[Swift-devel] VDS1 transfer executable

[Ian Foster]

Ben:

I feel strongly that we should be focusing our scarce development 
resources on problems that we have documented via user experience. That 
means we need that performance monitoring infrastructure in Swift ...

I do think that data movement and caching are likely to become important 
issues. But it would be good to know when/how exactly they do.

Mike mentioned that he thought Nika's MolDyn code had some workaround in 
it to reduce data movement, introduced because of a lack of caching 
support. Does anyone know about that?

Ian.

Ben Clifford wrote:
> Not numerical data. I just recall it being something that we ISI sysadmin 
> people used to laugh about as people had VDS moving data all over the 
> place unnecessarily within the ISI network whilst they complained that 
> there wasn't enough space in one particular space or that ftp servers 
> weren't coping.
>
> On Tue, 24 Jul 2007, Ian Foster wrote:
>
>   
>> Do we have data that show this to be a problem?
>>
>> Ben Clifford wrote:
>>     
>>> On Tue, 24 Jul 2007, Mihael Hategan wrote:
>>>
>>>   
>>>       
>>>> Can you be more specific on what bottlenecks we're trying to avoid?
>>>>     
>>>>         
>>> pumping all the data for the workflow through one ethernet card and CPU.
>>>
>>>   
>>>       
>>     
>
>   

-- 

   Ian Foster, Director, Computation Institute
Argonne National Laboratory & University of Chicago
Argonne: MCS/221, 9700 S. Cass Ave, Argonne, IL 60439
Chicago: Rm 405, 5640 S. Ellis Ave, Chicago, IL 60637
Tel: +1 630 252 4619.  Web: www.ci.uchicago.edu.
      Globus Alliance: www.globus.org.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/swift-devel/attachments/20070724/095040e1/attachment.html>