Skip to content

Derive balanced CRT node selection and precision from the physical source before invoking the numerical producer (no new kappa) - #7154

Open
CubeJerry wants to merge 1 commit into
CrocSwap:mainfrom
CubeJerry:research/continuous-20261110-crt-control
Open

CubeJerry wants to merge 1 commit into
CrocSwap:mainfrom
CubeJerry:research/continuous-20261110-crt-control

Conversation

@CubeJerry

Copy link
Copy Markdown
Contributor

Exact result and novelty

A tape-driven frontend derives every balanced CRT node, width and precision from the physical axis stream, invoking unchanged #5219/#5053 numerical code on the #5505 bank. It uses 78 stable owned roles with linear-use result leases, O(p^4) source operations and O(p^2) owned backing plus the supplied source history. Its raw frontend actions have bound 2^16(M+2)^2; complete control retains an unspecified fixed constant.

Reproduction and independent checks

Four physical fixtures execute 14 actual numerical nodes; an independent 63-node/78-role review and negative controls passed. Replay: python3 -B scripts/research_package.py --package research/encoded-balanced-crt-entry

Limitations and bound status

No new kappa. Row/ROM selection, source updates, three-epoch transport, binary lowering, and whole-iterator restart remain open. Selected conditional kappa remains 0.00363005817; the five framework assumptions remain.

Attribution

Prepared for @CubeJerry with substantial OpenAI assistance; credit #5353, #5219, #5053, #5505 and earlier CRT construction.

Complete review packet

Full PR description, proof and source pins

Base: e54523517d0f03e58352e647c403a5af3e525a57. No upstream PR is opened by this link.

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.

1 participant