EXP-061NEGATIVE
Canonical GPU engineering / SHA-256d

Gray-order σ₀ reuse over nonce blocks

Traversing aligned blocks of 32 nonces in Gray order can reuse σ₀ through XOR linearity and improve throughput by at least 1%.

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
HardwareNVIDIA GeForce RTX 5080 · exact SHA-256d verifier
Run scopePreregistered campaign with sealed result and audit
ReproducibilityAudit bijection, equal hash coverage and exact hits, then repeat balanced timing with the frozen binaries.
01 / Question & hypothesis

What was tested?

Traversing aligned blocks of 32 nonces in Gray order can reuse σ₀ through XOR linearity and improve throughput by at least 1%.

02 / Scientific basis

Why the test is meaningful

The identity σ₀(base XOR gray)=σ₀(base) XOR σ₀(gray) is exact, but materializing the traversal may consume its theoretical savings.

σ₀(x⊕δ)=σ₀(x)⊕σ₀(δ)coverage must be bijectivegate: geometric speedup≥1.01
Gray-order σ₀ reuse over nonce blocksVisual reading of the published metrics and gates for EXP-061; it summarizes the registered result, not a mining advantage.EXP-061 / GRAY-ORDER Σ₀ REUSE OVER NONCE BLOCKSKERNEL RATIO1.000626SASS EXPANSION3.95×GATEFAILED
FIGURE / RESULT READINGVisual reading of the published metrics and gates for EXP-061; it summarizes the registered result, not a mining advantage.
03 / Method

How it was tested

Compare an unrolled Gray traversal with the frozen external baseline on four unseen headers and 32 balanced processes, auditing 393,216 nonces for bijection.

04 / Observed result

What happened

Kernel ratio1.000626
SASS expansion3.95×
GateFAILED

Coverage was bijective and hashes exact. The candidate won 4/4 headers but only by 1.000626× with 95% interval 0.999949–1.001304. SASS expanded from 2,507 to 9,895 instructions.

05 / Validation

Exactness and statistical controls

A valid B32 hit was recovered by both variants with zero discrepancies and spills. The interval crosses one and the registered 1% gate failed.

06 / Interpretation

What the result means

The identity is exact, but this unrolled implementation trades away the saving through code expansion.

Limitations

  • The result applies to the unrolled implementation.
  • A compact recurrence requires a separate test.
  • A directional 4/4 result below gate is not promoted.
07 / Reproduction

Evidence trail

Audit bijection, equal hash coverage and exact hits, then repeat balanced timing with the frozen binaries.

Canonical variants

CANONICAL-EXP-137PREREGISTERED-CAMPAIGNSEALED-AUDIT

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