SimFactory’s functionality is divided into a number of separate commands. Each command can take both mandatory arguments as well as options.
The essential SimFactory commands are:
build | build cactus either locally or remotely |
create | create a simulation |
create-submit | create and submit a simulation |
purge | purge a simulation |
setup | interactively configure simfactory |
stop | stop an active simulation |
submit | submit a simulation |
sync | sync cactus sourcetree to a remote machine |
The following commands are also available:
archive | archive a simulation or an individual restart |
checkout | checkout thorns using GetComponents |
cleanup | clean up one or more simulations |
create-run | create and run a simulation |
execute | execute a command |
get-archived-simulation | retrieve an archived simulation by simulationid |
get-output-dir | get the output directory for a given simulation/restart |
interactive | initiate an interactive session on a compute node |
list-archived-simulations | list archived simulations |
list-conf | |
list-configurations | list simfactory cactus configurations |
list-machines | list all the machines in the machine database |
list-sim | |
list-simulations | list simfactory cactus simulations |
login | launch an interactive shell on a remote machine |
print-mdb | print parsable mdb |
print-mdb-entry | list information about a single machine |
remove-submitscript | remove a submit script from a configuration |
run | launch a simulation |
run-debug | launch simulation using an interactive debugger |
setup-silent | silently create a set of reasonable defaults |
show-output | show the output of a simulation |
whoami | what is the name of the machine simfactory is running on |
Todo
Not all commands are documented
All SimFactory commands take the following options. These are the important ones:
Execute command on a remote machine instead of the local machine
and these are the remaining ones:
CPU allocation to use
Additional Arguments
Base directory for simulations
INI file to use as the config database
Output to console as opposed to writing to a log file
Force cleanup even though a simulation may still be running.
Machine hostname
Switch to directory
Force the usage of a certain machine entry
INI file to use as the machine database
Number of SMT threads (aka hyperthreads) to use
Number of threads per process to use
Cores per node to request
Number of cores per node to use
Total number of threads to use
Queue to use for job submission
the Cactus path to use on the remote system
see –remote
run script for job execution
Scratch directory
Base directory for source tree
Submit script for job submission
Be verbose with output
Maximum run time for your simulation
Todo
These don’t all apply to all commands. Why are they in the common.ini file?
Todo
Include arguments and default values