Skip to content

Bit word R480: totals on dead registers, c = 0.001574753352953808650916 (+0.595% over #493) - #503

Merged
Mister-Meeseeks merged 3 commits into
CrocSwap:mainfrom
DreamingOfClouds:e8-r480-dead-totals
Oct 12, 2026
Merged

Mister-Meeseeks merged 3 commits into
CrocSwap:mainfrom
DreamingOfClouds:e8-r480-dead-totals

Conversation

@DreamingOfClouds

@DreamingOfClouds DreamingOfClouds commented Oct 11, 2026 •

Copy link
Copy Markdown
Contributor

One change to #493's E8 bit word. In R489, the nine decoder totals each get a fresh register. Here each total is
formed on a register that is already dead, and the host is the last eligible register of the total's combination.
#493's generator already supported dead hosts but took the first register, and that word fails the official finite
prefix bill. Everything else is #493's word, gate for gate. The result is R = 480, the R ≥ 480 bound #493's
README argues for its family (an analysis under its stated hypotheses: this partition, three branches per source,
totals in phase A; not machine-checked).

R five-stage payload (limit: < 2^104) bit c (#472 eleven-stage whole-residual pipeline, as in #493)
#493 (fresh totals) 489 100 bits 0.001565442026981656908596
#493's generator, dead hosts, first register 480 129 bits, refused by columns_h –
this PR: dead hosts, last register 480 102 bits, PASS 393688338238452162729/250000000000000000000000 = 0.001574753352953808650916 (+0.5948%)

κ does not change, because the side that binds is the complex side. With #484's complex b (Lemmas A and G) κ =
662515748707749/500000000000000000 = 0.001325031497415498. With #471's it is 0.001303530097714685, and with #458's
0.001295523322678044. The bit gain is headroom for any complex supplier above #493's c.

Exploration outside the package (not run by verify.py): over needer swaps and first, last and random total hosts at
R = 480, every variant we priced gave the same c.

What is in research/e8-r480-dead-totals/

BOOST_INCLUDE=/path/to/boost python research/e8-r480-dead-totals/verify.py /absolute/new/output --light   # skips #493's verify.sh
BOOST_INCLUDE=/path/to/boost python research/e8-r480-dead-totals/verify.py /absolute/new/output

Scope

The scope is #493's, unchanged: all of #472's eleven-stage composition interfaces and #425's mutable-source interface
(checked by #493's interface.py). Complex savings are cited from their receipts, not rebuilt. No unconditional
theorem is claimed.

Credits

eumemic (#493: the R489 carrier-skeleton word, the generator with its dead-register option, the relaxed checker, the
interface checks and the R ≥ 480 family argument). Through #493: Zhihao Chen (#425, #386), DreamingOfClouds (#463, #315,
#484), DaysSky (#442, #381), chafreaky (#445, #451, #472, #458), Dylan Crow (#370), icekylinx (#433), Dugongue (#471,
#483), Jacob Sussman (the E8 unit).

Prepared by DreamingOfClouds with Claude Opus 5.5 (Anthropic), via Claude Code.

🤖 Generated with Claude Code

eumemic and others added 3 commits October 11, 2026 13:09
…97714685

A geometry-first bit word under CrocSwap#425's mutable-basis interface: 30 orthogonal
4-frames from 5 complementary D4-block pairs, sources reaching their cells
directly or through two carriers, 60 D4 quads, and D6-triple tops. R = 489
(best previous bit word: 715). Priced by CrocSwap#472's eleven-stage whole-residual
hosted pipeline: c = 0.001565442026981656908596. Complex-bound: kappa =
0.001303530097714685 with CrocSwap#471's complex supplier, 0.001325031497415498 with
CrocSwap#484's draft supplier. Includes CrocSwap#318's joint-dual pin repair.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…16 (+0.595% over CrocSwap#493)

PR CrocSwap#493's R489 word with its nine decoder totals formed on dead registers, each
hosted by the last eligible register of its combination (the parent generator
takes the first, whose five-stage payload is 129 bits > 104). Payload 102 bits;
R = 480, CrocSwap#493's own family lower bound. Same eleven-stage pipeline, audit and
assembly as CrocSwap#493; kappa unchanged (complex-bound).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Scope CrocSwap#493's R >= 480 family bound as an argued analysis under its
hypotheses (not machine-checked); note the adapted audit scripts in
NOTICE; regenerate MANIFEST.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@DreamingOfClouds DreamingOfClouds changed the title [Draft] Bit word R480: totals on dead registers, c = 0.001574753352953808650916 (+0.595% over #493) Bit word R480: totals on dead registers, c = 0.001574753352953808650916 (+0.595% over #493) Oct 11, 2026
@DreamingOfClouds
DreamingOfClouds marked this pull request as ready for review October 11, 2026 22:11
@Mister-Meeseeks
Mister-Meeseeks merged commit 3f27147 into CrocSwap:main Oct 12, 2026
58 checks passed
Mister-Meeseeks pushed a commit that referenced this pull request Oct 12, 2026
…over #493)

The R480 word of #503 (PR #493's word, totals on dead registers), priced in
the nine-stage whole-residual hosted pipeline (32 replicas, m = 64; #443/#444
template and emitter with #442 hosting, vendored byte-identical from #463).
All 138,240 helper residuals hosted whole, no banks. Nine-stage formal columns,
template, parent and hosted audits, invoice, both moment engines and #493's
47-inequality outer assembly pass; kappa unchanged (complex-bound).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@Mister-Meeseeks

Copy link
Copy Markdown
Contributor

Integrated on main in 9449300, with original commit history and individual credit retained.

Your R480 dead-register totals are integrated and used by the selected #535/#510 composition. The changed word and eleven-stage pipeline were replayed. A maintainer portability correction checks exact feasibility and residual histogram when HiGHS chooses a different tied allocation; the frozen physical witness still passes the unchanged owner audit.

Maintainer review and exact scope · Contributor record. The multiplication claim remains conditional on the inherited all-size/realization/analytic interfaces; these checks do not establish a complete unconditional theorem.

Thank you for contributing this work.

Mister-Meeseeks pushed a commit that referenced this pull request Oct 12, 2026
…bit c = 0.00165947204563779652541

Adds research/e8-punctured111-star: an E8 bit word with 111 live ports (9 of the 120
labelled lines punctured), nine rank-7 pseudo-target totals (decoder constant 63),
R = 489 helpers, and the parameterisation of PR #493's hosted-audit chain that prices it
(55 exact string edits, one configuration derived from the word). Bit c is 4.54% above
the live bit c of PR #510; with PR #535's complex b, kappa = 0.001587467473762886
(complex-bound, +0.16% over #535). Regression: PR #503's R480 word is byte-identical through
the stock and the adapted chain. Eight mutation controls, all rejected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants