Skip to contents

When using MPI, we save the sample immediately into a file, each rank saves to its own file. This function collects all of these smaller samples into one. The samples should be saved with saveRDS().

Usage

loadSubSample_mpi(
  files,
  size = NA,
  selection = NA,
  mc.cores = parallel::detectCores()
)

Arguments

files

the files where the individual samples are stored

size

sub sample size, if not set, the whole sample is returned

selection

integer index vector or logical vector indicating which temperatures to return: betaselection is returned, in decreasing order of beta.

mc.cores

defaults to the total number of cores, but can be reduced with this option.

Value

a list of matrices, by temperature, concatenated.