[FLASH-USERS] VisIt with many AMR levels

[Brian O'Shea]

Hi Sean,

I can't speak to the FLASH issues directly, but we ran into this with  
Enzo simulations and VisIt a couple of years ago.  The primary problem  
was that VisIt only kept track of position values using 32-bit  
precision, and that when using many levels of AMR the code can no  
longer adequately track the position of grids.  I'm not sure if this  
is still true about VisIt, but it certainly sounds like it is.   
Perhaps this particular issue would be resolved more quickly if you  
sent it to the VisIt mailing list?

Regards,
Brian

> Hello everyone,
>
> I am using VisIt to visualize FLASH 2.5 data with 31 levels of  
> refinement in 2D.  It seems this many levels is breaking VisIt (I am  
> using VisIt 1.11.1).  The plots are reminiscent of the 'over- 
> lapping' block levels problem of the past; whole sections are single- 
> valued, the high-level detail is absent, etc.  This plot pattern  
> will change if I re-open the data file or change which levels VisIt  
> is plotting.  I have reduced the number of AMR levels in some runs  
> and find that VisIt can handle up to around 20, but 31 seems too  
> much...
>
> Has anyone used VisIt with a large number of AMR levels?  Any ideas  
> as to what's going on, and more importantly, how to fix it?
>
> Cheers,
> Sean
>
>
> ---------------------------------------------
> Sean M. Couch
> PhD Candidate
> Department of Astronomy
> The University of Texas at Austin
> 1 University Station C1400
> RLM 17.312
> (512) 471-7418
> smc at astro.as.utexas.edu
> www.as.utexas.edu/~smc
>
>