Sungho Lee

Hi @vladotomov, When TMOP runs, I'd like the boundary nodes to move tangentially to the arbitrary shape of the left, right, top, and bottom boundaries. So, if the domain is...

Hi @vladotomov, It's been a long time since I started working on boundary-preserved remeshing on a deformed boundary. I realized that a two-step remeshing approach would resolve this issue. Figure...

@vladotomov, Yes yes, I've faced the negative Jacobian issue (https://github.com/mfem/mfem/issues/4047). This is strange as this process is technically stretching. However, things tangle eventually. I don't understand why, and how to...

@vladotomov, Yes, coding is my work. Addressing questions and offering general guidance would be sufficient. My plan is to use the existing structure in Laghos. First, I added a new...

@vladotomov, I've managed to solve the issue, and reshaping from vector to tensor had some problems. However, I've encountered another problem related to memory management in CUDA. I have a...

@vladotomov Yes, I did but it is over my head. Since `rhs ` is the result of `ForcePA->Mult(one, rhs)`, `rhs` is allocated in GPU's memory, isn't it? So I presumed...

@vladotomov Thank you for your comment, but I think I take my time to digest it. I just tested vector operations within MFEM. In Laghos, velocity is calculated component-wise, as...

Thank you for your comments. `B` printed out zero, but it printed out some values after using `HostRead()`. I ran a problem (consolidation due to body force) using CPU and...

@vladotomov I see, but I don't know the reason for this yet, and I'll let you know if I figure it out. By the way, I've tested laghos with CUDA...

I used the master version of Laghos with a few modifications to run high-order elements greater than 4. For example, in `laghos_assembly.cpp`, I modified the lines to include `{0x290,&ForceMult2D}` and...