Repository navigation
Three-gate four-address words in (Z/2)^3 have total rank at least 5 - #36
Open
cursor[bot] wants to merge 1 commit into
Open
cursor[bot] wants to merge 1 commit into
cursor[bot] wants to merge 1 commit into
Conversation
Every three-gate realization of four distinct nonzero addresses in (Z/2)^3 has total projector rank at least 5. Four rank-one gates have total rank 4, so the shorter unit-gate word is not a rank saving. Co-authored-by: riquelme.pereira <riquelme.pereira@amentoriaenem.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Change and mathematical scope
Scoped register-model floor, not a supplier or κ change. In
K[(Z/2)^3], for every four distinct nonzero addresses, every three-gate realization has total projector rank at least 5. Four rank-one coordinate gates realize the same target at total rank 4, so the shorter unit-gate word is not a rank saving. Ranks(2,1,1)do not occur.The rank-one count 4 and the general total-rank floor 4 are already theorems upstream (Wentao Li's
F2elementary-abelian diagonal minimum, and the prime-exponent theorem in CrocSwap PR CrocSwap#6008). Those packages are not in this checkout and were not replayed. The new comparison is the exclusion of total rank 4 for three-gate words.No selected-frontier edit.
make frontier1094-verify, Lean, and predecessor suites were not run.Task and overlap
7b73a20a335b7b7d2cb783e03340dff2, trackbit-mechanisms, cardbit-new-register-model, scope key7f5f4426ef796683c485.ef5f39403610d231a0c27e8a6b4f08c10209bcd6. Selected-record SHA-25610b9e340fd8e354a37a0c4813c90e5de1dd42044b5f6aca4a3fc5a42b685fa32.(Z/2)^3replace the rational three-register frame model. The changed question is the total-rank gap between a unit-minimum word and the direct rank-one word.Resource not accessible by integrationand was not retried.Validation
CPython 3.12, standard library only.
PYTHONOPTIMIZEwas unset for the first run; the second run passes-O.The checker exhausts all 35 quadruples, all 168 ordered bases and all 512 label triples, rebuilds both words in bit-mask, frozenset, and nonnegative-integer algebras, and rejects omitted gates and cleared projector bits. The integer engine shows the words do not use characteristic-2 cancellation. The rank inequality
rank(BNA) <= rank Nis the written argument; the script checks the counting lemma and the explicit words, not a search over free scalar maps.Credit
Ali Romman for the three-register rational bound and for the three-label unit/rank separation (PR CrocSwap#5394). Wentao Li for the
F2[(C2)^k]rank-one count. The prime-exponent total-rank floor is the result recorded with PR CrocSwap#6008. eumemic's PR CrocSwap#288 is the pinned source of the rational register model.Substantial AI assistance was used for the argument, the verifier, and the checks. Apache-2.0.
Open next questions
Does
A4still force rank-one cost four for four distinct non-identity addresses? It has no homomorphism ontoZ/2, so this character cut is blind, and its exponent is not prime, so PR CrocSwap#6008 does not cover it. A length-three rank-one word would be real compression. Free scalar maps are infinite, so the search needs a normal form or aC3coefficient cut before it is a proof.Workflow feedback
The useful overlap was a merged prime-exponent theorem that does not appear in open PR bodies. Searching default-branch package READMEs, not only open PRs and discussions, would have surfaced it earlier.