Validation is an important part of this work, particularly for the computational projects.
The RGE-256 generators were tested using established statistical test suites, including Dieharder.
Summary:
The PRNG work is used as an experimental probe of recursive entropy theory rather than as a cryptographic claim.
One PRNG paper was endorsed for submission to arXiv but did not pass moderation. Since then, revisions have incorporated feedback on clarity, framing, and presentation. The version history reflects that process.
This site is intended to document that progression openly rather than present a single static result.
The following checks were run locally in clean environments where possible.
pip install -e . succeeded.pytest -q completed with 73 passed, 2 skipped.pip install -e . now succeeds via added package metadata.librge256ctr.so is now optional with a Python fallback when native loading fails.pytest -q tests completed with 7 passed (runtime warnings only).make all succeeded.python3 tests/run_results.py succeeded and wrote RDT_RESULTS.txt.make benchmark-honest succeeded and regenerated benchmark artifacts in results/.pip install -e . succeeded in a torch-enabled venv.pytest -q passed (17 passed).run_demo(...) executed successfully on CPU.256x256, 200 steps, about 2397.82 steps/s on this machine (step(...) loop timing).