this *may* help, my setup is as follows:
I have 2 boxes, one a Windows Server 2008 instance. I installed R 2.10.1
and 2.11 on them, and used the Remote App function to create a remote
app "icon" that users (x 4 currently) put on their desktop. Yes, that
setup needs Terminal Services to work, but so far I haven't seen any
major complaints.
The second box is just running ubuntu, again for multiple users (x 3
currently) along with 2 Rserve daemon instances running on diff ports. I
haven't noticed anything wrong with running that as well.
The only thing I see occasionally is if all users are running a certain
section of some homegrown scripts that creates a SNOW cluster on the 8
core win08 server, there's 4 x 8 snow instances... takes a bit more time
to execute rather than just 1 or 2 users. I haven't quantified yet how
much more time exactly, but no complaints.
hth
c
On 5/24/2010 7:13 PM, dherrlin wrote:>
> I am planning to purchase a server to dedicate to running R processes. 4-5
> analysts will use the server and it must have capabilities to process large
> data sets fast. I am not sure if Windows or Linux based servers are better
> or what specs I should be focused on for optimum production. Also is Linux
> required in order to utilize multiple processors concurrently for a single
> process?
>
> Dan Herrlin