Issues Installing exafmm-t for BEMPP

[Psetty97]

I’ve successfully installed the necessary dependencies, including GNU Make, OpenBLAS, FFTW3, and GFortran. I’ve also updated my MSVC build tools to ensure support for C++14 standards. Despite these preparations, when I try to install exafmm using:

pip install git+https://github.com/exafmm/exafmm-t.git
I receive the following error:

---

error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for exafmm
Running setup.py clean for exafmm
Failed to build exafmm
ERROR: Failed to build installable wheels for some pyproject.toml based projects (exafmm)

---

I am unsure how to proceed as I believe my development environment is correctly set up with the necessary compiler and SDK. Has anyone faced a similar issue or can offer any guidance on how to successfully install exafmm for use with BEMPP? Any help would be greatly appreciated!

Thank you!

[mscroggs]

This doesn't really solve your installation issue, but as an alternative solution, you could run inside the docker image that Bempp provides:

docker pull bempp/cl-notebook:latest
docker run -p 8888:8888 bempp/cl-notebook

(more info about the docker image at https://bempp.com/installation.html)

[Psetty97]

Thank you for your suggestion. I have successfully employed the BEMPP Docker image within a Docker container running on Ubuntu via Windows Subsystem for Linux (WSL) and used fmm assembler.

I am working on calculating the sound pressure around a gearbox casing using a boundary element method (BEM). The approach involves applying velocities to each node of the gearbox and calculating the resulting pressure on an external sphere mesh. The computation is conducted with the following line of code:

---

p_total, info = gmres(double_layer - 0.5 * identity, 1j * omega * rho_0 * single_layer * u_total, tol=1E-5)
p_plotting = double_layer_plotting * p_total - 1j * omega * rho_0 * single_layer_plotting * u_total

---

However, I'm experiencing performance issues. The computation involves 105,000 degrees of freedom (DOF) for both the single and double layer spaces and takes over an hour to complete the calculations for 1 frquency. In contrast, a similar FEM model in Abaqus, which includes domain elements as well as boundary elements, completes in just 7 minutes.

Given this context, I am concerned there might be an issue with my system settings or the configuration of the Docker container that could be impacting performance. My Laptop is equipped with Windows 10, 32GB of memory and an Intel i7-2.30 GHz (10 cores) processor. I would appreciate any specific Docker or WSL configurations known to enhance BEMPP's performance, or if further tuning of my system's settings is recommended, I would be grateful for such specific suggestions.

[mscroggs]

I'm afraid I'm not very familiar with using Docker in windows. I might be worth posting this in the Bempp discourse (https://bempp.discourse.group/) where a lot of our users will see it and may be able to help.