From owner-chemistry@ccl.net Fri Nov 12 01:21:00 2010
From: "Simon Halstead joyjoyhappyjoy**yahoo.co.uk" <owner-chemistry=-=server.ccl.net>
To: CCL
Subject: CCL: Hardware suggestions for CPMD calculations
Message-Id: <-43151-101112001257-14354-2wsPTNjnNjOuiRYHUeIXmg=-=server.ccl.net>
X-Original-From: Simon Halstead <joyjoyhappyjoy(0)yahoo.co.uk>
Content-Type: text/plain; charset=us-ascii
Date: Fri, 12 Nov 2010 05:12:51 +0000 (GMT)
MIME-Version: 1.0


Sent to CCL by: Simon Halstead [joyjoyhappyjoy!A!yahoo.co.uk]
Dear CCLers,
 
I am starting a new research project which will require CPMD calculations. 
However, my current cluster is not especially powerful and I feel it will not be 
up to the task. So, I am looking to buy and build a new machine. I have very 
little experience with running CPMD, so I am not sure about what kind of 
hardware requirements it has. 

 
I am planning to run CPMD calculations with something on the order of 200 to 400 
relatively light atoms, primarily C, H and O. I do not need to run large 
quantities of simulations, and would be perfectly happy with only being able to 
run one simulation at a time as long as it is within a reasonable time scale. 
With this in mind, does anybody have any suggestions about what hardware I 
should be looking to use, especially which CPU? Also, would should be my primary 
concern with the cluster? I am currently assuming that the most important thing 
will be large quantities of RAM for each CPU...
 
Any help or advice is appreciated.
 
Regards,
 
Simon Halstead