lply: Powerful lapply function with a qsub(Torque/PBS)-cluster as...

In sorhawell/fastRditijuu: Easy cluster computing

Description

Powerful lapply function with a qsub(Torque/PBS)-cluster as backend

Usage

lply(X, FUN, conf = NULL, user = NULL, host = "login.gbar.dtu.dk",
  transferHost = NULL, keyPath = NULL, Rscript = T, packages = c(),
  max.nodes = 8, local = FALSE, globalVar = list(), async = T,
  nCores = 4, qsub.walltime = "00:09:59", qsub.proc = 4, qsub.nodes = 1,
  ...)

Arguments

X	object to map
FUN	mapper function
user	username for server
host	server adress (will be combined as user@host)
keyPath	path to local ssh key (mainly for windows)
Rscript	if true verbose, however only supported by gbar, nor compute.cluster
packages	required packages on server
max.nodes	maximum nodes reuired, do not set higher than 80 (79 if async=T)
local	should the lply run on local computer (only for debugging/testing)
globalVar	list of global variables
async	true/false, if to run lply loop asynchronously
nCores	number of cores(>1 will trigger parallel::mclapply)
qsub.walltime	change wall time qsub, will apply to both master and slaves
qsub.proc	how many processes to ask for per job, 1 unless using doParallel etc.
qsub.nodes	how many nodes to ask for per job, leave unchanged if in doubt
...

Value

list of results

sorhawell/fastRditijuu documentation built on May 30, 2019, 6:32 a.m.