EXP-035VALIDATED
Advantage Decomposition / SHA-256

K/P/G attribution at carry 20

The exact source of carry-21 regeneration can be separated into Generate at bit 20 and the product of Propagate with incoming carry, revealing which primitive restores x3 complexity.

Claim statusNO END-TO-END MINING ADVANTAGE DEMONSTRATED

Diagnostic work, exact algebra and local capabilities are not treated as end-to-end mining advantage.

Date2026-08-14
HardwareClassical exact algebraic verifier
Run scopeThree canonical variants
ReproducibilityDerive G20 and P20 from the exact operands, reconstruct c21 for every lane, verify equality with the direct oracle and repeat x3 attribution on the untouched fresh cohort.
01 / Question & hypothesis

What was tested?

The exact source of carry-21 regeneration can be separated into Generate at bit 20 and the product of Propagate with incoming carry, revealing which primitive restores x3 complexity.

02 / Scientific basis

Why the test is meaningful

The K/P/G identity is a causal decomposition of a carry bit. G20 creates carry independently; P20·c20 transmits prior complexity. Comparing these exact components is more informative than correlating only the final carry mask.

c21 = G20 ∨ (P20 ∧ c20)G20=u20∧v20P20=u20⊕v20
K/P/G attribution at carry 20Visual reading of the published metrics and gates for EXP-035; it summarizes the registered result, not a mining advantage.EXP-035 / K/P/G ATTRIBUTION AT CARRY 20K/P/G RECOMPOSITIONEXACTFRESH RECORDS300MINING CLAIMNO
FIGURE / RESULT READINGVisual reading of the published metrics and gates for EXP-035; it summarizes the registered result, not a mining advantage.
03 / Method

How it was tested

Compute exact masks for G20, P20, c20 and their product, attribute x3-containing monomials at each stage and repeat fixed survivor/non-survivor summaries on original and fresh cohorts.

04 / Observed result

What happened

K/P/G recompositionEXACT
Fresh records300
Mining claimNO

K/P/G recomposition was exact with zero semantic errors for 1,361 original and 300 fresh records. The atlas localized regeneration between the direct-generate and propagated-carry products but did not yield an independent selector or mining advantage.

05 / Validation

Exactness and statistical controls

The OR/XOR-equivalent carry truth table was checked lane by lane against direct addition. Original and fresh reports preserve every component, preventing post-hoc selection of whichever term looked favorable.

06 / Interpretation

What the result means

V35 identifies the arithmetic channel by which canceled structure can return. It supports mechanistic understanding and subsequent consumer tests, not faster hashing.

Limitations

  • Component attribution does not prove controllability.
  • Group means may vary with the registered context distribution.
  • The cost of computing K/P/G masks was not compared with native SHA-256.
07 / Reproduction

Evidence trail

Derive G20 and P20 from the exact operands, reconstruct c21 for every lane, verify equality with the direct oracle and repeat x3 attribution on the untouched fresh cohort.

Canonical variants

SUBENGINE-V35ASUBENGINE-V35BSUBENGINE-V35C

Source: internally audited canonical reports. Local filesystem structure, private headers and operational identifiers are excluded from publication.