Parallelize 'vegan' functions

+ =

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(vegan)

data(dune)
data(dune.env)
dune.mrpp <- with(dune.env, {
  mrpp(dune, Management) |> futurize()
})

Introduction

The vegan package provides methods for community and vegetation ecologists. Some of the functions have built-in support for parallelization, which futurize simplifies further.

Example: MRPP

Example adopted from help("mrpp", package = "vegan"):

library(futurize)
plan(multisession)
library(vegan)

data(dune)
data(dune.env)
dune.mrpp <- with(dune.env, {
  mrpp(dune, Management) |> futurize()
})

This will parallelize the computations, given that we have set up parallel workers, e.g.

plan(multisession)

The built-in multisession backend parallelizes on your local computer and 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)

Example: anova() for 'cca' objects

The anova() S3 method for 'cca' objects supports parallelization via the parallel argument. With futurize, you can parallelize this directly. Example adopted from help("anova.cca", package = "vegan"):

library(futurize)
plan(multisession)
library(vegan)

data(dune)
data(dune.env)
ord <- cca(dune ~ A1 + Management, data = dune.env)
res <- anova(ord, permutations = 99) |> futurize()

Supported Functions

The following vegan functions are supported by futurize():

• adonis()
• adonis2()
• anova() for 'cca'
• anosim()
• cascadeKM()
• estaccumR()
• mantel()
• mantel.partial()
• metaMDSiter()
• mrpp()
• oecosimu()
• ordiareatest()
• permutest() for 'betadisper', and 'cca'

Without futurize: Manual PSOCK cluster setup

For comparison, here is what it takes to parallelize mrpp() using the parallel package directly, without futurize:

library(vegan)
library(parallel)

data(dune)
data(dune.env)

## Set up a PSOCK cluster
ncpus <- 4L
cl <- makeCluster(ncpus)

## Run MRPP in parallel
dune.mrpp <- with(dune.env, {
  mrpp(dune, Management, parallel = cl)
})

## Tear down the cluster
stopCluster(cl)

This requires you to manually create and manage the cluster lifecycle. If you forget to call stopCluster(), or if your code errors out before reaching it, you leak background R processes. You also have to decide upfront how many CPUs to use and what cluster type to use. Switching to another parallel backend, e.g. a Slurm cluster, would require a completely different setup. With futurize, all of this is handled for you - just pipe to futurize() and control the backend with plan().