[Ncep.list.nems.announce] NEMS trunk commit

[Dusan Jovic]

Committed revision 45983.

On 09/22/2014 11:16 AM, Dusan Jovic wrote:
> Hi, Everyone,
>
>  I reintegrated my branch containing the restructured regression test 
> scripts with the trunk and I'm ready to commit it to the NEMS 
> repository. Full regression test passed on zeus (no changes to 
> baseline results). On wcoss it already passed full test on Saturday 
> but I made one small change (while running a full test on Zeus) which 
> I do not expect to change the outcome of the repeated test on wcoss 
> which I'm going to do as soon as tide comes back online. If that wcoss 
> test pass I would like to make a commit tomorrow afternoon or 
> Wednesday morning.
>
> My working copy on Zeus is 
> /scratch2/portfolios/NCEPDEV/meso/save/Dusan.Jovic/trunk_rt.
>
> Once I make the commit you will find new top-level directory named 
> tests and the new script named rt.sh. The new procedure for running 
> the regression test is as follows:
>
> 1) in order to run standard suite of regression tests you will need to 
> execute rt.sh with '-s' option:
>
> *./rt.sh -s*
>
> 2) to run full regression test run the option is '-f':
>
> *./rt.sh -f
> *
> 3) once you make a code change which changes the baseline results you 
> will need to, as before, first generate new set of outputs and then 
> rerun standard or full test using the newly created set of baseline 
> results. that is done by running
>
> *./rt.sh -c nmm
> *
> or
> *./rt.sh -c gfs
> *
> first which will create new results for 'nmm' or 'gfs'. Then you rerun 
> the tests using new result by running rt.sh with option '-m' in 
> addition to '-s' or '-f':
>
> *./rt.sh -m -f*
>
> The option '-m' tells the script to use 'modified' results (created in 
> the previous step) for comparisons.
>
> I usually execute these two steps this way:
>
> *nohup sh -c "( ./rt.sh -c nmm && ./rt.sh -m -f )" &*
>
> which submits both steps in the background. If you run this in the 
> evening by the next morning you will have new set of baseline results 
> and,  hopefully, full suite of regression tests passed successfully.
>
> The old job directory will be renamed to job_unused to indicate that 
> it is not used anymore and after some time we will delete it. I would 
> keep it at least until somebody adds all nuopc tests and few more 
> missing gfs tests. Two nmm tests are also currently commented out 
> (those that use gfs physics). I tried to run nuopc test on zeus but it 
> failed, probably because the results were not updated recently.
>
> Dusan
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lstsrv.ncep.noaa.gov/pipermail/ncep.list.nems.announce/attachments/20140924/a9966376/attachment.html