Parallel Bootstrapping with plssem

This vignette shows how to run bootstrap standard errors in parallel using boot.ncpus and boot.parallel.

Example model (ordered indicators)

m <- '
  X =~ x1 + x2 + x3
  Z =~ z1 + z2 + z3
  Y =~ y1 + y2 + y3

  Y ~ X + Z + X:Z
'

Windows: PSOCK cluster (`boot.parallel = "snow"`)

Use "snow" on Windows (PSOCK clusters; similar to a “multisession” approach).

fit_win <- pls(
  m,
  data          = oneIntOrdered,
  ordered       = colnames(oneIntOrdered),
  bootstrap     = TRUE,
  boot.R        = 500,
  boot.ncpus    = 2,
  boot.parallel = "snow",
  boot.iseed    = 123
)
summary(fit_win)

Linux: forked processes (`boot.parallel = "multicore"`)

Use "multicore" on Linux (forking; not available on Windows).

fit_linux <- pls(
  m,
  data          = oneIntOrdered,
  ordered       = colnames(oneIntOrdered),
  bootstrap     = TRUE,
  boot.R        = 500,
  boot.ncpus    = 2,
  boot.parallel = "multicore",
  boot.iseed    = 123
)
summary(fit_linux)