(this page is under development) |
| Neutral | Good for jet engines, not good for much else. Works, but there are no use cases for biology. It does interface with many of the other FEM programs, like Fenics and code-aster, I haven't found that specific need yet but it's good to know the capability exists. | | Neutral | Another C++ program, it's actually pretty powerful but it doesn't have a Windows version and on that basis it gets a neutral. It does have a Docker image though, and it works fine on Ubuntu, if you're using Linux you may want to check it out. | | Thumbs Up! | Installs perfectly on Windows, works right out of the box. Good stuff! Great tutorials. Has a wide selection of materials. Does not accept input in OBJ form. (Sparse input choices). VTK viewer, output saved in VTK form and directly importable into ParaView. This is a functional, very easy to use FEM simulator. It has its own scripting language. Much of it is written in Fortran, some in C++. There is a Python package, PyElmer. Not specifically for biology, but has a variety of useful functionalities. The solver script is human readable and editable. This is a pro system, the learning curve is steep if you want to build your own simulation from scratch. However the tutorials are excellent, there are 28 of them and you can get through them in about 3 days if you focus. They're how-to, they don't talk much about the science, and they don't talk much about the canned models, but they take you through the workflows and that's what counts. Elmer does not handle biological scenarios especially well, it has no biological materials and converting meshes from triangular form is painful. However there are workaround in all cases, and Elmer is great for testing small portions of larger models. If you're going to subdivide your forces into parts, Elmer is a great tool to have in your toolkit. | | Neutral | Fenics is a powerful tool that's been around a long time. The only reason it gets a neutral is because it has a peculiar way of solving the diff-eq's that doesn't necessarily lend itself well to neuroscience, and in some cases becomes dangerous because of the implicit assumptions. However it's an (open-source) industry standard for solid body modeling and it's nice in that it has a fully capable Python implementation. If you're engaged in FEM on a daily basis this is a great tool to have in your toolbox, you may not end up using it a lot but it's good for some particular things that others have a hard time with. Fenics uses the Dolfinx file format. | | Neutral | Does a lot, but doesn't handle fields very well. Not fully capable for biophysical work. It has its own scripting language, that looks a lot like C. There is a binary install for Windows, and there's a python wrapper for FreeFEM++ called pyFreeFem. I haven't tested the Python version, don't know if it works. | | Thumbs Up! | MFEM is a C++ library that's used in several biological applications. It has some cool particle abilities, and it scales pretty well. There is a Python wrapper called PyMFEM, it has additional dependencies. | | Thumbs Up! | MOOSE is not for the meek. It's a professional, fully capable finite element suite, it's very complicated to install and very complicated to use. BUT... there's nothing else like it. If you really need extreme capability, MOOSE may be the best option for you. It lives in the supercomputer world, and it's been down-ported for use with MPI and GPU's. It works, it's painful to configure on a Windows PC and there's some parts of it that can't be accessed on Windows, but by and large you'll get the same capability the supercomputers have. The Docker version is an attractive option if you don't want to clutter up your PC with dozens of little sub-applications. MOOSE uses VTK and it has many other dependencies, you'll definitely want to read through the installation guide before taking the plunge. MOOSE is being showcased at CNS 2026. Neuroscientists all over are realizing that FEM is the wave of the future, and biological FEM is more complex than other kinds, and they're going to have to jump in and get with the program. Elmer is a lot faster than MOOSE, but it also doesn't have biological materials. You can create them, but it's not easy. | | Neutral | OpenFoam is for computational fluid dynamics "only". It doesn't handle charge very well. It's works, and it's good at what it does, the only reason it gets a neutral rating is because it's not well suited for neuroscience. Even though computational fluid dynamics (CFD) is part of what we do, it's only a small part. Bringing simulations to life is a lot of work, so unless you specifically need parts of CFD the other simulators can't deliver, this tool is a nice-to-have but not essential. However it's definitely educational, it's worth running a simulation on it just so you can say you've done it. |
|