On Thu, 17 Aug 2017, Roy Stogner wrote:
On Thu, 17 Aug 2017, Renato Poli wrote:
One more question: ok, now I have the solution at the nodes. I read
in other posts that to evaluate the results in any (x,y) I must get
the shape functions of the element at (x,y), multiply by the
coefficients and accumulate. This is no big deal, but ... is there
any function in libmesh to help me with that?
The overload System::point_value(var_num, point, elem) is the simplest
function to use when you're evaluating the system's solution on a
local or algebraically ghosted element,
Oh, for others' reference: "simplest" still isn't "most efficient" if
you're evaluating lots of different points on a single element,
because in that case you want to reuse things like dof_indices, not
recalculate and tear down for each point. If you're evaluating every
system variable at multiple points within a local element then using
FEMContext should be relatively easy and efficient. If you're
evaluating only one or a few variables and doing so at several points
per element then I don't think we *have* any easy-and-efficient ways
to do so, and if you need efficiency then I guess you're stuck
managing dof_indices and FE objects and doing the arithmetic in user
code.
---
Roy
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users
