Need advice on cluster hardware

Eray Ozkural (exa) erayo at
Sat Jan 12 07:28:26 PST 2002

Hash: SHA1

On Friday 11 January 2002 03:35, Ron Choy wrote:
> My advisor has made me responsible (*gasp*) for purchasing a 8 node
> cluster, used mainly for computational linear algebra problems.  After some
> research I came up with the following configuration:
> Frontend (+file server):
> Asus A7M266-D (AMD 760MPX)
> Enermax EG365P-VE 350W PS (meets the 15A on 12V requirement of the board)
> 2 x Athlon MP 1800+ (1.533GHz)
> 2 x Thermaltake VOLCANO 6Cu+ heatsink
> 2 x Kingston PC2100 512MB ECC Registered
> Adaptec SCSI Controller
> 2 x Netgear GA622T gigabit nic
> ATI xpert Rage XL 8MB AGP vid
> Floppy drive
> Sony 52X cdrom drive
> Tower case
> Compute nodes:
> same as front end, minus 1 nic and SCSI adapter, and replace SCSI HD with

You shouldn't be needing cdrom drives at nodes. I don't know, but with those 
2 power-hogging cpu's you might be needing an extra fan somewhere. 

> Switch:
> Intel 410T 16 port 10/100 switch
> (* The reason why I have gigabit nics and 10/100 switch is that I don't
> know if bandwidth is going to be a limit on the computations so I would
> rather start out small and expand later.  (is this a good idea?) )

Linear Algebra problems are likely to require a lot of bisection bandwidth. 
Is that switch going to work with your 1000Base-TX (?) NIC's at all? I assume 
you'd be better off with a switch that suits your hardware. If you have the 
budget go for a gigabit switch.

Will you work on dense or sparse problems? Your requirements are likely to 
differ for the type of matrices you will use, and of course the kind of 
research you will make. If you are a computational scientist you'd like a 
faster network, if you are a computer scientist you might need a slower 
network to show that your algorithm is effective in low-bandwidth 

> Does this configuration looks reasonable?  Any known conflicts and driver
> issues?  I am going to use the 2.4.x kernel.

It does look reasonable. Usually there aren't many problems with ASUS boards 
and AMD cpus but of course testing is the best way to make sure.


- -- 
Eray Ozkural (exa) <erayo at>
Comp. Sci. Dept., Bilkent University, Ankara
GPG public key fingerprint: 360C 852F 88B0 A745 F31B  EA0F 7C07 AE16 874D 539C
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see


More information about the Beowulf mailing list