Re: [Libmesh-users] Element data

2012-07-02 Thread David Knezevic
Agreed. Someone (any volunteers?) should just make a Poisson problem with a per-element "conductivity" and put it in examples/introduction. David On 07/03/2012 12:36 AM, Derek Gaston wrote: > Man we get these questions a lot. Do we have any examples of using > explicit systems in this manner

Re: [Libmesh-users] Element data

2012-07-02 Thread Derek Gaston
Man we get these questions a lot. Do we have any examples of using explicit systems in this manner? We should _really_ make some... Derek On Mon, Jul 2, 2012 at 6:30 PM, David Knezevic wrote: > The standard advice is to set the variable type to "CONSTANT MONOMIAL" > so that you just store one

Re: [Libmesh-users] Element data

2012-07-02 Thread David Knezevic
The standard advice is to set the variable type to "CONSTANT MONOMIAL" so that you just store one value per element. David On 07/02/2012 07:48 PM, Andrew E Slaughter wrote: > Is there a proper way to handle element data? The way I am doing it now, > with an TransientExplicitSystem, computes a

[Libmesh-users] Element data

2012-07-02 Thread Andrew E Slaughter
Is there a proper way to handle element data? The way I am doing it now, with an TransientExplicitSystem, computes a value at every node, thus for a 4-node element the same value is computed 4 times. Obviously, things get really inefficient with 3D and higher-order elements. Any help is greatly ap