On Apr 29, 2016, at 2:02 PM, Kirk, Benjamin (JSC-EG311)
<benjamin.k...@nasa.gov<mailto:benjamin.k...@nasa.gov>> wrote:
On Apr 29, 2016, at 1:53 PM, Derek Gaston
<fried...@gmail.com<mailto:fried...@gmail.com>> wrote:
What kind of error do you get? I could try a quick problem…
No assertions or anything, just some crap values coming out of the first linear
solve after AMR. Interestingly looks like on the interface of the subdomain
restriction.
For example, if i manually force AMR in the middle of the ablation subdomain,
on one quad, i get 4 hanging nodes. I then get 4 crap values elsewhere, on the
ablator/insulator interface.
So I’m leaning to the app code, but can’t be sure. hence the question.
To assuage fears, we tracked this down to the app code.
DofMap::extract_local_vector was being used, except the wrong DofMap for the
input global vector! (lots of systems going on at once.) I looked into the
libMesh side of things to see if we could prevent this in the future with an
assert, but it’s not clean - the NumericVector has no notion of the DofMap, so
there is no clean way to assert the right DofMap is being used for the input
vector. Happily, the two different DofMaps that were erroneously being used
interchangeably were in fact identical until AMR+subdomain restricted variables
caused them to be different.
Subtle, but resolved.
-Ben
------------------------------------------------------------------------------
Mobile security can be enabling, not merely restricting. Employees who
bring their own devices (BYOD) to work are irked by the imposition of MDM
restrictions. Mobile Device Manager Plus allows you to control only the
apps on BYO-devices by containerizing them, leaving personal data untouched!
https://ad.doubleclick.net/ddm/clk/304595813;131938128;j
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
