inla.qstat {INLA} | R Documentation |
Control and view a remote inla-queue of submitted jobs
inla.qget(id, remove = TRUE) inla.qdel(id) inla.qstat(id) inla.qlog(id) inla.qnuke() ## S3 method for class 'inla.q' summary(object,...) ## S3 method for class 'inla.q' print(x,...)
id |
The job-id which is the output from |
remove |
Logical If FALSE, leave the job on the server after retrival, otherwise remove it (default). |
x |
An |
object |
An |
... |
other arguments. |
inla.qstat
show job(s) on the server,
inla.qget
fetch the results (and by default remove
the files on the server), inla.qdel
removes
a job on the server and inla.qnuke
remove all jobs on the server.
inla.qlog
fetches the logfile only.
The recommended procedure is to use r=inla(...,
inla.call="submit")
and then do r=inla.qget(r)
at a later
stage. If the job is not finished, then r
will not be
overwritten and this step can be repeated. The reason for this
procedure, is that some information usually stored in the result
object does not go through the remote server, hence have to be
appended to the results that are retrieved from the server. Hence
doing r=inla(..., inla.call="submit")
and then later retrive
it using r=inla.qget(1)
, say, then r
does not contain
all the usual information. All the main results are there, but
administrative information which is required to call
inla.hyperpar
or inla.rerun
are not there.
inla.qstat
returns an inla.q
-object with information about current jobs.
Havard Rue
## Not run: r = inla(y~1, data = data.frame(y=rnorm(10)), inla.call="submit") inla.qstat() r = inla.qget(r, remove=FALSE) inla.qdel(1) inla.qnuke() ## End(Not run)