% SPDX-License-Identifier: CC-BY-NC-4.0
Knot-Identification Pipeline — machine & environment
W.A.Stanggassinger in collaboration with Claude Code (Anthropic)
CC BY-NC 4.0 — https://creativecommons.org/licenses/by-nc/4.0/

This records the fixed, version-pinned environment on which the knots were
identified and the blind validation was run. The toolchain is held to recorded
versions so that results are reproducible and free of silent tool drift; the
no-lookup discipline (no catalogue lookups in the primary identification path)
is enforced in code (a pipeline seal).

HARDWARE
  CPU   AMD Ryzen 9 3900X   — 12 cores / 24 threads
  RAM   32 GB DDR4
  GPU   NVIDIA GeForce RTX 3070, 8 GB VRAM   (driver 580.126.09, CUDA 13.0)
  OS    Linux kernel 6.17.0

IDENTIFICATION TOOLCHAIN (pinned versions)
  Python        3.11.15 / 3.13.12   — pipeline env / regina env
  SnapPy        3.3.2               — hyperbolic structure, volume, symmetry group
  regina        7.4                 — diagram / topology
  SageMath      10.7 (2025-08-09)
  NumPy         2.4.3
  SciPy         1.16.3
  SymPy         1.14.0              — polynomial algebra (Alexander / Jones / HOMFLY)
  CuPy          14.0.1              — GPU curve scan
  KnotJob       build 2026-02-12    — Khovanov homology / s-invariant (D. Schuetz)
  Theta.sage    build 2026-06-02    — Theta invariant of Bar-Natan & van der Veen
  Blender       5.0.1               — centerline extraction / rendering

CROSS-CHECK CATALOGUES  (used only after the invariants are computed)
  KnotInfo · the SnapPy census · a numerical Theta-DB over the 59,937
  knots.sage entries.