Tobias Ribizel

Your example has a slight bug BTW, which causes the cuSPARSE transpose routine to do something slightly weird - you claim to have `2 * discretization_points` non-zeros, but only fill...

We should also add Conan to the package managers, even though we don't really do much work there. On the release workflow, I want to propose a slight change: If...

Implementation example: Replace ```cpp if (auto& obj = config.get("max_iters")) ``` by ```cpp std::set allowed_keys; auto get_config_value = [](const char* key) { allowed_keys.insert(key); return config.get(key); }; // ... if (auto& obj...

I likely just ran into the same issue, where my original project had the shortname `test`, the new project had the shortname `polymorphismtemplates`, and as far as I can tell,...

This can be ignored safely, CGS is somewhat unstable, and it's only exceeding the test tolerance (which we set somewhat arbitrarily) by a factor of 2.

Unfortunately no. If you want to patch this, replace `(r_mixed() * 1e2)` by `(r_mixed() * 3e2)` (2x with a bit of safety margin)

Unfortunately, we have no ARM systems to test on ourselves (yet)

The output logs are too big to attach here, but I can provide them if necessary. Here are only the lock acquire and release log lines: [locks-filtered.txt](https://github.com/user-attachments/files/20503608/locks-filtered.txt)

Interestingly enough, this also happens when pushing to a source cache while a build is running ``` => Adding package [email protected] to mirror ==> Warning: Error while fetching [email protected] Timed...

If you want a quick fix, you can add a `return()` after the warning message ``` message( WARNING "Disabling test: ${test_target_name}. Only ${MPIEXEC_MAX_NUMPROCS} " "MPI processes available but ${add_test_MPI_SIZE} processes...