Elmer Discussion Forum

Hi all,

I'm trying to get up to speed with the ParStokes solver using the information on the wiki. However, I'm having trouble getting a simple inclined slab test-case to run. When I try, I get the error "ERROR:: LUDecomp: Matrix is singluar.", followed by a segfault from within ParStokes. I suspect that I have something misconfigured (initial conditions?) but I'm not sure exactly what.

Does anybody have a working example of a simple test-case for ParStokes?

Cheers,
Martin

Hi,

The following sif-file should give a starting point for writing a working case definition.

Best regards,
Mika

Thanks, Mika - I managed to figure it out on my own, but hopefully this will help others in the future.

Also, am I right in thinking that the ParStokes solver is 3d-only? I haven't had any success with 2d cases.

Checking the code, it seems that the dimension of the problem is being taken care of - i.e., I found no reason why it shouldn't work in 2D. Question is - do you really need it in a 2D problem or are you equally happy with the standard Navier-Stokes Solver and a direct solver?

Regards,

Thomas

No, I'm fine with just 3d. I was only trying 2d for a quick test comparison between Navier-Stokes and ParStokes.