Re: [Libmesh-users] variables at node

2013-03-12 Thread Subramanya Gautam Sadasiva
2, 2013 5:10:59 PM Subject: Re: [Libmesh-users] variables at node On 03/12/2013 05:09 PM, Roy Stogner wrote: > > On Tue, 12 Mar 2013, David Knezevic wrote: > >> Quick question though, because I'm not sure I understand how this >> works when we're numbering in "var

Re: [Libmesh-users] variables at node

2013-03-12 Thread David Knezevic
On 03/12/2013 05:09 PM, Roy Stogner wrote: > > On Tue, 12 Mar 2013, David Knezevic wrote: > >> Quick question though, because I'm not sure I understand how this >> works when we're numbering in "var major" order. Suppose we have >> variables "u", "v" and "p", then the dof indices corresponding to

Re: [Libmesh-users] variables at node

2013-03-12 Thread Roy Stogner
On Tue, 12 Mar 2013, David Knezevic wrote: > Quick question though, because I'm not sure I understand how this works when > we're numbering in "var major" order. Suppose we have variables "u", "v" and > "p", then the dof indices corresponding to any particular node would > typically be very fa

Re: [Libmesh-users] variables at node

2013-03-12 Thread Roy Stogner
On Tue, 12 Mar 2013, Manav Bhatia wrote: > Using a functor was my original intent, but then realizing that the > calculated values at a point (for the post-processed variables) > would depend on the solution from another system (the original > conservative variables), I was not sure of how to pas

Re: [Libmesh-users] variables at node

2013-03-12 Thread Derek Gaston
It's funny because MOOSE completely relies on being able to loop over local_active_nodes and set degrees of freedom for Lagrange based variables. If that ever went away we would be screwed ;-) Derek On Tue, Mar 12, 2013 at 11:43 AM, Roy Stogner wrote: > > On Tue, 12 Mar 2013, John Peterson wro

Re: [Libmesh-users] variables at node

2013-03-12 Thread Roy Stogner
On Tue, 12 Mar 2013, John Peterson wrote: > On Tue, Mar 12, 2013 at 11:31 AM, Roy Stogner > wrote: > > > For problems with lots of variables it should be a tiny bit more > > efficient to just test the DofObject::processor_id(); we never put > > dofs on a processor other than the one where thei

Re: [Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
Thanks Roy. Using a functor was my original intent, but then realizing that the calculated values at a point (for the post-processed variables) would depend on the solution from another system (the original conservative variables), I was not sure of how to pass the latter to the functor. So,

Re: [Libmesh-users] variables at node

2013-03-12 Thread John Peterson
On Tue, Mar 12, 2013 at 11:31 AM, Roy Stogner wrote: > > On Tue, 12 Mar 2013, John Peterson wrote: > > On Tue, Mar 12, 2013 at 11:13 AM, Manav Bhatia > >wrote: >> >> Should I be iterating over the pid_nodes or the active_nodes if my >>> intention is to set the values of dof on each node for o

Re: [Libmesh-users] variables at node

2013-03-12 Thread Roy Stogner
On Tue, 12 Mar 2013, John Peterson wrote: > On Tue, Mar 12, 2013 at 11:13 AM, Manav Bhatia wrote: > >>Should I be iterating over the pid_nodes or the active_nodes if my >> intention is to set the values of dof on each node for output. > > Neither. You should be checking to see if a given dof

Re: [Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
But what should be the outer for loop? I am currently doing this iterate over local nodes iterate over vars on node get dof_id of var and set value in system.solution Are you suggesting that I do the following on each processor iterate over ALL nodes iterate over vars on node

Re: [Libmesh-users] variables at node

2013-03-12 Thread John Peterson
On Tue, Mar 12, 2013 at 11:13 AM, Manav Bhatia wrote: > Hi John, > > Question: > >Should I be iterating over the pid_nodes or the active_nodes if my > intention is to set the values of dof on each node for output. > Neither. You should be checking to see if a given dof lies between the first

Re: [Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
Hi John, Question: Should I be iterating over the pid_nodes or the active_nodes if my intention is to set the values of dof on each node for output. Thanks, Manav On Mar 12, 2013, at 12:11 PM, John Peterson wrote: > > > > On Tue, Mar 12, 2013 at 9:36 AM, Manav Bhatia wrote: > Ok, s

Re: [Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
Not for initial conditions. I am adding some post-processing variables (Mach, cp, p, T, etc.) after solution of Euler equations. I want to also use the VTKIO to write out the post-processed variables. So, I ended up adding another system to my equation systems, which only defines these additi

Re: [Libmesh-users] variables at node

2013-03-12 Thread John Peterson
On Tue, Mar 12, 2013 at 9:36 AM, Manav Bhatia wrote: > Ok, so if I iterate over the pid_nodes and set the nodal variable values > in system.solution, should I call any function (localization, etc.) on the > solution vector before I can use it (for example, write it out using > VTKIO)? > Are you

Re: [Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
Ok, so if I iterate over the pid_nodes and set the nodal variable values in system.solution, should I call any function (localization, etc.) on the solution vector before I can use it (for example, write it out using VTKIO)? Thanks, Manav On Mar 12, 2013, at 11:24 AM, John Peterson wrote: >

Re: [Libmesh-users] variables at node

2013-03-12 Thread John Peterson
On Tue, Mar 12, 2013 at 9:14 AM, Manav Bhatia wrote: > Hi, > >Is there a function to initialize a dense vector to the variable values > at a node? Not that I know of. > Or do I need to iterate over each variable and get the dof values > independently? > Yes, and if you want it to work fo

[Libmesh-users] variables at node

2013-03-12 Thread Manav Bhatia
Hi, Is there a function to initialize a dense vector to the variable values at a node? Or do I need to iterate over each variable and get the dof values independently? Thanks Manav -- Symantec Endpoint Protection