parallelMap
and allows to set a
parallelMap
with a matching level are
parallelized.parallelStart(mode = "local", cpus, ...,
level = as.character(NA), log = NULL)
character(1)
]
Which parallel mode
should be used: integer(1)
]
Number of used cpus.
Default is mpi.universe.size
for
snowfall/MPI and 1 otherwise.sfInit
currently.character(1)
]
You can set this so
only calls to parallelMap
are parallelized
that have the same level specified. Default is NA
which means all calls are parallelized.character(1)
]
Path to an existing
directory where a log files for each job is stored via
sink
. Note that all nodes must have write
access to exactly this path. Files are named
"sfStop
,
sfSetMaxCPUs
,
sfInit
,
sfClusterSetupRNG
are called in
this order.