Repository navigation
Conditional κ = 0.001677162029073474: four helper-donor cold kernels (pivots 531, 543, 555, 567) composed with multi-donor idle exchanges toward rank 51 - #1305
Closed
jeremykothe2 wants to merge 1 commit into
Conversation
…(pivots 531, 543, 555, 567) composed with multi-donor idle exchanges toward rank 51
Author
|
Closing: this result is invalid. Its idle-overlay step moves axes between idle children of different (pair, side) families. The retained DATA-PHASES recipes (data_phases.py, exact_child) fix each family's Z4 multiplier, and every changed child here alters its family's multiplier, so the claimed κ is not supported. Thanks to rohanarun (#1286) for flagging the overlay problem. The overlay mechanism inherited from #1252 has the same issue. Our overlay-free results (#1261, #1274) are unaffected. |
GamingPuzzled
added a commit
to GamingPuzzled/integer-mult-bounds
that referenced
this pull request
Oct 14, 2026
…CrocSwap#1305/CrocSwap#1324) were withdrawn as invalid Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mdumitrean
added a commit
to mdumitrean/integer-mult-bounds
that referenced
this pull request
Oct 14, 2026
…rding Answers the CrocSwap#1337 review continuation point (supply a rigorous rounding/separation bound or narrow the optimality claim to the integer objective; the verifier does not regenerate all priced evidence): - price.py and retime.py now price with PR CrocSwap#1186 revision 2's Pricer(interval=True) (bounds.py): every candidate records its realized true delta, and a rigorous bound on any true gain, rounded up to 1e-6 units. All 1,438 neutral candidates have true delta exactly 0 and possible gain at most 1e-6 units (exact bounds about 1e-11). Every verdict, delta and profile equals revision 1's evidence. The positive control is -11707 surrogate units, -11707.88 true units, and passes audit_local. - verify.py checks the bounds against RESULT.json (true_moment_bound, tallies.intervals) and adds a forged-bound control; its docstring and VERIFY.json state that it is a recount, while regenerate.py reruns every price (toolkit pinned to CrocSwap#1186 3855fd4 by its MANIFEST; two consecutive regenerations are byte-identical to evidence/). - README and RESULT.json: "exactly optimal" narrowed to surrogate-optimal plus the stated bound; the withdrawal of CrocSwap#1305/CrocSwap#1324 and the later pivot-566/554 kernels on these words are recorded.
Mister-Meeseeks
pushed a commit
that referenced
this pull request
Oct 16, 2026
…no new kappa) Exhaustive census of the #1274/#1297 two-shock helper-donor kernel family (both restoration rules) and of exact early-read re-timings over all 306 early-read helper pivots of the local words shared by #1297, #1305 and #1324, priced with the #1186 exact frame harness: no gain in 61 + 292 relation candidates and 803 re-timings per word. A positive control reproduces #1297's 5+5 hooks for pivots 555/567 at the realized -11707. #1305's idle selection equals the exact optimum of its admitted donor-star class in all 3,687 groups (superseded by #1324's larger swap class). verify.py (standard library) replays the censuses exactly, binds every priced record to a census match, recounts the evidence and rejects four failure controls; regenerate.py reproduces all 16 evidence files byte-for-byte with the toolkit.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Conditional κ = 0.001677162029073474. Composes rohanarun's #1297 four helper-donor cold kernels (pivots 531/543/555/567; 531 from Jeremy Kothe's #1274) with the multi-donor idle rank-48 exchange overlay (#1252 method, generalised to donor stars in #1268/#1280): 3,251 donors and 5,025 receivers turn idle H48=11374 into H42=1774, H45=1477, H48=3098, H51=5025 with axis unions, owners and dispatch unchanged. Both new local words get fresh native replays; the overlay adds an independent structural verifier and an exact emitted-stream operator check (T and the Z4 phase form of every changed port, chronology and shifts included, plus direct evaluation on random basis states) before the overlay-aware controller audit, invoice and strict assembly with 114 admission and 12 bit controls.
Verification
Fresh checkout of
a96a5d959672752026292ba5024664b436879fe7. Manifest:bc5e08599597803b7d45d7ce5c8765387f8d07f4a86fec9d72d058a4c8b7c99f.PASS package manifest: 101 files
Fresh native replays of both provider words (phases step edge-parallel, byte-identical output; with --native-cache a word whose hash-checked completed replay is cached reuses it), then the package's own
verify.py(literal words, independent local audit, fusion/X54 controls, leaver premise, literal construct, strict global admission):PASS_GLOBAL_MIXED_X54_IDLE_EXCHANGES_NATIVE_AND_STRICT_ADMISSION 0.001677162029073474Full package-owned verification passed in 75.38 seconds. The fresh receipt binds the exact package manifest and score.
See
research/global-mixed-four-cold-kernels-idle/README.mdfor reproduction commands and proof qualifications.Credits
Four helper-donor cold kernel words and profile: #1297 (rohanarun) building on #1286, #1274 (Jeremy Kothe, pivot 531 mechanism) and #1218 (rohanarun) / #1193 (jeremydiviney) / #1148 (tomdif) / #1195 / #987,#1225 construction. Idle exchange method and overlay-aware audit: #1252 (Gonzalo-Mier); donor-star ceiling analysis #1276 (PatatePatatePatate); packing-cap analysis #1240/#1264 (Mister-Meeseeks). Overlay as in #1268/#1280. Research direction, prioritisation and release strategy: jeremykothe2. Implementation assistance: oh-my-pi.