Parallelize 'crossmap' functions

Henrik Bengtsson

+ =

The futurize package allows you to easily turn sequential code into parallel code by piping the sequential code to the futurize() function. Easy!

TL;DR

library(futurize)
plan(multisession)
library(crossmap)

xs <- list(1:5, 1:5)
ys <- xmap(xs, ~ .y * .x) |> futurize()

Introduction

The crossmap package adds to the [purrr]-set of functions. For example, xmap() can apply a function to every combination of elements in a list, e.g.

library(crossmap)

# Multiply the 15 combinations of values in 1:3 and -2:2
xs <- list(1:3, -2:2)
ys <- xmap(xs, function(x, y) x * y) |> futurize()

Here xmap() evaluates sequentially over each combination of (.y, .x) elements. The crossmap package provides its own future-counterpart functions, e.g. there is a future_xmap() that mimics xmap(). The futurize() transpiles xmap() into future_xmap(), meaning you can do:

library(futurize)

# Multiply the 15 combinations of values in 1:3 and -2:2
xs <- list(1:3, -2:2)
ys <- xmap(xs, function(x, y) x * y) |> futurize()

to process this xmap() call concurrently, which allows you to execute it on a set parallel workers, e.g.

plan(multisession)

The built-in multisession backend parallelizes on your local computer and it works on all operating systems. There are other parallel backends to choose from, including alternatives to parallelize locally as well as distributed across remote machines, e.g.

plan(future.mirai::mirai_multisession)

and

plan(future.batchtools::batchtools_slurm)

Supported Functions

The futurize() function supports parallelization of the following crossmap functions:

• imap_vec(), map_vec(), map2_vec(), pmap_vec(), xmap_vec()
• xmap()
• xmap_chr(), xmap_dbl(), xmap_int(), xmap_lgl(), xmap_raw()
• xmap_dfc(), xmap_dfr()
• xmap_mat(), xmap_arr()
• xwalk()