scatter-method {pbdMPI}

A Rank Scatter Objects to Every Rank

Description

This method lets a rank scatter objects to every rank in the same communicator. The default input is a list of length equal to ‘comm size’ and the default return is an element of the list.

Usage

scatter(x, x.buffer = NULL, x.count = NULL, displs = NULL,
        rank.source = .pbd_env$SPMD.CT$rank.source,
        comm = .pbd_env$SPMD.CT$comm)

Arguments

x

an object of length ‘comm size’ to be scattered to all ranks.

x.buffer

a buffer to hold the return object which probably has 'size of element of x' with the same type of the element of x.

x.count

a vector of length ‘comm size’ containing all object lengths.

displs

c(0L, cumsum(x.count)) by default.

rank.source

a rank of source where elements of x scatter from.

comm

a communicator number.

Details

All elements of x are likely presumed to have the same size and type.

x.buffer, x.count, and displs can be NULL or unspecified. If specified, the type should be one of integer, double, or raw specified correctly according to the type of x.

If x.count is specified, then the spmd.scatterv.*() is called.

Value

An element of x is returned according to the rank id.

Methods

For calling spmd.scatter.*():

signature(x = "ANY", x.buffer = "missing", x.count = "missing")

signature(x = "integer", x.buffer = "integer", x.count = "missing")

signature(x = "numeric", x.buffer = "numeric", x.count = "missing")

signature(x = "raw", x.buffer = "raw", x.count = "missing")

For calling spmd.scatterv.*():

signature(x = "ANY", x.buffer = "missing", x.count = "integer")

signature(x = "ANY", x.buffer = "ANY", x.count = "integer")

signature(x = "integer", x.buffer = "integer", x.count = "integer")

signature(x = "numeric", x.buffer = "numeric", x.count = "integer")

signature(x = "raw", x.buffer = "raw", x.count = "integer")

Author(s)

Wei-Chen Chen [e-mail], George Ostrouchov, Drew Schmidt, Pragneshkumar Patel, and Hao Yu.

References

Programming with Big Data in R Website: https://pbdr.org/

See Also

bcast().

Examples

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

spmd.code <- "
### Initialize
suppressMessages(library(pbdMPI, quietly = TRUE))
.comm.size <- comm.size()
.comm.rank <- comm.rank()

### Examples.
N <- 5
x <- split(1:(N * .comm.size), rep(1:.comm.size, N))
y <- scatter(lapply(x, matrix, nrow = 1))
comm.print(y)
y <- scatter(x, double(N))
comm.print(y)

### Finish.
finalize()
"
pbdMPI::execmpi(spmd.code, nranks = 2L)