Repository navigation
bench(adr-001): witness audit — full residual vs closure-restricted - #43
Merged
Merged
Conversation
Empirical receipt for the witness primitive shipped in PR #41. Compares two ways to verify a SubLinear delta-solve output: - full_residual: ‖b - A·x_new‖_∞ over the whole n-vector. Linear. - closure_audit: verify_sparse_solution restricted to closure entries. SubLinear in n. Smoke results (cargo bench -- --quick witness_audit): n full_residual closure_audit 64 221 ns 549 ns 256 885 ns 602 ns 1024 3.57 µs 906 ns full_residual grows ~4x per 4x size (Linear). closure_audit stays nearly flat (549 → 602 → 906 ns across 16x growth in n). Crossover at n ≈ 200; by n=1024 the closure-restricted audit is ~4x faster and the gap widens linearly with n. This is the architectural match for the SubLinear orchestrator: auditing the output costs the same complexity class as the solve itself, so trust-but-verify gating in production stays cheap. Co-Authored-By: claude-flow <ruv@ruv.net>
1 of 2 tasks
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Empirical receipt for the witness primitive shipped in PR #41. Compares two ways to verify a SubLinear delta-solve output:
Smoke results (`cargo bench -- --quick witness_audit`)
Architectural property holds:
This is the architectural match for the SubLinear orchestrator: auditing the output costs the same complexity class as the solve itself, so trust-but-verify gating in production stays cheap.
Test plan
🤖 Generated with claude-flow