Modifications to allow running molecular dynamics / minimzations of
systems where one segment is nonpolarizable and the other polarizable (within
the CHarge EQuilibration formalism, CHEQ) are added. Initial applications of
this code involved phospholipid bilayer systems where the solvent is treated
as the polarizable TIP4P-FQ water model, and the lipid bilayer segment as the
nonpolarizable CHARMM param27 model.
Files Modified:
source/dynamc/dynamc.src
source/nbonds/ewald.src
New Testcase:
test/c33test/nawat.inp
..............................................................................
2.1.3 Optimized CMAP and atomic input radii for GBSW
The backbone phi/psi cross-term (CMAP) and the atomic input radii have
been re-optimized specifically for GBSW implicit solvent, to balance the
solvation and intramolecular interactions and to reproduce experimental
conformational equilibra of both helical peptides and beta-hairpins. The
re-optimized GBSW force field also successfully folds both trpzip2 and
trp-cage, indicating that it is quite robust and transferable.
Files Added:
toppar/gbsw/par_all22_prot_gbsw.inp
toppar/gbsw/radius_gbsw.str
Modified Documentation:
doc/gbsw.doc
New/Modified Testcase:
test/c33test/gbsw_ff05.inp
test/c33test/gbsw_images.inp
test/data/top_all22_prot_cmap.inp
test/data/par_all22_prot_gbsw.inp
test/data/radius_gbsw.str
..............................................................................
2.1.4 Constant pH molecular dynamics
PHMD is a lambda-dynamics based extended Hamiltonian method that allows
implicit solvent molecular dynamics to be performed simultaneously with the
titration of ionizable residues under specified pH condition.
[Note] New pref key PHMD, dependent on GBSW
Files Added:
source/energy/phmd.src
source/fcm/phmd.fcm
doc/phmd.doc
Files Modified:
source/charmm/charmm_main.src
source/charmm/iniall.src
source/energy/energy.src
source/energy/eutil.src
source/energy/printe.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/dynamc/dynamc4.src
source/dynamc/dynamcv.src
source/dynamc/dynio.src
source/fcm/energy.fcm
New Testcase:
test/c33test/phmd/inp
test/data/top_all22_prot_cmap_phmd.inp
test/data/par_all22_prot_chen_cmap_phmd.inp
test/data/radius_gbsw.str
test/data/asp_h.psf
test/data/asp_h.pdb
test/data/phmd_asp.in
..............................................................................
2.1.5 RUSH: all atom short-range force field
RUSH is a simple implicit-solvent force-field that adds terms to the
bonded portion (bond + angle + dihe + impr + urey) of the all-atom CHARMM22
force field to account for volume-exclusion (_R_epulsion), the hydrophobic
effect (_U_nburied _S_urface), and intra-molecular and protein-solvent
hydrogen-bonding (_H_ydrogen-bonding) (hence _R_ _U_ _S_ _H_).
This is a simplified all atom force field based on implicitly short-raged
interactions and a rapid surface burial calculation. The parameters are
optimized to fold small proteins like trpzip and trpcage.
[Note] New pref key RUSH, no dependencies
Files Added:
source/energy/rush.src
source/fcm/rush.fcm
toppar/rush/par_rush_058.inp
toppar/rush/par_rush_058_a.cmap
toppar/rush/top_rush_058.inp
doc/rush.doc
Files Modified:
install.com
source/charmm/charmm_main.src
source/charmm/iniall.src
source/energy/energy.src
source/energy/eutil.src
source/energy/printe.src
source/fcm/energy.fcm
New Testcase:
test/c33test/rush.inp
test/data/top_rush_058.inp
test/data/par_rush_058.inp
test/data/par_rush_058_a.cmap
..............................................................................
2.1.6 Added needed ##IF X pref control for RPATH, AFM and HQBM
Corrected incorrectly coded modules for RPATH, NEB, AFM and HQBM to allow
pref keyword control of inclusion/exclusion of code.
[Note] New pref keywords introduced: AFM, HQBM, RPATH added as default in
install.com.
Files Modified:
source/charmm/charmm_main.src
source/charmm/iniall.src
source/dynamc/dynio.src
source/energy/afm.src
source/energy/ecnstr.src
source/energy/energy.src
source/energy/epath.src
source/energy/eutil.src
source/energy/hqbm.src
source/energy/printe.src
source/fcm/afm.fcm
source/fcm/energy.fcm
source/fcm/hqbm.fcm
source/fcm/neb.fcm
source/fcm/path.fcm
source/fcm/pderiv.fcm
source/gukint/gukini.src
source/machdep/paral1.src
source/mbond/mbdyn.src
source/mbond/mbmodes.src
source/mc/mcener.src
source/minmiz/abner.src
source/minmiz/egrad1.src
source/minmiz/steepd.src
source/nbonds/nbndcc.src
source/nbonds/nbonds.src
source/pert/epert.src
..............................................................................
2.1.7 Bugfixes
- New install.com targets itanium and em64t
- new ifort keyword for gnu
- improved ibm makefiles
- Lots of timer fixes to solve overcounting problem
- Timer debugging code added new keyword DEBUG_TIMER
- small binary io enforcing of 32 bit ints
- sgi SGIF90 modifications (very small)
- some bandaid fixes for memory hog routines, just a start with a big problem
- Increase node limit for parallel pme
- small mods for size HUGE
Files involved:
install.com - New targets for itanium and em64t to relieve some of the
gnu target confusion
- getting the tpp7/tpp2 compile optimization right for ifort
- ifort keyword makes ifc/efc keywords vestigial
- mpif90 keyword added to enable using mpif90 instead
of having to find the include files and libs for mpi.
build/UNX/Makefile_itanium
build/UNX/Makefile_em64t
build/UNX/Makefile_ibmaix
build/UNX/Makefile_ibmaixmp
- uses xlf95_r and mpxlf95_r with better arch support
build/UNX/Makefile_gnu
- ifort added, tpp7/2 selection
source/cff/parrdr_cff.src - 32 bit integer binary io enforced
source/correl/solana.src - SGIF90 file opening fixed
source/dynamc/cveloci.src - memory management bandaid
source/dynamc/dynamc.src - timer fixes
source/dynamc/dynio.src - 64 bit traj reading enabled to convert to 32 bit
source/dynamc/mts.src - memory hog bandaid
source/energy/energy.src - timer fixes
source/energy/eutil.src - timer fix
source/fcm/cveloci.fcm - memory management
source/new_time.fcm - cosmetic
source/pme_par.fcm - increase parallel pme node limit
source/stack.fcm - size HUGE additions
source/image/upimage - message for exceeding max sizes
source/io/psfres.src - 32/64 bit int binary io fix
source/machdep/machutil.src - itanium title, ia64_mpi code, mpif90 includes
source/machdep/paral1.src - mpif90, IA64_MPI for 64 bit idiosyncracies
source/machdep/paral2.src - same
source/machdep/paral3.src - same
source/machdep/startup.src - same
source/manip/shake.src - same
source/misc/surfac.src - size HUGE added
source/nbonds/enbond.src - timer fix
source/nbonds/ewald.src - cosmetic
source/nbonds/nbutil.src - memory management
source/nbonds/pmeutil.src - mpif90 include, node limit for parallel increased
source/quantum/mpiqub.src - END statement fix
source/quantum/qub.src - time routine workaround
source/util/new_timer.src - DEBUG_TIMER code added
source/util/title.src - ##IF tags, purely cosmetic and optional
..............................................................................
2.1.8 Testcase modified
The following testcases are modified to make them run easily and compared
well with benchmarks.
Files Modified:
test/compare.awk
test/c20test/h2otst.inp
test/c20test/maatest.inp
test/c20test/vibpafl.inp
test/c22test/block2.inp
test/c22test/fshake1.inp
test/c22test/pert.inp
test/c22test/umbrella.inp
test/c23test/nmrtest1.inp
test/c24test/dimb1.inp
test/c24test/pert2.inp
test/c26test/pathint.inp
test/c27test/adumb-enum.inp
test/c27test/adumb-phichi.inp
test/c27test/mctest02.inp
test/c27test/testeef1.inp
test/c28test/hmcm.inp
test/c29test/eef1_2ndderiv.inp
test/c30test/cftiangl.inp
test/c30test/cftidihe.inp
test/c30test/cftidist.inp
test/c30test/cftmala10.inp
test/c30test/cftmtst1.inp
test/c30test/cftmtst2.inp
test/c30test/cftmtst3.inp
test/c30test/cftmtst4.inp
test/c30test/cftmtst5.inp
test/c30test/entropylj.inp
test/c31test/afm.inp
test/c31test/nebdiala.inp
test/c31test/off_path_simulation.inp
test/c31test/scpismtest.inp
test/c31test/testimm1.inp
test/c32test/dmolvib.inp
test/c33test/ensemble_rex.inp
test/c33test/hqbm_rc3_ens.inp
test/c33test/hqbm_rc4_ens.inp
test/c33test/hqbm_rc8.inp
test/c33test/hqbm_single.inp
test/c33test/neb_abnr.inp
test/c33test/neb_noener.inp
test/c33test/neb_sd.inp
test/c33test/nmrtest2.inp
test/c33test/repdstrio.inp
------------------------------------------------------------------------------
2.2 Developments at the National Institute of Health
Developers : Jhih-Wei Chu, H. Lee Woodcock, Xiongwu Wu and
Bernard R. Brooks
Email Address : chu@hec.utah.edu, hlwood@nih.gov, wuxw@nhlbi.nih.gov,
brb@nih.gov
Institution : Laboratory of Biophysical Chemistry, NHLBI, NIH
Date : December 20, 2005
2.2.1 EMAP IO and command fixes
The CCP4 map head data format is corrected. Map operation bugs are fixed
Files involved:
source/emap/emapio.c - EMAP output format corrected to the CCP4 standard.
source/emap/emapop.src - Command line parsing and trajectory io bugs fixed.
Operations, DIFF, SCORE, are added.
source/emap/emapsubs.src - Command line parsing, trajectory io, and correlation
calculation bugs are fixed.
Subroutines for DIFF, SCORE, are added.
source/emap/emapdock.src - Unnecessary lines are cleaned.
source/fcm/emap.fcm - Variables for SCORe calculation are added.
Benchmark files:
test/output/emaptest.out
test/output/dockemap.out
..............................................................................
2.2.2 Bug fix for guiding temperature calculation
Guiding temperature calcuation is corrected based on friction costant,
local average time, and local average of guiding force products.
Files involved:
source/dynamc/dcntrl.src - TEMPSG input and initialtion
source/dynamc/dynamc.src - Arguments in subroutines SGMDW and SGLDW.
source/dynamc/sglds.src - Changes are made for TEMPSG calculation.
source/fcm/sgld.fcm - Changes are made for TEMPSG calculation.
test/c31test/sgld.inp - Modified to check the current SGLD functions
doc/sgld.doc - Added documentation for setting TEMPSG, ISGSTA, ISGEND, TSGAVP
Benchmark files:
test/output/sgld.out
..............................................................................
2.2.3 Q-Chem QM/MM parallel/parallel Replica Path
Add a fix to get the node names and assign spawned QM jobs based on Milan
Hodoscek's PARHOST additions.
Files Modified:
source/gukint/gukini.src
..............................................................................
2.2.4 Bugfixes for energy computation when using "cons fix"
The problem was that when using 'cons fix' some atoms non-bonded
interactions (with other fixed atoms) were still computed even when they
are fixed.
Files Modified:
source/image/nbondm.src
source/nbonds/nbonda.src
..............................................................................
2.2.5 NEB minimization improvements
Added steepd NEB fixes (from Paul Sherwood and Jhih-Wei Chu). These were
added by creating a separate steepd routine, only called when doing NEB.
Adjusted ABNR printing when running NEB (now printing off-path gradient too).
Files Modified:
source/minmiz/abner.src
source/minmiz/minmiz.src
source/minmiz/steepd.src
..............................................................................
2.2.6 Minimization improvements
Saddle point / transition state optimization option added to the Newton-
Raphson (NRAP) minimizer. This works by following a positive eigenvector
rather than following a negative one, which is standard method for finding
local minima. The code searches in the opposite direction of the minimzation
(i.e. uphill) until the nsadd=X condition is satisfied. This code is controled
by setting "NSADD" to 1,2,... to find a first(transition state), second,...
order saddle point.
New Testcase:
test/c33test/ts_opt_cyclohexane.inp
------------------------------------------------------------------------------
2.3 Developments at NIC-Slovenia
Names : Milan Hodoscek
Email Address : milan@cmm.ki.si
Institution : National Institutes of Chemistry, Ljubljana, Slovenia
Date : November, 2005
2.3.1 New PARALLEL INFO command implemented
Support for PARAllel INFO command was added. It lists the hostnames of
the machines assigned to parallel processes. It also fills the PARHOST(MAXNODE)
array in COMMON/INTLC/ which is usefull in preparing parallel/parallel QM/MM
jobs for programs that are not directly linked into CHARMM, ie Q-Chem.
Files modified:
source/charmm/iniall.src
source/fcm/parallel.fcm
source/machdep/cstuff.c
source/machdep/paral1.src
source/machdep/startup.src
doc/parallel.doc
..............................................................................
2.3.2 GAMESS/CHARMM interface
New version of GAMESS 27 JUN 2005 R3 is supported with the CHARMM program.
New features include support for larger number of QM atoms, basis sets up to I
orbitals, etc.
Files modified:
source/fcm/gamess.fcm
source/gamint/blur.src
source/gamint/ddi.src
source/gukint/gukini.src
source/machdep/cstuff.c
source/machdep/paral1.src
doc/gamess.doc
..............................................................................
2.3.3 Support for the new fortran compiler from GCC (Gnu Compiler Collection)
Support added for gfortran, the replacement for g77 compiler. This is
different than g95 development. Gfortran comes with the standard gcc compilers,
as was the case for g77, which is not maintaned anymore and is also not
available in newer versions of gcc. Due to some problems the GFORTRAN pref.dat
keyword was introduced and is used in 3 files(psfsum,startup,qub).
Except nmrtest2 and hba2 all the testcases pass with this compiler.
This compiler supports INTEGER*8 with the compiler flag, so it can be used
also on x86_64 type machines.
Files modified:
install.com
build/UNX/Makefile_gnu
source/dynamc/dcntrl.src
source/gener/psfsum.src
source/machdep/startup.src
source/quantum/qub.src
..............................................................................
2.3.4 Fixes for REPDSTR
Added the authors into the replica distributed development.
Files modified:
source/gener/repdstr.src
doc/repdstr.doc
test/c33test/repdstrio.inp
..............................................................................
2.3.5 64 bit compile problems
Some changes are needed to compile CHARMM on 64 bit machines.
Files modified:
source/io/parmio.src
source/machdep/paral1.src
source/util/util.src
..............................................................................
2.3.6 More fixes to support non-standard pref.dat keywords
Files modified:
source/misc/gsbp.src
source/util/util.src
source/quantum/mpiqub.src
------------------------------------------------------------------------------
2.4 Karolinska Institute Fixes and Enhancements
Developer : Lennart Nilsson
Email Address : Lennart.Nilsson@biosci.ki.se
Institution : Karolinska Institute
Date : December 30, 2005
2.4.1 Protein secondary structure analysis
The Kabsch&Sander secondary structure (hbond-based) definition has been
coded and can be accessed as COOR SECStructure. Code is setup to allow it to
be added to correl as well, if the need arises.
Files added:
source/manip/secstr.src
test/data/3adk-a.crd
test/c33test/secstr.inp
test/output/secstr.out
Files modified:
source/charmm/iniall.src
source/manip/corman.src
source/manip/hbanal.src
source/fcm/hbanal.fcm
doc/corman.doc
doc/subst.doc
..............................................................................
2.4.2 Dimension fixes
The default number of residues and donors/acceptors in dimens.fcm is
on the small side for big versions.
MAXRES changed to MAXA/3
MAXPAD changed to SIZE
Files involved:
source/fcm/dimens.fcm
..............................................................................
2.4.3 64-bit space allocation
From Bogdan's report.
Files involved:
source/machdep/space.src
..............................................................................
2.4.4 Reduced stacksize for big versions.
Memory footprint is an issue for big version (at least on 32-bit systems);
on GNU/linux systems a dynamic link only leaves the address space below 1GB
for static data. Static linking does not have this limitation (dynamic
libraries are loaded at 1GB), but it is not always possible to link
static. Eg SCALI MPI Connect only supports dynamic linking.
Files involved:
source/fcm/stack.fcm
..............................................................................
2.4.5 Big_endian
Modification of Makefile_gnu to use the native binary format as default,
but to allow the user to force big_endian or little_endian from the install.com
command lined for compilers that allow this (eg ifort or pgi).
Files involved:
build/UNX/Makefile_gnu
..............................................................................
2.4.6 Water-water lookup routine
For large systems the coding was not optimal, and has been re-written.
Now also uses a smaller group-based non-bond list. For this to work with
newer list-generators these also have to be made at least "group-aware".
Also fixed a previously reported bug related to this routine in enbaexp.src.
This version does a gather/scatter for coordinates and forces, which however
does not really improve performance. This may be cachesize dependent.
A non-gathering version of the file is available upon request.
Changed the single precision value of RSMALL (to 0.0001) in number.fcm since
some comparisons made in single precision fail if 10**-10 is used.
Files involved:
source/nbonds/ewwnb.src
source/nbonds/nbonds.src
source/image/upimag.src
source/nbonds/enbaexp.src
source/fcm/number.fcm
test/c32test/nbsolv.inp
..............................................................................
2.4.7 Bugfixes
1) The DONOr/ACCEptor commands did not work correctly in ADD mode.
Also clarified documentation.
Files involved:
source/gener/hbonds.src
doc/hbonds.doc
2) The nmr module was non-functional after the last release. I have put
back my version (w/o M. Buck's modifications). Als clarified the documentation.
Files involved:
source/misc/nmr.src
doc/nmr.doc
3) The code to correct inadvertent reordering of the minor inertial axes in
CORREL timeseries INERtia ALL has been corrected.
Files involved:
source/correl/anacor.src
4) Fix the workings of selections and orientating structures in CORREL
timeserise RMS
Files involved:
doc/correl.doc
5) Fast SHAKE bugfix (charmm.org)
Files involved:
source/manip/fsshake.src
6) PERT SSBP long range correction bugfix (charmm.org)
Files involved:
source/pert/pert.src
7) MPIQUB preflx directives. Problems when not compiled with MPI
Files involved:
source/quantum/mpiqub.src
8) Document IF-THEN-ELSE to miscom.doc
Files involved:
doc/miscom.doc
------------------------------------------------------------------------------
2.5 University of Minnesota Developments
Developers : Kwangho Nam, Dan T. Major, Darrin M. York, and Jiali Gao
Email Address : gao@chem.umn.edu; nam@chem.umn.edu; major@chem.umn.edu
Email Address : pu@comp.chem.umn.edu
Institution : University of Minnesota
Date : December 29, 2005
2.5.1 Semiempirical QM/MM updates
1) Non-bond list updating for QM-MM interaction
The non-bond list generation routines of the three semiempirical modules
(QUANTUM, MNDO97, and SQUANTM) have been merged into one routine. Thus,
it shares the same procedure to update non-bond list for QM-MM interactions.
This was done separately for the QUANTUM module before based on group-group
cutoff scheme, and it was updated at the same time when MM non-bond list is
generated.
2) Image handling for QM-MM interaction
The image atoms for QM-MM interaction pairs are updated separately
from IMAGe update and IMAGE list. At the time of QM-MM non-bond list
generation, the new routine runs over all the image transformations and
find the minimum distance, and we keep the image transformation information
for the QM-MM interaction evaluations. This is swapped with primary
cell atoms at the beginning of QM/MM energy evaluation.
Files added:
source/mndint/mndene.src
source/mndint/mndewald.src
source/quantum/qmevdw.src
Files modified:
source/fcm/mndo97.fcm
source/fcm/nbndqm.fcm
source/fcm/quantm.fcm
source/fcm/qubpi.fcm
source/energy/energy.src
source/energy/hqbm.src
source/gener/update.src
source/mndint/mndini.src
source/mndint/mndgho.src
source/mndint/nbonds.src
source/mndint/mndo97q/97/FOCK.f
source/mndint/mndo97q/97/FOCKX.f
source/mndint/mndo97q/97/QMMDER.f
source/mndint/mndo97q/97/KEYCHK.f
source/mndint/mndo97q/97/KEYGET.f
source/quantum/qmene.src
source/quantum/qmjunc.src
source/quantum/qmnbnd.src
source/quantum/qmpac.src
source/quantum/qmset.src
source/quantum/qub.src
source/quantum/qubene.src
source/quantum/mpiqub.src
build/UNX/mndint.mk
build/UNX/quantum.mk
doc/mndo97.doc
New Testcase:
test/data/mndo_wat.inp
test/cquantumtest/qmmm_ewald.inp
test/cquantumtest/mndo97_ewald.inp
..............................................................................
2.5.2 New semiempirical QM/MM code: SQUANTUM package
Developers : Kwangho Nam (UMN), Ross Walker (TSRI), Mike Crowley (TSRI),
Charles L. Brooks, III (TSRI), David A. Case (TSRI),
Darrin M. York (UMN), and Jiali Gao (UMN)
Email Address : gao@chem.umn.edu; york@chem.umn.edu; brooks@scripps.edu;
nam@chem.umn.edu; rcw@sripps.edu; crowley@sripps.edu
Institution : University of Minnesota (UMN) and
The Scripps Research Institute (TSRI)
Date : December 29, 2005
A semiempirical module written in fortran 90 that includes MNDO, AM1,
PM3, PDDG/PM3, and PDDG/MNDO types of model Hamiltonians, initially
written for the AMBER package by Ross Walker and Mike Crowley, has
been incorporated into CHARMM. This implementation speeds up the
computations employing QM/MM potential considerably. The current
implementation is separate from the MOPAC version of the semiemprical
QM/MM module. In addition, the generalized hybrid orbital (GHO)
method and Switching function for QM-MM interaction are incorporated
into the QM/MM interface by Kwangho Nam, while the QM/MM-Ewald method
was implemented by the collaborated work by Ross Walker, Mike Crowley,
and Kwangho Nam.
Files Added:
build/UNX/squantm.mk_altix
build/UNX/squantm.mk_gnu
build/UNX/squantm.mk_ibmaix
build/UNX/squantm.mk_ibmaix64
source/fcm/squantm.fcm
source/squantm/assert.h
source/squantm/dprec.h
source/squantm/constants.h
source/squantm/conversions.h
source/squantm/qm2_array_locations.h
source/squantm/qm2_elements.h
source/squantm/qm2_parameters.h
source/squantm/sqnt_setup.src
source/squantm/sqnt_nbnd.src
source/squantm/sqnt_ene.src
source/squantm/sqnt_qmmm_module.src90
source/squantm/sqnt_qmmm_util.src90
source/squantm/sqnt_qmmm_int.src90
source/squantm/sqnt_qm2_energy.src90
source/squantm/sqnt_qm2_grad.src90
source/squantm/sqnt_qm2_mopac.src90
source/squantm/sqnt_qm2_ewald.src90
source/squantm/sqnt_qm2_gho.src90
Files Modified:
install.com
build/UNX/Makefile_ibmaix
build/UNX/Makefile_ibmaix64
build/UNX/Makefile_ibmaixmp
build/UNX/Makefile_ibmaixmp64
build/UNX/Makefile_ibmlnxmp
build/UNX/Makefile_ibmlnxmp_so
source/fcm/nbndqm.fcm
source/fcm/gamess.fcm
source/charmm/charmm_main.src
source/charmm/iniall.src
source/energy/energy.src
source/flucq/flucq.src
source/flucq/fluqqmmm.src
source/gener/update.src
source/minmiz/tnpack.src
source/nbonds/enbaexp.src
source/nbonds/enbfast.src
source/nbonds/enbfs8p.src
source/nbonds/enbonda.src
source/nbonds/enbondg.src
source/nbonds/ewaldf.src
source/nbonds/ewald.src
source/nbonds/nbonda.src
source/nbonds/nbonds.src
source/nbonds/pme.src
source/quantum/addlnat.src
source/quantum/qmevdw.src (New file on JG1_qmmm.form)
source/quantum/qmnbnd.src
doc/qmmm.doc
New Testcase:
test/cquantumtest/but0_squantm.inp
test/cquantumtest/but1_squantm.inp
test/cquantumtest/but3_squantm.inp
test/cquantumtest/but4_squantm.inp
test/cquantumtest/squantm1.inp
test/cquantumtest/qmmm_ewald_squantm.inp
------------------------------------------------------------------------------
2.6 HQBM Restraints and Ensemble Codes
Developers : Robert Best and Emanuele Paci
Email Address : best@helix.nih.gov
Institution : National Institutes of Health
Date : December 29, 2005
Updates and bugfixes to hqbm and ensemble modules - most significant
being the changes to the replica exchange code.
- merged c33a1 install.com changes into install.nickname.com
- included preflx3 in install.nickname.com so only one file needs
to be distributed
- added symbolic linking to mpi for ensemble, so it works with
rest of charmm mpi code.
- updated doc/ensemble.doc
- updated doc/hqbm.doc
- replaced machdep/ensemble.src
- replaced fcm/ensemble.fcm
- put in include guards in dynamc/dcntrl.src
- updated dynamc/dynamc.src
- modified test cases for ensemble to remove eef1 output
- renamed ensemble_rex.inp to rex_ens.inp
- renamed hqbm_rc8.inp to hqbm_rc8_ens.inp
- modified misc/eef1.src to avoid slaves printing output
Files Modified:
install.com
install.nickname.com
doc/ensemble.doc
source/machdep/ensemble.src
source/dynamc/dcntrl.src
source/dynamc/dynamc.src
source/fcm/ensemble.fcm
source/misc/eef1.src
doc/ensemble.doc
doc/afm.doc
New Testcase:
test/benchmark/hqbm_rc3_ens.out
test/benchmark/hqbm_rc4_ens.out
test/benchmark/hqbm_rc8_ens.out
test/benchmark/rex_ens.out
test/c33test/rex_ens.inp replace ensemble_rex.inp
test/c33test/hqbm_rc8_ens.inp replace hqbm_rc8.inp
------------------------------------------------------------------------------
2.7 University of Maryland, School of Pharmacy Developments
Developers : Alexander D. MacKerell Jr., Igor Vorobyov, Victor Anisimov
Olgun Guvench
Email Address : alex@outerbanks.umaryland.edu
victor@outerbanks.umaryland.edu
Institution : University of Maryland, School of Pharmacy
Date : December 31, 2005
2.7.1 Eight-character atom-type-names
The length of atom-type-names (in the MASS section of the topology file)
can now be up to eight-characters long; previously the limit was four
characters.
Files Modified:
source/cff/parrdr_cff.src
source/fcm/dimens.fcm
source/fcm/param.fcm
source/gener/update.src
source/io/parmio.src
source/io/rtfio.src
[Note] Force field parameters for the developmental polarizable model
based on a classical Drude oscillator have been added. Currently,
only parameters for the SWM4-DP water models, alkanes and ions are
available. Toppar files along with example inputs are in the
toppar/drude directory.
2.7.2 Fitcharge enhancements
Previous version of fitcharge was capable of doing charge fitting for
the Drude polarizable model only. In this release the fitting is extended to
support the additive model and Lone-Pairs. Charges on LPs if present will be
determined from fitting to the QM electrostatic map.
Files Modified:
source/misc/fitcharge.src
doc/fitcharge.doc
2.7.3 MOLVIB for lone-pairs
A support of LPs has been added to Molvib. No modification in the Molvib
instruction set is required. The LP support is implemented by removing LP
blocks from the Hessian matrix and by not passing LP coordinates to the Molvib
analysis, because only real atoms are allowed in the Molvib analysis section.
Note, Hessian calculation is performed at the beginning of the Molvib module
prior to the Molvib analysis.
Files Modified:
source/molvib/molvco.src
2.7.4 Z-matrix format
Control the number of characters in atom names printed in CHARMM generated
Gaussian file. NCHAr specifies the number of characters of the atom type that
will be output. The default is one (NCHAr=1) such that, for example, for that
atom type HG1, only the character H will be printed in the output file.
Files Modified:
source/misc/zmat.src
doc/miscom.doc
The following are enhancements and bugfixes of the development period
from February 15, 2006 through August 15, 2006. The developments are
based on CHARMM version c33a2.
3.1 TSRI Fixes and Enhancements
Developers : Mike Crowley and Charles L. Brooks, III
Email Address : crowley@sripps.edu, brooks@scripps.edu
Institution : The Scripps Research Institute
Date : July 10, 2006
- Compiler upgrades in the makefiles and install.com
- Ability to read psf's with CHEQ information with executable without CHEQ
- Flush for xlf compilers
- Include file specs for mpif.h fixed and enhanced
- Excess output from parallel runs with fast shake fixed
- Fixed use of noninteger loop indices and incorrect array indices
- Timer formatting changed to allow very long times to be printed
- Preprocessor fix for gfortran compiles
- Missing CMAP entry for HSP in toppar/top_all22_prot_cmap.inp
Files Modified:
install.com
build/UNX/Makefile_gnu
build/UNX/Makefile_osx
source/energy/ecnstr.src
source/fcm/path.fcm
source/io/psfres.src
source/machdep/machio.src
source/machdep/paral1.src
source/machdep/startup.src
source/manip/fsshake.src
source/manip/shake.src
source/manip/secstr.src
source/minmiz/minmiz.src
source/minmiz/steepd.src
source/misc/quicka.src
source/nbonds/nbutil.src
source/quantum/mpiqub.src
source/util/new_timer.src
source/util/string.src
tool/preflx
toppar/top_all22_prot_cmap.inp
------------------------------------------------------------------------------
3.2 Developments at University of Chicago (Benoit Roux)
Developers : Edward Harder, Hyung-June Woo,
Nilesh Banavali, Yuqing Deng, Benoit Roux
Email Address : edh2004@med.cornell.edu, woo@unr.edu,
banavali@uchicago.edu, roux@uchicago.edu
Institution : Weill Cornell Medical College
Date : June 27, 2005
3.2.1 Constrained RMSD Enhancements
The flat bottom potential has been added to the force-const-spect as
specified by the keyword BOFFset in force-const-spec of the CONS command.
The potential functional form is
E = Sum_i KFORCE_i * [Abs(RMSD - OFFSET_i) - BOFFset_i]**2
if Abs(RMSD-OFFSET_i) > BOFFset_i and zero otherwise, where RMSD is either
the single RMSD, or RMSD(Main)-RMSD(COMP) with the RELAtive keyword.
The BOFFset keyword is the bottom offset distance (default: 0.0).
The dynamic memory allocation requirement has been improved. The memory
is now allocated for a pre-declared maximum "number" of rmsd constraints,
specified by the keyword MAXNrmsd (default: 10), which replaces the
MAXRmsd (maximum number of atoms constrained) of the previous implementation.
A warning message is added to indicate the change if MAXRMSD is invoked in
input file. The code WRNDIEs if CONS RMSD is invoked more than MAXN times
without CONS RMSD RESEt (One RELAtive constraint counts twice). The actual
memory spaces for atoms are allocated dynamically within the atom SELEction
parts of each CONS RMSD commands.
The zeta potential added in CHARMM32 is commented out since it fails
TEST FIRST.
Files Modified:
source/fcm/rmsd.fcm
source/energy/ecnstr.src
source/energy/rmsd.src
source/manip/cstran.src
New Testcase:
test/c33test/test_rmsd3.inp
..............................................................................
3.2.2 PDB File I/O Enhancements
Added the option OFFI to pdb read and write. The OFFI option enforces
the segid to be one character in length.
Files Modified:
source/io/coorio.src
Modified Documentation:
doc/io.doc
..............................................................................
3.2.3 GSBP and SSBP Warning Messages
Added warning messages if the nonbond interaction options are not
compatible with GSBP and SSBP. Added a steep potential in the GSBP
cavity potential at the GSBP boundary to prevent atoms from escaping
the simulation sphere.
Files Modified:
source/fcm/ssbp.fcm
source/machdep/ensemble.src
source/misc/gsbp.src
source/misc/nmr.src
source/misc/fitcharge.src
source/nbonds/nbutil.src
..............................................................................
3.2.4 DYNAMVV2 Bugfix
"floating point exception error" when running the dynamvv2 integrator
on alpha machines is fixed by initializing HEAP(XNEW).
Files involved:
source/dynamc/dcntrl.src
------------------------------------------------------------------------------
3.3 University of Chicago Fixes
Developers : Aaron R. Dinner
Email Address : dinner@uchicago.edu
Institution : The University of Chicago
Date : June 30, 2006
(1) Failure to de-reference pointer to the FREEAT array in TPSWRI call.
FREEAT replaced with STACK(FREEAT) in TPSWRI call.
Files involved:
source/dynamc/dcntrl.src
(2) GC arrays are not properly updated following global moves such as
HMC and MC-minimization. GCMCUP call is forced after acceptance of a global
move.
Files involved:
source/mc/mc.src
------------------------------------------------------------------------------
3.4 University of Vienna Bugfixes
Developers : Tibor Rudas, Stefan Boresch and Othmar Steinhauser
Email Address : tibi@mdy.univie.ac.at, stefan@mdy.univie.ac.at
Institution : University of Vienna
Date : June 30, 2006
(1) CMAP fails in combination with PERT, but the problem is CMAP's.
c30test/cmaptest.inp does NOT get run since CMAPSET is not set in iniall.src.
If PERT is active, no CMAP lambda=0 get computed. The real culprit is an error
in image/images.src related to the calculation of NCRTT
Files involved:
source/charmm/iniall.src
source/image/images.src
(2) Several PERT related problems are observed.
- Wrong IMAGE Lennard-Jones energy with FAST ON / FAST GENERIC.
- PERT does not work with the fast BYCBim nonbonded list routine. Ideally,
attempting to do so leads to a segfault. However, the Intel compiler manages
to produce numbers which even appear reasonable, which doesn't prevent them
from being utterly wrong.
- The soft core code (PERT/PSSP) does not handle the combination of group
based cut-off with Ewald. Instead of bombing, incorrect values are reported.
- The WCA/softcore code sets the substition variable SSBPLRCSD. This is fine
for the fact that these variable names get truncated to 8 characters. Thus,
each attempt to set this variable sets a new variable SSBPLRCS (note the
missing D !!!). If one goes crazy and creates a situation where this code
is run fairly often, one exceeds the maximum number of substitution
variables and bombs. Interestingly, c32test/repul.inp does not trigger the
problem, but the testcase should be modified to reflect a correct
8-character substitution variable name.
Files involved:
source/image/upimag.src
source/nbonds/nbonds.src
source/pert/pert.src
source/pert/epert.src
test/c32test/repul.inp
(3) FAST GENERIC to correct (V)ATOM VFSWIT EWALD energies. Upon setting FAST
GENERIC (old fast routines), the nonbonded option combination VFSWIT EWALD
is not handled correctly and leads to wrong energies (and forces, presumably).
In the old fast scheme (FAST GENERIC), REWALD was expected to handle
all ATOM / Ewald based functionality. However, it does not support
VFSWit. So far, so good. When REWALD "fails" (cannot accomplish its
task), EVDW should take over. Because of a faulty IF statement,
however, ENBFS8 (which has no clue about EWALD) takes over, leading to
an interesting mix of electrostatic energy terms.
Files involved:
source/nbonds/enbond.src
------------------------------------------------------------------------------
3.5 Multiconformational Charge Fitting with Lone-Pairs
Developers : Victor Anisimov and Alex D. MacKerell
Email Address : victor@outerbanks.umaryland.edu
alex@outerbanks.umaryland.edu
Institution : University of Maryland, School of Pharmacy
Date : June 30, 2006
Previous version of fitcharge did not support multiconformational charge
fitting in presence of Lone pairs. This is fixed in the current version
along with various small bugs. An extended output of CHARMM MESP and its
difference from QM MESP is added for the "TEST" mode of FITCHARGE.
Files Modified:
source/misc/fitcharge.src
doc/fitcharge.doc
------------------------------------------------------------------------------
3.6 TREK Support Files
Name : Fabian Ille, Frank Noe, Stefan Fischer
Email Address : stefan.fischer@iwr.uni-heidelberg.de
Institution : University of Heidelberg, Germany
Date : June-29, 2006
A method for generating an initial path when setting-up a CPR refinement
of a large conformational transition in a protein is described in detail in
Proteins 59, p.534-544, 2005.
The initial path (i.e., a series of transition intermediates) that is to
be refined by CPR is generated by interpolation between the Internal
Coordinates (IC) of the reactant and product end-states
(rather than the default interpolation in Cartesian coordinates).
Interpolation in IC describes well the torsional transitions
between rotameric states of the sidechains. However, when IC-interpolation is
applied to the backbone-torsion angles, it can lead to severe disruption of
the backbone fold. In contrast, Cartesian interpolation approximately preserves
the backbone fold in most cases (if the protein remains compact during the
transition), but often leads to extreme deformation of the sidechains.
Therefore, the two interpolation methods are combined here:
First, the backbone atoms are interpolated in Cartesian coordinates
so as to preserve the backbone fold, and then the sidechain atoms are
built onto the interpolated backbone, using IC that are interpolated between
the values of the end-states.
The initial path build by this "combined interpolation" can be further
improved, so as to avoid unrealistic events such as bond-crossing
or ring-penetration. These can be effectively avoided by shrinking all
the sidechains in each intermediate point of the initial path.
This is achieved by reducing all bond lengths of the sidechains
(for ex. to half the original size) before building them onto the backbone.
At first glance, this approach would seem to be unphysical and
would result in energetically unfavorable paths. However, the minimization
process applied to the path-points during the CPR procedure rapidly
restores the shrunken sidechains to their normal size, while undesirable
events no longer occur in the resulting path.
Files Added:
support/trek/initial_path/00readme
support/trek/initial_path/comb_interpol.inp
support/trek/initial_path/make_traj.str
support/trek/initial_path/mini_interpol.str
support/trek/initial_path/user_def.str
support/trek/initial_path/vector.pl
support/trek/initial_path/frames/interpol_comb_4_steps.dcd
support/trek/initial_path/frames/interpol_comb_4_steps.dcd.pdb
Note that initial_path is put into support/trek/initial_path.tar.gz.
Modified Documentation:
doc/trek.doc
------------------------------------------------------------------------------
3.7 Developments at NIC-Slovenia
Names : Milan Hodoscek
Email Address : milan@cmm.ki.si
Institutions : National Institutes of Chemistry, Ljubljana, Slovenia
and National Institute of Health, Bethesda, MD, USA
Date : June 30, 2006
3.7.1 Further support for 64 bit machines
Some fixes necessary to support 64 bit machines. Mostly
regarding parallel ab initio QM/MM methods and some other too.
Files Modified:
source/gamint/ddi.src
source/parallel/paral1.src
source/parallel/paral2.src
source/parallel/paral3.src
.............................................................................
3.7.2 pref.dat fixes
Some fixes for ##IF statement
Files Modified:
source/energy/afm.src (missing label)
source/fcm/mndgho.fcm (file was empty ???)
source/mndint/mndene.src (GHO and other fixes)
source/mndint/mndgho.src (GHO fix)
source/mndint/mndini.src (GHO fix)
source/quantum/qmevdw.src (Intel fortran doesn't like empty source files)
source/util/string.src (MMFF, CFF fix)
source/nbonds/nbutil.src (LRVDW,PBEQ dependency problem)
.............................................................................
3.7.3 Other bug fixes
(1) Replacing 0.0 or 0.0d0 with the ZERO from number.fcm
Files Modified:
source/energy/ecnstr.src
source/util/vector.src
(2) cstuff.c support for glibc-2.4
Files Modified:
source/machdep/cstuff.c
(3) New gcc support fixes (gfortran)
Files Modified:
source/machdep/startup.src
(4) Subroutine POWELL name clash with GAMESS
Files Modified:
source/minmiz/minmiz.src
source/minmiz/powell.src
(5) CMAP parameters files don't work with MEDIUM CHARMM
Files Modified:
source/fcm/dimens.fcm
(6) Parallel fix for EEF1
Files Modified:
source/misc/eef1.src
test/c27test/testeef1.inp
..............................................................................
3.7.4 All occurrences of mpi/mpif.h changed to mpif.h
All the occurences of the include 'mpi/mpif.h' statements were
simplified and generalized for all the platforms. From now on use just
include 'mpif.h', since install.com and Makefile_ take care of
the problem.
Files Modified:
install.com
source/dynamc/dynamc.src
source/energy/energy.src
source/machdep/machutil.src
source/machdep/paral1.src
source/machdep/paral2.src
source/machdep/paral3.src
source/machdep/startup.src
source/manip/shake.src
source/nbonds/nbndcc.src
source/nbonds/pblist.src
source/nbonds/pmeutil.src
source/quantum/mpiqub.src
source/quantum/qmene.src
source/quantum/qmjunc.src
source/quantum/qubene.src
source/vibran/vibran.src
------------------------------------------------------------------------------
3.8 Developments at the National Institute of Health
Developers : Jhih-Wei Chu, H. Lee Woodcock, Xiongwu Wu,
George Stan and Bernard R. Brooks
Email Address : chu@hec.utah.edu, hlwood@nih.gov, wuxw@nhlbi.nih.gov,
brb@nih.gov
Institution : Laboratory of Computational Biology, NHLBI, NIH
Date : July 1, 2006
3.8.1 EMAP bugfix and document update
Bug fix in coorelation fucntion calculation and update in io.doc and
emap.doc about direct read segment id from PDB files.
Files involved:
source/emap/emapsubs.src
doc/io.doc
doc/emap.doc
..............................................................................
3.8.2 Bugfix: QCHEM keyword added the QM/MM section of non-bond files
Files Modified:
source/nbonds/enbaexp.src
source/nbonds/enbfast.src
source/nbonds/enbfs8p.src
source/nbonds/enbonda.src
source/nbonds/enbondg.src
source/nbonds/ewald.src
..............................................................................
3.8.3 Bugfix: Protected debug write statement in Replica Path Code
Files Modified:
source/energy/epath.src
..............................................................................
3.8.4 Bugfix: Fixed install.com bug to allow proper SCCDFTB libraries to be
added to Makefiles
Files Modified:
install.com
..............................................................................
3.8.5 SCPISM Code Optimized: Runs 30% faster, Uses Less Memory
Developer : Peter Steinbach
Email Address : steinbac@helix.nih.gov
Institution : CMM/DCB/CIT, National Institutes of Health
Date : August 9, 2006
The implementation of the SCP continuum solvent model was simplified,
optimized, and tested. It now runs 30% faster and uses memory more
efficiently. A few large two-dimensional arrays are no longer used.
Energies and forces are unchanged, to within roundoff errors.
Files Modified:
source/fcm/scpism.fcm
source/misc/scpism.src
------------------------------------------------------------------------------
3.9 CHARMM Version 33 Beta Release
The CHARMM meeting, Harvard University, July 21-23, 2006, decided to
test further the following features.
3.9.1 SPASIBA SPASIBA force field code
Files Removed:
source/energy/ebaspas.src
source/fcm/spa_code.fcm
source/fcm/spacod.fcm
source/fcm/spapar.fcm
source/gener/setup_spa.src
test/c29test/spasibatest.inp
Files Modified:
source/charmm/iniall.src
source/emap/emapsubs.src
source/energy/ecnstr.src
source/energy/ediff.src
source/energy/eintern.src
source/energy/energy.src
source/energy/intere.src
source/fcm/dimens.fcm
source/fcm/exfunc.fcm
source/fcm/param.fcm
source/fcm/psf.fcm
source/gener/update.src
source/image/eimage.src
source/io/parmio.src
source/manip/intcor2.src
doc/energy.doc
..............................................................................
3.9.2 PBCUBES Periodic boundary cubic list builder
Files Removed:
source/nbonds/pblist.src
source/fcm/pbcubes.fcm
Files Modified:
source/cff/enbfast_cff.src
source/cff/ewald_cff.src
source/charmm/iniall.src
source/energy/energy.src
source/energy/energy.src
source/energy/gbim.src
source/energy/gbmv.src
source/energy/genborn.src
source/energy/intere.src
source/fcm/deflts.fcm
source/fcm/image.fcm
source/fcm/inbnd.fcm
source/image/images.src
source/image/nbndgcm.src
source/image/upimag.src
source/image/xtlfrq.src
source/io/parmio.src
source/mbond/mbdiag.src
source/misc/hbuild.src
source/mmff/enbscalar_mm.src
source/nbonds/ace.src
source/nbonds/enbfast.src
source/nbonds/enbonda.src
source/nbonds/enbond.src
source/nbonds/ewald.src
source/nbonds/ewald.src
source/nbonds/ewwnb.src
source/nbonds/nbonds.src
source/nbonds/nbutil.src
..............................................................................
3.9.3 RUSH all atom short-range force field
Files Removed:
source/energy/rush.src
source/fcm/rush.fcm
toppar/rush/par_rush_058.inp
toppar/rush/par_rush_058_a.cmap
toppar/rush/top_rush_058.inp
doc/rush.doc
test/c33test/rush.inp
test/data/top_rush_058.inp
test/data/par_rush_058.inp
test/data/par_rush_058_a.cmap
Files Modified:
source/charmm/charmm_main.src
source/charmm/iniall.src
source/energy/energy.src
source/energy/eutil.src
source/energy/printe.src
source/fcm/energy.fcm
doc/commands.doc
..............................................................................
3.9.4 SPACDEC Spatially distributed parallel method
Files Removed:
source/fcm/spacdec.fcm
Files Modified:
source/charmm/iniall.src
source/dynamc/dcntrl.src
source/dynamc/dynamc4.src
source/dynamc/dynamc.src
source/dynamc/dynamvv2.src
source/dynamc/dynamvv.src
source/dynamc/dynlng.src
source/dynamc/mts.src
source/dynamc/prssre.src
source/dynamc/tmd.src
source/energy/ecnstr.src
source/energy/ediff.src
source/energy/eintern.src
source/energy/enefscal.src
source/energy/energy.src
source/energy/epath.src
source/gener/update.src
source/image/upimag.src
source/machdep/paral1.src
source/manip/holonom.src
source/manip/shake.src
source/misc/aspenermb.src
source/misc/aspener.src
source/nbonds/ewaldf.src
source/nbonds/exelec.src
source/nbonds/heurist.src
source/nbonds/nbndcc.src
source/nbonds/nbonda.src
source/nbonds/nbondg.src
source/nbonds/nbonds.src
source/nbonds/pme.src
source/pert/epert.src
source/quantum/qmene.src
source/quantum/qmevdw.src
source/quantum/qmjunc.src
source/quantum/qmpac.src
source/quantum/qubene.src
source/util/util.src
..............................................................................
3.9.5 LNTAB1 Partial Table Lookup
SLVSLV SOLVENT-SOLVENT interaction routine
Files Removed:
source/nbonds/ewwnb.src
source/fcm/wwnb.fcm
source/fcm/wwnb1.fcm
test/c32test/nbsolv.inp
Files Modified:
source/charmm/iniall.src
source/charmm/miscom.src
source/energy/energy.src
source/image/eimage.src
source/image/upimag.src
source/nbonds/enbaexp.src (ln_c32a1.dff)
source/nbonds/nbonds.src
source/nbonds/nbutil.src
doc/nbonds.doc
..............................................................................
3.9.6 QUB Path-Integral simulations with QM/MM potentials.
Files Removed:
source/quantum/qub.src
source/quantum/mpiqub.src
source/quantum/qubene.src
source/fcm/qubpi.fcm
doc/qub.doc
test/cquantumtest/qmpathint1.inp
test/cquantumtest/qmpathint2.inp
Files Modified:
doc/commands.doc
source/charmm/charmm_main.src
..............................................................................
3.9.7 PIPF Implementation of Polarizable Intermolecular Potential Functions
Files Removed:
source/pipf/dpfimg.src
source/pipf/dpipf.src
source/pipf/epfimg.src
source/pipf/epipf.src
source/pipf/pfdyn.src
source/fcm/pipf.fcm
doc/pipf.doc
test/c33test/pipf_test1.inp
test/c33test/pipf_test2.inp
test/c33test/pipf_test3.inp
test/data/pol2_125.crd
build/UNX/pipf.mk
Files Modified:
source/charmm/charmm_main.src
source/dynamc/dcntrl.src
source/dynamc/dynamc4.src
source/dynamc/dynamc.src
source/dynamc/dynamvv.src
source/dynamc/dynio.src
source/energy/energy.src
source/energy/eutil.src
source/energy/printe.src
source/fcm/energy.fcm
source/image/upimag.src
source/io/coorio.src
source/mbond/mbback.src
source/mc/mcio.src
source/nbonds/enbond.src
install.com
doc/commands.doc
doc/energy.doc
..............................................................................
3.9.8 SQUANTM SQUANTUM package interface
Files Removed:
build/UNX/squantm.mk_altix
build/UNX/squantm.mk_gnu
build/UNX/squantm.mk_ibmaix
build/UNX/squantm.mk_ibmaix64
source/fcm/squantm.fcm
source/squantm/assert.h
source/squantm/constants.h
source/squantm/conversions.h
source/squantm/dprec.h
source/squantm/qm2_array_locations.h
source/squantm/qm2_elements.h
source/squantm/qm2_parameters.h
source/squantm/sqnt_ene.src
source/squantm/sqnt_nbnd.src
source/squantm/sqnt_qm2_energy.src90
source/squantm/sqnt_qm2_ewald.src90
source/squantm/sqnt_qm2_gho.src90
source/squantm/sqnt_qm2_grad.src90
source/squantm/sqnt_qm2_mopac.src90
source/squantm/sqnt_qmmm_int.src90
source/squantm/sqnt_qmmm_module.src90
source/squantm/sqnt_qmmm_util.src90
source/squantm/sqnt_setup.src
test/cquantumtest/but0_squantm.inp
test/cquantumtest/but1_squantm.inp
test/cquantumtest/but3_squantm.inp
test/cquantumtest/but4_squantm.inp
test/cquantumtest/squantm1.inp
test/cquantumtest/qmmm_ewald_squantm.inp
Files Modified:
install.com
doc/qmmm.doc
source/charmm/iniall.src
source/energy/energy.src
source/fcm/gamess.fcm
source/fcm/nbndqm.fcm
source/flucq/flucq.src
source/flucq/fluqqmmm.src
source/gener/update.src
source/minmiz/tnpack.src
source/nbonds/enbaexp.src
source/nbonds/enbfast.src
source/nbonds/enbfs8p.src
source/nbonds/enbonda.src
source/nbonds/enbondg.src
source/nbonds/ewaldf.src
source/nbonds/ewald.src
source/nbonds/nbonda.src
source/nbonds/nbonds.src
source/nbonds/pme.src
source/quantum/addlnat.src
source/quantum/qmevdw.src
source/quantum/qmnbnd.src
The following are enhancements and bugfixes of the development period
of August 15, 2006 through February 15, 2007. The developments are
based on CHARMM version c33b1.
4.1 TSRI Developments
Developers : Mike Crowley and Charles L. Brooks, III
Email Address : crowley@sripps.edu, brooks@scripps.edu
Institution : The Scripps Research Institute
Date : January 1, 2007
4.1.1 Bugfixes and necessary changes for compiling and testing
- OSX Makefile improvements for new compilers and mpi
- double selection enabled in test second
- g95 added
- Parallel cheq fixed
- Output protected by prnlev and iolev
- solana bugfix for nint(t) when t not defined
- nmdimb bugfix for wrong local subroutine call
- afm and scpism pref label incorrect fixed
- parallel phmd fixed
- new_timer bugfix for incorrect array boundaries
Modified Files:
install.com
build/UNX/Makefile_osx
doc/test.doc
source/cheq/fqener.src
source/correl/clustr.src
source/correl/shell.src
source/correl/solana.src
source/dimb/nmdimb.src
source/energy/afm.src
source/energy/energy.src
source/energy/gbmv.src
source/energy/gbsw.src
source/energy/phmd.src
source/energy/rmsd.src
source/fcm/new_time.fcm
source/gener/replica.src
source/misc/olap.src
source/misc/pbeq2.src
source/misc/pbeq.src
source/misc/scpism.src
source/misc/testch.src
source/nbonds/ewald.src
source/nbonds/nbutil.src
source/pert/wham.src
source/util/util.src
source/vibran/vibcom.src
..............................................................................
4.1.2 Testcase clean-up
Test cases are fixed and modified to make success or failure
detection possible
- Sanity checks with exits for tests which may not have the code compiled.
- Checks for exiting tests that should not be run in parallel.
- Correct message echoed for tests not run
- Remove langevin dynamics and abnr minimization for tests that do not require
these features so the results are reproducible and testable
- parallel output fixed with prnlev @pnode
where pnode is set in datadir.def
- tests that can run in parallel even though not parallel code included
- old integrator not used unless essential so tests can run in parallel
- Sections defined with appropriate skipping of sections for features not
compiled or not parallel for parallel runs
- Unnecessarily long dynamics, minimization, MC or other iterative proceedures
reduced to manageable but still meaningful lengths
- Some bugfixes
- Example test added c34test/gbsw_example.inp
following the prescription for tests header comments development
This test is called gbsw_example.inp and resides in c34a1/test/c34test.
The philosophy is to divide the test case into functional sections,
tag each section with using a specific syntax and self-test within
each section to detect correct execution of that section. Also, several
"dos" and "don'ts" are clear and obvious.
New/Modified Testcase:
test/c20test/ictest.inp
test/c20test/maatest.inp
test/c23test/nmrtest1.inp
test/c24test/mmfptest3.inp
test/c24test/pert2.inp
test/c25test/cortst25.inp
test/c25test/fastest.inp
test/c26test/block4.inp
test/c26test/luptst.inp
test/c26test/pathint.inp
test/c26test/whamtest.inp
test/c27test/adumb-enum.inp
test/c27test/adumb-phichi.inp
test/c27test/gb_cff95_na-test.inp
test/c27test/gb_cff95_prot-test.inp
test/c27test/gb_cff-data.inp
test/c27test/gb_mmff94_na-test.inp
test/c27test/gb_mmff94_prot-test.inp
test/c27test/gb_mmff94_test.inp
test/c27test/genborn19.inp
test/c27test/lnx_img.inp
test/c27test/testeef1imag.inp
test/c27test/tsoft.inp
test/c28test/gb_fep.inp
test/c28test/mcmd_rst2.inp
test/c28test/mcmd_rst.inp
test/c28test/mcmd_rst_slow.inp
test/c29test/ace-adumb.inp
test/c29test/adumbcorr.inp
test/c29test/rmsdyn.inp
test/c30test/cftiangl.inp
test/c30test/cmaptest.inp
test/c30test/cq1ab1gbdyn.inp
test/c30test/cq1fsvwat.inp
test/c30test/entropylj.inp
test/c30test/entropyosc.inp
test/c30test/test_rmsd2.inp
test/c31test/gbsaim.inp
test/c31test/off_path_simulation.inp
test/c31test/pert-mmfp.inp
test/c32test/argc_cbias.inp
test/c32test/argc_grid.inp
test/c32test/argc_rndm.inp
test/c32test/attr.inp
test/c32test/cons_rmsd_zeta.inp
test/c32test/dscfvib.inp
test/c32test/ips_tip216.inp
test/c32test/mpnoe.inp
test/c32test/repul.inp
test/c33test/gbsw_ff05.inp
test/c33test/hdgb.inp
test/c33test/hqbm_single.inp
test/c33test/nmrtest2.inp
test/c33test/phmd.inp
test/c33test/test_rmsd3.inp
test/compare.awk
test/datadir.def
test/test.com
------------------------------------------------------------------------------
4.2 Developments at NIC-Slovenia
Names : Milan Hodoscek
Email Address : milan@cmm.ki.si
Institutions : National Institutes of Chemistry, Ljubljana, Slovenia
and National Institute of Health, Bethesda, MD, USA
Date : December 30, 2006
4.2.1 GAMESS/CHARMM QM/MM interface: support for the latest GAMESS
Support for latest version of GAMESS (7 Sep 2006 R4) released on
December 15, 2006 was added to CHARMM. New development in this version
includes: parallel CC methods, Nuclear-Electron Orbital (NEO). Also
version 3 of DDI communication library is supported.
Files Modified:
doc/gamess.doc
source/gukint/gukini.src
source/gamint/ddi.src
source/fcm/gamess.fcm
source/fcm/consta.fcm
install.com
build/UNX/gamess.mk
build/UNX/gmscomp
tool/actvte.f
.............................................................................
4.2.2 pref.dat and other fixes
Some fixes for ##IF statement (for example install.com LITE
doesn't compile)
Files Modified:
source/correl/solana.src (wrong variable name used)
source/machdep/cstuff.c (error message when hostname not available)
source/mc/moveln.src (typo in routine name)
source/misc/scpism.src (missing ##ENDIF at the end)
source/nbonds/ewald.src (##CHEQ problem)
source/util/util.src (everyone needs ASSINT routine!)
source/io/psfres.src (NGRPX,NST2X undefined - please check with
original developer)
source/quantum/qmleps.src (VC12 undefined - please check with
original developer)
------------------------------------------------------------------------------
4.3 Parameter Updates
Developers : Alex D. MacKerell
Email Address : alex@outerbanks.umaryland.edu
Institution : University of Maryland, School of Pharmacy
Date : January 3, 2007
1) CMAP term for protonated histidine (RESI HSP) was omitted from
the topology files containing proteins. This was corrected in the
c33b1 release.
2) Special parameters for gbsw model added. See gbsw subdirectory
and parmfile.doc for details (TSRI/Brooks group addition).
Files Modified:
toppar/top_all22_prot_cmap.inp
toppar/top_all27_prot_lipid.rtf
toppar/top_all27_prot_na.rtf
------------------------------------------------------------------------------
4.4 Bugfixes
1) Image centering with PSF modified.
When PSF is modified, IMAGE structure is cleaned and image centering
is nullified.
Files fixed:
doc/image.doc
source/image/images.src
2) CMAP parallel code
PSND4 call for CTPA array is fixed
File fixed:
source/io/parmio.src
3) Doubly defined subroutine
FLUSH is defined twice. One is removed.
File Fixed:
soruce/machdep/machio.src
4) VELOcity option for trajectory IO
VELOcity can be written with VELOcity option to the TRAJectory
command.
Files Fixed:
doc/dynamc.doc
source/dynamc/dynio.src
5) PRES format fix
MKPRES writing format is fixed for PRES generation.
File Fixed:
source/gener/mkpres.src
6) Parallel code fix
EEF1 returns with broadcasting data
File Fixed:
source/misc/eef1.src
7) Offset error when copying time series
When copying a time series B to A, there is an offset of -1 in the
copied series A, and the last value of A is junk. Fixed by initializing
IF1 to zero in subroutine mantim (mantim.src).
File Fixed:
source/correl/mantim.src
8) Preprocessor error
Change mis-spelled "ENSEMBEL" to "ENSEMBLE" in dynamc/dynio.src
File Fixed:
source/dynamc/dynio.src
9) Sun64 Port
SUN 64 bit compilation problem fixed
Files Fixed:
install.com
build/UNX/Makefile_sun64
