[FLASH-USERS] AMR Prolong Error

[Klaus Weide]

On Mon, 25 Jan 2010, William Gray wrote:

> Hello all,
> 
> I am running FLASH3.1.1 where I am using multiple species in a chemistry 
> and cooling network. After my simulation runs for 5686 steps, I run into 
> the error: [amr_long_gen_unk1_fun] PE= 190, ivar= 23, value=******. 
> Trying to convert non-zero mass specific variable to per-volume form in 
> monotonic mesh interpolation, but dens is zero! ivar=23 in this case is 
> HD+ (ionized hydrogen deuteride). This also occurs right after a 
> derefinement.
> 
> I have seen the other thread on this issue 
> (http://flash.uchicago.edu/pipermail/flash-users/2009-April/002815.html). 
> In my case, I impose a floor on my abundances after my chemistry routine 
> to something much greater than smlrho (smlrho=1.0E-40, hd+ = 1.0E-19).
> 
> Is there any knowledge on this error and how to fix it? Any help is 
> appreciated.

Hello William,

Does the problem you described above still exist?  Or did you find a 
solution, or at least a workaround?

The message for which you included a URL started with a similar problem, 
but unfortunately it got never definitely resolved AFAIK. The last message 
in that thread that i could find, 
<URL:http://flash.uchicago.edu/pipermail/flash-users/2009-May/002825.html>, 
states: "It is looking more and more that this was an issue of memory 
corruption of some sort."

If your problem persists:

The message (and abort) could be generated from either a call 
to Grid_fillGuardCells, or during an invocation of amr_prolong
(via gr_updateRefinement). It could be important to understand which
is the case.

It would be important to know whether this message occurs while filling
data for a leaf block, or for a parent or ancestor block.

Try to put in some checks to make sure there are no blocks with a density 
<= 0.0 in any leaf blocks before the relevant calls (i.e., to either
Grid_fillGuardCells or Grid_updateRefinement).  Or perhaps do the same
checks for both leaf and parent blocks (i.e., ACTIVE_BLKS argument to
Grid_getListOfBlocks).

These are some general ideas, more details about your setup (starting 
with a list of the units included, log file and stdout excerpts, etc.) 
will be required to get a handle on this problem.


Klaus