[CIG-SHORT] Re: GeoFEST runs for strike-slip model (Jiangning L?)

[Charles D. Norton]

There is most likely an issue in the input file. I'll try to look at it  
today...

Charles

On Sep 5, 2006, at 1:10 PM, Jiangning Lü wrote:

> I tried to run the 500m case in GeoFEST sequential version on my pc,
> error also
> occurred:
> Number of nodes = 117649
> nsd = 3  ;  ndof = 3  ;  nrates = 0  ;  save_shape = 1  ;  solver = 2
> Start reading boundary info . . .
> Correct node tally in Processor = 0
> Global b.c. map generated...
> Decomposed node and equation totals:
> IAM: 0   local numnp = 117649  local neq = 320494
> Equation numbering complete...
> Coordinate generation complete...
> Displacement boundary value generation complete...
> Velocity boundary rates generation complete...
> Starting element generation...
> IAM: 0   local numel = 663552
> No surface tractions to process . . .
> Split nodes all processed . . .
>
> Generation of element group #1 complete.
> Doing shape functions...
> Element generation complete.  Setting up linear algebra...
> Equation setup completed.
> Time data  generation complete...
> CPU time = 16.570000
> Starting Elastic Soln At: Starting elastic solution...
> Calling SOLVER
> Bus error
>
> What does it mean by "bus error"?
> When I googled, I got this:
> If the program displays this message:
> 	Bus error
> or
> 	Segmentation fault
> then the program was trying to access a memory location outside its
> address space.  The computer detected this problem and sent a signal  
> to your
> program, which caused it to abort.
> Things that cause bus errors and segmentation violations are typically
> out-of-bounds array references and/or references through uninitialized  
> or
> mangled pointers.
>
> Any hints?
>
> Quoting Jiangning Lü <Jiangning at MIT.EDU>:
>
>> Hello,
>> The input files in 1km, 500m,250m resolutions can be found here:
>> http://www.geodynamics.org:8080/cig/workinggroups/eq/workarea/ 
>> benchmarks/benchmark-strikeslip/geofest-input
>>
>> Thanks a lot!
>> Jiangning
>>
>> Quoting Greg Lyzenga <lyzenga at hmc.edu>:
>>
>>> I have this response from Charles Norton:
>>>
>>>> 	From: 	  Charles.D.Norton at jpl.nasa.gov
>>>> 	Subject: 	Re: forwarded from the cig listserv
>>>> 	Date: 	September 4, 2006 5:22:25 PM PDT
>>>> 	To: 	  lyzenga at HMC.Edu, Jiangning at MIT.EDU
>>>> 	Cc: 	  jwp at cobra.jpl.nasa.gov
>>>>
>>>> Hello,
>>>>
>>>> If I can get a link to the GeoFEST input file I'll check it out and
>>>>  run GFMeshParse myself. As long
>>>> as the mesh description in the GeoFEST input file is okay we
>>>> generally have not had any problems
>>>> with the gfmeshparse conversion to the Pyramid input file. A
>>>> segfault here generally indicates that
>>>> there may be a problem in the GeoFEST input file. I can convert the
>>>>  file to tecplot format to check it
>>>> out...
>>>>
>>>> Thanks, Charles
>>>>
>>>
>>>
>>> -------------------------------------
>>> Gregory A. Lyzenga  <lyzenga at hmc.edu>           ***     (909)  
>>> 621-8378
>>> Dept. of Physics, Harvey Mudd College           *** fax (909)  
>>> 621-8887
>>> Claremont, CA 91711-5990
>>> <http://www.physics.hmc.edu/faculty/Lyzenga/index.html>
>>>
>>>
>>>
>>
>>
>>
>> ---
>> Jiangning Lü
>> 54-521     x3-6389
>> Geophysics Program
>> Earth, Atmospheric and Planetary Sciencs
>> Massachusetts Institute of Technology
>> johnnylu at mit.edu
>>
>> _______________________________________________
>> CIG-SHORT mailing list
>> CIG-SHORT at geodynamics.org
>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>>
>