future (version 1.2.0)

cluster: Create a cluster future whose value will be resolved asynchroneously in a parallel process

Description

A cluster future is a future that uses cluster evaluation, which means that its value is computed and resolved in parallel in another process.

Usage

cluster(expr, envir = parent.frame(), substitute = TRUE, globals = TRUE, persistent = FALSE, workers = NULL, user = NULL, revtunnel = TRUE, homogeneous = TRUE, gc = FALSE, earlySignal = FALSE, label = NULL, ...)

Arguments

expr

An R expression.

envir

The environment from where global objects should be identified. Depending on "evaluator", it may also be the environment in which the expression is evaluated.

substitute

If TRUE, argument expr is substitute():ed, otherwise not.

globals

A logical, a character vector, or a named list for controlling how globals are handled. For details, see below section. This argument can be specified via the ... arguments for future() and futureCall().

persistent

If FALSE, the evaluation environment is cleared from objects prior to the evaluation of the future.

workers

A cluster object created by makeCluster().

user

(optional) The user name to be used when communicating with another host.

revtunnel

If TRUE, reverse SSH tunneling is used for the PSOCK cluster nodes to connect back to the master R process. This avoids the hassle of firewalls, port forwarding and having to know the internal / public IP address of the master R session.

homogeneous

If TRUE, all cluster nodes is assumed to use the same path to ‘Rscript’ as the main R session. If FALSE, the it is assumed to be on the PATH for each node.

gc

If TRUE, the garbage collector run (in the process that evaluated the future) after the value of the future is collected.

earlySignal

Specified whether conditions should be signaled as soon as possible or not.

label

An optional character string label attached to the future.

...

Additional arguments passed to the "evaluator".

Value

A ClusterFuture.

Details

This function will block if all available R cluster nodes are occupied and will be unblocked as soon as one of the already running cluster futures is resolved.

The preferred way to create an cluster future is not to call this function directly, but to register it via plan(cluster) such that it becomes the default mechanism for all futures. After this future() and %<-% will create cluster futures.

Examples

## Cluster futures gives an error on R CMD check on
## Windows (but not Linux or OS X) for unknown reasons.
## The same code works in package tests.


## Use cluster futures
cl <- parallel::makeCluster(2L)
plan(cluster, workers=cl)

## A global variable
a <- 0

## Create multicore future (explicitly)
f <- future({
  b <- 3
  c <- 2
  a * b * c
})

## A cluster future is evaluated in a separate process.
## Changing the value of a global variable will not
## affect the result of the future.
a <- 7
print(a)

v <- value(f)
print(v)
stopifnot(v == 0)

## CLEANUP
parallel::stopCluster(cl)