[FLASH-USERS] problem implementing bcTypes, in particular "user"

[Ken Dere]

On 03/02/2010 04:51 PM, Klaus Weide wrote:
> On Mon, 1 Mar 2010, Ken Dere wrote:
>   
>> I am using the usm MHD driver in Flash 3.2 for a 2D problem.
>>
>> I am trying to migrate an 8wave problems to usm and am having problems
>> just specifying the boundary conditions.
>>
>> In my 8wave problem, I used xl,xr = 'periodic' and yl, yr =
>> 'hydrostatic-f2+nvrefl'
>>
>> I wanted to use the yl, yr = 'user' or 'user-defined' but when I do, the
>> simulation crashes with little output (something like p0_27722: 
>> p4_error: interrupt SIGSEGV: 11)
>>
>> So, I tried 'hydrostatic-f2+nvrefl' but that seems to get implemented
>> in  Grid_bcApplyToRegion.  I see it coming into 
>> Grid_bcApplyToRegionSpecialized but I can't really trap it there and
>> implement it.  I had hoped it would get applied in gr_applyFlash2HSEBC
>> but that did not happen.
>>
>> I have checked the constants.h file and am sure of my spelling.
>>
>> As far as I can see, there is something strange going on.  Any advice
>> would be appreciated.
>>     
> Hello Ken,
>
> Do you still have this problem?  I would like to help, but haven't 
> completely understood your setup.  Are you using your own 
> Grid_bcAppyToRegionSpecialized.F90 ? If yes, it could be helpful for me to 
> look at it.  What units are you including, and what are other relevant 
> lines from your simulation's Config file?
>
> If the problem is that you want to use the FLASH implementation of 
> 'hydrostatic-f2+nvrefl' together with your own behavior at some other
> boundary, we should be able to find a way to do that.  But it seems
> your problem was originally not about this but just about using a "user"
> boundary condition.  That should be even easier.
>
> Perhaps you could compile your code with debuggning enabled and run
> on just 1 processor under gdb, so you will be able to see exactly
> which source line triggered the SIGSEGV?
>
> Klaus
>   
Hi Klaus,

If I  run my problem with the standard BCs such as outflow, reflecting,
etc it works -  well, it produces reasonable output but does not do the
problem I am interested in.

Then I set
xl_boundary_type="user"

I then ran

mpirun -gdb flash3

at the first stop, I type c to continue and then I get the output from
some print statements which indicate that
 Grid_bcApplyToRegionSpecialized was called and went through one series
of applying the bc's to the lower boundary.

then an error occurred in glibc

I have attached the output from this point.  The error is fatal and
trying to continue from there is useless.

Ken

-- 
Kenneth P. Dere
Research Professor of Solar Physics
Dept. of Computational and Data Sciences
MS 6A2
George Mason University
4400 University Drive
Fairfax VA 22030
703-993-4555
703-993-1992 FAX
kdere at gmu.edu

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: gdb1.txt
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20100308/cfe5aac5/attachment.txt>