ChangeLog.c23
| | |
| CHARMM 22 |
| c22.0.b | Release |
April 22, 1991 |
| c22.0.b1 | Release |
September 30, 1991 |
| c22 | Release |
January 1, 1992 |
| c22g1 | Release |
February 15, 1992 |
| c22g2 | Release |
July 7, 1992 |
|
CHARMM 23 |
| c23a1 |
Developmental |
August 15, 1992 |
| c23a2 |
Developmental |
October 25, 1992 |
| c23f |
Developmental |
March 1, 1993 |
| c23f1 |
Developmental |
March 15, 1993 |
| c23f2 |
Developmental |
August 15, 1993 |
| c23f3 |
Release |
February 1, 1994 |
| c23f4 |
Release |
August 15, 1994 |
| c23f5 |
Release |
March 15, 1995 |
|
| | |
|
About CHARMM23 Development |
|
|
|
As an on-going project, CHARMM development has been carried out with
CHARMM version 23 series. Since we branched out from the c22g2
release version, we have made two alpha versions (c23a1 and c23a2) and
five FORTRAN versions (c23f and c23f1 through c23f4).
(1) As the "f" in c23f stands for FORTRAN version, we converted FLECS
source into FORTRAN. The conversion task had been completed by
c23f2. Now CHARMM is written in full FORTRAN except several
machine dependant codes in C.
(2) We maintain all source and support files under CVS (Concurrent
Versions System) control. The ROOT repository is
tammy.harvard.edu:/prog/chmgr/CVS. CHARMM23 is stored in
/prog/chmgr/CVS/c23f. A particular version is retrieved with
the version name as the rivision tag (e.g., c23f3)
(3) We require all developers to understand CHARMM coding standard and
checkin rules. Please, consult doc/developer.doc and doc/install.doc
for details. There are three forms you can get through
anonymous-ftp to tammy.harvard.edu:pub/charmm/problem.form,
bugfix.form and project.form. Use problem.form and bugfix.form to
report problems and bugfixes, respectively. These should be sent
to charmm-bugs@tammy.harvard.edu.
We also require a written description of source code modification
(done or planned) from each developer. Please consult project.form
to report your development status and plan. It will help the
manager to schedule integration and to provide better idea of code
changes on going. All such project descriptions will be sent to
other developers.
The following lists the modifications along the development line from
c23a1 through the current version.
CHARMM c22g2 (July 7, 1992) to c23a1 (August 15, 1992)
1.1 CVS/CDE Support
- A new directory cde_defs is added and contains CDE specific
information. You should modify any of its contents ONLY when you
add/remove files in the received snapshot.
- install.com, module makefiles (~/build/{machine_type}/*.mk) and
~/tool/makemod are modified to meed the requirements of CDE
init/get command. Also, mainmake (makemod for ~/source/main) is
added in ~/tool and should be used when you generate main module
makefiles. ~/tool/unix directory is removed.
- Note that when source file structure is modified, the modification
should be reflected in the corresponding .mk files and ~/tool/vaxvms/
vmsinstall.com. In general, any file structure alteration
(including addition and deletion) must be reflected in cde_defs
files.
1.2 Stardent Version Support
- There are problems on Kubota Computer (Stardent) platforms. Make
command does not function as we expect on UNIX machines. In order
to work around the problem, some stardent specific sections are
added in install.com. A proper set of module makefiles is found in
~/build/stardent.
- Note that not all testcases are checked out. Especially,
c20test/rtftest (associated with io/rtfio.src), testsel2
(associated with util/string.src and/or util/parse.src)
c22test/icpert, mewtest.
1.3 Molecular Simulations Inc. Modifications
===========================================================================
CHMGR Note: 05-Aug-92
All the modifications described are integrated with the tag msi9207 in c23a1
===========================================================================
Name : Ryszard Czerminski
Email Address : ryszard@connection.prospect.com
Institution : Molecular Simulations Inc.
Date : August 1, 1992
===========================================================================
(1) Cray Porting
main/miscom.src - SPECIFY defined for CRAY
energy/enefst2.src - CRAY port
energy/enefst2q.src - CRAY port
energy/enefvp.src - CRAY port
image/imnbfp.src - CRAY port
machdep/space.src - bugfix: VEHEAP on CRAY was allocating four times
- to much space
manip/fshake.src - CRAY port
nbonds/enbf2.src - CRAY port
nbonds/enbf3.src - CRAY port
nbonds/enbf4.src - CRAY port
nbonds/enbf5.src - CRAY port
nbonds/evdwf.src - CRAY port
nbonds/ewald.src - CRAY port
nbonds/ewaldf.src - CRAY port
nbonds/ewaldf2.src - CRAY port
nbonds/nbexcl.src - CRAY port
nbonds/nbndf2p.src - CRAY port
nbonds/nbndfp.src - CRAY port
(2) Frequency Based Crystal Update
The most extensive modifications was done to allow for automated,
frequency based, crystal update. New variable (IXTFRQ) was introduced
which controls frequency of the crystal update.
main/iniall.src - CUTIM moved here,
fcm/image.fcm - new CUTXTL & IXTFRQ
gener/update.src - new IXTFRQ
image/crystal.src - crystal update implemented,
- XBUILD split into XBUILD and XBUILD2;
- PRNXTLD modified, PRINTX removed
- CRYSTAL FREE added
image/images.src - CUTIM removed
nbonds/heurist.src - call to XBUILD2 added
(3) Replace calls to DIEWRN and DIE by calls to WRNDIE
The idea is to replace all calls to DIEWRN and direct calls to DIE by
appropriate calls to WRNDIE which gives more information to the output
and communicates error messages to QUANTA. Eventually, when all calls
to it are replaced, DIEWRN should be removed.
image/upimag.src - DIE replaced by WRNDIE
machdep/machio.src - DIEWRN replaced by WRNDIE
quantum/qmjunc.src - DIE(-5) replaced by WRNDIE, EXIT replaced by WRNDIE
quantum/qmpac.src - EXIT replaced by WRNDIE
quantum/qmset.src - DIE(-5) replaced by WRNDIE
(4) Dimension Increase in dimens.fcm
MAXSYM increased to 192 (was 24)
MXRTA increased to 2250 (was 2000)
MXRTB increased to 2250 (was 2000)
MXTBL increased to 3000 (was 2000)
remove the following three lines for ##IF QUANTA
##IF QUANTA
PARAMETER (MAXATC = 500, MAXCB = 1500, MAXCH = 1600, MAXCI = 600,
& MAXCP = 700, MAXCT = 5500, MAXCN = 7000)
(5) HP port (HPUX)
machdep/cstuff.c - HP port
machdep/machio.src - HP port; DIEWRN replaced by WRNDIE
machdep/startup.src - HP port
(6) Ability to Linearly Increase/Decrease Pressure
The goal was to allow for linear increase (decrease) of the pressure
during single dynamic run. New variables/keywords were introduced
(PIXX - initial value of XX component of pressure tensor, PFXX - final
value etc... for other components)
dynamc/dcntrl.src - new variable NPRIVOLD; new REFPI and REFPF
(initial and final reference pressure tensors)
- new keywords : PIXX,PIYY etc...
PFXX,PFYY etc...
for initial and final pressure tensor
format modified for initial and final
dynamc/dynamc.src - PRNXTLD modified
fcm/reawri.fcm - QCNSTPL logical flag for linear pressure ramping
(7) Minimizers & Communication with QUANTA
- All CHARMM/QUANTA communications has been removed from minimizers and
placed in PRINTE subroutine.
- attempt has been made to further standardize different minimizers
printout
- attempt has been made (not complete yet e.g. for Newton-Raphson)
to make minimizers to return the lowest energy
(and corresponding coordinates) as a final
energy/printe.src - QUANTA communication with minimizers moved here;
- format fixed;
- EPROP(GRMS) calculation removed for dynamics
minmiz/egrad1.src - QUANTA communication moved to PRINTE;
- PRINTX replaced by PRNXTLD
minmiz/abner.src - QUANTA removed
minmiz/conjug.src - QUANTA removed
minmiz/nraph.src - QUANTA removed
minmiz/powell.src - QUANTA removed
minmiz/steepd.src - QUANTA removed
- bugfix: now this minimizer can converge too!
(8) PRNLEV Modified
The goal of this modification is to make PRNLEV=0 the default
(similarly to other *LEV keywords e.g. BOMLEV). PRNLEV=0 would give
minimum but sufficient information. PRNLEV > 0 would be used mostly
for testing and debugging and of course for CHARMM output lovers.
PRNLEV < 0 would be use to cut output even further (e.g. for very long
dynamics run) What exactly PRNLEV should be in each particular place
will be establish most probably by trial and error self-convergent
(hopefully) procedure.
dynamc/dynamc.src - PRNLEV modified
image/crystal.src - PRNLEV modified
minmiz/abner.src - PRNLEV modified
minmiz/conjug.src - PRNLEV modified
minmiz/nraph.src - PRNLEV modified
minmiz/powell.src - PRNLEV modified
minmiz/steepd.src - PRNLEV modified
nbonds/heurist.src - PRNLEV modified
util/parse.src - PRNLEV modified
(9) QUANTUM Code Problem Fix
main/iniall.src - QMMMINIT moved from ENERGY
energy/energy.src - QMMMINIT moved to INIALL
fcm/quantm.fcm - QMDATA rearranged (REAL*8 first etc...)
quantum/qmpac.src - declaration of HELECT fixed
(10) Miscellaneous
image/nbondm.src - loops rearranged in ##IF PERT
util/util.src - QUANTA communication
1.4 Cray Fast Code and Problem Fix
===========================================================================
CHMGR Note: 07-Aug-92
All the modifications described are integrated with the tag djt9207 in c23a1
===========================================================================
SUBJECT: Cray version
PEOPLE INVOLVED: tobias@sun0.chem.upenn.edu, brooks@cmchem.chem.cmu.edu
BACKGROUND INFORMATION
The code was written by myself and John Mertz at Cray Research and
incorporated first into CHARMM 19. We later implemented similar code
into Polygen's CHARMm 21. Given the large differences between CHARMM
22 and previous versions, it was quite a task to port the code to the
first version of CHARMM 22. I completed that just as version c22g1
was released and then completed the port to c22g1 just as c22g2 was
released. Since myself and others were anxious to simply have a
working, optimized Cray version of CHARMM 22, the initial port, as you
will see, is very superficial. Given the large similarities between
the Convex and Cray vector/parallel code, it would be nice to use
mostly the same code on the two machines (and others with similar
architectures). But since I have been largely excluded from
discussions of plans for future development, I decided expedience and
simplicity on my part was in order and did not attempt to incorporate
Cray-specific modifications into the existing code. Rather, I crudely
"spliced" in our code. I would be happy to participate in
beautification of the Cray code and/or future developments of largely
machine-independent vector parallel code. I give below a list of the
files I modified and added with a brief description of what was done.
In addition to the code itself, I am planning to send a brief
document describing the algorithms and their implementation and usage,
and another prepared by Charlie that discusses some deviations between
the Convex and Cray test cases.
SUMMARY
Vector/parallel code for energy calculation, shake, and nonbonded
list generation on the Cray was implemented. Dynamic heap and
stack allocation on the Cray was added. Several bugs in the Cray
code were fixed. Overall, the changes to existing code are
superficial, the bare minimum necessary to incorporate the code,
and are therefore fairly mild but not very elegant. Several new
source files were added.
ROUTINES ADDED
fcm/cray.fcm -contains heap pointers for Cray fast shake
energy/cenerf.src -vector/parallel bonded energy routines for Cray
manip/cshake.src -vector/parallel shake routines for Cray
image/cnbndm.src -vector/parallel atom-based image nb list routine
nbonds/cenbf.src -vector/parallel nb energy routine
nbonds/cnbnd.src -vector/parallel atom-based nb list routine
ROUTINES CHANGED
install.com -added prefx PARVECT key for Cray
build/cray -added make of new files to several makefiles
build/cray/Makefile_cray-added compiler switches for parallel compilation
build/cray/image.mk -compile nbondm.f with optimization turned of
because of compiler bug
build/cray/nbonds.mk -compile ewald.f with optimization turned of
because of compiler bug
energy/energy.src -added crude interface to Cray bonded energy
routines
fcm/heap.fcm -declare pointer to heap for dynamic allocation
fcm/stack.fcm -declare pointer to stack for dynamic allocation
image/images.src -added compiler directives for vectorization in
TRANSO and TRANSI
image/upimag.src -modified some loops and added directives for
vectorization in MKIMAT
machdep/cstuff.c -added function to get # physical cpus from system
machdep/machutil.src -added routine to get NCPU from system
machdep/startup.src -adding parsing of heap and stack sizes to
command line parsing code
main/charmm.src -moved call to ARGUMT
main/charmm.src -made QFASTNB=.TRUE. default on Cray
main/miscom.src -set QFASTNB=.TRUE. when FASTER.GE.0 on Cray
manip/shake.src -added crude interface to Cray shake routines
nbonds/nbndgc.src -change RHEAP equivalence to pointer statement
nbonds/nbonds.src -added crude interface to Cray nb list routines
nbonds/enbond.src -added crude interface to Cray nb energy routine;
changed line IF(.NOT.LEWALD) to IF(ELECFG.AND.
.NOT.LEWALD) in EVDW to fix SKIPE ELEC problem
pert/icpert.src -bug fixes
rxncor/rxndef.src -change RHEAP equivalence to pointer statement
util/selcta.src -change LHEAP equivalence to pointer statement
util/util.src -added dynamic heap and stack allocation to
INITHP and INISTK routines
CHARMM c23a1 (August 15, 1992) to c23a2 (October 25, 1992)
2.1 NIH Development
===========================================================================
CHMGR Note: 22-Oct-92
All the modifications described are integrated either with the tag
nih9208 or nih9209 in c23a2
===========================================================================
Name : Bernard R. Brooks
Email Address : brbrooks@helix.nih.gov
Institution : NIH
Date : August 18, 1992 and September 18, 1992
===========================================================================
File(s) Removed : fcm/intel.fcm
File(s) Added : fcm/parallel.fcm
fcm/surface.fcm
machdep/parallel.src
minmiz/tndriv.src
minmiz/tnpack.src
misc/aspener.src
(1) BUGFIX: Changes in CHARMM to fix a specific problem.
File(s) modified: gener/genpsf.src
gener/modpsf.src
image/nbondm.src
io/coorio.src
nbonds/nbondg.src
quantum/qmene.src
quantum/qmset.src
Description of each bug:
...........................................................................
gener/genpsf.src
This ancient bug manifests itself with a one residue segment where
both the first residue patch and the last residue patch try to modify
the same atom. The code currently adds both atoms (with the same name).
This fix causes the atom in the last patch to take precedence for
charges, but the van der Waal type is taken from the first patch
(please don't ask why). This bug fix is messy and is interwoven with
the charge augment feature described below. I recommend not fixing
this bug in old versions, just beware of it.
...........................................................................
gener/modpsf.src
This is another ancient bug which causes problems when the
first residue of a PSF is patched. The fix is very simple. This
problem should be fixed in older versions of CHARMM.
A1177 DO (J=JS,IBLO(INVMAP(I)))
changed to
B1169 DO (J=IBLO(INVMAP(I)-1)+1,IBLO(INVMAP(I)))
...........................................................................
image/nbondm.src
Standard FORTRAN does not allow a jump into an IF block
although most compilers allow this. This fix is needed to correct
the problems for compilers where this is not allowed. This has
been acomplished by breaking the large IF block into two smaller
ones.
A601 ENDIF
A602 255 CONTINUE
A603 IF(QPERT) THEN
changed to
B596 255 CONTINUE
...........................................................................
io/coorio.src
This bug is present because the parsing for keywords in coordinate I/O
is done before the parsing for atom selections, thus the atom selection
token RESId is eaten by the keyword parser (as the RESI option). Other
keyword/token conflicts are possible as well. The quick fix is to
move the atom selection up by roughly 20 lines. This problem should
be fixed in older CHARMM versions.
A64 IMODE=0
A65 CALL SELRPN(COMLYN,COMLEN,ISLCT,NATOM,1,IMODE,
A66 2 .FALSE.,1,' ',0,RESID,RES,IBASE,SEGID,NICTOT,NSEG,
A67 4 .TRUE.,X,Y,Z,.TRUE.,1,WMAIN)
A68 IF(IMODE.NE.0) CRAP-OUT
A69 C
deleted before
B63 IFILE=GTRMI(COMLYN,COMLEN,'IFIL',1)
B84 IMODE=0
B85 CALL SELRPN(COMLYN,COMLEN,ISLCT,NATOM,1,IMODE,
B86 2 .FALSE.,1,' ',0,RESID,RES,IBASE,SEGID,NICTOT,NSEG,
B87 4 .TRUE.,X,Y,Z,.TRUE.,1,WMAIN)
B88 IF(IMODE.NE.0) CRAP-OUT
inserted before
A91 C
...........................................................................
nbonds/nbondg.src
Standard FORTRAN does not allow a jump into an IF block
although most compilers allow this. This fix is needed to correct
the problems for compilers where this is not allowed. This has
been acomplished by breaking the large IF block into two smaller
ones.
A755 ENDIF
A756 255 CONTINUE
A757 IF(QPERT) THEN
changed to
B748 255 CONTINUE
...........................................................................
quantum/qmene.src
This modification reverts this code to an older version. I don't
know why the LVDW flag was disabled in the call to EGROUP in the
routine EVDWQM, but it has to be wrong. This change had the effect
of deleting the subroutine altogether. I suspect that this was done
as a short term fix after the space allocation was mangled in qmset.src.
A1114 $ .false.,.true.,lcons,lshft,lfswt,
changed to
B1108 C $ .false.,.true.,lcons,lshft,lfswt,
B1109 $ .false.,.false.,lcons,lshft,lfswt,
...........................................................................
quantum/qmset.src
As with the changes in qmene.src, the changes here make no sense.
The space allocation was non-optimal in the older version, but this
recent modification is just wrong! This corrects the problem,
but there may be more to this than is obvious.
A1018 CC nqmlen=natqm*(natom-natqm)
A1019 nqmlen=natqm*natom
A1020 inblos=allstk(integ4(natom))
A1021 jnbs =allstk(integ4(nqmlen))
A1022 inbloq=allstk(integ4(natom))
changed to
B1006 nqmlen=natqm*natqm
B1007 inblos=allstk(integ4(nqmlen))
B1008 jnbs =allstk(integ4(nqmlen))
B1009 inbloq=allstk(integ4(nqmlen))
---------------------------------------------------------------------------
(2) STANDARDS
These changes are usually simple, but reflect the need for the code
to adhere to coding standards. Changes in this section include;
Declaring all variables
Removal of unused variables
Removal of unused include files
Insuring calling sequences match the subroutine or function
that is called
Ending source files with an END statement
Ending common include files with a comment line
Fixing variable type mismatches between calling sequnces and
subroutines
Adding an "!" after a ##ENDIF statement (or other ## statements)
before comments
e.g. ##ENDIF ! BLOCK - is correct
##ENDIF BLOCK - should be avoided
Fully implementing the ##IF SAVEFCM feature
Commenting all variables within include files
Put all compiler directives within ##IF blocks.
Some additional comments for code readability and clarity.
Though I have fixed a number of problems, I have not yet had the opportunity
to do a complete crosscheck on this code. Additional mismatch problems and
unused variable problems may be uncovered.
Files modified:
main/iniall.src
energy/cenerf.src, enefscal.src, enefvect.src, enefst2.src
fcm/am1parm.fcm, impnon.fcm, quanta.fcm, quantm.fcm, shake.fcm
image/cnbndm.src
io/rtfio.src
machdep/machutil.src, startup.src
manip/cshake.src
minmiz/egrad1.src, nraph.src, steepd.src
nbonds/cenbf.src, cnbnd.src, enbfast.src, ewaldf.src
quantum/qmdata.src, qmene.src, qmjunc.src, qmpac.src, qmset.src
rxncor/rxndef.src
---------------------------------------------------------------------------
(3) FORTRAN Conversion
Files modified:
main/iniall.src, usersb.src
correl/anacor.src
dynamc/dynamc.src
energy/energy.src
---------------------------------------------------------------------------
(4) PARALLEL
General code for support of CHARMM on MIMD machines is now complete.
This includes; Control of the I/O levels for all file I/O. For parallel
machines or workstation clusters, only node zero performs I/O and it
broadcasts data to other nodes. Control of all printing through
PRNLEV or WRNLEV.
All compuationally intensive code exercised in MD is now fully parallel
which includes: DYNAMC, ENERGY (and most subsections), SHAKE, PRSSRE, DYNLNG,
IMAGES,... Almost all comutationally intensive code in the first order
minimizers is fully parallel. Other usage of the energy routines are
parallel (such as the energy time series in CORREL).
Many of the changes were required to remove the Intel specific nature
of the initial implementation. Most of the remaining changes deal with I/O
control. Please include I/O control in all future code that is developed.
Files modified:
main/charmm.src, iniall.src, miscom.src
correl/anacor.src, corrio.src, solana.src
dynamc/dcntrl.src, dynamc.src, dynamcv.src, dynio.src, dynlng.src,
prssre.src
energy/ecnstr.src, eintern.src, enefscal.src, enefvect.src,
energy.src, eutil.src
fcm/dimens.fcm, heap.fcm, parallel.fcm(NEW), stack.fcm,
intel.fcm(REMOVED)
gener/hbonds.src, update.src
graphics/apodraw.src, drawit.src
image/crystal.src, eimage.src, imagio.src, nbondm.src, xtlfrq.src
io/coorio.src, mainio.src, parmio.src, psfres.src, rtfio.src, univio.src
machdep/machio.src, parallel.src(NEW), startup.src
manip/cstran.src, intcor.src, rmsdyn.src, shake.src, wrgaus.src
minmiz/egrad1.src, nraph.src
misc/drawsp.src, noe.src, sbound.src, testch.src, xray.src
molvib/molvco.src, molvio.src
nbonds/enbond.src, etable.src, heurist.src, nbondg.src, nbonds.src
pert/epert.src
rxncor/path.src, rxndef.src, travel.src
util/datstr.src, parse.src, title.src
vibran/rbquas.src, redbas.src, vibcom.src, vibio.src, vibsub.src
---------------------------------------------------------------------------
(5) ATOMPARSE
A new atom name parsing subroutine has been developed. All commands
in CHARMM which use atom parsing now use this new utility which provides a
consistency in the parsing of atom names. This is done in a fashion that
all existing input scripts will continue to function. This makes the code
simpler and facilitates further advancements in atom parsing. One new
feature allows an atom selection to be used to select a series of atoms.
This is very useful in CORREL for specifying clusters of atoms for analysis.
When the atom selection feature is used to specify 4 atoms of a dihedral,
the first 4 selected atoms will be chosen.
Files modified:
correl/correl.src
fcm/exfunc.fcm
manip/cstran.src, intcor.src
util/chutil.src
vibran/vibutil.src
---------------------------------------------------------------------------
(6) MSCNUM
New code for flexible miscellaneous command substitutions has been
fully incoporated. Additional types were needed to make this code more
flexible. Three types are supported, REAL(*8), INTEGER, CHARACTER.
There are three subroutines which can be called; integer (SETMSI),
character (SETMSC), and real (SETMSR) to specify a command substitution
variable. Now it is possible for ?NATOM to return an integer, ?RSM to
return a real number, and ?SEGID to return the segment identifier of the
first selected atom.
Files modified:
main/charmm.src, iniall.src, miscom.src
correl/corrio.src
fcm/mscpar.fcm
machdep/machutil.src
manip/corman.src, corman2.src, rgyr.src, rmsdyn.src, rotlsq.src,
scalar.src
misc/quicka.src
util/parse.src, selcta.src
vibran/thermo.src, vibio.src
---------------------------------------------------------------------------
(7) EULER Dynamics Integrator
The incorporation of of the Langevin/Implicit Euler dynamics
integrator has been achieved. This code is currently being evaluated for
suitability in biomolecular simulations and in free energy studies.
The effect is to remove the energy in the high frequency degrees of
freedom which eliminates the noise in free energy studies where bonds are
being modified. This effort is the first method of several that will be
explored (the next will eliminate not only high frequnecy energy and high
frequency heat capacity as well, and still allow a flexible system).
To support the Implicit Euler integration, a Truncated Newton Minimizer
has been added. This minimizer may be used directly using the MINI TN
command. The minimizer is not yet fully implemented (it works, but is
not as efficient as it will be), but it is already very competitive
relative to existing minimization methods. Do not use TN with SHAKE.
This code has been developed by Tamar Schlick at NYU. It has been
integrated within CHARMM with some modifications. Testing and evaluation
are ongoing. Further modifications will be forthcoming.
Files modified:
main/iniall.src
dynamc/dcntrl.src, dynamc.src
energy/eutil.src, printe.src
fcm/energy.fcm, euler.fcm, number.fcm
minmiz/minmiz.src, tndriv.src, tnpack.src
pert/epert.src
---------------------------------------------------------------------------
(8) Histogram Option in CORREL
A histogram option to time series manipulation has been developed.
This is executed by the command;
MANTime time-series-name HISTogram min-value max-value num-steps
The selected time series is replaced with a histogram which contains
the probability of finding the time series within a given value range.
Files modified:
correl/correl.src, mantim.src
---------------------------------------------------------------------------
(9) EHFC: High Freequency Correction
The leap-frog dynamics integrator has been modified to have an
improved high frequency correction (HFC) term. With the old term,
energy was conserved within a harmonic degree of freedom, but total energy
would drift as energy exchanged between high and low frequency degrees of
freedom. The new code avoids this problem. As per request from several
people, the total energy and kinetic energy that is printed in the first
line of dynamics energy printout has reverted to the standard Verlet
energies, and these match the output of the old integrator. The HFC terms
(total energy, and kinetic energy) are now printed on the second line.
The fluctuation of the HFC total energy is usually an order of magnitude
smaller than that of the total energy. The HCF total energy is a good
indicator of problems with NVE dynamics because small changes in total
energy are not lost in the noise of high frequency oscillations.
Files modified:
dynamc/dynamc.src
energy/eutil.src, printe.src
fcm/energy.fcm
---------------------------------------------------------------------------
(10) QUANTUM - Changes to the QUANTUM code
The QUANTUM code has been modified extensively (also see bug report
above) to bring the code into CHARMM standards. Several hundred variables
were declared and the IMPLICIT statements were removed. These variables
were declared with minimal checking of variable names for spelling errors.
Naming conflicts involving the variable NALPHA were corrected.
Files modified:
energy/energy.src
fcm/quantm.fcm
quantum/qmdata.src, qmene.src, qmjunc.src, qmpac.src, qmset.src
---------------------------------------------------------------------------
(11) SURFACE
New analytic surface area code and energy terms for ASP (Atomic
Solvation Parameters) energy and forces have been fully integrated
(and parallelized for multi-machines). This has been achieved by the
incorporation and adaptation of the code from Wesson and Eisenberg.
The default for the COOR SURFace command is now the analytic surface
area. The older Lee and Richards algorithm is still available if a
nonzero accuracy is requested.
Files modified:
main/iniall.src
energy/energy.src, eutil.src, printe.src
fcm/energy.fcm, surface.fcm
io/mainio.src
misc/aspener.src, surfac.src
---------------------------------------------------------------------------
(12) CRAYVEC Preprocessor Keyword
The CRAY code that has been recently incorporated has been worked on.
In addition to fixing several problems, the code has been made general for
use on non-CRAY architectures and all features are tested with the NIH test
cases that check all possible nonbond options with all of the FAST versions.
The new keyword CRAYVEC has been added which controls the compilation of
the optimized CRAY code. The required changes to energy/energy.src,
nbonds/enbond.src, and nbonds/nbonds.src in order to support the optimized
CRAY code has been significantly simplified and localized. The variable
LMACH=3 now selects the fast CRAY code. Flags (## type) which control
FAST usage (in order of priority subject to change) now include;
SCALAR
VECTOR
PARVECT
CRAYVEC
All four sets of fast options are supported in a general manner. It will
now be possible to test and compare these methods for performance on a wider
variety of machines.
The code has also been brought up to CHARMM standards. Specific
problems fixed included;
Fix number of argument mismatches (2 occurances: one in energy.src
and one in cenerf.src)
Fix argument mismatch (2 occurances of REAL*8 to REAL in cnbnd.src)
Declaration of several undeclared variables.
Incorporation of ##IF UNICOS blocks around cray compiler directives
Addition of generic code for non-cray architectures to simulate
boolean operations.
Addition of generic subroutines for non-cray architectures to simulate
cray system calls.
Additional problems have not been dealt with (such as the replicated
computation of R**2 in cnbnd.src, once in subroutine CPAIRS and once in
main routine).
Files modified:
main/iniall.src, miscom.src,
energy/cenerf.src, energy.src
image/cnbndm.src, upimag.src
nbonds/cenbf.src, cnbnd.src, enbond.src, nbonds.src
---------------------------------------------------------------------------
(13) EFFICIENCY
A few changes to enhance efficiency were made. Changes to the image
code were made for speed. The changes to io/psfres.src were made to save
a considerable amount of memory.
Files modified:
image/nbondm.src, upimag.src
io/psfres.src
---------------------------------------------------------------------------
(14) GRMS
The gradient root mean squared (GRMS) calculation that was recently
added to energy.src has been removed. Instead, two GRMS sections have been
added to the old integrator which accomplishes the same task without the
cost for other uses of the energy routine.
Files modified:
dynamc/dynamcv.src
energy/energy.src
---------------------------------------------------------------------------
(15) ##IFN NOXX Preprocessor Keys
The full implementation of the ##IFN NOxxx was not completed in the
earlier version. Several problems have been encountered and fixed as new
combinations have been tried. The ##IFN NOxxx feature has been implemented
to save memory on non-virtual memory machines so that larger molecular
systems can be employed.
Files modified:
main/charmm.src
dynamc/dynio.src
gener/psfsum.src
io/mainio.src
minmiz/egrad1.src
misc/testch.src
---------------------------------------------------------------------------
(16) PRNLEV
In some of the recent changes, the definition of the PRNLEV variable
was modified, but not for all of CHARMM. The use of PRNLEV has been reset
to its previous definition for now. The redefinition of PRNLEV will be made
as the code gets converted to FORTRAN for the entire program.
The definitions are;
OLD NEW
no print version 0 -1
very brief version 2 0
brief version 3 1
inner loop default 4 4
default 5 5
verbose 6 6
very verbose 7+ 7+
The new PRNLEV changes have not yet been incorporated.
Files modified:
main/iniall.src
dynamc/dynamc.src
image/crystal.src
machdep/parallel.src
minmiz/conjug.src, nraph.src, powell.src, steepd.src
nbonds/heurist.src
util/parse.src
---------------------------------------------------------------------------
(17) QAUGMENT
It is desirable for a patch to be able to augment the charge of an atom.
The current code could only set a charge. The new code can add or subtract
a value from the charge. This is done by using a patch charge value near
100.0. For example, a charge of 100.15 will add 0.15 to the current charge.
A charge value of -101.0 will subtract 1.0 from the current charge. Charge
values less than -90.0 or larger than 90.0 are no longer allowed for generate
or patch without charge augment. This feature is coupled with the bugfix in
genpsf described above. It allows more flexible patches to be developed where
the prior charge on modified atoms need not be known.
Files modified:
gener/genpsf.src, modpsf.src
---------------------------------------------------------------------------
(18) SPACEALLOC: Space Allocation
A few modification have been made with the intent of reducing the
amount of data structure resizing. These are straightforward.
Files modified:
image/upimag.src
nbonds/nbexcl.src
---------------------------------------------------------------------------
(19) VACUUM_OP: COOR SEARCH Subcommand
The ability to manipulate pixel bitmaps generated from the COOR SEARCH
command has been developed. The new syntax for the COOR SEARCH command is;
COOR SEARch {PRINt [UNIT int]} { } {[VACUum]} {[RESEt]} [SAVE]
{[NOPRint] } {[RCUT real]} { FILLed } { AND }
{[RBUFf real]} { HOLES } { OR }
{ XOR }
The new keywords are;
SAVE - save the resultant bitmap for subsequent operations
AND - logical AND the new bitmap with the previously saved map
OR - logical OR the new bitmap with the previously saved map
XOR - logical XOR the new bitmap with the previously saved map
HOLES - search for holes (vacuum points surrounded by filled points)
Files modified:
manip/corman.src, corman2.src
---------------------------------------------------------------------------
(20) VERNUM: Version Number
The version number (as a float value) and version name have been added
to version.fcm. This allows subversions to be specified and defined simply.
Files modified:
main/iniall.src
fcm/version.fcm
===========================================================================
(21) OTHER MINOR CHANGES:
This section deals with simple changes which involve just
one file and which don't fit into one of the above categories.
...........................................................................
dynamc/dcntrl.src - (mod to QORIG)(new NPREv keyword)
The new NPREv keyword allows an MD run trajectory to start on other
than step 1. This is useful if one wants to do a "restart" without using
a CHARMM restart file.
The modifications to the logical flag QORIG now specify the new
leap-frog integrator as the integrator to use when one or more of the
following conditions is met;
The NIH keyword is used during compilation
The PARALLEL keyword is used during compilation
The PERT keyword is used during compilation
The OLDDYN keyword is not used during compilation
The LEAP keyword is used (leap-frog)
The NEW keyword is used
The CPT keyword is used (Constant Temperature and/or Pressure)
The EULEr keyword is used (Implicit Euler integration)
The new integrator should be the general default, but I'll let someone
else make that push.
...........................................................................
machdep/space.src - (don't declare library routines as external with UNIX)
The library routines FMALLOC and IADDR do not need to be declared
as external with UNIX. On some systems, the explicit declaration disables
the usage of the library routines.
...........................................................................
machdep/machutil.src - (timing calls modified)
The variable ELATM (not to be confused with ELATIM) has been added
to give correct elapsed times. The current code has a minor problem.
...........................................................................
manip/intcor.src - (new delete or keep for impropers)
A new keyword, IMPRoper, has been added to the IC KEEP and
IC DELEte commands. This allows internal coordinate entries
to be selected based on whether it is an improper type of IC entry.
This keyword works in conjuntion with atom selections.
...........................................................................
manip/scalar.src - (SCALar STAT format)
The format for print in the SCALar STAT command has been
changed from (F12.6) to (1PG12.6) so that very large and very
small property values can be displayed.
...........................................................................
graphics/graphx.src - (Titles for CHARMM graphics)
A general titling facility has been added to CHARMM graphics.
This graphics facility is fully supported on Apollo, partially
supported on SGI, and will soon be supported on IBM-RS and
HPUX (HP730s).
...........................................................................
pert/epert.src - (partial ST2 implem.)
I have started to implement the ST2 water model with the PERT
code. This implementation is not complete, but the new code is
nonintrusive.
...........................................................................
util/selcta.src - (new key numbers and names)(new tokens)
One new operator; .BYGROUP.
And one new token; IGROup :
have been added to allow the selection of atoms based on electrostatic
groupings.
Several keynames have been added to allow the query of the
characterstics of selected atoms;
?SELATOM - number of first atom selected
?SELIRES - number of first residue selected
?SELISEG - number of first segment selected
?SELTYPE - name of first atom selected
?SELRESI - resid of first residue selected
?SELSEGI - segid of first residue selected
?SELRESN - residue type of first atom selected
?SELCHEM - chemical type of first atom selected
These new keywords are in addition to the existing keyword;
?NSEL - Number of atoms selected
...........................................................................
vibran/vibio.src - (card format NMD files)
A card NMD (normal mode) file is now supported. This allows
normal mode files to be moved more easily between machine types.
...........................................................................
io/coorio.src - (preselection before data conversion)
When reading coordinates in card format in CREAD, the values of
the coordinates for a trial atom are now not decoded until the target
atom has been identified. This allows coordinate files with a few
very poorly placed atoms (e.g. '**********') to be read as long as the
offending atom is not selected.
===========================================================================
2.2 MSI Development
===========================================================================
CHMGR Note: 22-Oct-92
All the modifications described are integrated with the tag msi9210
===========================================================================
Name : Ryszard Czerminski
Email Address : ryszard@connection.prospect.com
Institution : Molecular Simulations, Inc.
Date : October 10, 1992
===========================================================================
(1) Bugfixes
correl/solana.src
- READV changed to CREADV
gener/update.src
- WRNLEV decreased to 0 for missing parameters
images/images.src
- in IMATOM call to SPLITI introduced this fixes the bug not
allowing to use more then single digid in image patches atom names;
now the only limitation is caused by the fact that atom names are CHAR*6
(with two characters for number, this is why also MXNPA=100 is replaced
by MXNPA=99
- IPRES(9),IMPRES(9) replaced by IPRES(MXNPA),IMPRES(MXNPA)
- MXNPA=100 replaced by MXNPA=99 (we have only 2 characters for number
before atom name)
util/string.src
- ports (ardent, hpux, ibm)
input/vibran.inp
- second OPEN WRITE CARD UNIT 7 NAME @9FOR007.DAT commented. It is
producing "OPEN" not possible - error message on hp.
(2) FORTRAN conversion
dynamc/dynio.src
image/crystal.src, eimage.src, images.src, imagio.src
io/mainio.src, parmio.src, psfres.src, rtfio.src, trajio.src, univio.src
nbonds/heurist.src
Folowing files are waiting to resolve differences with Bernie:
dynamc/prssre.src
image/upimag.src
io/coorio.src
(3) COMPILATION PROBLEMS:
quantum code has to be compiled with -static on sgi and -K on hp
===========================================================================
2.3 Hanyang University Development
===========================================================================
CHMGR Note: 22-Oct-92
All the modifications described are integrated with the tag hyu9210
===========================================================================
Name : Youngdo Won
Email Address : won@gaya.kreonet.re.kr
Institution : Hanyang University
Date : October 22, 1992
===========================================================================
(1) FORTRAN Conversion
dynamc/prssre.src, trnphi.src
energy/ecnstr.src, enst2.src, eutil.src, intere.src, printe.src
machdep/machio.src, space.src, startup.src
main/miscom.src
manip/pucker.src, rgyr.src, scalar.src, shake.src
misc/drawsp.src, hbuild.src, sbound.src, surfac.src
util/array.src, datstr.src, parse.src, random.src, util.src
(2) HPUX Porting
The build/hpux directory is added. install.com and test22.com are
modified. ar option -ruvco is changed to -ruvc in .mk files and
install.com. The makemod script is modified to generate the modified .mk
files. Makefile_hpux is written. Not all testcases are checked out;
especially note that maatest runs into an infinite loop.
|
| | |
|
3. About c23f Developmental Version |
|
|
|
Most of FLECS source files are converted to FORTRAN. Files before and
after conversion are gathered in tammy.harvard.edu:/prog/chmgr/fortraned
directory, which contain the same subdirectories as charmm/source.
'before' files have an extension designating the version number (e.g.,
g2, a1, etc.) and 'after' files are taged with the developer's initial
(brb for Bernie Brooks, rcz for Ryszard Czerminski, ydw for Youngdo Won,
sf for Stefan Fischer, djt for Doug. Tobias, td for Tom Ngo, etc.)
As of the c23f release, graphics/apodraw.src, apograph.src, drawit.src,
graphx.src and nbonds/nbndgc.src are still in FLECS.
Graphics routines will be replaced with a FORTRAN version SGI modules.
Tom Ngo has deposited nbndgc.src in FORTRAN, however it contains new
files and will be put into c23f1. At the February 13th CHARMM meeting,
we decided to keep the charmm/tool/flexfort.f translator for the time
being.
All DEC/CMS headers are removed from source files and the INCLUDE
directory designator charmm_22 is replaced by charmm_23. We put
"CHARMM Element ..." on the first line of each each file.
CHARMM c23a2 (October 25, 1992) to c23f (March 1, 1993) and
c23f1 (March 15, 1993)
4.1 New .mk files and Makefiles
When new modules are added, we must add the corresponding module
make files (such as solvation.mk). In this version, we simplify
machine specific makefiles by using RANLIB macro definition in
Makefile_{machine_type}. ranlib is needed by convex, ibmrs, sun
and stardent platforms and not needed by cray and sgi machines.
For hpux, it is no longer required in recent OS versions.
4.2 Tom Ngo's BYCUbe Nonbond List Generation Method
source/nbond/nbndgc.src is converted to FORTRAN. During the
conversion Tom made an enhancement in speed by using ZTBL.
Files Modified:
doc/nbonds.doc
- Added docs for BYCUbes and ZTBL representation.
source/fcm/dimens.fcm
- Added variables to support ZTBL representation of exclusion list.
The ZTBL representation of the exclusion list is a data structure
that allows you to find out whether a given pair of atoms is
1-2, 1-3, or 1-4, in a manner that is much more efficient than
traversal of the exclusion list as it is currently implemented
(JNB/INBLO).
Dimens.fcm now contains two extra constants, MAXZ14 and MXZ14G,
the maximum sizes of the atom-atom and group-group ZTBL lists,
respectively.
source/fcm/psf.fcm
- I added Z14 and Z14G, which are the ZTBL forms of the atom-atom and
group-group exclusion lists, respectively. Additionally, there
are a number of supporting variables with names that start with
Z14 and Z14G. The contents of these tables are generated in
MAKINB and MAKGRP.
source/machdep/machutil.src
- Added Convex-specific code and corrected some spellings.
The Convex library routines etime() and dtime(), at least as of
ConvexOS 9.1, contain bugs that cause them to add 0.25s to the
apparent CPU time, sporadically. I added Convex-specific code that
calls the library routine cputime() instead.
source/nbonds/enbond.src
- Corrected a bug in calculation of RVAL.
The old code for estimating the size of the group-group nonbonded
interaction list (JNBG) gave figures that are overestimates by a
factor of (NATOM/NGRP)**4. On any machine, this wasted memory, and
one some machines, this caused out-of-memory errors. In consultation
with Bernie Brooks, I rewrote the line, adding an extra buffer that
accounts for the the non-infinitesimal size of groups.
source/nbonds/nbexcl.src
- MAKINB and MAKGRP now generate the ZTBL rep'ns of the exclusion list.
Added some ZTBL-specific arguments to the end of the MAKINB and MAKGRP
argument lists. This code also produces some user feedback that
shows the degree of compression achieved, if PRNLEV.GE.6.
source/nbonds/nbndgc.src
- Converted to FORTRAN and overhauled to use ZTBL representation.
This version of NBNDGC.SRC is quite different from the old one.
It is all in FORTRAN, and it should be somewhat faster. Its speed
can easily be doubled by removing a Quicksort phase, which
unfortunately is required by the present energy routines.
File Added:
source/util/ztbl.src
- Brand-new file. Support for ZTBL representation.
This file contains routines that construct the ZTBL representation of
the exclusion list and access it.
4.3 Georgios Archontis' RISM integration
===========================================================================
Name : Georgios Archontis
Email Address : georgios@tammy.harvard.edu
Date : February 11, 1993
===========================================================================
Integration of the RISM module into CHARMM
The RISM module enables the calculation of the site-site
pair correlation functions c(r) and radial distribution
functions g(r) for a multi-component molecular system. It
can also calculate the chemical potential of solvation of
a solute species into a solvent and decompose this potential
into the energy and entropy of solvation.
The module is invoked from a standard CHARMM input file
using the command "RISM". This command transfers control to
the subroutine RISM (work/rism.src). The subroutine RISM
uses its own character*4 parser to recognize two kinds of
commands:
1) the familiar CHARMM commands GOTO, LABEL, OPEN, CLOSE,
STREAM, RETURN, SET, DECR, INCR and IF [EQ NE GT GE LT LE]
Currently only the commands OPEN CLOSE STREAM RETURN call
the appropriate CHARMM routines, whereas the other commands
call routines in the directory solvation. The main CHARMM
parser can be used for all the above commands, and this is
to be done in the near future.
2) the RISM-specific commands READ, WRITE, SETUP, EDTZM,
STATE, ITERATE, DERIVATIVE, SOLVATION, ANALYSIS
Exiting the RISM module is achieved by the command STOP
==========================================================================
Files modified
--------------
main/charmm.src - call to RISM added
main/iniall.src - initialization of RISM-specific variables added
Files added
-----------
solvation/rism.src - controls the execution of the RISM-related
commands by calling the appropriate routines.
solvation/charli.src - manipulates character strings.
solvation/coorman.src - generates the intramolecular matrix and the
structure.
solvation/cycles.src - controls the iteration cycle that solves
the integral equation
solvation/deriv.src - solvent density response
solvation/fft.src - controls the fourier transform of the
integral equation.
solvation/imsld.src - FFT transformation routines (for other
machines than CONVEX) and matrix inversions.
solvation/rismio.src - reads/writes RISM related input and output
and generates pointers for the arrays.
solvation/soluu.src - solute-solute calculation
solvation/soluv.src - solute-solvent calculation
solvation/solvv.src - solvent-solvent calculation
solvation/solvation.src - calculates thermodynamics
solvation/state.src - dielectric correction, analysis of the pair
correlation functions.
fcm/charli.fcm - declares RISM-related functions
fcm/control.fcm - iteration variables
fcm/distri.fcm - RISM related arrays for the distribution functions
and pointers
fcm/fft.fcm - FFT variables
fcm/rism.fcm - dimensions of RISM arrays
fcm/rismio.fcm - character variables
fcm/struc.fcm - internal parameter arrays, residue and segment
names, coordinates.
doc/rism.doc - RISM documentation
test/c23test/rism.inp - Input test file for RISM calculations.
===========================================================================
4.4 Masa Watanabe's Dynamics Code Development
===========================================================================
Name : Masa Watanabe
Email Address : watanabe@tammy.harvard.edu
Date : February 16, 1993
===========================================================================
1. VELOCITY VERLET INTEGRATOR
Two integrator (Verlet and Leap-frog) methods presented in CHARMM have
own flavors, but Verlet method are handling velocities rather awkward,
and the form of the integrator may introduce some numerical imprecision.
On the other hand Leap-frog minimizes loss of precision on a computer,
but it do not handle the velocities in a completely satisfactory manner.
Velocity Verlet integrator can store positions, velocities, and
accelerations all at the same time and minimizes round-off error.
Files modified:
dynamc/dcntrl.src
Files added:
fcm/tbmts.fcm
dynamc/dynamvv.src : Velocity Verlet method
----------------------------------------------------------------------------
2. NOSE-HOOVER CONSTANT TEMPERATURE METHOD
[Based on S. Nose, JCP 81, 511 (1984) and W.G. Hoover, Phy. Rev. A 31,
1695 (1985)]
This is an another type of constant temperature method, but
an equilibration time in the vicinity of the desired temperature is
faster than other routines which are available in CHARMM. Also
multi-temperature controls are also developed. Different parts or
molecules in the different heat baths in order to equilibrate the system
faster and keep the system in the desired temperature well. This method
can now works in Verlet and Velocity Verlet integrators.
There are two different way to call Nose-Hoover method during dynamics
run. One way to call it simply types:
DYNA NOSE QREF desired heat bath dumping TREF desired temperature -
Files modified:
dynamc/dcntrl.src
dynamc/dynamcv.src
dynamc/dynamvv.src
dynamc/dynio.src
main/charmm.src
main/iniall.src
Files added:
fcm/nose.fcm
dynamc/nose.src
-----------------------------------------------------------------------------
3. MULTIPLE TIME-SCALED METHOD
[Based on Tuckerman, Berne, Martyna, JCP 97, 1990 (1992)]
Tuckerman et al proposed a reversible RESPA algorithm recently. Previous
MTS methods had the disadvantages lied in the loss of accuracy due to
the approximation of holding the slow variables fixed while integrating
the equations for the fast variables. But in this reversible RESPA
equations of motions were derived from Liouville operators and Trotter
theorem. This method give more accurate dynamics than previous methods.
In this method, you can specify up to three different time steps in
dynamic simulation run.
Files modified:
fcm/bases.fcm
fcm/tbmts.fcm
dynamc/dcntrl.src
dynamc/dynamvv.src
image/nbondm.src
image/upimag.src
nbonds/nbondg.src
nbonds/nbonds.src
main/charmm.src
main/iniall.src
Files added:
fcm/tbmtsp.fcm
energy/enemts.src
image/emts.src
===========================================================================
4.5 MMFP and NMR codes from Benoit Roux
MMFP (Miscellaneous Mean Field Potential) Code added
Files modified:
doc/commands.doc
source/main/charmm.src
source/energy/energy.src
source/energy/printe.src
File added:
doc/mmfp.doc
source/misc/mmfp.src
source/fcm/mmfp.fcm
NMR analysis commands added
Files modified:
doc/commands.doc
source/charmm.src
Files added:
doc/nmr.doc
source/misc/nmr.src
source/misc/zmat.src
===========================================================================
4.6 Bugfixes
(1) source/correl/correl.src
- Bernie R. Brooks
Unused lable 105 removed
(2) source/dynamc/prssre.src
- Bernie R. Brooks
VXY to VXZ in the EPRESS(PIZX) term
(3) source/gener/update.src
- Stefan Fischer
(IMGFRQ.GT.0) is fixed by (IBMGRQ.NE.0)
(4) source/io/trajio.src
- Bernie R. Brooks
##IFN NOIMAGES code fixed
(5) source/io/univio.src
- Bernie R. Brooks
Name translation bug fixed: .OR. to .AND.
(6) source/machdep/cstuff.c
- Youngdo Won
HPUX definition added to access getunamef
(7) source/manip/fshake.src
- Stefan Boresch
##IFN NOST2 code fixed
(8) source/manip/scalar.src
- Bernie R. Brooks
1PG14.6 replaces 1PG12.6
(9) source/manip/shake.src
- Doug. J. Tobias
SHKCRA called added (fixed) for Cray version
(10) source/manip/wrgaus.src
- Youngdo Won
FMT extraneous blank error fixed
(11) source/minmiz/conjug.src
- Ryszard Czerminski
Divided by zero bug fixed
(12) source/misc/aspener.src and source/fcm/surface.fcm
- Bernie R. Brooks
Replaced with the most recent NIH version
(13) source/misc/noe.src
- Bernie R. Brooks
Parsing bug fixed
(14) source/nbonds/enbfast.src
- Bernie R. Brooks
A nonstandard SKIP bug fixed
(15) source/pert/pert.src
- Stefan Boresch
##IF NOMISC code fixed
(16) source/rxncor/rxndef.src
- Stefan Boresch
##ENDIF moved
4.7 New testcase for CHARMM version 23 (test/c23test)
(1) test/c23test/mmfptest.inp
- Benoit Roux
Miscellaneous Mean Field Potential testcase
(2) test/c23test/mtsm1.inp
test/c23test/mtsm2.inp
- Masa Watanabe
Multiple Time Scale dynamics testcases
(3) test/c23test/nmrtest1.inp
test/c23test/nmrtest2.inp
test/c23test/nmrtest3.inp
test/c23test/zmat.inp
- Benoit Roux
NMR Utility testcases
(4) test/c23test/nose1.inp
test/c23test/nose2.inp
- Masa Watanabe
Nose thermostat testcase
(5) test/c23test/rism.inp
- Georgios Archontis
RISM calculation testcase
CHARMM c23f1 (March 15, 1993) to c23f2 (August 15, 1993)
Fixes cumulated during the period of March 15, 1993 and August 1,
1993 are incorporated into this release. In addition to bugfixes, we
add supports for DEC ALPHA running under OSF (identified by 'alpha' for
the installation), DEC stations running under Ultrix (dec), MacIntosh
System 7 (not installable by install.com), and PARALLEL architecture
platforms. Problems with RISM and MTS code have been fixed. We have
incorporated the current NIH PARALLEL code and Leo Caves' REPLICA
program into the current version. The following details the
modifications done over c23f1 version.
5.1 New .mk files and build procedure
There are files removed and added in this releases:
Files/Directories Removed:
build/cray/
build/convex/
build/hpux/
build/ibmrs/
build/sgi/
build/stardent/
build/sun/
source/fcm/charli.fcm
source/fcm/external.fcm
source/fcm/rismio.fcm
source/solvation/charli.src
source/solvation/imsld.src
Files Added:
build/UNX/
source/fcm/replica.fcm
source/gener/replica.src
source/graphics/grutil.src
source/graphics/iris_graphics.c
source/machdep/parcomm.src
source/machdep/socket.c
source/manip/corman3.src
source/quantum/addlnat.src
source/quantum/qmnbnd.src
test/c23test/replica.inp
Note that build/{platform} directories are replaced by build/UNX, which
is the generic UNIX build directory. UNX supports Cray, Convex, HP-UX,
IBM RS/6000 AIX, SGI, Stardent, DEC ALPHA OSF, DEC Ultrix, and SUN.
The installation procedure, install.com, is modified to handle the
unified build directory.
The module makefiles are adjusted for the removal/addition in the
source code. We use the following set of macros for certain source
files. For all other files, $(FC2) is used.
UNX/energy.mk
$(FCR) enefst2.f
$(FCR) enefst2q.f
$(FC3) enefvect.f
UNX/image.mk
$(FCR) imnbf2p.f
$(FC3) imnbfp.f
$(FC0) nbondm.f
UNX/manip.mk
$(FC3) fshake.f
$(FCR) fshake2.f
UNX/nbonds.mk
$(FCR) enbf2.f
$(FCR) enbf3.f
$(FCR) enbf4.f
$(FCR) enbf5.f
$(FC3) evdwf.f
$(FC3) ewaldf.f
$(FCR) ewaldf2.f
$(FCR) nbndf2p.f
$(FC3) nbndfp.f
UNX/quantum.mk
$(FC0) addlnat.f
$(FC0) qmdata.f
$(FC0) qmene.f
$(FC0) qmjunc.f
$(FC0) qmnbnd.f
$(FC0) qmpac.f
$(FC0) qmset.f
5.2 ALPHA and DEC versions
- Xiaoling Liang (Harvard) and Mary E. Karpen (CMU)
c22f1 is ported to a Digital ALPHA platform 3000/500 running OSF/1
V1.2 and Dec FORTRAN 3.3.1 and to a DEC 5000/240 station running under
Ultrix.
Files modified:
fcm/number.fcm
machdep/machutil.src, machio.src, startup.src, space.src
Files added:
build/UNX/Makefile_alpha, build/UNX/Makefile_dec
5.3 NIH Developments
- Bernard R. Brooks
Email Address : brbrooks@helix.nih.gov
Institution : National Institutes of Health
Date : April 23, 1993 and August 5, 1993
This report covers corrections of c23f1 and further software
development for the period of March 15, 1993 to August 5, 1993.
(1) Macintosh (System 7) code added
Files modified:
dynamc/dynamc.src
dynamc/dynamcv.src
dynamc/dynamvv.src
fcm/iochan.fcm
fcm/number.fcm
machdep/machio.src
machdep/machutil.src
machdep/space.src
machdep/startup.src
quantum/qmdata.src
quantum/qmset.src
(2) QUANTUM Code Corrections
1) April 23, 1993
There were several problems with the quantum code that have been
fixed. The QM/MM code has been worked on to clean up several
remaining major problem. New skip flags for energy subcomponents in
the QM and QM/MM potential have been introduced. We added the Bohr
radius (needed for consistency in QM code and to provide agreement
with MOPAC). We fixed the problems in the calls to MECI, to SS, and
the variable miscmatch problems for CUC. This code still gives
different answers on different machines (the problems are in the QM
section). This code is being replaced in the near future, but the
fixes here are an definite improvement. Note: when using TRAVEL code
with the QM/MM code, make sure you increase the DELTa value.
Files modified:
main/charmm.src
main/iniall.src
energy/eutil.src
energy/printe.src
quantum/addlnat.src
quantum/qmdata.src
quantum/qmene.src
quantum/qmjunc.src
quantum/qmpac.src
quantum/qmset.src
fcm/consta.fcm
fcm/quantm.fcm
2) August 5, 1993
A major bug in the QM/MM code has been found and corrected. The
van der Waal group nonbond list was missing elements due to an improper
interpretation of the group-group exclusion list in CHARMM (It's a two
state list, not a 3 state as in the atom-atom exclusion list). All
vdw interactions between QM and MM group where any QM atom had an
exclusion or a 1-4 interaction with any MM atom was not computed.
This caused major problems in certain situations where there was a
strong electrostatic attraction with no compensating vdw interaction.
The bug was found when dealing with the HIV protease where we found
energy changes of several hundred Kcal/mol during a minimization
procedure. New code to add link and place link atoms has been written.
Files modified:
main/charmm.src
quantum/qmset.src
util/chutil.src
nbonds/nbutil.src
Files added:
quantum/addlnat.src
quantum/qmnbnd.src
(3) Graphics
Graphics code converted to FORTRAN and overhauled. Versions that
work with Xwindows and GL are in progress. All remaining FLECS code
removed. A new preflx keyword, NODISPLAY, builds a version which
produces HPGL, PLUTO FDAT, and LIGHT .atm files without requiring any
screen display capabilities. The SG (IRIS) code incorporation is
relatively untested. Postscript file output similar to HPGL (but much
nicer looking, hopefully) is also planned. The X11 development will
begin shortly, using a model similar to original Apollo graphics.
Files modified:
graphics/apodraw.src APOLLO
graphics/apograph.src APOLLO
graphics/inter.src APOLLO
graphics/ugap.src APOLLO
graphics/drawit.src
graphics/graphx.src
fcm/graph.fcm (referenced in main/iniall.src, manip/corman.src)
main/charmm.src
dynamc/dynsub.src
Files added:
graphics/grutil.src (new file)
graphics/iris_graphics.c (new file) IRIS
(4) Leap-Frog Integrator
I have continued to develop the leap-frog integrator in CHARMM. I
have modified Berendsen's method so that it will work for very small
systems and for very weak coupling constants. I have added a langevin
implicit Euler scheme (from Tamar Schlick), I have corrected the high
frequence correction code printout. Now it is possible to use SHAKE
with CPT and get correct pressures and temperatures.
Another change is that I now calculate the change in potential
energy due to the constant pressure algorithm. The energy lost due to
the changes in box size is now added to the kinetic energy during the
constant temperature procedure. This allows the constant presure
code to nearly conserve energy and allows the constant temperature
code to be used with weak coupling times. This correction was made
when we found that water box simulations with the Berendsen's method
were running about 10 degrees too cold when both temperature and
pressure coupling times of 1ps were used. Now the correct target
temperature is achieved, even in the limit of very weak couplings.
Files modified:
dynamc/prssre.src
dynamc/dynamc.src
dynamc/dcntrl.src
main/iniall.src
(5) New COOR DIST command
The COOR DISTance command has been overhauled and has additional
features. One such feature is the ability to get g(r) plots from
trajectory files using atom selections. It has several other
features. The new syntax is:
COOR DISTance
{ WEIGhting vector-spec atom-selection }
{ }
{ [UNIT int] [CUT real] [ENERGy [CLOSe]] 2X(atom-selection) - }
{ [Nonbonds] } { [NO14exclusions] } { [NOEXclusions] } -
{ NONOnbonds } { 14EXclusions } { EXCLusions }
[TRIAngle] [ HISTogram HMIN real HMAX real HNUM integer -
[HSAVe] [HPRInt] [HNORm real] [HDENsity real] ]
For a more detailed description, see the documentation.
Files modified:
manip/corman.src
manip/corman2.src
File added:
manip/corman3.src
(6) New Parsing Options
The IF command will be expanded to allow commands such as:
IF ?ENER .GT. ?VDW THEN GOTO label
or
IF ?NSEL .LT. 8 THEN GOTO label
The old syntax continues to work. Other parsing enhancements are also
new.
Files modified:
main/miscom.src
fcm/string.fcm
util/parse.src
util/string.src
(7) NOE Overhaul
The NOE restraints now work on groups of atoms and with the
flexibly defined distance restraint.
Files modified:
fcm/noe.fcm
fcm/pert.fcm
fcm/dimens.fcm
main/iniall.src
quantum/addlnat.src
gener/modpsf.src
energy/enemts.src
energy/energy.src
pert/epert.src
pert/pert.src
misc/noe.src
(8) PARALLEL Code
The parallel code in C23f1 is old and it has not been updated. A
new version has been introduced.
Files modified:
main/iniall.src
main/miscom.src
correl/anacor.src
dynamc/prssre.src
energy/energy.src
energy/eutil.src
fcm/parallel.fcm
gener/update.src
image/crystl.src
image/xtlfrq.src
machdep/cstuff.c
machdep/machio.src
machdep/machutil.src
machdep/parallel.src
machdep/parcomm.src
machdep/startup.src
minmiz/abner.src
minmiz/egrad1.src
minmiz/egrad2.src
minmiz/nraph.src
minmiz/steepd.src
misc/surfac.src
molvib/molvco.src
nbonds/nbonds.src
pert/epert.src
rxncor/path.src
util/title.src
vibran/redbas.src
vibran/vibcom.src
vibran/vibio.src
vibran/vibsub.src
dynamc/dynamc.src
energy/enefscal.src
energy/enemts.src
nbonds/enbfast.src
nbonds/etable.src
nbonds/heurist.src
nbonds/nbondg.src
Files added:
machdep/parcomm.src
machdep/socket.c
(9) CID and CPD Declaration Correction
The array CID was declared as REAL*8 in several places (it never was).
CPD was placed in the wrong PARAM common block.
Files modified:
energy/cenerf.src
energy/enefvp.src
fcm/param.fcm
(10) Calling Argument Mismatch Fix: ENBF5NV
The call and the routine ENBF5NV did not match. This routine is
not tested and rarely used (I have only one test case that runs it,
and I'll bet that no one has ever used it...). I have fixed the
problem, but routines like this should be removed from CHARMM.
Files modified:
nbonds/enbf5.src
nbonds/evdwf.src
(11) Removal of DFLOAT
DFLOAT is not a standard function. Its use made no difference in
any of the places where it was used. All uses of DFLOAT have been
removed.
Files modified:
dynamc/dynamcv.src
dynamc/dynamvv.src
(12) STAR code removed
I removed the ##IF STAR code (obsolete).
File modified:
energy/enefvect.src
(13) Atom limit exceeded warning.
An atom limit warning (fatal) has been added when the GETE command
is used with images and the COMP option, where NATOMT exceed MAXA.
File modified:
energy/eutil.src
(14) Use of PRNLEV, WRNLEV and IOLEV
Several print lines have been added without reference to WRNLEV or
PRNLEV. I have fixed these problems where I find them. This is VERY
important for parallel CHARMM and it is a reasonable coding standard.
Likewise, IOLEV must be used for any I/O other than printing. I will
fix these as I find them.
File modified:
nbonds/nbexcl.src
(15) CORDYN Bugfix
I introduced a bug in CORDYN in converting to Fortran. I fixed it.
File modified:
correl/cordyn.src
(16) Group boundaries in GENERATE command
The group boundaries were not being properly set in
gener/genpsf.src for fisrt and last residue patching. This problem
has been fixed. Some PSF's may change and energies using the GROUP
methods may be different. This makes the method similar to what was
obtained with CHARMM22, but there are cases where the boundaries are
different than what CHARMM22 provided. This algorithm is an
improvement and makes more intuitive sense.
File modified:
gener/genpsf.src
(17) Atom selection overhauled
The atom selection routine had problems. I did a major overhaul
and tested it well. We have been using the tested routine for several
months at NIH. There were several bugs in operator priority and in
the use of the .around. operator. I vectorized most atom loops, and
removed copy procedure during push and pop.
File modified:
util/selcta.src
(18) String print bug.
RMV found a discrepancy in the how the last two arguments are
passed to PRNTST, particularly from SELRPN; He noticed it when he
omitted the END keyword from an atom selection while testing the all
Fortran graphics code; the consequence is an infinte output loop,
requiring a keyboard interrupt. I fixed this by modifying how PRNTST
handles its input values (-1 now means: put the full string on one line.
File modified:
util/string.src
(19) MAKGRP call bugfix
The calls to the routine MAKGRP have the wrong number of arguments.
I fixed this by adding a flag and making all of the calls consistent.
Files modified:
image/upimag.src
nbonds/nbexcl.src
(20) Function call update
The old function SETMSC had been replaced with SETMSR. The calls
in misc/mmfp.src and rxncor/travel have been brought up to date.
Files modified:
misc/mmfp.src
rxncor/travel.src
(21) Error in DIEWRN call.
In string.src, there was a bad call to DIEWRN. This has been
replaced with a call to WRNDIE.
File modified:
util/string.src
(22) Duplicate energy term addresses in fcm/energy.fcm
Both EHARM and GEO were given the same address. The addresses for
GEO and MDIP were increased by one. Note that this modification
changes the dynamic restart files.
File modified:
fcm/energy.fcm
(23) Branch into IF block
Branch into IF block code in misc/testch.src has been rewritten.
File modified:
misc/testch.src
(24) NOE restraint remapping
Added remapping code for NOE restraints when atoms are added or
deleted. Now when atoms are added or deleted, the NOE restraints
remain correct. Restraints involving deleted atoms are removed.
Files modified:
gener/modpsf.src
quantum/addlnat.src
(25) FORMAT statement moved
Move format line 100 in nbonds/heurist to outside conditionally
compiled code.
File modified:
nbonds/heurist.src
(26) Group-group list print
The routines that print the nonbond list have been improved so
that the group-group lists are also printed. In the past, only
atom-atom lists were considered. This enhancement was needed to debug
the QM/MM problem, but is generally useful so I include it as a
permanent feature.
Files modified:
nbonds/nbutil.src
util/chutil.src
(27) Print control bugfix
Minor print control bug fixed.
File modified:
dynamc/dynamcv.src
(28) ##IF - Code fix
There were several problems with ##IF code. Since some linkers
require zero undefined globals, these problems must be fixed.
1) The NOSE integration code called two of the BLOCK routines (BLASGN
and I4VAL), but there is no check to see if the BLOCK code has been
compiled. I moved the routines BLASGN and I4VAL to outside the ##IF
BLOCK conditional code in block.src. Also, a ##IFN NOIMAGES was added
before the call to IMHBON.
File modified:
pert/block.src
2) The routine INCRYS was inside a ##IFN NOIMAGES, but the routine
is need in INIALL for all cases. I moved it.
image/crystal.src
3) There were calls to XTLAXS and XTLLAT in READCV that were not
blocked with NOIMAGES. A call to IMFILL in READYN. I added
appropriate ## control code to fix this.
File modified:
dynamc/dynio.src
4) The arrays REACLS and PRODLS were used in NBNDGC without testing
the "##IF TSM" flag. I fixed it.
File modified:
nbonds/nbndgc.src
5) The use of QBLOCK in nbonds/evdwf needed to be sectioned off with
##IF QBLOCK
File modified:
nbonds/evdwf.src
(29) JOIN/RENUMBER developed.
A "JOIN segid RENUMBER" feature will be added in the JOIN command.
This allows resid's to be made sequential within a single segment.
File modified:
gener/modpsf.src
(30) SCALAR command enhanced.
The scalar commands have been enhanced. The ASP arrays (IGNOre,
ASPV, and VDWS) are now accessible. There is a sort option for the
SHOW command. There is a new MASS keyword for the STATistics and
AVERage commands
File modified:
manip/scalar.src
(31) PERT/FLAGS
The energy term flags will be properly used in PERT. Thus it will
be possible to multiply separate energy terms by a scale factor.
Files modified:
pert/pert.src
pert/epert.src
(32) IMPLICIT NONE
The IMPLICIT statement is not in a standard FORTRAN, which is
replaced with ##INCLUDE '~/charmm_23/impnon.fcm'
Files modified:
misc/nmr.src
misc/zmat.src
util/ztbl.src
5.4 CMU Bugfixes
- Charles L. Brooks, Chiachin Tsoo, Mary E. Karpen, Bill Young
Email Address : brooks@chem.cmu.edu; tsoo@chem.cmu.edu
Institution : Carnegie Mellon University
Date : May 3, 1993
(1) CSHAKE problem with fixed atoms
Correct expressions for hmassi & hmassj when there are fixed atoms.
File modified:
manip/cshake.src
(2) Improper intrinsic function argument type
The original code, SQRT( DBLE(N) ) is translated to fortran code
on cray as SQRT( (N) ), which gives improper intrinsic function
argument type. It is fixed to SQRT( DBLE( FLOAT(N) ) ).
File modified:
minmiz/tnpack.src
(3) Machine(Cray) dependent arithmatic error
Subroutine DCNTRL control the dynamic runs from heating to
equilibrating by comparing the variable J to I=(FINALT-FIRSTT)/TEMINC.
However, this expression doesn't always give the expected values.
This is a problem of number representation on the cray. Cray people
said "cray uses different number representation than other
machines(convex, sgi, sun, ibm and dec use IEEE)". One way to get
around with this problem might be:
instead of
I=(FINALT-FIRSTT)/TEMINC,
let
I=(FINALT-FIRSTT)/TEMINC*(1.+1.d-13)
File modified:
dynamc/dcntrl.src
(4) Illegal option ls -1
Use -c instead of -1 for ls as in "set mdlist = `/bin/ls -c | sed
'/fcm/d' | sed '/main/d'"
File modified:
install.com
(5) Angle correlation bugfix
The following modifications to correl/anacor.src (SUBROUTINE
ANACOR) and corrio.src (SUBROUTINE SHOWTS) will correct the average &
rms fluctuations reported by the correl facility for angle data.
Files modified:
correl/anacor.src
correl/corrio.src (SUBROUTINE SHOWTS)
(6) Fixed atom coordinates in dynamics trajectory reading
Fixed atoms are not read in when BEGIN starts the trajectory read
after the first frame. The fixed atoms are normally defined only in
the first frame of a dynamics trajectory. When the first frame is not
asked for by the user (i.e., BEGIN points to a trajectory frame later
than the first frame), the fixed atom data is never read, and the atom
coordinates are left undefined. This is due to the RDFST logical
always being defined as .false. This variable appears to have been
defined as .true. under certain conditions in charmm19, but this line
is dropped in charmm22 & charmm23.
File modified:
dynamc/dynio.src
(7) EHOUSS arguments corrected
The subroutine EHOUSS returns the incorrect eigenvalues on some
machines. The original subroutine EHOUSS is called from the
subroutine EIGRS with the following statement:
CALL EHOUSS(A,N,D,WK(ND),WK(ND))
The EHOUSS subroutine has arguments:
SUBROUTINE EHOUSS(A,N,D,E,E2)
E2 is computed within EHOUSS as the square of E. On return, there is
a machine-dependent ambiguity in the values stored in WK. WK is
subsequently used in the routine EQRT2S, which requires EHOUSS's E
vector (not E2) as input.
File modified:
util/matrix.src
(8) Record overflow
A write statement in the PERTAN subroutine caused a record
overflow on some machines, causing a core dump. Format statement #43
was changed to avoid record overflows. NOTE: upon the request from
Bernie, it is fixed as in the following
##IF LONGLINE
43 FORMAT(' PERTRES> LSTART=',F12.5,' LSTOP=',F12.5,
* ' EPRTOT=',F12.5,' EFORWARD=',F12.5,' EPREF=',
* F12.5,' DIFAVE=',F12.5,' DIFFLC=',F12.5)
##ELSE
43 FORMAT(' PERTRES> LSTART =',F12.5,' LSTOP=',F12.5,
* ' EPRTOT=',F12.5,/,10x,'EFORWARD=',F12.5,' EPREF=',
* F12.5,' DIFAVE=',F12.5/10x,'DIFFLC =',F12.5)
##ENDIF
File modified:
pert/pert.src
(9) CPU time on a DEC Alpha machine
Added an ALPHA section to the beginning of TIMRB. This section
does not contain the "EXTERNAL SECNDS" line, since the SECNDS function
as written does not run on the alphas (ITIME writes an integer*8 array
rather than an integer*4 array). The secnds function used is thus the
one provided in the alpha's libfor.a library. Add the following in
the subroutine TIMBR.
##ELIF ALPHA
REAL SECNDS
REAL CPUREF
REAL CPUT,ZEROT
REAL TARRAY(2), TREF, ETIME
EXTERNAL ETIME
SAVE CPUREF,TREF,ZEROT
100 CONTINUE
TREF = ETIME(TARRAY)
File modified:
machdep/machutil.src
(10) 4-byte integer sort bugfix
The cshkin2 routine (in cshake.src) incorrectly sorts the 'constr'
vector on machines that use four-byte integers. This is because the
util/sort.src routine SORT sorts only integer vectors, but the cshkin2
passes two integer vectors and one real*8 vector to SORT. When
integers are defined as eight-bytes (as on the Cray Y/MP), the real
vector is sorted correctly along w/ the two integer vectors. With
four-byte integers, however, the sort splits each real*8 value, giving
wrong answers.
An integer vector (ktemp) containing the original indices
(1,2,...,nconst) has been defined & is sent in to SORT rather than the
real vector (rtemp). This vector is then used to fill the constr
vector in the proper order via indirect addressing. This is
summarized in brief below:
rtemp(i) = constr(i), i = 1, nconst
ktemp(i) = i , i = 1, nconst
call sort (itemp,jtemp,ktemp)
constr(i) = rtemp(ktemp(i)), i = 1, nconst
File modified:
manip/cshake.src (cshkini and cshkin2 routines)
5.5 Bugfixes and Enhancements
The following are the records of cumulated bugfixes and minor
enhancements done at Harvard and by requests from users.
(1) Floating Point Error: Division by Zero Fix
- Leo Caves, reported
File modified:
minmiz/abner.src
(2) CUTIM < CUTNB Case Fix
- Chandra Verma, reported
When CUTNB is reset to a value less than CUTIM, reset CUTIM to the new
CUTNB value and inform it to the user.
File modified:
nbonds/nbutil.src
(3) Heuristic Update Problem Fix on Cray Version
- Stefan Fischer
Bug occurs in CRAY executables built with the UNICOS pre-compiler
directive. Even if heuristic non-bond updating is requested, i.e.
INBFRQ is set to -1 , the non-bond list update is done at EVERY energy
call, whatever the non-bond cut-off buffer zone or the next set of
coordinates are.
One of the UNICOS pre-compiler directives in NBONDS.SRC causes the
execution of
CALL NBNDF(BNBND, LNBND, MAXJNB, MXJNBG, X, Y, Z, ATSX,
& ATSY, ATSZ, CMPLTD)
to be skipped. The NBNDF() subroutine does the job of copying the
latest X,Y,Z coordinate set into the ATSX,Y,Z array, which is used to
determine how much the coordinates change from one energy call to the
next and whether an update should be done.
Adding code ( that copies the X,Y,Z coordinates into ATSX,Y,Z ) to
the section of code that is executed only if the UNICOS pre-compiler
directive is set.
File modified:
nbonds/nbonds.src
(4) A Problem in RXNDEF Fix
- Jeffry D. Madura
Email Address : madura@hobbes.chem.uh.edu
Institution : University of South Alabama
Date : Aug. 6, 1992
When defining the reaction coordinate distance in the rxncor
module the above error message was reported. After looking at the
source code I found that the if statement was not properly
constructed. I made the change indicated below, recompiled and ran
the job. Everything appeared to work correctly. I have NOT be able
to fully test the rest of the reaction coordinate code since I am not
familiar with this module. I will be running a few more tests to
determine if all is working well.
To fix the error message I changed line 247 in the file
RXNDEF.SRC
if (ind1.ne.80) call wrndie(-3, '',
to
if (deftyp(ind1).ne.80) call wrndie(-3, '',
(5) Branch into IF block
- Masakatsu Watanabe
Branch back in MTS code in dynamvv.src to label 2222 jumps into an
IF block. Also IMTS is used outside of an MTS conditional compile
section in dcntrl.src.
Files modified:
dynamc/dcntrl.src
dynamc/mts.src
(6) TRAVEL Improvement
- Stefan Fischer
"Floating point exception" on Cray upon first call to sub-routine
DBRENT is due to the fact that the REAL*8 variable LASTU (not explicitly
initialized in DBRENT) is given a CHARACTER value by the Cray compiler
upon program startup. This problem is fixed.
CPR (alias REFINE) in TRAVEL module is improved to handle the case
when convergence reaches the limits of numerical precision.
File Modified:
rxncor/travel.src
(7) Extension of Solanal ANALysis command:
- Arnaud Blondel
A command -CROSs- was added to allow a cross analysis on two
selected subsets of atoms. For the moment the exclusion of the couple
of atoms belonging to the same SEGId is not implemented. The keyword
CROSs cannot be selected with the following options: WATer, SITE,
IKIRkg, ISDIst, IFDBf. IVAC, IMSD, IFMIn were not tested with CROSs.
Files modified:
correl/solana.src
doc/corman.doc
test/c22test/solanal.inp
(8) RISM code overhauled
- Georgios Archontis and Youngdo Won
The solvation module (RISM program) was introduced into CHARMM
c23f and had its own parser and utility routines. Once initiated by
the command "RISM", it runs by itself without communicating with other
module programs. It takes up about 25 Mbyte space of memory over the
c23f version.
We have fixed all the problems that have been reported. Name
conflict of CHECK and CONSTR is resolved. IMSL routines are moved to
util/imsl.src from solvation/imsld.src and MISC and other parsing
routines are replaced by RDCMND and appropriate string functions. For
memory management, all big arrays are allocated on HEAP.
Files modified:
fcm/distri.fcm
fcm/exfunc.fcm
fcm/struc.fcm
solvation/coorman.src
solvation/cycles.src
solvation/deriv.src
solvation/fft.src
solvation/rism.src
solvation/rismio.src
solvation/solvation.src
solvation/state.src
util/imsl.src
util/parse.src
util/string.src
Files removed:
fcm/charli.fcm
fcm/external.fcm
fcm/rismio.fcm
solvation/charli.src
solvation/imsld.src
(9) REPLica: Support for replicas added.
- Leo Caves
Tool to support LES and MCSS calculations. Performs replication of
arbitrary regions of PSF. Data structure interfaces to non-bond list
generation routines, to perform appropriate exclusions. In association
with BLOCK can provide appropriate energy/force normalizations for
various classes of methods employing replicas. For details see
replica.doc.
Introduced REPLICA and REPDEB preprocessor directives. Code for
cray multi-tasking list generation routine used inference and has not
been tested. Convex parallel code works fine. Added miscellaneous
parameters to report number of atom/group pairs from non-bonded
routines: ?NNBA, ?NNBG, ?NNBI for atom/group/images respectively. For
replica-based exclusions from the list there are ?NRXA and ?NRXG for
atom and group exclusions.
Files added:
gener/replica.src
fcm/replica.fcm
c23test/replica.inp
doc/replica.doc
Files modified:
main/charmm.src:
main/iniall.src:
nbonds/cnbnd.src:
nbonds/nbexcl.src:
nbonds/nbndf2p.src:
nbonds/nbndfp.src:
nbonds/nbndgc.src:
nbonds/nbondg.src:
nbonds/nbonds.src:
manip/scalar.src:
util/selcta.src:
gener/modpsf.src:
CHARMM c23f2 (August 15, 1993) to c23f3 (February 1, 1994)
6.1 NIH Modifications and Enhancements
- Bernard R. Brooks
Email Address : brbrooks@helix.nih.gov
Institution : National Institutes of Health
Date : January 15, 1994
(1) Dihedral time series in CORREL.
Fixed problems with recent diheral code in correl to account for
torsional timeseries. The correct fluctuation is now determined.
The extra processing has been removed from the SHOW command because
the data may no longer be valid for this processing when MANTIME
commands are present in a script. A new command option "MANTime
CONTinuous-dihedral" has been added to allow a dihedral timeseries to
be unfolded to a continuous function.
Files modified:
correl/anacor.src
correl/mantim.src
correl/corrio.src
..............................................................................
(2) Bug in energy time series in CORREL.
Fixed bug in NEXTE in anacor regarding step number. The incorrect
nonbond list was being used to determine the energy due to an
off-by-one error.
Files modified:
correl/anacor.src
..............................................................................
(3) New CORREL MANTIME options.
New options (RATIo and KMULt) added to the CORREL MANTIME command.
Files modified:
correl/correl.src
correl/mantim.src
..............................................................................
(4) Bug in the CORREL SHOW command.
An incorrect pointer was used when multiple time series were present.
Files modified:
correl/corrio.src
..............................................................................
(5) Bug in the CORREL ENTER VECT command.
The conversion to FORTRAN was suboptimal.
Files modified:
correl/correl.src
..............................................................................
(6) Analytic default for COOR SURF.
Recent modifications to the COOR SURF command have been removed.
The default for the COOR SURF command has reverted to the analytic
answer. The anaylitic answer is less expensive and more accurate.
The older Lee and Richard's algorithm may still be invoked by specifying
a nonzero RPRObe value. The maximum number of contacts that a sphere may
have has been increased from 15 to 35.
Files modified:
misc/surfac.src
..............................................................................
(7) SCALAR READ command.
A new SCALAR READ option has been added. It allows values to be
entered from a file. The use is:
OPEN READ CARD UNIT 12 NAME file.dat
SCALar WMAIn READ 12 SELE ... END
which will read selected entries to the weighting array.
Files modified:
manip/scalar.src
..............................................................................
(8) NODISPLAY keyword made functional
There were problems with NODISPLAY keyword.
- in graphics/graphx.src EOF was not inititalized
- in graphics/drawit.src VECMMULN index K fixed
- prefx.f changes all INTEGER*2 to INTEGER on CRAY.
Changed everything except apollo system calls to INTEGER, so graphics
code now works on the CRAY as well. Fixed a bug in the graphics
transformation routine for non Apollo. The SHELL command option now
functional for IRIS and UNIX machines.
Files modified:
fcm/graph.fcm
fcm/graphdep.fcm
graphics/drawit.src
graphics/grutil.src
graphics/apodraw.src
graphics/apograph.src
..............................................................................
(9) Completion of X windows CHARMM graphics.
New files to support X display and PostScript output for CHARMM
graphics. machdep/xdisp.c has to be modified for port to other than
HPUX machines.
Bug in graphics:
If apodraw.src is not compiled with -K than stereo feature does
not work. For now we just compile it with -K, but this is 20% slower
executable!
files modified:
graphics/apograph.src
graphics/graphx.src
graphics/grutil.src
graphics/apodraw.src
graphics/psdraw.src
graphics/drawit.src
machdep/xdisp.c (new)
graphic/psdraw.src (new)
Changed name of subroutine READV because it is the name of system
routine used by Xlib.
files modified:
correl/solana.src
..............................................................................
(10) SGI Graphics
Put #ifdef sgi in the first line. Also change cc line in
build/UNX/Makefile_sgi to cc -cckr -Dsgi. This prevents other machines
without gl include files to compile this file.
files modified:
graphics/iris_graphics.c
..............................................................................
(11) Setting environment variables in CHARMM.
A new command: ENVIronment
It does the same as specifying setenv in csh. It is useful for
defining X display within charmm scripts before running the graphics
command and it is used for the GAMESS interface.
files modified:
main/miscom.src
..............................................................................
(12) Numbers added in number.fcm
Some reorganization of number.fcm was performed. Numbers are
placed in sequential order (to find them with less effort. The CRAY
and ALPHA keywords were removed. The number THR60 (360.D0) was
renamed to THR6TY to make it consistent with the rest of the 3 digit
numbers.
Files modified:
fcm/number.fcm
pert/icfix.src
pert/tsmp.src
solvation/fft.src
..............................................................................
(13) Some ## control bugs fixed.
##IF NOMISC and ##IF NOVIBRAN code modified
Files modified:
gener/modpsf.src
image/crystal.src
misc/zmat.src
quantum/addlnat.src
..............................................................................
(14) Bug in the "IC DELEte IMPRoper ..." fixed
There was a logic error for the deleting improper IC table
entires. For current executables, the code allows all impropers to be
deleted with a null atom selection (IC DELEte IMPRoper SELEct NONE
END). This correction allows the selection of improper entries to
delete, as implied in the documentation. The same problem was also
corrected in the IC KEEP command.
Files modified:
manip/intcor.src
..............................................................................
(15) Resolve Possible Name Conflicts with GAMESS
Problems with SGI compiler which doesn't like the same name for
SUBROUTINE or FUNCTION and COMMON block. It was easier to change this
in CHARMM than in GAMESS.
Names and files modified:
VIRIAL -> VIRAL : dynamc/prssre.src
energy/enemts.src
energy/energy.src
pert/epert.src
GAUSS -> GAUSSI : dynamc/dynutil.src
GAMMA -> GAMA : misc/nmr.src
CONNEC -> CONECT : misc/testch.src
..............................................................................
(16) Parallel code enhanced.
Problems were encountered when running SBOUND in parallel. Needed
to change QCMPRS to .FALSE. in call to RDCMND. A global combine
option has been added to TRANSI for printing the total forces and torques.
Files modified:
misc/sbound.src
image/images.src
machdep/parallel.src
Changed CALL PARUPDATE and fixed the subroutine
Files modified:
gener/update.src
Added new keyword (PARAGON) for Intel Paragon machine.
Files modified:
machdep/machio.src
machdep/parallel.src
..............................................................................
(17) DBLE usage removed
The use of the DBLE function has been removed. Its use was superfluous.
Files modified:
minmiz/tnpack.src
gener/replica.src
misc/nmr.src
..............................................................................
(18) Single precision overflow conditions corrected.
Arbitrary large values have been replaced with RBIG in DCNTRL and
in NOSECT. The order of operations for calculating a constant has
been modified so that intermediate values do not overflow in NMRRAT.
Files modified:
dynamc/dcntrl.src
misc/nmr.src
dynamc/nose.src
..............................................................................
(19) FREHP added
Missing FREHP statement for array IPAX in manip/corman.src
Files Modified:
manip/corman.src
..............................................................................
(20) Nonstandard Intrinsic function DNINT replaced.
The HPUX compiler doesn't accept DNINT as a function name so it
was changed to its generic name ANINT
Files modified:
dynamc/dcntrl.src
..............................................................................
(21) Add link atom bug.
Off-by-one bug fixed in the generation of link atoms in the middle
of a PSF.
Files modified:
quantum/addlnat.src
..............................................................................
(22) A variable added for TNPACK.
Files modified:
fcm/contrl.fcm
..............................................................................
(23) Some errors in solvation code fixed.
The routines EDTZM2, MKGR, MKWR, and CLOSURE were called with the
wrong number of arguments. I fixed the call to EDTZM2, but the other
are still wrong (the needed correction was not transparent).
Files modified:
solvation/coorman.out
solvation/soluu.out
solvation/soluv.out
..............................................................................
(24) Floating point numbers in calling sequences removed.
Double precision constants are removed from calling sequences.
The common, fcm/number.fcm, is employed in a few places. In general,
numbers such as 3.0D0 should not be found in the code outside of a
##IFN SINGLE block. The routines in quantum/qmpac.src had some major
cleaning due to lots of "D0" in intrinsic functions. These changes
are needed for Compiling on the IBM RS6K.
Files modified:
quantum/qmpac.src
quantum/qmset.src
rxncor/adiab.src
rxncor/travel.src
minmiz/tnpack.src
rxncor/adiab.src
solvation/fft.src
solvation/soluu.src
solvation/soluv.src
solvation/solvation.src
solvation/solvv.src
..............................................................................
(25) More TABS removed.
Files modified:
solvation/solvation.src
solvation/rismio.src
..............................................................................
(26) Redunant compilation keywords removed.
The keywords IBMVM and VM were all replaced by IBMVM.
The keywords IBMMVS and MVS were all replaced by IBMMVS.
Files modified:
machdep/machutil.src
..............................................................................
(27) Unused routines removed:
The obsolete routine, SHAKES, has been removed.
Files modified:
manip/shake.src
..............................................................................
(28) Unused variables removed:
Old unused variables have been removed. A check was made to
ensure that they are not contained within ##IF blocks.
Files modified:
correl/anacor.src
energy/cenerf.src
energy/enefst2.src
energy/enefst2q.src
energy/enefvp.src
gener/makphi.src
gener/psfsum.src
image/cnbndm.src
image/imnbf2p.src
manip/cstran.src
manip/fshake.src
manip/fshake2.src
manip/rgyr.src
manip/shake.src
minmiz/conjug.src
molvib/molinp.src
molvib/molvio.src
molvib/molvsb.src
nbonds/cenbf.src
nbonds/enbf2.src
nbonds/enbf3.src
nbonds/enbf4.src
nbonds/ewald.src
nbonds/ewaldf.src
nbonds/ewaldf2.src
nbonds/nbexcl.src
nbonds/nbndf2p.src
pert/icfix.src
pert/icpert.src
pert/tsme.src
pert/tsmp.src
pert/tsms.src
..............................................................................
(29) PREFX.SRC overhauled.
The PREFX program has been overhauled. The new code has the
following features:
- It allows "!" comments at the end of valid fortran statements.
- Conversion to single precision is performed ONLY if the SINGLE
keyword is present.
- It allows the use of identifier comments in ## statements.
For example:
##IF PERT (pertprint)
...
##ELSE (pertprint)
...
##ENDIF (pertprint)
This makes the code easier to read and allows ##ENDIF statements to be
uniquely identified. A fatal error is flagged if the identifiers do
not match. Most of the features of the previous prefx have been
retained.
Files modified:
prefx.f to prefx.src
------------------------------------------------------------------------------
6.2 KI/CSB Bugfixes
- Lennart Nilsson
Email Address : ln@pike.csb.ki.se
Institution : Departement of Medical Biophysics, Karolinska Institute
Date : December 13, 1993
(1) Long Filename Support
CHARMM crashes at first coordineate write with filenames >32
characters used.
Files modified:
machdep/machio.src
(2) Nbond List Generation
crash due to integer overflow when natoms > 65536 (2**16). With
large systems (NATOMS>65536) the estimate of space allocation in
nbonds causes an integer overflow (natoms*natoms then is too large for
an integer*4 variable)
Files modified:
nbonds/nbonds.src
(3) CORREL Update Problem
correl calls update, which is usually not necessary when you
analyze trajectories, and consequently crashes if parameters and
coordinates are not present. The PSF and trajectory should be
sufficient.
Files midified:
correl/correl.src
(4) CORREL PUCKER Timeseries
The timeseries is a 2dim vector, currently returns same value in
both positions, I will also change the pucker calculation to use the
Altona&Sundaralingam definitions instead of currently used
Cremer&Pople; the latter is more general (applicable also to rings w/
> 5 members), but not as commonly used as the former. (A&S and C&P
give very similar results)
Files modified:
correl/correl.src
(5) Date Format Fix
On all other machines dates are reported in the month/day/year
order (this is in misc. titles & banner etc.) - to be consistent and
use the same order thruout (and to stick to the historically given
m/d/y order from the good old vax/vms days where the routine IDATE had
its arguments in that order) the wrapper routine idate has to be
changed sligthly.
Files modified:
machdep/machutil.src
(6) Parameter Checking in SURFAC
Added check that parameters are actually present.
Files Modified:
misc/surfac.src:
(7) Rotation Matrix
Added COOR ROTAte MATRix variant (also updated corman.doc)
Files Mofified:
manip/corman.src
(8) GETUNAMEF
Get username via getunamef on DEC/ALPHA as on our other platforms.
Files Modified:
machdep/machutil.src
(9) Miscellaneous Problems in SOLANA
Get DTC from icntrl array in trajectory and place FREEAT and TEMP
on HEAP. Correct calls to corfuncb heap(VX) --> heap(VZ).
Files Modified:
correl/solana.src
------------------------------------------------------------------------------
6.3 Extension of Command Line Parameter Handling
- Leo Caves
Email Address : caves@tammy.harvard.edu
Institution : Harvard University
Date : January 18, 1994
A command line parameter token can now be a string rather than
just one of the single characters 0-9,a-z,A-Z. For substitution a
token is indicated by the use of the @ character as before. The token
is end-delimited by any non-alphanumeric character. In the case that
the token is not found in the parameter table, a check is made to see
if the first character of the token is itself a token in the parameter
table. If this single character token is in the table, the
corresponding value is substituted -- this is the necessary scheme to
allow backwards compatibilty with the old parameter substitution,
which allowed parameters embedded in strings. For unambiguous token
detection, "protect" the token with brackets {} --- this allows for
the use of non alphanumerics in tokens such as -,_. To test whether a
token is in the parameter table, use @?token. This will substitute 1
if token is in the table, 0 if not. This is useful (in conjuction with
the IF command) for setting defaults. (Note thar @? takes precedence
over any of the built-in parameters such as ?ENER etc. --- it is
parsed first). eg. if @?maxTemp .eq. 0 set maxTemp = 300.
At present the parameter table is dimensioned as follows (set in
cmdpar.fcm):
Maximum number of parameters: 256
Maximum token length: 32
Maxiumum value length: 128
For current sizes use command: SHOW PARAmeters VERBose.
A extension to the parameter manipulation commands such as
INCR,DECR,MULT etc. has been made in the form of a rudimentary
expression evaluator, accessible through the LET command (one of the
miscallaneous commands). It has a VERY SIMPLE syntax definition, it is
not a recursive descent parser! It supports compound assignment
operators +=,-=,*= etc.., It has some functions SQRT, SIN, MAX, MIN,
ACOS, INT etc... It supports the usual binary operators, +,-,*,/ and %
(MOD). There are no precedence rules, so you have to split the
expressions appropriately. eg. The following expression:
a = (( @a + max( @b, 4) ) ** sqrt(@area) ) / acos(-1)
needs to be dealt with as follows:
SET tmp ! all variables on LHS of a LET statement need to be defined.
LET a += max @b 4
LET tmp = sqrt @area
LET a = @a ** @tmp
LET a /= acos -1
OK, so its not very sophisticated ! But it is robust to argument,operator
and range errors.
All of the above information is contained in miscom.doc. The
testcase test/c23test/cmdpar.inp gives examples of usage.
Files added:
util/cmdpar.src
test/c23test/cmdpar.inp
Files modified:
fcm/cmdpar.fcm
fcm/string.fcm
fcm/exfunc.fcm
util/string.src
util/parse.src
main/miscom.src
machdep/startup.src
doc/miscom.doc
------------------------------------------------------------------------------
6.4 Modifications of Multiple Time Step (MTS) Method
- Masa Watanabe
Email Address : watanabe@tammy.harvard.edu
Institution : Harvard University
Date : January 20, 1994
MTS energy routines were modified in order to unify with standard
routines. MTS energy and image energy routines were converted into
standard energy and image energy routines. Both enemts.src and
emts.src were deleted.
Files Modified:
fcm/bases.fcm
fcm/tbmts.fcm
fcm/tbmtsp.fcm
dynamc/mts.src
dynamc/dcntrl.src
dynamc/dynamvv.src
main/charmm.src
main/iniall.src
energy/energy.src
image/eimage.src
nbonds/nbonds.src
Files Added:
fcm/tbmtsp1.fcm
Files Deleted:
energy/enemts.src
image/emts.src
------------------------------------------------------------------------------
6.5 CMU Development
- Charles L Brooks, III, Mary E. Karpen and W. S. Young
Email Address : brooks@chem.cmu.edu
Institution : Carnegie-Mellon University
Date : January 18, 1994
(1) Clustr code integrated into CORREL
The CLUSTER command clusters time series data obtained within the
CORREL facility. The data are grouped into sets with similar time
series values, using euclidean distance as the dissimilarity measure
between different time frames of a set of time series. It is useful,
for example, for grouping together similar conformations or energy
levels.
Files Modified:
correl/correl.src
doc/correl.doc
Files Added:
correl/clustr.src
test/c23test/clustst.inp
test/data/clustst.hex
(2) STRING common block renamed.
Changed common name STRING to STRNGC to avoid compiler errors.
Files Modified:
fcm/string.fcm
(3) Generic intrinsic function name used
Changed DABS and DSQRT to generic functions ABS and SQRT, needed
when not double precision, i.e., Cray and ALPHA machines.
Files Modified:
dynacm/dynamvv.src
(4) Compiler directive correction
Changed cray compiler directive from CDIR$ SCALAR to CDIR$ NOVECTOR.
SCALAR is not a valid directive.
Files Modified:
nbonds/nbndfp.src
(5) Dyanmics with NOE restraints
Dynamics fails when using noe restraint. DDF must be initialized
to zero.
Files Modified:
misc/noe.src
(6) FREHP calls fixed
incorrect FREHP statements after call to CNBNDM
Files Modified:
source/nbonds/nbonds.src
(7) Memory management problem fixed
code logic does not allow RSCSHK to release heap space allocated
in CSHKINI
Files Modified:
manip/shake.src
(8) FASTER CRAYVEC & FSHIFT
The code in cenbf.src (invoked w/ "FASTer CRAYvec") does not
currently support the force shift nonbonded option "FSHIFT".
A nonparallel routine (ENBFV8) is called instead. Vector/parallel
code is added to cenbf.src to implement the nonbonded electrostatic
option FSHIFT.
Files Modified:
nbonds/enbond.src
nbonds/cenbf.src
------------------------------------------------------------------------------
6.6 Other Harvard Fixes
- Xiaoling Liang (lia@tammy.harvard.edu)
Georgios Archontis (georgios@tammy.harvard.edu)
Masa Watanabe (watanabe@tammy.harvard.edu)
Institution : Harvard University
Date : January 26, 1994
(1) IASVEL=0 correction to the Verlet integrator
- Xiaoling Liang
The IASVEL 0 option does not assign comparison coordinates as
initial velocity.
Files Modified:
dynamc/dcntrl.src
(2) QMMMDATA block data
- Xiaoling Liang and Youngdo Won
Initial data assigned through data block QMMMDATA get lost on
ALPHA, DEC and SGI machines so quantum can not be performed. Unused
common block /VERSION/ removed from am1parm.fcm amd qmdata.src
Files Modified:
fcm/am1parm.fcm
quantum/qmset.src
quantum/qmdata.src
(3) BLOCK COMP CONT
- Xiaoling Liang
keyword CONT added for TI calculations in BLOCK which gives an
ensemble average as a function of time.
Files Modified:
pert/block.src
(4) FREHP call added
- Masa Watanabe
Charles L Brooks III reported the problem of missing FREHP
statement. It is fixed in the Nose dynamics code and the testcase is
modified accordingly.
Files Modified:
dynamc/dcntrl.src
test/c23test/nose2.inp
(5) RISM code cleaned
- Georgios Archontis
Adds three 0.0's in the calling argument of CLOSURE calls in
soluu.src and soluv.src. IWK(NN) changed to IWK(6*DVECT+150) in
fft.src. Remove unused variables (NU and NUU) in fcm/distri.fcm.
test/c23test/rism.inp is modified to show diatomic solute with two
different sites. doc/rism.doc is updated
Files Modified:
solvation/soluu.src
solvation/soluv.src
solvation/fft.src
fcm/distri.fcm
test/c23test/rism.inp
doc/rism.doc
CHARMM c23f3 (February 1, 1994) to c23f4 (August 15, 1994)
Bugfixes reported to charmm-bugs@tammy during the development cycle
are incorporated. The bug-reports are collected in ~ftp/charmm/Bugs.
(1) Stack allocation errors in ANACOR
- Mary E. Karpen (B940425.mek and B940526.mek)
Several errors in stack allocation in the call to TRNPHI from
ANACOR are fixed. Stack allocation for TRNPHI arrays are moved from
correl to ANACOR. Since only DIHE time series require these arrays,
the stack space in ANACOR is only allocated for these time series.
Also, rather than have dihedral angle transitions calculated as the
default, it makes more sense to calculate them by request. This is
done by adding the option TRANsition on the ENTE DIHE command line.
(doc/correl.doc is updated)
Files Modified:
/correl/correl.src
/anacor.src
(2) SOLANA bugs
- Ryszard Czerminski (B940523.rcz)
Calling argument list is corrected for the RDNEXCO call, and a
warning message is coorected.
- Jay Banks (B940317.jb)
Crystal test corrected from QCRYS=(IC11.EQ.1) to QCRYS=(IC11.NE.0)
File Modified:
/correl/solana.src
(3) Dynamics extra step
- Bernard R. Brooks (B940210.brb and B940601.brb)
When running a DYNA LEAP MD step with SHAKE and with equilibration /
heating (or with saving a restart file) determined by NCYCLE would
take an incorrect (extra) step. The DO I=ATFRST,ATLAST block is added
to fix the problem.
File Modified:
/dynamc/dynamc.src
(4) Parallel code addition/correction
- Bernard R. Brooks (B940210.brb and B940601.brb)
Parallel code added.
- Leo Caves and Arnaud Blondel
A bug in the addition is fixed by ##INCLUDE '~/charmm_23/stream.fcm'
File Modified:
/dynamc/dynio.src
(5) Correct virial summation
- Bernard R. Brooks (B940601.brb)
When running dynamics using the PARALLEL code, the VIRI and VIRKe
values reported during dynamics with SHAKE are incorrect. This is due
to shake corrections to the internal virial coefficient not being
globally summed before being added to EPROP(VIRI) in routine VIRSHK
(dynamc/prssre.src) called from DYNAMC (dynamc/dynamc.src). The same
problem occurs with the pressure tensor EPRESS(VIXX). The problem is
fixed to get the virial summed across processors. The fix needs to
include INTEGER RPXX,RPYY,RPZZ,RPXY,RPXZ,RPYZ in reawri.fcm
Files Modified:
/dynamc/prssre.src
/fcm/reawri.fcm
(6) Trajectory reading with crystal data
- Bernard R. Brooks (B940601.brb #9)
The routine CREAD has been updated to allow the reading of
trajectory files with crystal data. Also, coordinates with bad fields
will be ignored (instead of causing a fatal error).
File Modified:
/io/coorio.src
(7) Command line parsing
- Bernard R. Brooks (B940601.brb #10)
Command line parsing is fixed for multi-character command line
parameters.
File Modified:
/machdep/startup.src
(8) FREHP calls added
- Mary E. Karpen (B940712.mek #3)
Heap allocated in SHKCOM is not released on program exit. This
does not affect charmm job output in any way, except in the "SPACE
CURRENTLY IN USE" report at the end of the job. The space is freed,
so SPACE USED is zero at job end, allowing other potential problems
with heap allocation to be detected. The IF (QSHAKE) condition to
freeing the cray shake heap (RSCSHK) is added for consistency, though
this makes no difference in the current code logic.
Files Modified:
/main/iniall.src
/manip/shake.src
(9) ORINTC problem with parallel code
- Bernard R. Brooks (B940601.brb #11)
There was a bug in the routine ORINTC which caused CHARMM to fail on
parallel systems when the COOR ORIEnt command was executed.
File Modified:
/manip/corman2.src
(10) ATOMID correction
- Bernard R. Brooks (B940601.brb #12)
Atom identifier was wrong in printing dihedral restraints. Calls
to ATOMID are fixed. Only the printout was affected.
File Modified:
/manip/cstran.src
(11) PHASE value fix
- Arnaud Blondel (B940601.ab)
Wrong values (sign flip) can be returned if PseudoPhase is close
to 90 or -90 due to infinity of tan(PsPhase). For unknown reason the
same phenomenon occurs around 180.
File Modified:
/manip/pucker.src
(12) TRACE bugfix
- Bernard R. Brooks (B940601.brb #14)
There was a bug in COOR ORIE RMS command in computing the trace of
the coordinate Lagrangian matrix in FROTU. W(3) is changed to W(4).
File Modified:
/manip/rotlsq.src
(13) ING14P fix
- Jay Banks (B940603.jb)
In the IF (QPERT) block, the index into ING14P array should be
NXIP not NXI. Two lines are corrected; ING14P(NXI) to ING14P(NXIP).
File Modified:
/nbonds/nbondg.src
(14) BLOCK code cleaned
- Arnaud Blondel (B940302.ab and B940303.ab)
Coefficient matrix print out is corrected by checking QBLOCK upon
the BLOCK END command. BLOCK usage flags are removed upon the BLOCK
CLEAR command.
File Modified:
/pert/block.src
(15) Initialize ALLC and ALLZ arrays
- Ryszard Czerminski (B940514.rcz)
Floating point exception error is avoided by initializing ALLC and
ALLZ arrays.
File Modified:
/quantum/qmpac.src
(16) REAL/REAL*8 usage fix
- Ryszard Czerminski (B940412.rcz)
An inconsistent use of memory (real vs. real*8) is corrected.
real*8 wt(n) to real wt(n).
FIle Modified:
/rxbcor/rxnene.src
(17) Division by zero
- Bernard R. Brooks and Georgios Archontis
Division by zero problem is fixed in ADIE=(1.0+CDIE2*(3.0*YDIE-1.0))/
(3*YDIE*(CDIE2-1.0)).
File Modified:
/solvation/state.src
(18) Move IF test in parsing
- Bernard R. Brooks (B940601.brb)
The if test that was used to set the FFOUR variable was put in the
wrong place. The line IF(QCMPRS) FFOUR=COMLYN(1:4) is moved from l.
142 to l. 129.
File Modified:
/util/parse.src
(19) Infinite loop problem
- Ryszard Czerminski (B940209.rcz)
An infinite loop is created when CHCKHP calls WRNDIE. Call to
CHCKHP in PRINHP should be commented out.
File Modified:
/util/util.src
(20) Vectorization Error
- Mary E. Karpen (B940425.mek #5)
AMASST in subroutine VIBFLU is incorrectly calculated on the cray
c90. In the testcase vibran.inp, the command
FLUCT ATOM MODE 1 THRU 57 TEMP 300.0 SELE TYPE CA END
results in the variable AMASST in the vibran/vibsub.src routine VIBFLU
to be assigned 19 (= NATOM) rather than the correct amount (1). By
removing vectorization from the loop, AMASST is calculated correctly.
There is no obvious reason why vectorization would break this loop,
so there is potentially another bug lurking, which was not found.
File Modified:
/vibran/vibsub.src
(21) Non-standard continuation character
- Bernard R. Brooks
Non-standard continuation character # is replaced by $. Four such
lines are changed.
File Modified:
/minmiz/tnpack
(22) Declaration fix in REPLICA
- Mary E. Karpen (B940425.mek #1a)
Incorrect private/shared declarations for replica code are corrected.
File Modified:
/nbonds/cnbnd.src
(23) EXTERNAL XDIST
- Bernard R. Brooks (B940601.brb #15)
It appears that SGI's new compiler has a nonstandard definition of
what EXTERNAL means. EXTERNAL XDIST is encapsulated by ##IFN IRIS and
##ENDIF.
File Modified:
/nbonds/nbndgc.src
(24) Apollo graphics
- Bernard R. Brooks (B940601.brb)
An INTEGER declaration is fixed to INTEGER*2 for Apollo versions.
Files Modified:
/graphics/apodraw.src
/graphics/apograph.src
/fcm/graphdep.fcm
(25) Preprocessor keys for MTS code
- Ryszard Czerminski (B940223.rcz)
- Mary E. Karpen (B940425.mek #1b)
Problems with preprocessor directives for MTS is fixed.
Files Modified:
/dynamc/mts.src:
/image/eimage.src:
(26) PRNLEV fix
- Bernard R. Brooks (B940601.brb)
WRNLEV.GE.2 is corrected to PRNLEV.GE.3
File Modified:
/gener/modpsf.src
(27) Vector dependency
- Mary E. Karpen (B940425.mek #2)
Vector dependencies exist in TRANSI routine but CDIR$ IVDEP
specified on line 126. Comment out.
File Modified:
/image/images.src:
(28) UNIX command
- Bernard R. Brooks (B940601.brb)
The ENVIronment command is encapsulated by ##IF UNIX -- ##ENDIF.
File Modified:
/main/miscom.src
(29) Non-standard quote
- Michael E. Holder (B940531.mh)
Non-standard use of " fixed to '.
File Modified:
/util/cmdpar.src
(30) IMAGE/CRYSTAL keyword parsing bug
- Ryszard Czerminski
Nonbonded list is not set up properly upon NBOND command. LIMAGE
should be set .TRUE. in the argument list of the UPDATE call.
File modified:
/main/cahrmm.src
(31) Version number updated
- CHARMM Manager
Never used VERNMR is removed and VERNMC is updated.
File Modified:
/fcm/version.fcm
|
| | |
|
8. c23f5 Change Log |
|
|
|
CHARMM c23f4 (August 15, 1994) to c23f5 (March 15, 1995)
8.1 Convex Exemplar SPP-100 and generic PVM Ports
Developer : Charles L. Brooks, III and Stephen H. Fleischman
Email Address : brooks@scripps.edu
Institution : The Scripps Research Institute
Date : January 5, 1995
CHARMM Version: c23f4
Proj. Report : D950104.clb
A port of CHARMM version 24a2 to general PVM based parallelism using
existing parallel code as well as a port to the Convex parallel machine are
included. The original port is the work of Stephen H. Fleischman from
Convex Computer Corp. Charles L. Brooks, III simply implemented this port
in other versions of CHARMM.
Files Added:
build/UNX/Makefile_cspp
Files Modified:
install.com
doc/parallel.doc
source/charmm/iniall.src
fcm/parllel.fcm
machdep/machutil.src
parallel.src
parcomm.src
startup.src
util/string.src
------------------------------------------------------------------------------
8.2 Cray T3D Port
Name : Charles L. Brooks, III and Barry C. Bolding
Email Address : brooks@scripps.edu
Institution : The Scripps Research Institute
Date : January 6, 1995
CHARMM Version: c23f4
Proj. Report : D950106.clb
A port of CHARMM version 24a2 to the Cray T3D parallel computer using
existing parallel code. The original port is the work of Barry C. Bolding
from Cray Research Inc. in collaboration with Charles L. Brooks, III, who
implemented this port in other versions of CHARMM.
Files Added:
build/UNX/Make_t3d
Files Modified:
install.com
doc/parallel.doc
source/charmm/charmm_main.src
machdep/cstuff.c
machio.src
machutil.src
parallel.src
space.src
startup.src
fcm/dimens.fcm
heap.fcm
number.fcm
stack.fcm
------------------------------------------------------------------------------
8.3 Thinking Machine's CM5 Port
Name : Robert Nagle
Email Address : nagle@tammy.harvard.edu
Institution : Harvard
Date : August 24, 1994
CHARMM Version: c23f4
OS : CMOST 7.3_final.1-rev.3
Proj. Report : D940825.rn
1) Many I/O operations on the CM5 resulted in a system wide hang.
Since I/O was set up to be synchronous broadcast, all processors
should have been making the same I/O requests. However, in some
instances, branches were made based on either PRNLEV (more frequently)
or IOLEV. This resulted in node 0 doing an open, say, and all other
nodes continuing on their merry way. Result - deadlock.
2) Previous communication scheme was based on a simple send and
receive model. By using TMC's active message layer, communication
bandwith can be increased by anywhere from 50% to 5X.
3) DIE on one node could leave the others merrily trundling along.
4) Other minor improvements
Files Added:
build/UNX/Makefile_cm5
source/fcm/cm5.fcm
Files Modified:
install.com
doc/parallel.doc
source/machdep/machutil.src
parallel.src
startup.src
string.src
------------------------------------------------------------------------------
8.4 OS/2 Port
Name : Stefan Boresch
Email Address : boresch@tammy.harvard.edu
Institution : Harvard University (Karplus group)
Date : November 1994
CHARMM Version: c23f4
OS : S/2 2.x and higher
CHARMM (c23f4) has been ported to the OS/2 operating system,
version 2.x and higher. The Watcom Fortran compiler (v. 9.5,
patch-level (c)) has been used. The port is fairly complete, a
detailed list of limitations and known problems is given in the 'Other
Comments' section. The modifications should also be applicable to
c24a2 by applying the changes described below to the respective files.
(I can't hold two versions of CHARMM on my hard-drive at the same time;
I'm happy to work / check the modifications of c24a2/c24a3 once they
become available.)
A new pre-processor keyword, OS2, has been introduced, and all
OS/2 related changes hide behind the OS2 keyword. The port has revealed
several bugs of the compiler, one of them is annoying, and I hope it
will be fixed soon. On the other hand, as OS/2 / Watcom FORTRAN is not
yet another Unix platform, a few quirks of the code have become apparent
as well; one might consider changing the basic code in these instances...
While most Unix utilities have been ported to OS/2, I do not
want to rely on the user having these installed. Therefore, I shall
write an install script in REXX (the OS/2 built in scripting
language); I will also provide utilities to automatically generate
makefiles. (to be done a.s.a.p!) Unfortunately, WATCOM's make
utility doesn't support a number of features used in the Unix
makefiles for CHARMM, so the makefiles differ significantly. Most
likely one can write them much more elegant, but this needs to be done
as well...
Watcom FORTRAN allows direct access to any OS/2 system call.
Therefore, there is no need for C routines; all .c files are superfluous
for OS/2!
Currently, there is no free FORTRAN compiler for OS/2 (asides from
f2c, and I really didn't want to use that...) I'll watch out for the
GNU FORTRAN compiler which apparently is in alpha testing. The WATCOM
compiler has proven quite reliable, and to the best of my knowledge is
the cheapest option for academic institutions. The same package
also supports 32bit extended DOS, Windows 3.1 and Windows NT; i.e. if
someone is so inclined, porting to these platforms should be more
or less painless. However, I used the HPFS filesystem, so I didn't
worry about shortening filenames to DOS' 8.3 convention.
Files Modified (module/file.src):
source/charmm/iniall.src
/miscom.src
dynamc/dynio.src
fcm/impnon.fcm
/number.fcm
machdep/machio.src
/machutil.src
/space.src
/startup.src
manip/corman.src
rxncor/path.src
solvation/fft.src
Other Comments:
The user needs the Watcom Fortran compiler v. 9.5. Patchlevel
(c) was used for the port. The installation will work only on a HPFS
partition. It should be considered to include an OS/2
executable in the distribution package. (The executable itself should
run on both HPFS and FAT partitions)
As described above, patchlevel (c) contains an annoying bug
that requires a somewhat clumsy workaround and Charmm to be called through
a script file. I will provide patches as soon as the compiler is fixed.
A set of makefiles that will work for c23f4 is provided. I'm
working on a general install utility written in REXX (the built in
scripting language of OS/2). I shall also provide REXX scripts to generate
makefiles solely based on the list of source files.
The executable has been currently tested only when compiled
without optimization. I'll test which optimization levels are
possible a.s.a.p. I'm confident, however, that no major problems will
arise (I never had any optimization related problems with the c22gx
series of CHARMM).
KNOWN LIMITATIONS/PROBLEMS:
* Compiler bug related to REWIND of STDIN (clumsy workaround exists)
* QUANTUM code not supported: I didn't even try as I don't think that
a PC will be used for QM calculations anyways.
* Aside from the quantum testcase, two other test-cases fail:
rism.inp
xtlala2.inp
* OS/2 doesn't have an OS funtion that reports the CPU time consumed
by a program/process. Hence dummy routines are introduced, a CPU time
of 0.00 is reported.
------------------------------------------------------------------------------
8.5 Other Bugfixes and Enhancements
Bugfixes reported to charmm-bugs@tammy during the development cycle
are incorporated. The bug-reports are collected in ~ftp/charmm/Bugs.
(1) Initialize R2 in ehbond.src
- Ryszard Czerminski (B940822.rcz)
There is a logic flow in energy/ehbond.src (subr. EHBOND) resulting in
computation with undetermined values and in "floating point problems"...
If S.GT.C2OFHB and ANALYS.ne.3 R2 will not be calculated but used later on...
Files involved:
source/energy/ehbond.src
..............................................................................
(2) Bug in INQUIRE Function on ALPHA Machine
- Ryszard Czerminski (B940824.rcz)
Machine : DECALPHA
Operating System : OSF/1 V2.0 240
As a work around I have increased MAXLEN from 80 to 256.
At the same time I have eliminated MXFLNM and replaced it with MAXLEN.
Files involved:
source/util/title.src
..............................................................................
(3) Uses of CHARMM Variables in Parallel Environment
- Lennart Nilsson (B940825.ln)
The last argument is left off when invoking the slave processes.
No problem as long as only the master process uses the variable,
but hangs the whole thing if the slaves need to have the correct value.
In parcomm.src, two loops having the number of program arguments as upper
bound should use NPRARG, not NPRARG-1.
Files involved:
source/machdep/parcomm.src
..............................................................................
(4) Write Commands in Parallel CHARMM with Slave Proccesses
- Lennart Nilsson (B940907.ln)
The command "write ... name ..." is incorrectly parsed by the slave
processes in a parallel environment and they crash (or hang).
Files involved:
source/io/mainio.src
..............................................................................
(5) ENCODF of ".0"
- Arnaud Blondel (B941220.ab)
On some machines' Zero is written .0. This causes ENDCOF to return
empty parameter.
Files Modified:
source/util/string.src
..............................................................................
(6) NOE Bugfixes
- Lennart Nilsson (B941221.ln)
(6.1) The RESET command in the NOE module does not zero the variable NOENM2
so add
NOENM2=0
to the RESEt handling
(6.2) When reading old format NOE input files the following two lines are
needed in the OLD section of the ASSI parser
RNEXP=THREE
RNTCN=ZERO
Files Modified:
source/misc/noe.src
..............................................................................
(7) Verlet Dynamics Coordinate Update
- Lennart Nilsson (B941222.ln)
- Ryszard Czerminski (B941221.rcz)
Coordinates after dynamics (verlet) are propagated one extra step.
When using the standard verlet integrator (ORIG keyword to dynamics command)
the coordinates in charmm after a dynamics command are not the same as the
coordinates for at the last step of dynamics, as seen if an energy command is
issued immediately after the dynamics. This does not happen with the leap
integrator.
This is because in dynamcv x,y,z are updated after the energy call in
the main dynamics loop and before the ENDDO, whereas in dynamc x,y,z are
updated before the energy call, and thus after the loopback decision is made.
DYNAMC (leap):
loop
x=xcomp+xold
writec x
updeci x
energy x
printe x
xold=....
xcomp=x
endloop ! x is still what it was at the energy call in step i
DYNAMCV (verlet):
loop
writec x
updeci x
energy x
printe x
xold=x
x=xnew
endloop ! x is NOT what it was in step i, rather what it will be in i+1
In order to fix this, Ryszard added following lines at the very end of
DCNTRL routine
...
C
##IF OLDDYN
IF(ORIG) THEN ! bug#92242xfsq01
DO I=1,NATOM
X(I)=XOLD(I)
Y(I)=YOLD(I)
Z(I)=ZOLD(I)
ENDDO
ENDIF
##ENDIF ! OLDDYN
NOTE: This fix results in differences in dynamics benchmarks.
Files Modified:
source/dynamc/dcntrl.src
..............................................................................
(8) Reading Lower Case PDB Files
- Lennart Nilsson (B950104.ln)
Lower case characters in charmm/pdb coordinate files are not understood.
Call CNVTUC after reading lines from charmm/pdb formatted files. Also change
name of varaible LEN (conflict with intrinsic function) to SLEN.
Files Involved:
source/io/coorio.src
..............................................................................
(9) Cray Dependent Coding Fixed
- Ryszard Czerminski & Mike Mirman (B950105.rcz)
FORTRAN version : CRAY FORTRAN CFT77 6.0.4.2
The construct does not compile properly on CRAY
##IF IBM IBMRS CRAY
FUNCTION FUN(A)
REAL*8 FUN
##ELSE
REAL*8 FUNCTION FUN(A)
##ENDIF
and the CRAY key is removed in ##IF
Files Inveloved:
source/dimb/eispack.src
image/crystal.src
minmiz/tnpack.src
quantum/qmpac.src
util/random.src
string.src
util.src
vector.src
..............................................................................
(10) Branch into IF/DO Block Problem Fixed
- P950105.elm
Files Involved:
source/dynamc/dynamvv.src
misc/aspener.src
/surfac.src
vibran/vibran.src