Beowulf digest, Vol 1 #1249 - 10 msgs

Sanjeev verma sanjeev.verma at progression.com
Wed Mar 12 03:43:18 PST 2003


Hi,

I am on the way to build a Beowulf cluster by 30 machines using Dolphin card (Wulfkit). I am going to use Intel E7500 Chipset. I need to ask that whether Super Micro Motherboard based on E7500 chipset (Model Number. SMP4DME) is certified for Beowulf cluster or not. What is the risk factor involved into it. 


Regards
SKV

-----Original Message-----
From: beowulf-admin at beowulf.org [mailto:beowulf-admin at beowulf.org] On Behalf Of beowulf-request at beowulf.org
Sent: 11 मार्च 2003 22:31
To: beowulf at beowulf.org
Subject: Beowulf digest, Vol 1 #1249 - 10 msgs

Send Beowulf mailing list submissions to
	beowulf at beowulf.org

To subscribe or unsubscribe via the World Wide Web, visit
	http://www.beowulf.org/mailman/listinfo/beowulf
or, via email, send a message with subject or body 'help' to
	beowulf-request at beowulf.org

You can reach the person managing the list at
	beowulf-admin at beowulf.org

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Beowulf digest..."


Today's Topics:

   1. Re:[OT] Beowulf UK (Joseph Mack)
   2. RE:Building a beowulf with old computers (Washington, G. DR CHEM)
   3. Re:Building a beowulf with old computers (Rocky McGaugh)
   4. Re:[OT] Beowulf UK (Lyle Bickley)
   5. CCL:64-bit machines document - where is it? (fwd) (Eugen Leitl)
   6. RE:FW: [OT?] Using clusters of computers for redundancy (Cannon, Andrew)
   7. energy costs (Helms, Scott A)
   8. Re:kill a watt? (Robert G. Brown)
   9. Re:energy costs (Josip Loncaric)
  10. Re:energy costs (Dean Johnson)

--__--__--

Message: 1
Date: Mon, 10 Mar 2003 12:14:44 -0500
From: Joseph Mack <mack.joseph at epa.gov>
CC: beowulf at beowulf.org
Subject: Re: [OT] Beowulf UK

John Hearns wrote:

> just turned up this rather nice site  http://www.beowulf.co.uk

shouldn't this be beerwulf?

Joe
-- 
Joseph Mack PhD, Senior Systems Engineer, SAIC contractor 
to the National Environmental Supercomputer Center, 
ph# 919-541-0007, RTP, NC, USA. mailto:mack.joseph at epa.gov

--__--__--

Message: 2
From: "Washington, G. DR CHEM" <Gary.Washington at usma.edu>
To: beowulf at beowulf.org
Subject: RE: Building a beowulf with old computers
Date: Mon, 10 Mar 2003 10:42:01 -0500

> 
> Message: 4
> From: "D. Scott" <ds10025 at cam.ac.uk>
> To: Greg Lindahl <lindahl at keyresearch.com>
> Cc: beowulf <beowulf at beowulf.org>
> Subject: Building a beowulf with old computers
> Date: 09 Mar 2003 15:51:42 +0000
> 
> Hi Beowulf experts
> 
> I've got old computer that have 32MB RAM some with 1GB hard 
> disk and 512MB 
> hard disk and some with 16MB RAM. How best to build a 
> beowulf? I've looked 
> into OSCAR but that requires large about of disk & memory. 
> SCE & SMA also 
> require min 256MB RAM. Have anyone built a beowulf with low spec PCs?
> 
> 
> Thanks in advance
> 
> Dan


I have built several Beowulf clusters with low spec PC's. Given the
sometimes small hard disk and ram sizes I wanted a Linux distribution that
would not take up much space after it is installed including compilers (gcc,
g77), rsh, mpich and pvm. You don't have the resources for a smooth running
x-windows and it is not needed to run the cluster unless you want a copy on
the master node so save the hard disk space. I used Debian 3.0 (woody) and
it works great and with a fairly complete installation you are at around 250
to 300 MB of hard drive space and allowing for swap space you have room for
parallel applications. Several computers I had did not have CD-ROM's so I
used NFS to share the CD-ROM from the PC that had one during installation
and after installation to add additional packages. I would suggest you get a
crossover cable and after the nodes are built then a fast Ethernet switch
and I hope you have 10/100 base T Ethernet cards (good ones can be bought
for $5 - $10). I applied this approach to 4 Pentium II PC's (all desktops)
and 2 P1's (one desktop and one laptop). You can run the cluster from the
master node with one monitor and shut the nodes down through rsh. This is a
very streamlined cluster and all the bells and whistles are not there but it
is pretty fast. I know that if you have 25 or more PC's in the cluster the
extra monitoring stuff is essential but most people seem to have fewer than
10 nodes. I can provide more detail if you are interested. I have another
cluster that has 3 nodes running Debian (woody) and one node with Mandrake
8.1 and all is well.
If you decide to spend any money here after the Ethernet cards and switches
buy ram you will see the difference immediately.

Gary Washington


> 
> 
> -- __--__-- 
> 
> _______________________________________________
> Beowulf mailing list
> Beowulf at beowulf.org
> http://www.beowulf.org/mailman/listinfo/beowulf
> 
> 
> End of Beowulf Digest
> 

--__--__--

Message: 3
Date: Mon, 10 Mar 2003 10:16:08 -0600 (CST)
From: Rocky McGaugh <rocky at atipa.com>
To: "D. Scott" <ds10025 at cam.ac.uk>
cc: beowulf <beowulf at beowulf.org>
Subject: Re: Building a beowulf with old computers



With such limited resources, i'd probably look at Debian. It would 
probably have a bit larger learning curve. It is a "small install; grow
as you need" type system.

I'll assume this is just for "fun". To get started, all you really need is
SSH, mpich, and maybe Ganglia for monitoring.

-- 
Rocky McGaugh
Atipa Technologies
rocky at atipatechnologies.com
rmcgaugh at atipa.com
1-785-841-9513 x3110
http://1087800222/
perl -e 'print unpack(u, ".=W=W+F%T:7\!A+F-O;0H`");'



On 9 Mar 2003, D. Scott wrote:

> Hi Beowulf experts
> 
> I've got old computer that have 32MB RAM some with 1GB hard disk and 512MB 
> hard disk and some with 16MB RAM. How best to build a beowulf? I've looked 
> into OSCAR but that requires large about of disk & memory. SCE & SMA also 
> require min 256MB RAM. Have anyone built a beowulf with low spec PCs?
> 
> 
> Thanks in advance
> 
> Dan





--__--__--

Message: 4
From: Lyle Bickley <lbickley at bickleywest.com>
To: Joseph Mack <mack.joseph at epa.gov>
Subject: Re: [OT] Beowulf UK
Date: Mon, 10 Mar 2003 11:37:45 -0800
Cc: beowulf at beowulf.org

On Monday 10 March 2003 09:14, Joseph Mack wrote:
> John Hearns wrote:
>
> just turned up this rather nice site  http://www.beowulf.co.uk
>
> shouldn't this be beerwulf?

With a 300 node network, GNU class "independent" and "freehouse" software, 
they should be fine with using the generally accepted "Beowulf" terminology.

See quote below from their Site: 

"The Beowulf Brewing Company is Birmingham's only independent brewery. Founded 
in 1996 the micro brewery now supplies over three hundred freehouses 
nationwide."

Just my 2 cents worth. 

Lyle
-- 
Lyle Bickley
Bickley Consulting West Inc.
http://bickleywest.com
V 650-428-0621
"Black holes are where God is dividing by zero"

--__--__--

Message: 5
Date: Mon, 10 Mar 2003 22:38:35 +0100 (CET)
From: Eugen Leitl <eugen at leitl.org>
To: Beowulf at beowulf.org
Subject: CCL:64-bit machines document - where is it? (fwd)


User view on 64 bit machines.

---------- Forwarded message ----------
Date: Mon, 10 Mar 2003 20:03:36 +0100
From: Daniel Svozil <svozil at iveta.uochb.cas.cz>
To: chemistry at ccl.net
Subject: CCL:64-bit machines document - where is it?

Hi,

the PDF document didn't get into CCL list due to its size, it can be
downloaded from:

http://www.ccl.net/cca/documents/64-bit_in_CC_summary.pdf

Regards
  Dan

-- 
Daniel Svozil, PhD
Institute of Organic Chemistry and Biochemistry
and Center for Complex Molecular Systems and Biomolecules
http://www.molecular.cz/
Czech Republic

phone: +420-2-20 183 263

-= This is automatically added to each message by mailing script =-
CHEMISTRY at ccl.net -- To Everybody  | CHEMISTRY-REQUEST at ccl.net -- To Admins
Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan: jkl at osc.edu








--__--__--

Message: 6
From: "Cannon, Andrew" <Andrew.Cannon at nnc.co.uk>
To: "'beowulf at beowulf.org'" <beowulf at beowulf.org>
Subject: RE: FW: [OT?] Using clusters of computers for redundancy
Date: Tue, 11 Mar 2003 08:01:47 -0000

Thanks for all your messages. You've all certainly given me (and the IT
boss) some food for thought. I'll be looking into some of the suggestions in
more detail, so don't be surprised if a question suddenly pops into your
inbox...

Thanks again.

Andrew


***********************************************************************************
 NNC Limited
 Booths Hall
 Chelford Road
 Knutsford
 Cheshire
 WA16 8QZ
 
 Country of Registration: United Kingdom
 Registered Number: 1120437
 
 This e-mail and any files transmitted with it are confidential and 
 intended solely for the use of the individual or entity to whom they   
 are addressed. If you have received this e-mail in error please notify 
 the NNC system manager by e-mail at eadm at nnc.co.uk.
***********************************************************************************


--__--__--

Message: 7
From: "Helms, Scott A" <HelmsScottA at uams.edu>
To: "'beowulf at beowulf.org'" <beowulf at beowulf.org>
Subject: energy costs
Date: Mon, 10 Mar 2003 20:14:39 -0600

When I was getting hardware bids, I asked about kilowatts per gigaflops, and they pretty much seemed to think that I'd slipped a gear.  I have kept my number of nodes pathetically few intentionally because I don't want to have to feed a mammoth machine on my own nickel if I get kicked out of my lab for some reason.  I may go as high as ten nodes, but any advancement after that will be done by replacing motherboards/cpu's.  Just thought I'd throw in my two cents' worth.  That's about all I've got left after buying hardware. ;-)

--__--__--

Message: 8
Date: Tue, 11 Mar 2003 08:32:03 -0500 (EST)
From: "Robert G. Brown" <rgb at phy.duke.edu>
To: Mark Hahn <hahn at physics.mcmaster.ca>
Cc: beowulf at beowulf.org
Subject: Re: kill a watt?

On Sun, 9 Mar 2003, Mark Hahn wrote:

> the name is aweful, I know, but has anyone tried:
> 
> http://www.p3international.com/products/special/P4400/P4400-CE.html
> 
> it liiks like a nice (cheap) way to take the guesswork out of 
> how much power a cluster node is drawing.  admittedly, it would 
> be nicer if it had a serial/net/usb port, but...
> 
> given that it apparently measures V, f and PF, does that mean it'll
> be hip to the true-rms thing?

I don't know, but based on the information in this and the reply, I
ordered one from EIP yesterday.  For $45 shipped, I've spent more on
surge protectors and multimeters and if it works (even with a bit of
Kentucky Windage) it will fill a need I've had for years.

It could get here as early as tomorrow (they claim it shipped yesterday,
and shipping usually takes 2-3 days) so perhaps I'll be able to play and
report.

  rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu




--__--__--

Message: 9
Date: Tue, 11 Mar 2003 10:40:10 -0500
From: Josip Loncaric <josip at nianet.org>
Reply-To: josip at nianet.org
Organization: National Institute of Aerospace
To: "Helms, Scott A" <HelmsScottA at uams.edu>
CC: "'beowulf at beowulf.org'" <beowulf at beowulf.org>
Subject: Re: energy costs

Helms, Scott A wrote:
> [...] I asked about kilowatts per gigaflops [...]

Your concern about KW/Gflop is completely justified.  Both IBM and Cray 
salesmen will happily advertise their power efficiency.  Today's CPUs are 
primarily constrained by their thermal limits, and about half of the time 
spent in CPU design workshops is spent on CPU power management.  As a result, 
we get interesting "features" such as thermal throttling on CPUs, memory, etc.

In theory, one could use reversible physical processes for computing and thus 
recover all of the energy by reversing the computation at the end.  This kind 
of computer would require no net power, although some energy would be needed 
to reach the desired result.  However, this computer would not be allowed to 
forget any of the states it went through.  In some sense, power dissipation is 
the penalty we pay for forgetting -- and thanks to speculative evaluation, 
today's CPUs evaluate then forget even more than is commonly known.

I can't seem to find my KW/Gflop data right now, but I believe that in that 
regard a machine such as Cray X1 compares rather favorably with Beowulf 
clusters.  Both kinds of systems will improve power efficiency each year, but 
their relative merit will probably evolve quite slowly...

Sincerely,
Josip

-- 
Dr. Josip Loncaric, Research Fellow              mailto:josip at nianet.org
National Institute of Aerospace       http://research.nianet.org/~josip/
144 Research Drive                       mailto:j.loncaric at larc.nasa.gov
Hampton, VA 23666, USA         Tel. +1-757-766-1395  Fax +1-757-766-1401


--__--__--

Message: 10
Subject: Re: energy costs
From: Dean Johnson <dtj at uberh4x0r.org>
To: josip at nianet.org
Cc: "Helms, Scott A" <HelmsScottA at uams.edu>,
   "'beowulf at beowulf.org'" <beowulf at beowulf.org>
Date: 11 Mar 2003 10:30:38 -0600

On Tue, 2003-03-11 at 09:40, Josip Loncaric wrote:
> I can't seem to find my KW/Gflop data right now, but I believe that in that 
> regard a machine such as Cray X1 compares rather favorably with Beowulf 
> clusters.  Both kinds of systems will improve power efficiency each year, but 
> their relative merit will probably evolve quite slowly...
> 

Its rated at 13 GFLOPS/kw and Compute Density of 25 GFLOPS/ft^2

-- 
Dean Johnson <dtj at uberh4x0r.org>



--__--__--

_______________________________________________
Beowulf mailing list
Beowulf at beowulf.org
http://www.beowulf.org/mailman/listinfo/beowulf


End of Beowulf Digest



Privileged or confidential information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), please delete this message and kindly notify the sender by an emailed reply. Opinions, conclusions and other information in this message that do not relate to the official business of Progression and its associate entities shall be understood as neither given nor endorsed by them.

--------------------------------------------------------------
Progression Infonet Private Limited, Gurgaon (Haryana), India






More information about the Beowulf mailing list