Go to the previous, next section.
The commands described in this section are generally more simple
in nature than those of previous sections. Some are perhaps obsolete,
but included for the sake of completeness.
PARALLEL [CG ] [NCPU integer] [SCHED {GANG} ]
[CONGen ] [ {FREE} ]
[LOOPs ]
[OFF ]
The PARALLEL command is used to control the usr of parallel
processing. Presently, parallel processing can be used for either
the conformational search part of CONGEN, see
section Conformational Search, or
compute intensive loops in the Poisson-Boltzmann electrostatics calculations,
see section Poisson-Boltzmann Electrostatics. It cannot be used for both.
Parallel processing is only supported on multiprocessor Silicon Graphics
workstations or servers.
Parallel processing is not supported on any other machine
as yet.
The options, CG or CONGEN, are used to turn parallel processing
on for the conformational search. The option, LOOPS, is used to turn
on parallel processing for compute intensive loops. The option, OFF, is
used to turn parallel processing off.
If none of these options is specified, then CONGEN searching is run
in parallel.
The NCPU option is used to specify the number of CPU's to use for
the calculation. If this is not specified, then the program will check the following
in order:
-
The value of the environment variable, MP_SET_NUMTHREADS.
-
The value of the enrironment variable, NUM_THREADS.
-
The number of CPU's on the machine where CONGEN is running.
The maximum permitted value for NCPU is the minimum of 32 and
the number of CPU's on the system where CONGEN is running.
The SCHED option controls how process execution is scheduled. The
option, GANG, specifies that all processes working together be
scheduled as a unit. GANG is the default for compute intensive
loops. The option, FREE, specifies that all processes are
scheduled independently. This option is the default for conformational
search.
See section Conformational Search, for more information about parallel processing
in the conformational search.
If you are running a conformational search which uses Poisson-Boltzmann
elecstrostatics, it is more economical to parallelizes the PBE
calculation, and leave the search running serially because this
optimizes memory usage. Parallel searching requires one copy of the PBE
data structure per process, and this takes up a great deal of space.
For large memory jobs running in parallel on SGI workstations, it is
critically important to set the STACK resource limit down from the
default value. See section RLIMIT Command -- Set Resource Limits, for more information.
CODEs
The CODES command invokes the subroutine, CODES
,
which determines all the parameter type code indices for the internal
coordinates and hydrogen bonds. It is useful to call this function
before invoking the CONGEN command if energy values must be
computed in a conformational search, see section Conformational Search.
DELIm char
The DELIM command sets the default command delimiter for
command options which are specified as strings. The delimiter may be
only one character. The default value for the delimiter is a
dollar-sign, $.
DRAW derivative-factor-spec frame-spec
derivative-factor-spec ::= [DFACt real] [NOMO]
frame-spec ::= UNIT integer [DASH real] [FRAMe integer] [RETUrn integer]
The DRAW command (called directly from CONGEN, not to be
confused with the DRAW command found under the ANALYSIS
command) is useful for displaying small molecules. The output is a
command file that can be read by various displaying and plotting
programs such as PLT2. This command file can be edited for different
types of displaying. In addition to atom positions and bonds, velocity
and forces may also be displayed. The current keywords are:
- NOMO
- No molecule option (only velocities or derivatives).
- DFACT
- Derivative factor. The default is 0.0.
- DASH
- Spacing of dashed line used for hydrogen bonds. The default is
.01.
- FRAME
- Specifies that a frame tag will be written first. The default is not to
specify a frame.
- RETURN
- Specifies which stream the plotting program will return to after
plotting this section. The default is no change.
{ STOP }
{ EXIT }
The STOP or EXIT commands cause the program to
terminate and to ignore all command that follow this command. This is
useful for making temporary modifications to input files.
DISTance
The DISTANCE command will cause the distance for every
pair of atoms to be printed on a separate line. This command is not
recommended for systems with more than 50 atoms. It is a great way to
waste paper and there are other way to obtain this information such as
with the builder commands or the analysis commands.
USER
The USER command is described in greater detail in section Interfacing to CONGEN.
TIME integer
The TIME command sets the value of TIMER
in COMMON
/TIMER/
to the specified value. This variable is used to time
different functions in the program.
- 1
- will print out the time to evaluate ENERGY.
- 2
- will print out individual component times in
ENERGY
, and the
times for various components of the EXEL
nbonds update.
WRNLEV integer
The WRNLEV command sets the value of the WRNLEV
variable in
COMMON
/TIMER/
to the specified value. At the present
(17-Nov-1990) this variable is not widely used. Suggested values for
future use:
- 0
- (default) should print brief warning and error messages for conditions
that will affect outcome.
- 1
- more extensive information on errors and some information on normal
partial results and conditions.
- 2
- verbose error messages and more normal processing information for
debugging.
- 3
- all information that might be relevant to an error condition plus
checking results.
- 4, 5
- debugging levels for anything you might conceivably want.
- 10 or higher
- for term by term outputs from energy routines, or other tasks where huge
amounts of data useful only in debugging might be generated.
NOBOmb
The NOBOMB command prevents the program from bombing out if it
hits an unrecognized command in the main program. This is useful for
attempting interactive work.
DEBUg repeat(name int)
{ ALLHP }
{ ALLOC }
{ ALLSTK }
{ CGCONS }
{ CGEN }
{ CLSCHN }
{ CORE }
{ DIVZERO }
{ ESOLVE }
{ FPE }
{ GENIC }
{ GEPOL }
{ GRID }
{ GRIDSIG }
name ::= { INEXACT }
{ INVALID }
{ JCOUP }
{ MALLOC }
{ NOE }
{ OVERFLOW }
{ PARA }
{ PBE }
{ SEARCH_NEAR }
{ TLIMIT }
{ TREE }
{ UNDERFLOW }
{ XCONF }
The DEBUG command set the value for various debugging variables in
the system. Although you should check the source code for the exact
details, the following table gives an approximate idea of what the
variables do.
- ALLHP
- Displays details about heap management (see section Heaps). The settings
are as follows:
- 1
- Heaps are initialized when allocated.
- 2
- Messages are printed for all heap allocation
and freeing requests.
- 3
- The state of the heap is printed with every
allocation and freeing request.
- ALLOC
Attempts to debug allocation errors.
The value for this variable is interpreted in two parts, the units digit and
the tens digit. A "1" in the tens digit signifies messages about allocation
and freeing should be printed; a "0" means no messages. The units digit is
interpreted in a cumulative way (e.g., a setting of 2 implies the
actions of a setting of 1). Warning:, there is bug in
use of this debugging variable. Settings of 2 or higher are
incompatible with a setting of 1 or 0. Once you raise this
variable up over 1, do not lower it, otherwise, your storage
will be mutilated. Also, when the program exits, there is a chance
is will crash because some memory allocated before the first command
will not be freed correctly.
- 0
- No checking.
- 1
- All memory allocated through
cgalloc
is set to -1 before being
returned.
- 2
- The heap allocation routines are used to allocate memory for
cgalloc
,
and some checks are done when memory is freed using cgfree
.
- 3
- The memory allocated for
cgalloc
usage is checked everytime
cgalloc
and cgfree
are called.
- ALLSTK
Display details about stack management (see section Stacks).
- CGCONS
Controls display of information about dihedral angle and J coupling constraints
in conformational search. When set to 1, a list of atoms affected by these
constraints will be output. When set to 2, any sampling operation which is restricted
by a constraint will be displayed as well. This second setting can be voluminous.
- CGEN
Displays the course of a conformational search.
- CLSCHN
Displays details about chain closure.
- CORE
Controls whether core dumps are made on Unix systems when an error
occurs. By default, this variable is set to 1 which means core dumps are
taken. If set to 0, then core dumps are not taken.
- DIVZERO
Controls handling of division by zero errors. See FPE option for
meaning of the value. Normally set to abort, but only on the Iris.
- ESOLVE
Controls debugging display of simple equation solving. Default is 0, which
means no output.
- FPE
Controls handling of all floating point exceptions. Only the Iris
currently provides control of floating point exceptions. This option
effectively substitutes for DIVZERO, INEXACT, INVALID,
OVERFLOW, and UNDERFLOW. The value is interpreted as
follows:
- 0
- Continue execution, but substitute an appropriate operand when the error occurs.
- 1
- Print a count of errors on
stderr
at the end of execution.
- 2
- Provide a traceback each time the error occurs.
- 3
- Abort execution with an attempted core dump when the error occurs.
Note: on SGI Irix 5.3, there is a bug which prevents this code
from functioning. As a result, no exceptions are trapped at all.
On SGI Irix 6.0, libfpe is unavailable, so no exceptions are trapped
here either. Hopefully, future releases will take care of the problem.
- GENIC
Controls whether displays are made of atoms which are not found within
generated segments. Normally, topology files are set up so that some atoms
at the end of segments stick out the sides, and these generate errors.
- GEPOL
Turns on the print flag in the GEPOL code, see section GEPOL Command -- Set GEPOL Defaults,
and section Static Properties of Atoms.
- GRID
Displays details about the space grid.
- GRIDSIG
Displays a signature of the space grid before every node expansion.
This may be useful in diagnosing malloc errors or problems with serial vs
parallel execution.
- INEXACT
No effect on any machine. Although the IEEE floating standard provides
this exception for inexact arithmetic, it is not controllable on any machine.
See FPE option for
meaning of the value.
- INVALID
Controls handling of invalid floating point errors. See FPE option for
meaning of the value. Normally set to abort, but only on the Iris.
- JCOUP
When non-zero, display details about the J coupling NMR constraints.
- MALLOC
When non-zero, turns on debugging of
malloc
calls on some machines
(definitely, the Iris).
- NOE
Displays details about NOE constraint calculations.
A setting of 1 shows the energy calculations in detail.
A setting of 2 also display the NOE constraints prior to an energy calculation.
- OVERFLOW
Controls handling of overflow errors. See FPE option for
meaning of the value. Normally set to abort, but only on the Iris.
- PARA
Controls output of debugging messages from parallel processing routines.
- PBE
Controls debugging information from the Poisson-Boltzmann electrostatics code.
When set to 1, additional useful information about the calculation is output.
At 4, statistics about the potential and charge densities are displayed,
as well as other information. In the PBE TEST command, a setting of 3
will display the calculation of the terms in the spherical cavity test potential.
At 5, individual calculations of the potential are output (don't do this
unless you want to fill your disk!). At 11, the number of iterations for
the test cavity calculations will be put into the potential grid.
- SEARCH_NEAR
Displays details about searches near atoms.
- TLIMIT
Display information about atoms which have been retarded during
molecular dynamics using the TLIMIT option, see section Running Molecular Dynamics.
The options are interpreted as follows:
- 0
- All debugging output is turned off.
- 1
- The number of
atoms which have been retarded on each dynamics cycle is displayed.
- 2
- Each atom whose motion is slowed is displayed. This can be voluminous.
- TREE
Displays details about the conformational search tree.
- UNDERFLOW
Controls handling of underflow errors. See FPE option for
meaning of the value. Normally ignored, but only on the Iris.
- XCONF
Controls whether accessible surfaces are printed when the ESURF option
is used with the XCONF command.
WEIGHT repeat(name real) [END]
{ EB }
{ ET }
{ EP }
{ EI }
name ::= { ENB }
{ EEL }
{ EHB }
{ EC }
{ ENOE }
{ EJCP }
The WEIGHT command set the weight for each term in the potential energy function.
The default value is 1.0 if no weights are specified. Since the code for the nonbonded
and electrostatic terms are interdependent, these weights nust be the same. If not,
the weight for the electostatic term will default to the weight for the nonbonded term.
- EB
Changes the weight of the bond energy term.
- ET
Changes the weight of the angle energy term.
- EP
Changes the weight of the torsion angle energy term.
- EI
Changes the weight of the improper torsion angle energy term.
- ENB
Changes the weight of the van der Waals energy term.
- EEL
Changes the weight of the electrostatic energy term.
- EHB
Changes the weight of the hydrogen bond energy term.
- EC
Changes the weight of the harmonic atom and dihedral angle constraints, see section Constraints.
- ENOE
Changes the weight of the Nuclear Overhauser Enhancement constraints,
see section NMR Constraints.
- EJCP
Changes the weight of the NMR J coupling constraints,
see section NMR Constraints.
The GAUSSIAN command is used to invoke the Gaussian
program(22)
Currently, this interface may only be used to calculate partial
charges for fragments of the system. It uses the Gaussian 92 program to
calculate the wavefunction and electrostatic field for the fragment.
Four different methods are provided for calculating partial charges from
the wavefunction.
All of these methods determine partial charges by performing a
least squares fit of the potential generated by the partial charges to
the potential calculated using the wavefunction.
The fundamental difference between the methods is the layout of points
where the electrostatic potential is determined by the wavefunction, and
subsequently used for the least squares fit of atomic charges.
The first method, PDM, uses two programs written by Don Williams,
PDM88 and PDGRID.(23)(24) The PDGRID program
lays out a grid of points around the fragment where the potential will
be calculated, and the PDM88 program does the least squares fit to
determine the best values for the partial charges.(25)
The other methods have been incorporated directly into Gaussian 92, and
use different grid layouts. There is the scheme due to Merz, Singh, and
Kollman,(26)(27) identified by the keyword,
MK; the scheme due to Chirlian and Francl,(28)
identified by the keyword, CHELP; and the scheme to Breneman and
Wiberg,(29) identified by the keyword, CHELPG.
All of these schemes have their own values for van der Waals radii
encoded within them. However, the default in this interface is to use
the radii from the parameters in CONGEN. If you want to use the radii in
the external programs, use the EXTRADII keyword.
This command simplifies the use of these three programs. You specify the
atoms you want charges for, and the programs are invoked in turn to
calculate the charges. Remember that the time for the calculation
increases approximately with the fourth power of the number of
electrons. A number of files are generated when this command is executed.
Normally, these files are deleted after the command is complete, but
you can request that they be saved.
Note that the collection of atoms that you specify should be
a complete molecule including hydrogens. It does not make physical sense
to do anything different, although the program does not check
for completeness.
If you use this command to calculate results that are eventually
published, please ensure that both Gaussian 92 and the fitting
scheme you use is properly referenced.
GAUSsian CHARges {SELECT atom-selection END}
[BASIs word] [SCF word] [TOTAl real] [MEMOry int]
[UNIT unit] [PREFix word] [EXTRadii] [DIPOle]
[PDM [UNDEr real] [SHELl real] [SPACing real] ]
[MK ]
[CHELP ]
[CHELPG ]
[STEPs repeat(step-options) END]
[SAVE] [NORUn]
[MERGe atom-selection END]
repeat( AVERage atom-selection END )
[ ALL ]
[ NONE ]
[ [NO]CREAte ]
[ [NO]HF ]
step-option ::= [ [NO]GRID ]
[ [NO]POTEntial ]
[ [NO]FIT ]
[ [NO]SCAN ]
[ [NO]DELEte ]
See section Atom Selection, for the syntax of an atom-selection.
The GAUSSIAN command functions by writing a set of input files
for Gaussian, PDGRID, and PDM88; preparing a Bourne shell
script to execute each program in turn; executing the script; and reading
the results. The options are interpreted as follows:
- CHARGES
- This keyword must be specified. It is anticipated that other functions
of GAUSSIAN will be invoked in the future.
- SELECT
- The SELECT keyword is used to demark an atom-selection
that identifies the atoms to be selected. By default, no atoms are
selected, so you must specify something.
- BASIS
- The word which follows BASIS gives the basis set to be used
by Gaussian. The default is 6-31G.
- SCF
- The SCF option is used to add additional keywords to the SCF
command in Gaussian.
- TOTAL
- The TOTAL option specifies the total charge of the fragment. The default
is zero.
- MEMORY
- This specifies the number of words of memory to be allocated to Gaussian.
This option is used for the %mem keyword. The default is 5000000.
- UNIT
- This option is used to specify the Fortran unit for all I/O done by
this command. The default is 99.
- PREFIX
- This option specifies the file name prefix to use for all the
intermediate files generated by this command. The default is
`cgq_<pid>' where `<pid>' is the current process id.
- EXTRADII
- This option specifies that the atomic radii in the external programs
be used. Normally, CONGEN supplies the programs with the van der Waals radii
from the parameter file.
- PDM
- This keyword indicates that the PDM programs of Don Williams
should be used.
- UNDER
- This keyword specifies the distance under the van der Waals radii for
grid points to be placed. The default is 0.0, which means that no points
are placed under the van der Waals radius. This option is used only if
the PDM option is selected.
- SPACING
- This keyword specifies the spacing between grid points. The default is 0.8
Angstroms. This option is used only if the PDM
option is selected.
- SHELL
- This keyword specifies the maximum distance of any grid point to the van der Waals
surface of a molecule. This option is used only if the PDM
option is selected.
- MK
- This keyword specifies that the Singh, Besler, Merz and Kollman gridding scheme
should be used.
- CHELP
- This keyword specifies that the CHELP gridding scheme
should be used. Note that this keyword may not be abbreviated.
- CHELPG
- This keyword specifies that the CHELPG gridding scheme
should be used. Note that this keyword may not be abbreviated.
- DIPOLE
- For the MK, CHELP, and CHELPG schemes, this option specifies
that the fit of charges shall also consider the quantum mechanical calculation
of the dipole moment.
- MERGE
- This option requests that charges on the selected atoms be merged with the
atoms that they are bound to.
Only atoms which have exactly one bond can be treated this way, and
this option was created primarily for use with hydrogens.
- AVERAGE
- The AVERAGE options are used to average charges. They are most
appropriate when certain atoms are symmetric, but are exposed to
different electrostatic environments. You can specify as many
AVERAGE options as needed, but the sets of atoms may not overlap
within the system being analyzed. After the averaging and merging
steps are performed, the program adjusts all the charges to bring the
total to the value you have specified.
- STEPS
- The STEPS option allows detailed control over the execution of the
programs. CONGEN maintains a list of boolean variables which specify
whether certain steps will be executed, and the STEPS option
controls the setting of these variables. Each option is interpreted
sequentially to affect the variables. The presence of the string,
NO, preceding some of the keywords means that the variable should
be turned off. The options are interpreted as follows:
- ALL
- Turn on all steps.
- NONE
- Turn off all steps.
- CREATE
- Create all the input files and shell scripts necessary to run the charge
calculation. If a particular step is omitted from the operation, the
affected command in the shell script is commented out, so you can manually
enable it by editing the script. In order to use this command effectively,
you must use the same file prefix every time.
- HF
- Perform the initial Hartree-Fock single point calculation.
In all methods except PDM, this option also controls the
steps up to the FIT step.
- GRID
- Execute PDGRID which
lays out the grid where the electrostatic potential will be evaluated.
This applies only to the PDM option.
- POTENTIAL
- Calculate the electrostatic potential on the grid points from the
wave function.
This applies only to the PDM option.
- FIT
- Invoke PDM88 which fits the charges to the calculated potential.
This applies only to the PDM option.
- SCAN
- Scan the results from the calculations back into CONGEN.
- DELETE
- Delete intermediate files.
- SAVE
This option requests that all intermediate files be saved after the command runs.
It has the same effect as STEPS NODELETE END option.
Normally, all intermediate files are deleted.
- NORUN
This option requests that all initial input files be saved and not executed.
This is useful when you want to modify the Gaussian and PD input files created
by the program. Note that this option complete overrides the settings in the
STEPS option.
The following table gives the file types for all the intermediate files used:
- `.chk'
- Gaussian checkpoint file
- `.hfi'
- Hartree-Fock input
- `.hfo'
- Hartree-Fock output
- `.gri'
- PDGRID input file
- `.gro'
- PDGRID output file
- `.grd'
- PDGRID grid specification
- `.pti'
- Electrostatic potential calculation input
- `.pto'
- Electrostatic potential calculation output
- `.pdi'
- PDM88 input.
- `.pdo'
- PDM88 output.
- `.cho'
- Filtered charge calculation output.
- `.sh'
- Shell script to run everything.
- `.log'
- Output from shell.
The TEST command is used to test internal operations within CONGEN.
Currently, three test operations are provided; energy derivatives, calculation
of maximum contact distances for the VAVOID sidechain option when the
hydrogen bond energy replaces the van der Waals energy, and calculation
of torsion angle minima.
TEST [DERIV [DELTA real] [CUT real]]
[VAHB [MAXEVDW real]]
[PHIMIN [SGRID real] [SYMMETRY int]]
The energy derivative test is invoked by the DERIV keyword.
The code works by computing numerical
derivatives of energy as shown:
and comparing them against the analytic
derivatives computed by the code.
Statistics of the differences are
calculated and a histogram of the differences is displayed.
The keyword, DELTA, specifies the value of h in the above
expression. The default value is 0.001. For single precision
arithmetic, this value is close to optimum.(30) The keyword, CUT,
specifies a printing cutoff. Any difference in derivative whose
magnitude exceeds this value will be individually displayed.
The VAVOID hydrogen bond test checks the calculation of maximum distance for
a given value of MAXEVDW for the hydrogen bond potential. It is
invoked using the VAHB keyword. It performs the
calculation for all hydrogen bond parameters and shows the distance along with the
calculated energy at that distance.
The keyword, MAXEVDW, is used to set a value for the cutoff energy. The default
value is 20 kcal/mole.
The torsion energy minimum test is designed to check the code which finds
torsion angle values to use when performing a sidechain degree of freedom
in a conformational search, see section Sidechain Degree of Freedom. In order to see the
effect of this test, the CGEN debug variable, see section Set Debugging Variables -- DEBUG, must be
set to 2 or larger. The SYMMETRY keyword specifies the rotational symmetry
for the clump and defaults to 1. The SGRID keyword specifies the sidechain
grid to use. A value of -1 means use the minimum energy periodicity. Positive
values are interpreted as the grid in degrees. The default is -1.
The GEPOL command is used to set defaults for GEPOL surface
calculations(31) in the analysis facility. See section Static Properties of Atoms,
for more information about the GEPOL surfaces.
There is also an experimental capability to invoke the GEPOL_INCR
subroutine using the RUN option.
GEPOL [NDIV int] [OFAC real] [RMIN real]
[RSOL real] [RGRI real] [CAVIty real] [[NO]BULK]
[RUN run-options]
run-options ::= [WSURF] {INIT } [PEER <UNIT>] [PROP <UNIT>]
[ASURF] {PUSH atom-selection}
[ESURF] {POP }
The options in the GEPOL command control parameters used by the
GEPOL algorithm.(32)(33)(34) They have the following interpretation:
- NDIV
- NDIV specifies the division level for the triangles on the
surface. It may be a number between 1 and 5. The accuracy of the
calculation improves as NDIV rises, but the CPU time rises a lot
faster. The default for this command is 3.
- OFAC
- This parameter will be used only if the molecular surface is computed.
In the second field goes a real number that can take values
between 0.0 and 1.0. This parameter is the Overlapping FACtor.
The accuracy improves as the OFAC value increases. The default
value is 0.8.
- RMIN
- This parameter will be used only if molecular surface is computed.
In the second field goes a real number that can take values
larger than 0.0. This parameter is the radius of the smallest
sphere that can be created. The accuracy improves as the RMIN
value decreases. The default value is 0.50.
OFAC and RMIN are the parameters that control the creation
of new spheres.
- RSOL
- This parameter will be used only if accessible or molecular surface is computed.
In the second field goes a real number. It is the probe or
solvent radius. The default value is 1.4 Angstroms.
- RGRID
- This parameter sets the dimension of the space grid used to find neighbors when
GEPOL runs. The default setting of 2.5 Angstroms was optimal for running GEPOL
on a large protein. A value of 0.0 will result in the dimension being set
by code within `$CGS/gridc.c'.
- CAVITY
- This parameters sets the cavity energy term. If this value is non-zero,
then the conformational search command, see section Conformational Search,
will include a energy term equal to the molecular surface area calculated
by GEPOL multiplied by this cavity energy factor. The units for
the parameter are
A good value for
this parameter is 0.072.(35)
- BULK
- The BULK flag controls whether the BULK subroutine is invoked
as part of GEPOL. BULK adds spheres in the interior of large molecules
to speed up the molecular surface generation.
The RUN keyword specifies that the GEPOL_INCR subroutine is to
be run. The keywords; WSURF, ASURF, and ESURF; specify
van der Waals, accessible, and molecular surfaces, respectively.
The keywords; INIT, PUSH, and POP; specify
initialization, pushing the selected set of atoms, and popping the
last set of atoms, respectively. The PEER keyword specifies a unit
where all the spheres will be written as input to the peer program,
see section peer -- Interactive Molecular Display Program. The PROP keyword specifies a unit where the
atomic surfaces will be written as a property table suitable for use
in the Analysis Facility, see section Table Input and Output.
The RLIMIT command is used to set and display computer
resource limits. It is critically important to set the stacksize
limit (STACK) when running very large calculations in parallel
on an SGI workstation. The command is only implemented on a SGI
system.
RLIMIT repeat(limit [int ])
[INFinity]
[CORE ]
[CPU ]
[DATA ]
limit ::= [FSIZE ]
[NOFILE]
[STACK ]
[VMEM ]
[RSS ]
The options in the RLIMIT command set computer resource limits.
After the command finishes execution, the current limits are printed.
The keywords have the following meaning:
- CORE
- Sets the maximum core file size in bytes. A zero limit means
no core files.
- CPU
- Sets the maximum CPU utilization per process in seconds.
- DATA
- Specifies the maximum extent of the data segment in a process in
units of bytes.
- FSIZE
- The maximum size of a file in bytes. N.B. In Irix 5.3 (and probably 6.1),
this limit can only be changed downward, because of a bug in setting
this limit, where the system call requires both soft and hard limits
to be changed simultaneously.
- NOFILE
- Maximum number of open files.
- STACK
- Maximum stack size. For parallel execution of CONGEN,
this parameter is very important. The default value of the stack size,
64 MB, results in memory consumption of 1.28GB for a 20 processor
run. CONGEN can run very well using a stack size of 1MB, which
is recommended for these large jobs.
- VMEM
- Maximum amount of virtual memory allowed to a process. This is effectively
the sum of the data segment and stack memory.
- RSS
- Declared amount of physical memory required by the process. When
total system memory is limited, any process using more than this
amount of memory will have its pages written to the swap file first.
The SHELL executes a command under the Bourne shell for
Unix systems, or under DCL for VMS systems. The syntax is
SHELL string
where string is the command to be executed.
Go to the previous, next section.