Dear Zhenghuai Guo,

What version of GetFEM are you running? Is it possible to share your
crashing code with us or even better a minimal code that reproduces the
segmentation fault?

BR
Kostas

P.S. please send your messages only to the list address.

On Sun, May 5, 2019 at 4:27 PM Zhenghuai Guo <[email protected]>
wrote:

> Hi, Konstantinos and Yves  and Andriy,
>
> I use gmsh to make meshes. In case 1, I generate triangle elements. In
> case 2, I use the recombine function in
> gmsh to generate quadrangle elements - it is actually mixed quadrangle and
> triangle elements, but majority is quadrangle elements.
> This reduces the mesh size.
>
> I run the exact same code. In the code I only add linear_elasticity_brick:
>   getfem::add_isotropic_linearized_elasticity_brick(model, mim_1, "u_1",
> "lambda_1", "mu_1");
>   getfem::add_isotropic_linearized_elasticity_brick(model, mim_2, "u_2",
> "lambda_2", "mu_2");
>
> and Dirichlet_condition
>
>   model.add_initialized_fixed_size_data("Dirichlet_value",
> base_small_vector(0,0.));
>   getfem::add_Dirichlet_condition_with_multipliers(model, mim_1, "u_1",
> mf_u_1, total_bounday_1,"Dirichlet_value" );
>   getfem::add_Dirichlet_condition_with_multipliers(model, mim_2, "u_2",
> mf_u_2, total_bounday_2,"Dirichlet_value" );
>
> no source term, no mass term
>
> The code runs ok for meshes of the case 1 and case 2.
>
> But, if I add the contact_brick  as below, only the meshes of the case 1
> is ok.
>  ind = getfem::add_integral_large_sliding_contact_brick_raytracing(model,
> "1", release_dist, "0", "1", false); ///
>   getfem::add_contact_boundary_to_large_sliding_contact_brick(model,ind,
> mim_contact_1, total_bounday_1, true, true, "u_1", "pre_mflambda_1");
>   getfem::add_contact_boundary_to_large_sliding_contact_brick(model,ind,
> mim_contact_2, total_bounday_2, true, true, "u_2", "pre_mflambda_2"); ///
>
> When I use meshes of the case 2, it gives an error of
> Trace 2 in getfem_models.cc, line 2698: Global generic assembly RHS
> Segmentation fault (core dumped)
>
> Could you please advise about this?
>
> Thank you very much
> Regards
> Zhenghuai Guo
>
>
>
>
>
>
>
>

Reply via email to