[MITgcm-devel] Re: today testing

David Ferreira dfer at ocean.mit.edu
Wed Oct 18 18:54:32 EDT 2006


Ed Hill wrote:

>On Fri, 13 Oct 2006 10:14:54 -0400 David Ferreira <dfer at ocean.mit.edu>
>wrote:
>
>  
>
>>Hi Ed,
>>I have a testreport related question. I'm planning to
>>occasionally run a testreport on Columbia (this is
>>probably going to be a manually triggered one every
>>couple of weeks).
>>
>>When it's done, I was planning to send the report (under
>>some appropriate form, typically tr_columbia_$date_0.tar.gz
>>similar to what's done in your Aces scripts) to edhill at mitgcm.org
>>(or whatever it is going to be).
>>Will it be enough for the report to appear on the web site and/or to
>>be archived properly ?
>>    
>>
>
>Hi David,
>
>I'm cc-ing mitgcm-devel since others may want to see it.
>
>You need to do just four things:
>
> 1) put all the stuff you want to send into a tar file (see how 
>    testreport does it)
>
> 2) use the "mpack" command as contained in:
>
>      MITgcm/tools/mpack-1.6
>
>    to MIME-encode the tar-file within the email and send he email
>
> 3) send the email to either edhill at mitgcm.org or a soon-to-be-created
>    generic testing account that will do exactly the same thing
>
> 4) make sure the subject line is "MITgcm-test" since that's 
>    what the processing scripts look for -- again see testreport for
>    an example
>  
>
Hi Ed,

Thanks for the answer. The testreport actually ended up one the
web-site! But I have one more question.
I'd like to run 2 testreports, both with the same optfile, but with and 
without
the ieee option. I tried and sent the testreports as 
tr_machine-ieee.tar.gz and
tr_machine-noiee.tar.gz, hoping they would appear simultaneously on list,
but it failed ... (only the most recent one appears)

What' the trick to get them both ?? (similarly to aces-g77, aces-ifort, etc)
Thanks,
david




-- 

Ferreira David               Tel : 617 253 7967 
EAPS Room 54-1515,
Massachusetts Institute of Technology,
77 Massachusetts Avenue
Cambridge, MA, 02139




More information about the MITgcm-devel mailing list