[R] Working memory full

[Peter van Rijn]

I have a question concerning the working memory used by R.

I am running a simulation job (a function which replicates a simulation) which calls several other functions that I have written. Something like this:

name1<-function(number of replications){
	for(i in 1:replication)
		name2(parameters)
		name3()
		# get results and aggregate
		open..
		write(..., append=T)
	}
}

name2<-function(parameters){
	# simulate data
	...	
	# save data and parameters
	write...
}

name3<-function(){
	# get data and parameters
	open...
	# analyse data
	...
	# save results
	write..	
}

name1(100)


However, after a couple of replications working memory seems to be full.

Error: cannot allocate vector of size ...Kb
In addition: Warning message: 
Reached total allocation of 126Mb: see help(memory.size)

This puzzles me, because all arrays (Note: large arrays, i.e., 500x500) are defined locally (within each function) and redefined every replication (the meta-function). So my question is: Why is R eating up my working memory while I think I am redefining the same (local) arrays? Any thoughts on how this can be circumvented?

Thanks,

Peter.