Repository navigation
Build a six-T field decoder and a complete 72-T four-address pilot - #268
Merged
Merged
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The four-address field route now has a six-T, three-band decoder, replacing the retained 63-T, 36-band full decoder. Its complete direct indicator uses 72 T gates, at most 28 bands and 350 total gates with three returned dirty wires; the earlier field-route word used 300 T gates and at most 160 bands. This is a finite construction, not an optimality claim or a change to R96.
The change constructs a genuinely non-Clifford flat bank, its complete grading, a 12-T six-band physical copy, and a Clifford completion to a full literal field intertwiner (65 forward / 63 inverse gates). The shorter bank decoder supplies the cheaper indicator. All identities retain literal phases, arbitrary principal/dirty/reference inputs, actual inverses and work return.
The accompanying bounds have separate scopes: this finite seed has one-band copy error at least 1/(2 sqrt(2)); arbitrary Clifford embeddings of product THT-dagger seeds have a width-independent 3/32 floor when a selected output dual is Clifford; unique largest Pauli coefficients give a coefficient-gap floor. A full intertwiner below error one also needs axis pairing rank at least m−2 on the query-fixed Pauli space. That last condition is not imposed on the weaker one-probe decoder.
Internal proof and circuit audits are complete. Fifteen new exact tests cover the native seed/decoder/copy, all 512 direct-indicator input columns in both directions, full intertwinement, grading, literal signed-support constraints, fingerprints and the scoped rank filter. Earlier distinct proofs, anchors and native pilots are retained.
Validation on the reviewed final tree:
presentation-previews, ID12817869160, is 344,565,702 bytes and bound to the reviewed head; SHA-256f1d3de4ffbc5c9b733b02fea423f8e49ec9944b4f872d1788461fc54146743fe.d715072abfe0e3e3ba7baf69068f9a54c9237edahas the reviewed tree and parents matching the unchanged base plus final head.Exact review record:
c07ab3a1d7d0ea6892403689eadff476f45b511c.c2a71c8c5132fcff61624e22ed4baf7ace4bc546.68f0ce342bf105e4b3aa3b1a7c5d76c3f579998b.c843d987963bc4be1df46832a8b681137b13ec493e5fac17e30604ba178f6f32.a10a5dd7dc5138f189dabbcc11232184f57941f9461bc8618a85811c543d91f8.1bcb5a69a9e12f80fde7bab4b1a97bcceb477a4c84f1f5e83044fd5ccf93498f.98bdd7ee113849a7b8722cfdb84093346cb5f66d3609d19eacfdee80710e81cd.81334aa43f376a77477902c558636d1dc2350c4123a8d39287b9d84bb5d54e92.47b007a86c0ca28e2dceec60a6088278df396d0905e4f14b513ed03687045ca8.b544cf15eb5a7ab7f04f4c94fcd596afcdd463c7fda95e8cb4e5ad6fcb7813db.The next construction is a uniform decoder within polynomial discovery budgets and sublogarithmic T depth, allowing the complete lookup probe's commutant freedom. No manuscript or general-frontier promotion is made.