Repository navigation
Remove 60 native maps and 42 more ADDs after PR5350; kappa unchanged - #5403
Merged
Mister-Meeseeks merged 2 commits intoOct 24, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This exact circuit improvement removes 60 native-map records and 42 additional scalar ADD records per native call after PR #5350. The native-map count falls from 240 to 180. Scalar allowance falls by another 236 per call: forward 156904 → 156668, backward 156952 → 156716. These are circuit counts, not measured running times.
Kappa is unchanged. Every paid Gaussian record and all 408 independent helper boundaries are retained. The complete recursive moment and its binding main-transform cost are unchanged. The selected record is not edited.
Proof, deterministic result and event patch, source pins, and validation.
The pass tracks rational scalar factors and formal native-map words before comparing ADDs. It combines exact equal operators across commuting instructions and removes native maps that become an unused inverse pair. Source writes, target reads and resets remain barriers. The generated word has 193 fewer scalar records and 60 fewer native maps than the original final PR4610 word; the earlier 151 scalar deletions remain predecessor credit.
The base and all consumed source files are pinned to canonical main
1ec1119972c9aefe42525172d07b332fcff81cb4. The source manifest includes the exact constructor and both retained native interpreters. No predecessor file or frozen archive is changed.Run with Python 3.12 and the standard library:
The scientific command reports:
Passed locally: registered replay, isolated optimized replay, both retained interpreters on all 632 arbitrary persistent columns in both orientations, owned return, exact frame checks, complete local invoices, actual scalar-sign mutation rejection, committed-byte and export checks, and a fresh Git-free optimized replay with PATH=/nonexistent. The final commit adds validation text only. Hosted CI is pending at submission.
Existing absolute-phase recipes are reused; no new paid edge or monomial operator is introduced. The full phase suite, cold selected replay, complete controller/compiler/BIT/outer ledger and formal kernel were not rerun. The five selected framework assumptions remain; independent human review is open.
Credit Cao Yifan's PR #4610 native word, Cinna Davis's PR #4692 rewrites, ahrazzle's PR #5350 composition, and Chrisbryan17's earlier PR #5265 inverse-pair check. The retained contributor map records the native/helper/compiler lineage, including Dugongue, zxv and Jacob Alan Sussman. Both native interpreters are reused implementations, not new independent checkers. New work by Ali Romman with substantial OpenAI Codex assistance; Apache-2.0.