Re: Uses for a cluster

"Evan Turner" <[email protected]> Thu, 29 Jun 2006 09:04:12 -0500
Newsgroups gmane.linux.cluster.openmosix.general
Message-ID <[email protected]>
Lars, Chris,


Our largest production cluster is a ~1000 node dual-xeon blade system.
We never purchase failover hardware because we always want to get the
maximum system out there.  At any one time we might have 1 or 2 nodes
out for maintenance, but that does not impact our running. 

Even though we have a large system, we have a mandatory 24hr runtime
limit on our cluster systems because of stability.  If you run a 128
node job, the chances of a machine failing in between 24 hours becomes
128 * every other problem.  For the power5 and big iron machines we run
48-72 hour job limits (and mostly it's small so other users can get a
chance to get access to the machine).  We have a cray on the floor
that's about 7 years old and doesn't get rebooted but every 4 months or
so, and not because it needs it.. just because we feel like it may be a
good idea.

What you describe as buying a single machine to save on cooling & power
the industry buzz around here calls the 'price performance comparison'.
Basically it's a marketing spreadsheet that says FLOPS per dollar of the
entire solution of hardware/networking/storage/cooling/power (you
provide your own people usually.

It all depends on how much money you have, and what your trying to
accomplish, obviously.  We have a few sun quad socket, DC machines and
they start at 24k a piece!  The IBM, SGI, and SUN supermachine options
are nice, but you can easily bust a million dollars on one rack.  A
million dollars would buy a lot of commodity cluster nodes, and from
what everyone else is doing around here unless you have government money
thrown at you (DoD) you buy commodity cluster.

If you're going to stick to 'little' man you can get away with probably
about 12-16 nodes before you have to really worry about cooling.  Power
you'd worry about once you get past 5 machines, depending on if they
still take 110 voltage (which most servers don't anymore) you may need
to drop some power in the room.  GIGe is very attractive economically
until you get past around 16 processors, then you need to look into
myrinet or infiniband.

My advice is never just order the machines you're looking at straight
off the website, talk to the sales people first.  Since you're going to
order in bulk they will always get you a better deal.  All of the big
"three letter" companies have HPC sales divisions now.

-evan T.

-----Original Message-----
From: openmosix-general-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
[mailto:openmosix-general-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org] On Behalf Of
Chris Willard
Sent: Thursday, June 29, 2006 5:49 AM
To: openmosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [openMosix-general] Uses for a cluster

Hi Lars!

On Thu, 29 Jun 2006, Lars O. Grobe wrote:

> If you have 7 PCs running, I wonder how much energy you spend for 
> filtering spam and if it is worth it. Maybe instead of paying the 
> electricity bill, you could buy a new fast machine? Or a new 
> playstation, has exactly seven cores ;-)

4 Weeks!!! - That's scary.

I agree that it would probably be more cost effective to have 1 fast PC
but I wanted to learn about clusters so it's not a problem at the
moment. I was just looking for more uses for my cluster.

The Playstation sound interesting - don't let my kids know :)

Regards,

Chris

-- 
/*       _\|/_
         (o o)
 +----oOO-{_}-OOo---------------------------------+
 |Chris Willard <chris-1tO3AA1OMbbaIFlBGLnUhFpr/1R2p/[email protected]>         |
 |                                                |
 |Objects on screen are stupider than they appear.|
 |                                                |
 +-----------------------------------------------*/

Using Tomcat but need to do more? Need to support web services,
security?
Get stuff done quickly with pre-integrated technology to make your job
easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache
Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
openMosix-general mailing list
openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/openmosix-general

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642