Hi,

First of all, thanks for all the suggestions and comments! I'll sum up a few of the replies here.

It seems that most people would favor going for the route of having multiple, individual computers (iMacs, mac minis, etc) instead of having one workhorse. This mainly due to the fact that anything 3D would be needed to be run locally. This is indeed the way we have things setup currently, and it is working fine. However, as some of the computers are getting older (and our structures getting bigger), especially the number crunching takes a hit. My idea was then revolving around buying new laptops/deskops, or continue using the ones we have (which do work fine) and buy a singe workhorse (mac pro).
My initial idea was indeed using the Mac Pro as Jürgen described, thus mainly as a number crunching workhorse, with all building in Coot etc done locally on the existing laptops. This can be done just by using ssh and running the normal os x on the machine, and not using vnc (or similar). I believe this might work quite well in our lab, but I'll need to give some thought. 

Again, thanks for your suggestions.

Cheers,
Ronnie




On Dec 15, 2010, at 20:07, Bosch, Juergen wrote:

We have a slightly different approach to this.

Local machines are laptops if you want to do model building you can do it locally either 2d or connected to a Zalman in 3D.

I have one Mac mini hooked up to a Zalman as "permanent" 3D station. The dual HexaCore MacPro (16 GB RAM) is connected also to a Zalman but is mostly used remotely via ssh in X11. Since you most likely want to use the MacPro for number crunching the ssh connection is fast enough, also for bringing up the ccp4I interface it still works fine even from home. I have a second 8core MacPro in my office used the same way, although right now I only have a 2D display connected to it, but the cores are accessible for number crunching. We are not using OSX Server, multiple users can have simultaneous ssh sessions. I link to a central .bashrc script so that every user has access to all software as soon as they log in. This requires me when setting up a new user to change the local .bashrc file to instead read the /etc/.bashrc that's all.

So not sure if you really need OSX Server for your purposes.

Maybe if you could specify what types of things you want to run on those machines it would be easier to make suggestions.

I would not built in Coot over the network, that's just frustratingly slow and you don't get the benefit of graphics acceleration.

Here's an incomplete list of stuff on those MacPros:
USF stuff e.g Moleman etc.
Coot, Pymol
Docking software
CCP4 package
Phenix
XDS


Jürgen


-
Jürgen Bosch
Johns Hopkins Bloomberg School of Public Health
Department of Biochemistry & Molecular Biology
Johns Hopkins Malaria Research Institute
615 North Wolfe Street, W8708
Baltimore, MD 21205
Phone: +1-410-614-4742
Lab:      +1-410-614-4894
Fax:      +1-410-955-3655
http://web.mac.com/bosch_lab/

On Dec 15, 2010, at 12:51 PM, Francis E Reyes wrote:


Dear all,

We are currently considering buying a computer which can be used by  
multiple people, via our existing network, as a workstation for  
crystallography purposes. My thoughts are currently going towards a  
8-core Apple Pro (or 12-core) with a lot of RAM, with OS X Server,  
which in theory should be able to handle multiple (up to 4) users  
simultaneously running crystallography software. The idea would be  
to have the users access this computer using their own laptops  
(starting their own virtual sessions?) connected to the same network.

Does this sound like a viable strategy, or should it be setup in a  
different way? In that case how? Would it need advanced setup and  
maintenance, or would it be possible to jsut set up a number of user  
accounts in OS X Server, and let it run? I'm reasonably computer  
savvy, but haven't really done something like this before, so I  
would very much appreciate your advice or personal experiences  
regarding this matter.

I know that I could probably get a cheaper computer if I went for a  
pc with linux, but I have more experience with OS X, and would  
therefore want to stay with it.

Thank you in advance,
Ronnie Berntsson




----------
Ronnie Berntsson, PhD
PostDoctoral Fellow
Department of Biochemistry
Groningen Biomolecular Sciences and Biotechnology Institute
& Zernike Institute for Advanced Materials
University of Groningen
Nijenborgh 4, 9747 AG
Groningen, The Netherlands

---------------------------------------------
Francis E. Reyes M.Sc.
215 UCB
University of Colorado at Boulder

gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D

8AE2 F2F4 90F7 9640 28BC  686F 78FD 6669 67BA 8D5D