Repository navigation
DONE: P104 transport floor with added workspace lifetimes #2123
RossRichardson
started this conversation in
Ideas
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
DONE / HANDOFF / RELEASED — 2026-10-14.
PR2127, exact head
befebf0f4ae7a3f69fe9c5e136262559dcac125f, contains the completed scoped result inresearch/p104-storage-cut/. The claim is released.The 444-frame static graph splits into 222 source-reachable and 222 target-coreachable frames after removing the eight original C descents. This forgets every role, time and lifetime distinction and therefore grants arbitrary same-frame storage. The written finite-linear-program argument gives rank(X->Y)<=(8+c)d, so 104-array identity requires at least 96 added outward whole-array transports, even with arbitrarily many additional source-independent workspace lifetimes and arbitrary valid retiming of retained transports. Their nested positive-rank mass is at least 96; this is not an attaining construction.
Added rank mass is not automatically excess recursive cost. A concrete monotone 0->source-line->target-hyperplane->FULL path shows that additional helper trajectories can have productive progress. The inherited M(0)>=7659/7592 comparison applies only with unchanged old idle requirements and no new creditable endpoint progress. No supplier or new kappa is claimed.
Two separately implemented exact checkers agree on every transport and complete partition. Registered CPython 3.12.12 replay and six behavior groups pass normally and optimized, including actual cut-guard removal, source/receipt controls and read-only source-export replay. Selected-record integrity passes. No fresh full scalar/native/phase replay, global hosting, precision, all-size/outer proof or formal-kernel proof is supplied; hosted CI was queued, not passed.
Next: replace actual transport occurrences and test the new cut with a complete arbitrary-input/native bill, or realize and price a productive borrowed-helper trajectory under its actual coverage/owner contract. Discussion 2121's replacement joint-side producer is outside this retained-occurrence theorem.
Immediate credit: Jeremy Diviney/PR2113, David Arias/PR2082 and Adam Gall/PR2090/2083; channel/native/schedule credit Rohan Arun/PR1972, Chafik Boukhalfa/PR1991 and Adam Gall/PR1940, with geometry and coverage predecessors retained in the README. Substantial OpenAI Codex assistance, including complex-supplier and exact-verification specialists; no independent human review.
Base
0fdd7d82de03e43c9e5a1000dfbcd380d1b687d2, priority revision 4, selector 2323fe64f7a4189b7512b9541840f389. Submission overlap covered 381 open PRs and 789 Discussion records at 20:09-20:11 UTC with complete pagination/no warnings; text matching is not a novelty guarantee. Reproduce:python3.12 scripts/research_package.py --package research/p104-storage-cut.All reactions