[CIG-MC] CitcomS: adding input parameter to pyre version

Magali Billen mibillen at ucdavis.edu
Tue Jun 22 19:43:52 PDT 2010


Hello Eh,

That's easier than I expected...
I also had to move the line in my input file from under the  
[CitcomS.solver.ic] to [CitcomS.solver.bc] (aha!)
and it works!

I think I'm finally starting to understand how the pyre pieces fit  
together with the C program and the input file...

Thank you,
Magali

On Jun 22, 2010, at 7:15 PM, Eh Tan wrote:

> Hi Magali,
>
> You also need to add your input parameter on the python side, in  
> CitcomS/Components/BC.py.
>
> Eh
>
>
> On 6/23/10 6:46 AM, Magali Billen wrote:
>>
>> Hello,
>>
>> I have a question about a parameter that can be read from the input  
>> file in the standard version of CitcomS,
>> but is not recognized in the pyre version. I'd like to know what  
>> changes to make so that it gets read in.
>>
>> This is actually a more general question than just this one  
>> variable, as I'd like to make some other modifications to CitcomS,  
>> which
>> require creating a few new parameters that need to get read in and  
>> I'd like to use nested solvers, so I need to learn how to do this  
>> for the pyre version.
>>
>> Here are the details:
>> I'd like to use the variable E->convection.blob_bc_persist which  
>> exists in the non pyre version of CitcomS:
>>
>> [java:CitcomS/CitcomS-3.1.1/lib] billen% grep blob_bc_persist *.c
>> Initial_temperature.c:      input_boolean("blob_bc_persist",&(E- 
>> >convection.blob_bc_persist),"off",m);
>> Initial_temperature.c:        if(E->convection.blob_bc_persist){
>>
>> but does not have a binding
>>
>> [java:CitcomS/CitcomS-3.1.1/module] billen% grep blob_bc_persist *.c
>>
>> The other blob related variables are listed in setProperties.c
>>
>> [java:CitcomS/CitcomS-3.1.1/module] billen% grep blob *.c
>> setProperties.c:    getFloatVectorProperty(properties,  
>> "blob_center", E->convection.blob_center, 3, fp);
>> setProperties.c:    if( E->convection.blob_center[0] == -999.0 && E- 
>> >convection.blob_center[1] == -999.0 && E- 
>> >convection.blob_center[2] == -999.0 ) {
>> setProperties.c:        E->convection.blob_center[0] = 0.5*(E- 
>> >control.theta_min+E->control.theta_max);
>> setProperties.c:        E->convection.blob_center[1] = 0.5*(E- 
>> >control.fi_min+E->control.fi_max);
>> setProperties.c:        E->convection.blob_center[2] = 0.5*(E- 
>> >sphere.ri+E->sphere.ro);
>> setProperties.c:    getFloatProperty(properties, "blob_radius", E- 
>> >convection.blob_radius, fp);
>> setProperties.c:    getFloatProperty(properties, "blob_dT", E- 
>> >convection.blob_dT, fp);
>>
>> So, I added the following line to setProperties.c
>>
>> getIntProperty(properties, "blob_bc_persist", E- 
>> >convection.blob_bc_persist, fp);
>>
>> re-ran make (it did recompile/link setProperties), and then reran  
>> my model, but I still get the following error:
>>
>> >> inplume1.cfg:39:
>>  -- pyre.inventory(error)
>>  -- regional.ic.blob_bc_persist <- '1'
>>  -- unrecognized property 'regional.ic.blob_bc_persist'
>> usage: citcoms [--<property>=<value>] [-- 
>> <facility>.<property>=<value>] [FILE.cfg] ...
>> component 'CitcomS'
>>     properties: controller, help, help-components, help- 
>> persistence, help-properties, job, launcher, nodes, sc
>> heduler, solver, steps, typos, weaver
>>     facilities: controller,job,launcher,scheduler,solver,weaver
>> For more information:
>>   --help-properties: prints details about user settable properties
>>   --help-components: prints details about user settable facilities  
>> and components
>> CitcomS: configuration error(s)
>>
>> Can you tell me what else I have to do to make this get read in by  
>> the Pyre version?
>>
>> Thanks,
>> Magali
>>
>>
>> _______________________________________________
>> CIG-MC mailing list
>> CIG-MC at geodynamics.org
>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-mc
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/cig-mc/attachments/20100622/233a4b44/attachment-0001.htm 


More information about the CIG-MC mailing list