Beowulf digest, Vol 1 #540 - 16 msgs

tlioner tlioner at 263.net
Fri Aug 24 05:02:45 PDT 2001


Send Beowulf mailing list submissions to
	beowulf at beowulf.org

To subscribe or unsubscribe via the 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:Need to do something Useful (Eric T. Miller)
  2. Channel Bonding (Michael Sowka)
  3. Re:Channel Bonding (Rocky McGaugh)
  4. Re:Need to do something Useful (Pedro =?iso-8859-1?q?D=EDaz=20Jim=E9nez?=)
  5. RE:Need to do something Useful (Kim Branson)
  6. MPICH on RedHat 7.1 (William R. Pearson)
  7. Re:Anyone use Alpha 500a to form a beowulf cluster ? (Ron Chen)
  8. Re:Anyone use Alpha 500a to form a beowulf cluster ? (Patrick Geoffray)
  9. Socket Migration (Amber Palekar)
  10. Network RAM for Beowulf (Amber Palekar)
  11. Re:Network RAM for Beowulf (Felix Rauch)
  12. Re:Network RAM for Beowulf (Horatio B. Bogbindero)
  13. RE:Need to do something Useful (Luc Vereecken)
  14. Re:Network RAM for Beowulf (Robert G. Brown)
  15. Re:Network RAM for Beowulf (Amber Palekar)
  16. Re:Channel Bonding (Jared Hodge)

--__--__--

Message: 1
From: "Eric T. Miller" <emiller at techskills.com>
To: "Lambe, Dave" <dave.lambe at targacept.com>,
"Beowulf \(E-mail\)" <beowulf at beowulf.org>
Subject: RE: Need to do something Useful
Date: Wed, 15 Aug 2001 13:13:08 -0400
charset="iso-8859-1"


>About 2 months ago, I setup a small cluster using Scyld. The
>developers/programmers have been keyholed into another project for the time
>being. I would like to do something with the cluster as it's just using
>electricity & creating heat.
>Is there a way to fire up dnetc (or similar) on all the nodes? I apologize
>for my lack of *nix knowledge (I'm a hardware/setup guy). TIA
>
>
>Dave


Yes! I had a similar question about a week ago that didn't get much
response.  I too am new to clusters, and I just want to do something useful
and interesting with my new creation. It is currently just a very
intelligent space heater.


--__--__--

Message: 2
Date: Wed, 15 Aug 2001 14:00:20 -0400
From: Michael Sowka <msowka at doe.carleton.ca>
To: beowulf at beowulf.org
Subject: Channel Bonding

Hello,
I was hoping someone could point me in the right direction with some 
question I have about channel bonding. Linux kernel side seems to be 
easy enough with the CONFIG_BONDING=Y and other "various" /etc stuff... 
but what about the switches?! In the kernel Configuration Documentation 
it mentions something like a Cisco 5500 switch... I looked it up and it 
seems that the 5000 series switches are the "standard" for what Cisco 
calls Etherchannel (Channel bonding = Etherchannel). Now I sorta' assume 
that's an expensive piece of equipment, and the goal of our little 
project here is to build a very VERY "Class I" system (read: cheap, VERY 
off the shelf equipment)... won't a simple 300$ 8-port switch do??? 
Thanks for any comment and pointers,
Mike

-- 
/************************************************************************\
|      Mike Sowka                       o      _     _         _         |
| An Aspiring Engi"Nerd"        _o     /\_   _ \\o  (_)\__/o  (_)        |
|  Carleton University        _< \_   _>(_) (_)/<_    \_| \   _|/' \/    |
| msowka at doe.carleton.ca     (_)>(_) (_)        (_)   (_)    (_)'  _\o_  |
| (home msowka at home.com)                                                 |
\************************************************************************/



--__--__--

Message: 3
Date: Wed, 15 Aug 2001 13:37:16 -0500 (CDT)
From: Rocky McGaugh <rmcgaugh at atipa.com>
To: <beowulf at beowulf.org>
Subject: Re: Channel Bonding

On Wed, 15 Aug 2001, Michael Sowka wrote:

> Hello,
> I was hoping someone could point me in the right direction with some
> question I have about channel bonding. Linux kernel side seems to be
> easy enough with the CONFIG_BONDING=Y and other "various" /etc stuff...
> but what about the switches?! In the kernel Configuration Documentation
> it mentions something like a Cisco 5500 switch... I looked it up and it
> seems that the 5000 series switches are the "standard" for what Cisco
> calls Etherchannel (Channel bonding = Etherchannel). Now I sorta' assume
> that's an expensive piece of equipment, and the goal of our little
> project here is to build a very VERY "Class I" system (read: cheap, VERY
> off the shelf equipment)... won't a simple 300$ 8-port switch do???
> Thanks for any comment and pointers,
> Mike
>
>

Actually, 2 of the 8-port switches will work for 8 nodes, or 2 four
ports for 4 machines. Main consideration is that both nics will present
the same MAC address to the switch, and you cant have 2 nics with the same
MAC address on the same wire without special support. If you put all of
eth0 on one switch, and all of eth1's on the other, then this will work.

-- 
Rocky McGaugh
Atipa Technologies
rocky at atipatechnologies.com
rmcgaugh at atipa.com
1-800-360-4346 x3110
http://1087800222/
perl -e 'print unpack(u, ".=W=W+F%T:7!A+F-O;0H`");'


--__--__--

Message: 4
charset="iso-8859-1"
From: Pedro =?iso-8859-1?q?D=EDaz=20Jim=E9nez?= <pdiaz88 at terra.es>
To: "Eric T. Miller" <emiller at techskills.com>,
"Lambe, Dave" <dave.lambe at targacept.com>,
"Beowulf \(E-mail\)" <beowulf at beowulf.org>
Subject: Re: Need to do something Useful
Date: Wed, 15 Aug 2001 21:10:26 +0000

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Some ideas:

Interested in chess?
	-> distributed chess engine
Interested in crypto?  
	-> Crack RSA keys. 
	-> Search collisions in MACs 
	-> Help distributed.net 
	-> Do your own effort to brute-force DES (If you have time :-)
	-> Create your very own parallel version of GnuPG
	-> Implement a parallel pseudorandom number generator
Interested in multimedia?
	-> Distributed audio/video encoder
Interested in simulation?
	-> Implement a parallel nbody simulator
	-> Have a look at your physics books, sure you'll find something you would 		
	want to do in parallel
	-> Write a parallel implementation of Conway game of life
Interested in maths?
	-> Write a parallel fractal renderer
	-> Find primes
	-> help GIMPS
	-> Have a look at your math books, sure you'll find something you would want 
	   to do in parallel


Are you at a university?. Take a look at what other departments are doing. 
Get involved if possible. Be curious. Buy a good book and learn about 
parallel algorithms.

Just some thoughts 

Pedro

P.S.: I you achieve something interesting, drop me an email! :-)

On Wednesday 15 August 2001 17:13, Eric T. Miller wrote:
> >About 2 months ago, I setup a small cluster using Scyld. The
> >developers/programmers have been keyholed into another project for the
> > time being. I would like to do something with the cluster as it's just
> > using electricity & creating heat.
> >Is there a way to fire up dnetc (or similar) on all the nodes? I apologize
> >for my lack of *nix knowledge (I'm a hardware/setup guy). TIA
> >
> >
> >Dave
>
> Yes! I had a similar question about a week ago that didn't get much
> response.  I too am new to clusters, and I just want to do something useful
> and interesting with my new creation. It is currently just a very
> intelligent space heater.
>
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf

- -- 

/*
 * Pedro Diaz Jimenez: pdiaz88 at terra.es, pdiaz at acm.asoc.fi.upm.es 
 *                                                  
 * GPG KeyID: E118C651                              
 * Fingerprint: 1FD9 163B 649C DDDC 422D  5E82 9EEE 777D E118 C65
 *                                                  
 * http://planetcluster.org                         
 * Clustering & H.P.C. news and documentation       
 *
 * "Impossible is the adjective of the idiots"
 *      Napoleon Bonaparte
 */

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7euVInu53feEYxlERAm+aAJoCFdCzBzShU5VIE+3GsGPCSsomwACfT3YV
raWZBra2FoZrS/oHDymXW3c=
=Pcny
-----END PGP SIGNATURE-----

--__--__--

Message: 5
Date: Thu, 16 Aug 2001 09:18:14 +1000 (EST)
From: Kim Branson <bra369 at pp.molsci.csiro.au>
To: "Eric T. Miller" <emiller at techskills.com>
cc: "Lambe, Dave" <dave.lambe at targacept.com>,
"Beowulf (E-mail)" <beowulf at beowulf.org>
Subject: RE: Need to do something Useful



Of you like you could install globus on your machine and we can add it to
our global grid for drug design applications. 

see http://www.csse.monash.edu.au/~rajkumar/vlab/index.html
for more details. 

We are interested in designing small molecule atagonists for malaria and
leishmania. 

failing that theres always seti or something....

kim
______________________________________________________________________ 

Kim Branson
Structural Biology
Walter and Eliza Hall Institute
Royal Parade, Parkville, Melbourne, Victoria
Ph 61 03 9662 7136
Email kbranson at wehi.edu.au

______________________________________________________________________ 


On Wed, 15 Aug 2001, Eric T. Miller wrote:

> 
> >About 2 months ago, I setup a small cluster using Scyld. The
> >developers/programmers have been keyholed into another project for the time
> >being. I would like to do something with the cluster as it's just using
> >electricity & creating heat.
> >Is there a way to fire up dnetc (or similar) on all the nodes? I apologize
> >for my lack of *nix knowledge (I'm a hardware/setup guy). TIA
> >
> >
> >Dave
> 
> 
> Yes! I had a similar question about a week ago that didn't get much
> response.  I too am new to clusters, and I just want to do something useful
> and interesting with my new creation. It is currently just a very
> intelligent space heater.
> 
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
> 


--__--__--

Message: 6
Date: Wed, 15 Aug 2001 19:45:23 -0400 (EDT)
From: "William R. Pearson" <wrp at alpha0.bioch.virginia.edu>
To: beowulf at beowulf.org
Subject: MPICH on RedHat 7.1


Last week, I posted a question asking whether others had had problems
with MPICH 1.2.1 on RedHat 7.1.  Thanks to help from the MPICH folks
(mpi-bugs at anl.gov), I was able to identify a misconfiguration in my
/etc/hosts file (all localhosts had the name of the machine I cloned
from).  Fixing the misconfiguration fixed the problem, and my MPICH
programs run fine under RH7.1

Bill Pearson

--__--__--

Message: 7
Date: Wed, 15 Aug 2001 18:14:48 -0700 (PDT)
From: Ron Chen <ron_chen_123 at yahoo.com>
Subject: Re: Anyone use Alpha 500a to form a beowulf cluster ?
To: MICHAELCHEN <Yiping.Cheng at etatung.com.tw>, beowulf at beowulf.org

What problems did you encounter?

There are many research labs using Alpha Linux for
their beowulf clusters. See: http://www.beowulf.org/
(e.g. University of Wisconsin-Milwaukee, Viscoelastic
Group at Universite' catholique de Louvain, Center for
Atomic-Scale Materials Physics)

The question is, what kind of jobs do you run?
If all you want is simple batch jobs, you don't even
need to modify the kernel, all you need is a batch
system, like PBS (Portable Batch System), or SGE (Sun
Grid Engine). The batch system will glue all the
machines together: when a machine is free, the daemons
of the batch system will dispatch a job to the
machine. So all the machine in your company can be
used at night for batch jobs while the users get their
machine back in the morning. (This is ideal for
simulations for EDA, DNA sequencing, etc)


SGE:
http://gridengine.sunsource.net/project/gridengine/download.html
http://www.sun.com/gridware

PBS:
http://www.openpbs.com
http://www.pbspro.com

Both PBS and SGE are free, opensource, and support
Alpha Linux.

If you want to run parallel jobs (like PVM or MPICH),
then you can also use the batch systems to schedule
the jobs. However, TCP/IP and Ethernet is not fast
enough for message passing, so you may need to install
SCORE and Myrinet.

Tell us want you need :-)

-Ron






--- MICHAELCHEN <Yiping.Cheng at etatung.com.tw> wrote:
> Hi ! 
> I am in trouble .
> my HW : Alpha 500a 
> OS : RedHat 7.0
> MPI : MPICH 1.2.0
> Compiler : gcc ,  f77
> 
> any body use this cnfiguration ?
> 
> can you share the experience with me ?
> 
> tks.
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or
> unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf


__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

--__--__--

Message: 8
Date: Thu, 16 Aug 2001 00:27:35 -0400
From: Patrick Geoffray <patrick at myri.com>
Reply-To: patrick at myri.com
Organization: Myricom Inc
To: Ron Chen <ron_chen_123 at yahoo.com>
CC: MICHAELCHEN <Yiping.Cheng at etatung.com.tw>, beowulf at beowulf.org
Subject: Re: Anyone use Alpha 500a to form a beowulf cluster ?

Ron Chen wrote:

> If all you want is simple batch jobs, you don't even
> need to modify the kernel, all you need is a batch
> system, like PBS (Portable Batch System), or SGE (Sun
> Grid Engine). 

Can you imagine a cluster without a batch queue system ? 
You seem to present it as the most important component.
I mean a batch queue may be nice and usefull, but I 
know a lot of machines without one. All of you 


> If you want to run parallel jobs (like PVM or MPICH),
> then you can also use the batch systems to schedule
> the jobs. However, TCP/IP and Ethernet is not fast
> enough for message passing, so you may need to install
> SCORE and Myrinet.

TCP/IP and Ethernet are good enough for a lot of usage, 
including message passing. It's certainely the cheapest 
and easiest way to implement message passing.
BTW, you can use Myrinet without Score, or even Score 
without Myrinet.

Patrick

-----------------------------------------------------------
|   Patrick Geoffray, Ph.D.      patrick at myri.com         |
|   Myricom, Inc.                http://www.myri.com      |
|   Cell: 865-389-8852           325 N. Santa Anita Ave.  |
|   Fax:  865-974-1950           Arcadia, CA 91006        |
-----------------------------------------------------------

--__--__--

Message: 9
Date: Thu, 16 Aug 2001 02:11:38 -0700 (PDT)
From: Amber Palekar <amber_palekar at yahoo.com>
Subject: Socket Migration
To: beowulf at beowulf.org

Hi,
  I , along with my three friends , am planning to
implement TCP/IP Socket Migration for a distributed
system ( MOSIX in particular and thats why a kernel
level implementation of the same ). We have been
looking for papers for the same .But it seems that not
much papers are out as yet . So any body with any
information or pointers for the same plz let me know
ASAP . that would really be of great help to us.
 Thanks in advance.
        Amber 

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

--__--__--

Message: 10
Date: Thu, 16 Aug 2001 02:34:31 -0700 (PDT)
From: Amber Palekar <amber_palekar at yahoo.com>
Subject: Network RAM for Beowulf
To: beowulf at beowulf.org

hi,
   we as a group of four students are *also* thinking
of implementing Network RAM for a beowulf cluster
(assuming 100Mbps Ethernet ) whereby each node in the
cluster will donate some part of their RAM to be used
by all other nodes.so we will basically be mapping
this shared RAM to the address space of the current
node.One of the uses that we're thinking of is for
Journaling(as in file systems ).We'll be maintaining
the journals on the Network RAM instead of writing
them to the local disks.As we are completely new to
this , it is very difficult for us to determine the
statistics like :- the overhead in writing to Network
RAM . Any info or pointers to these stats would be
highly appreciated .
   TIA
    Amber

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

--__--__--

Message: 11
Date: Thu, 16 Aug 2001 12:07:13 +0200 (CEST)
From: Felix Rauch <rauch at inf.ethz.ch>
To: Amber Palekar <amber_palekar at yahoo.com>
cc: beowulf at beowulf.org
Subject: Re: Network RAM for Beowulf

On Thu, 16 Aug 2001, Amber Palekar wrote:
>    we as a group of four students are *also* thinking
> of implementing Network RAM for a beowulf cluster
> (assuming 100Mbps Ethernet ) whereby each node in the
> cluster will donate some part of their RAM to be used
> by all other nodes.so we will basically be mapping
> this shared RAM to the address space of the current
> node.

If I understand you right, then you want to do Distributed Shared
Memory (DSM). A similar project is TreadMarks, so [1] might be
interesting for you.

Regards,
Felix

[1] http://www.cs.rice.edu/~willy/TreadMarks/overview.html
-- 
Felix Rauch                      | Email: rauch at inf.ethz.ch
Institute for Computer Systems   | Homepage: http://www.cs.inf.ethz.ch/~rauch/
ETH Zentrum / RZ H18             | Phone: ++41 1 632 7489
CH - 8092 Zuerich / Switzerland  | Fax:   ++41 1 632 1307


--__--__--

Message: 12
Date: Thu, 16 Aug 2001 18:45:39 +0800
From: "Horatio B. Bogbindero" <wyy at cersa.admu.edu.ph>
To: Amber Palekar <amber_palekar at yahoo.com>
Cc: beowulf at beowulf.org
Subject: Re: Network RAM for Beowulf
Reply-To: wyy at admu.edu.ph
protocol="application/pgp-signature"; boundary="Qxx1br4bt0+wmkIi"
Organization: Ateneo Cervini-Eliazo Networks


--Qxx1br4bt0+wmkIi
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Aug 16, 2001 at 02:34:31AM -0700, Amber Palekar wrote (wyy sez):

>    we as a group of four students are *also* thinking
> of implementing Network RAM for a beowulf cluster
> (assuming 100Mbps Ethernet ) whereby each node in the
> cluster will donate some part of their RAM to be used
> by all other nodes.so we will basically be mapping
> this shared RAM to the address space of the current
> node.One of the uses that we're thinking of is for
> Journaling(as in file systems ).We'll be maintaining
> the journals on the Network RAM instead of writing
> them to the local disks.As we are completely new to
> this , it is very difficult for us to determine the
> statistics like :- the overhead in writing to Network
> RAM . Any info or pointers to these stats would be
> highly appreciated .

sounds like a very good project. but, i have heard of projects
that are trying to do distributed shared memory. i think the mosix
people have something going on http://www.mosix.org/.

but, in anycase, i would also like know about this. i am interested
in using a system with distributed shared memory.=20

--------------------------------------
William Emmanuel S. Yu
Ateneo Cervini-Eliazo Networks (ACENT)
email  :  william.s.yu at ieee.org
web    :  http://cersa.admu.edu.ph
phone  :  63(2)4266001-5925/5904
GPG    :  http://sysads.ateneo.net/wyu/wyy.pgp
=20
One man tells a falsehood, a hundred repeat it as true.
=20

--Qxx1br4bt0+wmkIi
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.5 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7e6RSOgIOlr0CsAERAovHAKCmYBMzuLbfqpSXAoNjKVq3FzKrnACgzcK9
okdVb0b3LArbWtrBEZpPe1A=
=oZRm
-----END PGP SIGNATURE-----

--Qxx1br4bt0+wmkIi--

--__--__--

Message: 13
Date: Thu, 16 Aug 2001 14:09:33 +0200
To: Kim Branson <bra369 at pp.molsci.csiro.au>,
"Eric T. Miller" <emiller at techskills.com>
From: Luc Vereecken <Luc.Vereecken at chem.kuleuven.ac.be>
Subject: RE: Need to do something Useful
Cc: "Lambe, Dave" <dave.lambe at targacept.com>,
"Beowulf (E-mail)" <beowulf at beowulf.org>
csiro.au>


For distributed computing, check out the list of projects given on Kirk
Pearson's webpage : 
http://www.aspenleaf.com/distributed/
Everything from SETI over Dnet to golem at home and UD-cancer is listed there. 

Luc Vereecken






--__--__--

Message: 14
Date: Thu, 16 Aug 2001 08:38:05 -0400 (EDT)
From: "Robert G. Brown" <rgb at phy.duke.edu>
To: Amber Palekar <amber_palekar at yahoo.com>
Cc: <beowulf at beowulf.org>
Subject: Re: Network RAM for Beowulf

On Thu, 16 Aug 2001, Amber Palekar wrote:

> hi,
>    we as a group of four students are *also* thinking
> of implementing Network RAM for a beowulf cluster
> (assuming 100Mbps Ethernet ) whereby each node in the
> cluster will donate some part of their RAM to be used
> by all other nodes.so we will basically be mapping
> this shared RAM to the address space of the current
> node.One of the uses that we're thinking of is for
> Journaling(as in file systems ).We'll be maintaining
> the journals on the Network RAM instead of writing
> them to the local disks.As we are completely new to
> this , it is very difficult for us to determine the
> statistics like :- the overhead in writing to Network
> RAM . Any info or pointers to these stats would be
> highly appreciated .

Check out the Trapeze project at Duke:

  http://www.cs.duke.edu/ari/trapeze/

This is a high end version of the project you are suggesting.  I suspect
that you can implement a simpler (but slower) version of this out of
component parts with existing kernels (or almost so).  If current
kernels support swap over NFS, for example, you can build a large
ramdisk on each node (and otherwise leave them just enough memory to
function comfortably without swapping) and export them all to a central
node for use as swap.  This would effectively extend the size of VM on
the central node, but swapping would occur at network-limited speeds
instead of disk-hardware limited speeds.  And of course you can
definitely export a set of ramdisks and NFS mount them for regular
"network bound" file I/O, and might even be able to glue them together
into a big striped filesystem (I don't think md will to this, but you
MIGHT be able to hack it so that it would).

Disk bandwidth, of course, has gotten much better over the years that
this might not make sense from a raw BW point of view.  For something
doing a lot of random accesses to disk, though, where the performance is
dominated by latency, might well benefit as the combination of memory
latency on the nodes plust network latency plus NFS latency will still
likely be an order of magnitude less than the seek time on a disk (which
requires the physical movement of big chunks of mass).  Order of
milliseconds for a disk seek, order of 100 microseconds for the
network+memory hit, a rough factor of ten improvement.  Note that even
if you hack the kernel and eliminate all the kludginess from this
approach (NFS swap?), you're still going to be limited by raw socket
latency and will therefore probably not improve on this estimate by as
much as a factor of 2.

Of course this works another order of magnitude better with Myrinet
(used in the Duke project) with latency less than 10 microseconds and
even the BW compares decently with local disk.  This is the primary
motivation for this project -- creating very large network-distributed
ramdisk-like storage with a very low level (and hence efficient)
implementation.

The PRIMARY place where this is useful is in projects that require far
more memory and/or faster (lower latency) disk than one can physically
add to a system, as it will almost always be cheaper and better to add
memory to a local system than to build a network of virtual memory IF
you can get to the desired memory regime by adding sticks to your own
box.  This is especially true now with memory prices in free fall (512MB
PC2100 DDR only $168 as of this morning on pricewatch.com -- it was over
$600 at the beginning of the summer -- and 512MB PC133 only >>$30<<
ditto).  Ditto with disk BW -- high end disk storage units now can
provide quite a lot of BW (with terrible latency of course).

Building a system with (say) 10 GB of network/virtual memory is a lot
more challenging.  To start with, this is more than the unhacked kernel
can address, I believe.  So there you'd need to build BOTH the
socket-based memory subsystem AND hack the kernel so that it could
somehow address it.  You should definitely look over the Trapeze site to
see how they attempt to finesse this problem via a higher level API (if
I understand their papers).  That is, they don't tamper with the
existing VM so much as to graft a set of hooks into a special library so
that the remote "memory" can be accessed via a file interface.  Or
something like that.  You'll need to read about it yourself, and might
even want to talk to the primary researchers as they'd likely have some
very sensible suggestions and direction to give you.

   rgb

>    TIA
>     Amber
>
> __________________________________________________
> Do You Yahoo!?
> Make international calls for as low as $.04/minute with Yahoo! Messenger
> http://phonecard.yahoo.com/
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>

-- 
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: 15
Date: Thu, 16 Aug 2001 06:48:36 -0700 (PDT)
From: Amber Palekar <amber_palekar at yahoo.com>
Subject: Re: Network RAM for Beowulf
To: beowulf at beowulf.org


--- Felix Rauch <rauch at inf.ethz.ch> wrote:
 
> If I understand you right, then you want to do
> Distributed Shared
> Memory (DSM). 
   
  yes i am talkin about DSM . we are also thinking
about implementing Network RAM for introducing
asynchronous writes ( i.e. converting a synchronous
write on a secondary storage device to an async write
on the Network RAM but thats too vague ) also we are
thinking of using Network RAM for swap space , for
process migration but that too are very raw ideas as
of now . still we will be working on it. The problem
we are facing is getting the stats related to Network
RAM . so plz HELP !!!! :)

   Amber

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

--__--__--

Message: 16
Date: Thu, 16 Aug 2001 09:19:19 -0500
From: "Jared Hodge" <jared_hodge at iat.utexas.edu>
To: Rocky McGaugh <rmcgaugh at atipa.com>
CC: beowulf at beowulf.org
Subject: Re: Channel Bonding

I thought the MAC was determined by the NIC (which would mean it would
vary from NIC to NIC even if in the same machine).  I may be wrong.

Rocky McGaugh wrote:
> 
> On Wed, 15 Aug 2001, Michael Sowka wrote:
> 
> > Hello,
> > I was hoping someone could point me in the right direction with some
> > question I have about channel bonding. Linux kernel side seems to be
> > easy enough with the CONFIG_BONDING=Y and other "various" /etc stuff...
> > but what about the switches?! In the kernel Configuration Documentation
> > it mentions something like a Cisco 5500 switch... I looked it up and it
> > seems that the 5000 series switches are the "standard" for what Cisco
> > calls Etherchannel (Channel bonding = Etherchannel). Now I sorta' assume
> > that's an expensive piece of equipment, and the goal of our little
> > project here is to build a very VERY "Class I" system (read: cheap, VERY
> > off the shelf equipment)... won't a simple 300$ 8-port switch do???
> > Thanks for any comment and pointers,
> > Mike
> >
> >
> 
> Actually, 2 of the 8-port switches will work for 8 nodes, or 2 four
> ports for 4 machines. Main consideration is that both nics will present
> the same MAC address to the switch, and you cant have 2 nics with the same
> MAC address on the same wire without special support. If you put all of
> eth0 on one switch, and all of eth1's on the other, then this will work.
> 
> --
> Rocky McGaugh
> Atipa Technologies
> rocky at atipatechnologies.com
> rmcgaugh at atipa.com
> 1-800-360-4346 x3110
> http://1087800222/
> perl -e 'print unpack(u, ".=W=W+F%T:7!A+F-O;0H`");'
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf

-- 
Jared Hodge
Institute for Advanced Technology
The University of Texas at Austin
3925 W. Braker Lane, Suite 400
Austin, Texas 78759

Phone: 512-232-4460
Fax: 512-471-9096
Email: Jared_Hodge at iat.utexas.edu



--__--__--

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


End of Beowulf Digest





More information about the Beowulf mailing list