NRTidalv2 self-spin paper-D component receipt — 2026-09-27
Scope: Dietrich et al. (2019) printed Eqs. (26)–(29) and Table II for positive \(\Lambda\), aligned spins, CPU FP64, 2PN+3PN D self-spin. The separate 3.5PN component is tested as an equation, not included in the published D composition. No pinned-binary or physical-data claim follows from this gate.
Source: arXiv:1905.06011v2, Eqs. (26)–(29), Table II. The coefficient strings are transcribed at printed precision into an independent mpmath reference. Torch/JAX do not import it. See the math specification for conventions and cases.
Frozen gate: 30, 100, 300, 800 Hz, four mass/spin/deformability cases, including body exchange and zero spin. Maximum component absolute error \(10^{-12}\); pre-taper complex quotient absolute error \(10^{-10}\) without alignment; spin derivative absolute error \(10^{-10}\). Both coalescence and 30 Hz reference-frequency adapters are checked. The latter is a native convention.
Run: host len, cwd
/home/xangma/tmp/diffgw-nrtidalv2-self-spin-20260927, command
PYTHONPATH=src:tests /home/xangma/.cache/diffgw-taylorf2-hm-venv/bin/python
-m pytest -q tests/test_nrtidalv2_self_spin_math_reference.py
tests/test_nrtidal_reference.py tests/test_nrtidalv2_phase_math_reference.py
tests/test_nrtidalv2_amplitude_math_reference_native.py tests/test_nrtidal.py
tests/parity/test_nrtidal_parity.py. Result: 116 passed, 1 skipped in
40.89 s. Versions:
Python 3.12.3, Torch 2.14.0+cpu, JAX 0.11.2, pytest 9.1.1, mpmath 1.3.0.
The skip is an existing comparison requiring LALSimulation or PyCBC.
Interpretation: The printed equations and chosen D composition agree with both native frontends on the frozen slice. This does not verify the whole NRTidalv2 waveform or binary/NR parity. The paper's Fig. 7 text reports \((C_Q,C_{Oc})=(4.30,7.28)\) at \(\Lambda=350\), whereas its printed Eqs. (28)–(29) yield \((5.29659,10.49832)\). The discrepancy remains unresolved; this gate deliberately follows the numbered equations and records the conflict. Historical binary-compatible CQ coefficients, Lambda clamps, reference-frequency anchoring, merger taper, and full spinning-waveform comparisons remain separate scopes.