[Beowulf] [EXTERNAL] Re: First cluster in 20 years - questions about today

Lux, Jim (US 337K) james.p.lux at jpl.nasa.gov
Sun Feb 2 13:08:30 PST 2020


How old is the old cluster?  You might actually spend more time trying to get the nodes all working than you’d save by having them as a compute element.  I’ve looked at piles of computers in my garage and thought “hey, I should cluster them” and then, I realize that the discount laptop I can buy for a few hundred bucks will blow the combination away.

So, unless you’re “learning how to build a cluster”, I wouldn’t think that’s the way to go.  And for the “how to bring up a cluster tinkering”, a batch of rPi or beagles and a cheap switch is probably cheaper and more reflective of modern distros.

OTOH, if your three old nodes are a year old, then have at it.


From: Beowulf <beowulf-bounces at beowulf.org> on behalf of "jaquilina at eagleeyet.net" <jaquilina at eagleeyet.net>
Date: Saturday, February 1, 2020 at 10:45 PM
To: Mark Kosmowski <mark.kosmowski at gmail.com>, "beowulf at beowulf.org" <beowulf at beowulf.org>
Subject: [EXTERNAL] Re: [Beowulf] First cluster in 20 years - questions about today

Hi Mark,

So you are going to revive your old 3 node cluster and expand that? I would suggest if you are looking to expand the cluster I would look at the ryzen epyc rome 2 the second generation of these chips is quite impressive in the sense you can do double the amount of 2 single intel chips. They vary from 8 core 16 threads up to 64 core 128 threads. Also you don’t have all these issues that intel are currently facing with 7nm process as well as the vulnerabilities. I just moved my gaming pc from a 6th gen i7 to a ryzen 5 3600 6 core 12 thread machine and im seeing a huge difference in performance.


Regards,
Jonathan Aquilina

EagleEyeT
Phone +356 20330099
Sales – sales at eagleeyet.net<mailto:sales at eagleeyet.net>
Support – support at eagleeyet.net

___________________________________________________________________________________________

From: Beowulf <beowulf-bounces at beowulf.org<mailto:beowulf-bounces at beowulf.org>> On Behalf Of Mark Kosmowski
Sent: Sunday, 2 February 2020 04:21
To: beowulf at beowulf.org<mailto:beowulf at beowulf.org>
Subject: [Beowulf] First cluster in 20 years - questions about today

I've been out of computation for about 20 years since my master degree.  I'm getting into the game again as a private individual.  When I was active Opteron was just launched - I was an early adopter of amd64 because I needed the RAM (maybe more accurately I needed to thoroughly thrash my swap drives).  I never needed any cluster management software with my 3 node, dual socket, single core little baby Beowulf.  (My planned domain is computational chemistry and I'm hoping to get to a point where I can do ab initio catalyst surface reaction modeling of small molecules (not biomolecules).)

I'm planning to add a few nodes and it will end up being fairly heterogenous.  My initial plan is to add two or three multi-socket, multi-core nodes as well as a 48 port gigabit switch.  How should I assess whether to have one big heterogenous cluster vs. two smaller quasi-homogenous clusters?

Will it be worthwhile to learn a cluster management software?  If so, suggestions?

Should I consider Solaris or illumos?  I do plan on using ZFS, especially for the data node, but I want as much redundancy as I can get, since I'm going to be using used hardware.  Will the fancy Solaris cluster tools be useful?

Also, once I get running, while I'm getting current with theory and software may I inquire here about taking on a small, low priority academic project to make sure the cluster side is working good?

Thank you all for still being here!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://beowulf.org/pipermail/beowulf/attachments/20200202/15293ef3/attachment.html>


More information about the Beowulf mailing list