Folks, I have some computations in a function that create some large matrices. I have been in the habit in these circumstances to call "null out" a matrix once used and call gc(). Some pseudo code: theFunction<-function(x, y, z, 100000000) { myMatrix <- black.box.function(x, y, z, 100000000) interesting.data<-myFunc(myMatrix) myMatrix<-NULL gc() .... interesting.data/pi } Would it be different if I chose to replace myMatrix<-NULL with rm(myMatrix)? The gc docs are fine but I was wondering if there were any other tips or links on the subject. Thanks for your time, KW --