[no subject]

I want to minimize traffic over the network.  In a lowcost cluster with only 
one switched network for both filesystem traffic and message passing, I'd like 
to store as much data as possible on the compute node in a ramdisk.  Only when 
you absolutely have to would you writeback to the master node's filesystem.

I may be spouting smoke since in molecular dynamics has to update the 
velocities, forces, and I think the coordinates on all the atoms in the 
simulation for each step.  So only the frames you specify to be saved are sent 
back to the master node.  This means that your message passing traffic is 
always heavy in a molecular dynamics simulation.

But, every little bit helps, since you could be calculating a Monte Carlo run 
on a few nodes and a molecular dynamics run on others.  Then, the Monte Carlo 
would not interfere with the MPI run except when writing output structures at 
the end of the job.
Dow


Danny Cox wrote:
> Dow,
> 
> On Fri, 2004-03-12 at 18:49, Dow Hurst wrote:
> 
>>Can this be the writeable file system of a diskless client?  Where data being 
>>worked on is stored in a HPCluster situation?
> 
> 
> 	Sorry, I don't understand.
> 
> 	Yes, it can be the writeable file system of a diskless client.
> 
> 	The HPCluster part is what I don't understand.  If the data is stored
> there, then why have it in a local FS?  Or do you copy it, modify it,
> and copy it back?  Is it NFS, FTP, or something else?
> 

-- 
__________________________________________________________
Dow Hurst                  Office: 770-499-3428            *
Systems Support Specialist    Fax: 770-423-6744            *
1000 Chastain Rd. Bldg. 12                                 *
Chemistry Department SC428  Email:   dhurst at kennesaw.edu   *
Kennesaw State University         Dow.Hurst at mindspring.com *
Kennesaw, GA 30144                                         *
************************************************************
This message (including any attachments) contains          *
confidential information intended for a specific individual*
and purpose, and is protected by law.  If you are not the  *
intended recipient, you should delete this message and are *
hereby notified that any disclosure, copying, distribution *
of this message, or the taking of any action based on it,  *
is strictly prohibited.                                    *
************************************************************


</pre>
<!--X-Body-of-Message-End-->
<!--X-MsgBody-End-->
<!--X-Follow-Ups-->
<hr>
<!--X-Follow-Ups-End-->
<!--X-References-->
<ul><li><strong>References</strong>:
<ul>
<li><strong><a name="00396" href="msg00396.html">[ale] ramdisk howto</a></strong>
<ul><li><em>From:</em> John.Armsby at motorola.com (Armsby John-G16665)</li></ul></li>
<li><strong><a name="00421" href="msg00421.html">[ale] ramdisk howto</a></strong>
<ul><li><em>From:</em> danscox at mindspring.com (Danny Cox)</li></ul></li>
<li><strong><a name="00427" href="msg00427.html">[ale] ramdisk howto</a></strong>
<ul><li><em>From:</em> dhurst at kennesaw.edu (Dow Hurst)</li></ul></li>
<li><strong><a name="00456" href="msg00456.html">[ale] ramdisk howto</a></strong>
<ul><li><em>From:</em> danscox at mindspring.com (Danny Cox)</li></ul></li>
</ul></li></ul>
<!--X-References-End-->
<!--X-BotPNI-->
<ul>
<li>Prev by Date:
<strong><a href="msg00525.html">[ale] Tomcat install</a></strong>
</li>
<li>Next by Date:
<strong><a href="msg00527.html">[ale] 1and1 Was:Webhosting + domainname</a></strong>
</li>
<li>Previous by thread:
<strong><a href="msg00456.html">[ale] ramdisk howto</a></strong>
</li>
<li>Next by thread:
<strong><a href="msg00397.html">[ale] OT: X10 weirdness</a></strong>
</li>
<li>Index(es):
<ul>
<li><a href="maillist.html#00526"><strong>Date</strong></a></li>
<li><a href="threads.html#00526"><strong>Thread</strong></a></li>
</ul>
</li>
</ul>

<!--X-BotPNI-End-->
<!--X-User-Footer-->
<!--X-User-Footer-End-->
</body>
</html>