Manuscript XVI-M@1 · XVI-M.tex · lemma · 7346c909
Lemma: Reported computation for M.4
M.4 Contact collapse and Helmholtz conditions · lines 1191–1200
XVI-M@1#lem:M4comp
contact_collapse.py, helmholtz_check.py and matching_probe.py exit with status .
Kernels are computed exactly over in each block of ; there are blocks, of maximal size .
The procedure has three steps: rows independent mod are selected; the exact nullspace of the selected rows is computed; and every candidate kernel vector is verified exactly against all rows.
Since , verified equality is exact.
The adjoint implementation is unit-tested against Euler derivatives of random quadratic Lagrangians computed by the independent route.
Archives: collapse_kernel_certificates.txt, helmholtz_certificates.txt, matching_certificates.txt.
The full collapse matrices are regenerated deterministically by the scripts and are not stored; they have about (formal) and (adjoint-defect) nonzero entries.
Nearby in this section
- definition: Definition: Contact collapse
- remark: Remark: Single switch; preferred presentation
- lemma: Lemma: Joint derivative budget
- proposition: Proposition: Collapse rank
- definition: Definition: Formal adjoint and self-adjoint subspace
- lemma: Lemma: Helmholtz lemma for linear operators
- section: M.4 Contact collapse and Helmholtz conditions
- theorem: Theorem: Variational subspace of the contact collapse