R in a netbook - system requirements for using R

Question

I know it's not a programming question but I'm in a hurry to choose a netbook like this and I haven't been able to find the minimum system requirements for an R installation (e.g. minimum RAM). I am interested in a small netbook so as to be able to use it in class. Has anybody used R in a netbook that would recommend for that use?

Answer 1

When I start an R process, and run ps axuw, my computer tells me

unutbu    6253  0.1  1.5  22660 16112 pts/3    Ss+  14:23   0:00 /usr/lib/R/bin/exec/R --no-readline

From this I conclude that R (version 2.9.2 on Ubuntu) requires about 23 MB of memory.

So I think a netbook with 768MB of RAM should be plenty to run R, unless you plan to deal with fairly large data sets.

Answer 2

I have found it helpful to use Emacs with ess when dealing with R, that way if I am doing a large computation, I can ssh to another (more powerful than a netbook) machine, start R, and attach to that machine with ess-remote. For example, I was recently doing a simulation in R, and tested about 100 iterations on my laptop (core 2 duo), then ssh'd to an 8 core machine to run 5000 iterations.

If you use iterators it is very easy to make your code able to switch from single to multiple core. You would only need to change 1 line of your code, from registerDoSeq() to registerDoMC(). With this workflow, it is easy to use a very low power machine for development, and easily use the resources of a larger machine for big simulations.

Answer 3

I'm runing R under Win XP on my Asus eee 1000HE with 2G of RAM. I mostly pull subsets of large dataset from Postgres through RODBC package so they arent bigger than 2-3k rows and up to a dozen of columns.

So far havent had any major problems and computation times are acceptable. Hovever, if you generate lots of graphics the bottleneck can most likely appear in the graphic card. I managed to push couple of large ones through ggplot2 but then you might be forced to wait little bit..

PS Under Win you can try Notepad++ and NppToR combo, which I found great to work with.