Search found 4812 matches

by raback
24 Aug 2011, 14:00
Forum: ElmerSolver
Topic: Element coupling
Replies: 1
Views: 1997

Re: Element coupling

Hi Alex First of all. We might distinguish explicit and implicit coupling. Implicit coupling of different equations is rather hard since then there needs to be a monolithic matrix equation that includes all the unknowns. The legacy strategy of Elmer, and the one still mainly used, is to have a separ...
by raback
24 Aug 2011, 13:40
Forum: Installation & compilation
Topic: ElmerSolver_mpi and MPICH2 v1.4
Replies: 3
Views: 4735

Re: ElmerSolver_mpi and MPICH2 v1.4

Hi The old mpi version of Elmer available at sf.net is certainly outdated. It takes some effort to compile the parallel version under mingw. At least the most developers use Elmer mainly on Unix platforms so this is not of high priority. As an altenative for testing parallel stuff on Windows platfor...
by raback
22 Aug 2011, 15:19
Forum: ElmerSolver
Topic: 2 instances of SaveScalars in sif possible?
Replies: 3
Views: 2222

Re: 2 instances of SaveScalars in sif possible?

Hi

The SaveData may have some static settings. So to be totally sure that things are going without problems you could make a copy of the shared library of SaveData under different name, and use that in the Procedure statement.

-Peter
by raback
19 Aug 2011, 09:55
Forum: ElmerSolver
Topic: Segregated and coupled algorithm
Replies: 3
Views: 3571

Re: Segregated and coupled algorithm

Hi cupper Yes, you're right. There are different pros and cons for monolithic and segregated systems. Monolithic systems tend to consume lots of memory and this is amplified by the denser matrix topology of FE when compared to FV. Now segragated schemes split the problem into a number of simpler tas...
by raback
19 Aug 2011, 09:39
Forum: ElmerSolver
Topic: monolithic fsi coding
Replies: 2
Views: 1645

Re: monolithic fsi coding

Hi To have everything as a monolithic system is often optimal for the convergence for the coupled nonlinear system but you tend to loose in that the linear system is more difficult to solve than the individual linear systems would be. Still, monolithic approach is sometimes desirable. Now the steps ...
by raback
19 Aug 2011, 09:30
Forum: ElmerSolver
Topic: Transient analysis as a series of steady-state problems
Replies: 3
Views: 3102

Re: Transient analysis as a series of steady-state problems

Hi Tibi

ElmerSolver Manual, Chapter "Parallel Runs" (currently Ch. 10).

-Peter
by raback
18 Aug 2011, 22:41
Forum: ElmerSolver
Topic: Transient analysis as a series of steady-state problems
Replies: 3
Views: 3102

Re: Transient analysis as a series of steady-state problems

Hi Sorry, you cannot parallellize the problem like that. The previous timestep has to be resolved before going further and hence the time-integration is doomed to be sequential. However, transient problems tend to parallellize rather nicely (mass matrix dominates i.e. preconditioning is easy) using ...
by raback
18 Aug 2011, 22:35
Forum: ElmerSolver
Topic: Avoid writing .ep file
Replies: 2
Views: 1815

Re: Avoid writing .ep file

Hi,

Just omit the keyword "Post File" and none will be written.

-Peter
by raback
18 Aug 2011, 18:48
Forum: Updates
Topic: Elmer group in LinkedIn
Replies: 7
Views: 13779

Elmer group in LinkedIn

Dear Elmer Users, This is just to inform that there is since some time a group "Elmer finite element software" in the LinkedIn professional network. http://www.linkedin.com -> Groups -> Search "Elmer" The group is open to all users and followers of Elmer finite element software. ...
by raback
16 Aug 2011, 14:49
Forum: Events
Topic: Elmer Intensive Course, Sep 7th 2011
Replies: 0
Views: 6507

Elmer Intensive Course, Sep 7th 2011

Dear Elmer users, There will be an Elmer Intensive Course taking place on 7th of September 2011. Elmer Intensive Course is intended for people who want to dwell quickly into to the world of Elmer. The participants should preferably have some background in the solution of partial differential equatio...