The future package provides a generic API for using futures in R. A future is a simple yet powerful mechanism to evaluate an R expression and retrieve its value at some point in time. Futures can be resolved in many different ways depending on which strategy is used. There are various types of synchronous and asynchronous futures to choose from in the future package.
This package, future.callr, provides a type of futures that utilizes the callr package.
For example,
> library("future.callr")
> plan(callr)
>
> x %<-% { Sys.sleep(5); 3.14 }
> y %<-% { Sys.sleep(5); 2.71 }
> x + y
[1] 5.85
This is obviously a toy example to illustrate what futures look like and how to work with them. For further examples on how to use futures, see the vignettes of the future package as well as those of future.apply and doFuture.
The future.callr package implements a future wrapper for callr.
| Backend | Description | Alternative in future package |
|---|---|---|
callr |
parallel evaluation in a separate R process (on current machine) | plan(multisession) |
When using callr futures, each future is resolved in a fresh background R session which ends as soon as the value of the future has been collected. In contrast, multisession futures are resolved in background R worker sessions that serve multiple futures over their life spans. The advantage with using a new R process for each future is that it is that the R environment is guaranteed not to be contaminated by previous futures, e.g. memory allocations, finalizers, modified options, and loaded and attached packages. The disadvantage, is an added overhead of launching a new R process.
(At the moment, I am neither aware of formal benchmarking of this extra overhead nor of performance comparisons of callr to alternative future backends.)
Another advantage with callr futures compared to multisession futures is that they do not communicate via R (socket) connections. This avoids the limitation in the number of parallel futures that can be active at any time that multisession futures and cluster futures in general have, which they inherit from SOCKcluster clusters as defined by the parallel package. The number of parallel futures these can serve is limited by the maximum number of open connections in R, which currently is 125 (excluding the three reserved by R itself). Note that these 125 slots have to be shared with file connections etc. To increase this limit, R has to be rebuilt from source. However, since callr futures rely on the callr package which does not make use of R-specific connections, there is no limit in the number of background R processes that can be used simultaneously.
A third advantage with callr futures, is that there is not risk for port-clashing with other processes on the system when clusters are set up (**), because callr does not rely on ports. Furthermore, on Windows, the firewall triggers an alert that the user needs to approve whenever a not-previously-approved port is requested by R - which happens also for local, non-public ports that are used by SOCKcluster:s. When using callr futures, no sockets and therefore no ports are involved.
(**) To lower the risk for such clashes SOCKcluster:s (of the parallel package) request random ports, but clashes still occur at times.
The future package provides a demo using futures for calculating a set of Mandelbrot planes. The demo does not assume anything about what type of futures are used. The user has full control of how futures are evaluated. For instance, to use callr futures, run the demo as:
library("future.callr")
plan(callr)
demo("mandelbrot", package = "future", ask = FALSE)