Learn R Programming

pbdMPI (version 0.2-3)

SPMD Control: Sets of controls in pbdMPI.

Description

These sets of controls are used to provide default values in this package.

Arguments

format

Objects contain several parameters for communicators and methods.

Details

The elements of .SPMD.CT are default values for controls including lcl{ Elements Default Usage comm 0L a communicator index intercomm 2L an inter communicator index info 0L an info index newcomm 1L a new communicator index op "sum" an operation port.name "spmdport" an operation print.all.rank FALSE if all ranks print message print.quiet FALSE if print/cat rank information rank.root 0L a rank of root rank.source 0L a rank of source rank.dest 1L a rank of destination request 0L a request index serv.name "spmdserv" a service name status 0L a status index tag 0L a tag number unlist FALSE if unlist returning divide.method "block" a way to divide jobs or data mpi.finalize TRUE if shutdown MPI quit TRUE if quit when errors occur msg.flush TRUE if flush message for comm.cat/comm.print msg.barrier TRUE if barrier message for comm.cat/comm.print Rprof.all.rank FALSE if call Rprof on all ranks lazy.check TRUEE if use lazy check on all ranks }

The elements of .SPMD.OP list the implemented operations for reduce() and allreduce(). Currently, four operations are implemented "sum", "prod", "max", and "min".

The elements of .SPMD.IO are default values for input and output including lcl{ Elements Default Usage max.read.size 5.2e6 max of reading size (5 MB) max.test.lines 500 max of testing lines read.method "gbd" default reading method balance.method "block" default load balance method } where balance.method is only used for "gbd" reading method when nrows = -1 and skip = 0 are set.

The elements of .SPMD.TP are default values mainly for task pull including lcl{ Elements Default Usage bcast FALSE if bcase() objects to all ranks try TRUEE if use try() in works try.silent FALSE if silent the try() message See task.pull() for details. }

References

Programming with Big Data in R Website: http://r-pbd.org/