Refactor benchmark solver setup

[sylvesterkaczmarek]

PR Type

• Refactoring (no functional changes)
• Documentation content changes
• Tests

Description

Closes #1002.

Consolidate benchmark solver construction in coreax.benchmark_util, while preserving benchmark-specific calibration. Reuse the calibrated Stein kernel within one benchmark registry, make visualiser output paths independent of the current working directory, and correct the blobs benchmark so the random seed is not passed as the thinning failure probability.

The contributor guide now documents the benchmark workflow and the settings that must remain comparable across runs.

How Has This Been Tested?

• 21 focused benchmark tests passed.
• Full unit suite: 1,753 passed, 19 skipped.
• Pre-commit checks passed, including Ruff and Pyright.
• Pylint code score: 10.00/10.
• Sphinx documentation build passed with -n -W.

Does this PR introduce a breaking change?

No. Existing public solver interfaces are unchanged.

Checklist before requesting a review

• I have made sure that my PR is not a duplicate.
• My code follows the style guidelines of this project.
• I have ensured my code is easy to understand, including docstrings and comments where necessary.
• I have performed a self-review of my code.
• I have made corresponding changes to the documentation.
• New and existing unit tests pass locally with my changes.
• I have updated CHANGELOG.md.