pbdMPI (version 0.3-0)

global print and cat: Global Print and Cat Functions

Description

The functions globally print or cat a variable from specified processors, by default messages is shown on screen.

Usage

comm.print(x, all.rank = .pbd_env$SPMD.CT$print.all.rank,
           rank.print = .pbd_env$SPMD.CT$rank.source,
           comm = .pbd_env$SPMD.CT$comm,
           quiet = .pbd_env$SPMD.CT$print.quiet,
           flush = .pbd_env$SPMD.CT$msg.flush,
           barrier = .pbd_env$SPMD.CT$msg.barrier,
           con = stdout(), ...)
comm.cat(..., all.rank = .pbd_env$SPMD.CT$print.all.rank,
         rank.print = .pbd_env$SPMD.CT$rank.source,
         comm = .pbd_env$SPMD.CT$comm,
         quiet = .pbd_env$SPMD.CT$print.quiet, sep = " ", fill = FALSE,
         labels = NULL, append = FALSE, flush = .pbd_env$SPMD.CT$msg.flush,
         barrier = .pbd_env$SPMD.CT$msg.barrier, con = stdout())

Arguments

x

a variable to be printed.

...

variables to be cat.

all.rank

if all ranks print (default = FALSE).

rank.print

rank for printing if not all ranks print (default = 0).

comm

communicator for printing (default = 1).

quiet

FALSE for printing rank number.

sep

sep argument as in the cat() function.

fill

fill argument as in the cat() function.

labels

labels argument as in the cat() function.

append

labels argument as in the cat() function.

flush

if flush con.

barrier

if barrier con.

con

stdout() is the default to print message.

Value

• A print() or cat() is called for the specified processors and the messages of the input variables is shown on screen by default.

Details

Warning: These two functions use barrier() to make sure the well printing process on screen, so should be called by all processors to avoid a deadlock. A typical misuse is called inside a condition check, such as if(.comm.rank == 0) comm.cat(...).

rank.print can be a integer vector containing the ranks of processors which print messages.

References

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

Examples

### Save code in a file "demo.r" and run with 2 processors by
### SHELL> mpiexec -np 2 Rscript demo.r

### Initial.
suppressMessages(library(pbdMPI, quietly = TRUE))
init()

### Example.
comm.print(comm.rank(), rank.print = 1)

### Finish.
finalize()