Pre-scaling weight monitoring with new QR filter

[KyleDavisSA]

Main changes of this PR

This PR closes #1152 due to being too old. The reason for this PR are copied from #1152.

1. Automatically deletes the first column in time step 4 if an initial sub-vector from at least 1 solver is zero, measured by the L2 norm.
2. Monitors the residual sum preconditioner weights and only updates them if they change by a specified factor. This is required for faster QR filtering.

Motivation and additional information

Currently the preconditioner weights are updated in every iteration. This is unnecessary if the weights change by a small value, but the entire QR decomposition is performed again. Vast improvements in runtime can be found by not always updating the preconditioner weights.

Author's checklist

• I used the pre-commit hook to prevent dirty commits and used pre-commit run --all to format old commits.
• I added a changelog file with make changelog if there are user-observable changes since the last release.
• I added a test to cover the proposed changes in our test suite.
• I sticked to C++17 features.
• I sticked to CMake version 3.16.3.
• I squashed / am about to squash all commits that should be seen as one.

Reviewers' checklist

• Does the changelog entry make sense? Is it formatted correctly?
• Do you understand the code changes?

• Allow preconditioner weight monitoring to be turned on/off via the config file.
• Allow column deletion to be turned on/off via the config file.
• Include the fast QR filter when preconditioner weight monitoring is applied.
• Include tests for QR filters
• Update user documentation

[davidscn]

Closing #1152 in order to keep track of the relevant PRs.