[astro] conjugate gradient descent stuff

[Travis Desell]

I've figured out the problem with the CGD not converging as fast as it 
should, and it's a problem with the
line search function.  I have sum of squares working in salsa, and it's 
finishing in under 30 iterations every
time, while gradient descent takes 70-100.  Once we can figure out why 
the astronomy stuff in the MPI
version is hanging i can start to get some results on the bluegene.

--Travis



Nathan Cole wrote:
> yes I can still use the machines in the wcl, just not the cluster machines.
>
> ~Nate~
>
> ==============Original message text===============
> On Thu, 21 Jun 2007 13:41:30 EDT Heidi Newberg wrote:
>
>
> Nate,
>
>        Are you still able to work on the 16 machines?
>
> Heidi
>
> Nathan Cole wrote:
>   
>> ==============Original message text===============
>> From: NATS <labstaff at cs.rpi.edu>
>> Date: Thu, 21 Jun 2007 13:07:29 EDT
>> Subject: [TICKET #15849] Can't log into CVS
>>
>> This is an automated email notification of an updated ticket:
>>
>>
>> From: lindss2 at cs.rpi.edu, 2007-06-21 13:07:28.637707
>> The machines in the CII will be down for about a week due to cooling work 
>> being done.  We don't have a firm date for when they'll be back in service 
>> but they'll let us know when they're done.
>>
>> Steven Lindsey               Research Programmer / Systems Administrator
>> lindss2 at cs.rpi.edu           Department of Computer Science
>> Lally 301                    Rensselaer Polytechnic Institute
>>
>>
>>
>>
>>  - labstaff at cs.rpi.edu
>>
>>
>>   
>>     
>
>   


-- 
----------------------------------------------------------------------------------------------------------
Travis Desell
<deselt @ cs.rpi.edu>
1-518-878-7424
Worldwide Computing Laboratory ( http://wcl.cs.rpi.edu/ )
Rensselaer Polytechnic Institute, 110 8th Street, Troy NY 12180, USA
----------------------------------------------------------------------------------------------------------