Hi,

  It sounds like you've figured a lot of this out already.

  Stata/MP is limited to two CPUs by our license - most of our machines are dual-CPU machines, so we only licensed Stata/MP for 2 CPUs.

  It is OK to run Stata on the main node, but you will be sharing CPUs and memory with other users.  Your options depends on the details of what you want to do.  If you have all of your Stata commands in a text file (what Stata calls a do file) you can run this as a batch job that will execute all of the commands, and generate any plots, etc. that you have specified.  If you want to run Stata interactively, you should connect to the interactive node on the cluster.

  The interactive node has 8 CPUs and 16 GB of memory which you may be sharing with other users, but typically only one or two.  To connect to the interactive node, login to the cluster and then run the command "ssh c-0-0" - your prompt should change from "[fisherj@mulnx37 ~]$" to "[fisherj@compute-0-0 ~]$" to indicate you are working on the interactive node.  All of your files, etc. are available on the interactive node, and you use the same commands to run Stata.

  If you have a do file and are interested in running batch jobs to do your analysis, let me know and I can provide details.

  For your question about monitoring the progress of Stata, the only thing I'm aware of is the log files that Stata writes.

Dave

-----Original Message-----
From: Research Computing Support [mailto:[log in to unmask]] On Behalf Of Fisher, James Mr.
Sent: Saturday, April 05, 2008 3:19 PM
To: [log in to unmask]
Subject: Stata usage...

Hey,

I'm trying to figure out how I can run the MP version of Stata on the cluster.  To give you a bit of background I'm trying to look at the roll of skills (intelligence, dexterity, etc) in earnings inequality; to do so, I need to run quantile regressions, problem being that I have around 300k observations and on a desktop that would take forever to fit and bootstrap.

At this point, I've figured out how to login to the main node (I think) and load up the mp command line version Stata; but it's limited to two cores and I'm not sure if I should be running so intensive on the main node.  Naturally leading me to ask:  how do I move what I need to do onto one of the distributed clients (I have no idea); can I get Stata to access more than two cores; and, is there a way to monitor the progress of Stata as I run it?

Many thanks, James