scivision

i see this problem with GCC 12.2, Cuda 12.2 and instant-ngp commit 9f6ab886

I don't see this on macOS with Apple Silicon. I wonder if it's a bug with HDF5 itself for Apple PowerPC?

I cherry-picked individual commits into main until the tests started failing. I then switched back to your branch and rebased on main--still failing tests, but at least consistent with main....

The MacOS CI is added in https://github.com/Reference-ScaLAPACK/scalapack/pull/77 Due to MacOS-specific Scalapack code issues, I made MacOS a separate job

Thanks, I had to fix other bugs first related to xarray changes. Thanks for your changes and endorsement!

the pymap3d doesn't convert velocities for ECEF2ECI

The design goal of --reconfigure is to work as you say: https://github.com/mesonbuild/meson/pull/4077 Except, I'm not sure about `c_flags` which is a reserved variable. What I found for Fortran was that:...

I think have a Readme for each, perhaps linking to the respective repos for more details. The main place where difference occur is in adding functions, since that works so...

This is broken in v5.2.1 as well. If you wish to build with CMake, consider https://github.com/scivision/metis

Just to add another CMake perspective, with details of how CMakes implement Fortran modules and submodules and thoughts how they'd do C++ https://mathstuf.fedorapeople.org/fortran-modules/fortran-modules.html