EXP-057NEGATIVE
Canonical GPU engineering / SHA-256d

External-kernel launch geometry autotune

A different threads-per-block and nonces-per-thread geometry can improve the frozen external kernel 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
ReproducibilityRebuild all nine binaries, preserve the upstream arithmetic, repeat the balanced four-header matrix and enforce the 1% gate.
01 / Question & hypothesis

What was tested?

A different threads-per-block and nonces-per-thread geometry can improve the frozen external kernel by at least 1%.

02 / Scientific basis

Why the test is meaningful

Geometry changes can alter occupancy, scheduling and amortization without changing SHA semantics; selection must still clear a preregistered effect gate on every dataset.

G(T,N)=geomean throughput(T,N)/throughput(512,32)promote iff G≥1.01 and datasets≥4/4hash(candidate)=hash(baseline)
External-kernel launch geometry autotuneVisual reading of the published metrics and gates for EXP-057; it summarizes the registered result, not a mining advantage.EXP-057 / EXTERNAL-KERNEL LAUNCH GEOMETRY AUTOTUNEBEST OBSERVED RATIO1.002607DATASETS ≥1%0 / 4SPILLS0
FIGURE / RESULT READINGVisual reading of the published metrics and gates for EXP-057; it summarizes the registered result, not a mining advantage.
03 / Method

How it was tested

Test TPB 128/256/512 crossed with 16/32/64 nonces per thread on four fresh headers, two repeats per combination and 72 measured runs.

04 / Observed result

What happened

Best observed ratio1.002607
Datasets ≥1%0 / 4
Spills0

T256-N16 was best at 6.348510 GH/s versus 6.332001 GH/s, ratio 1.002607. It failed the 1.01 gate and improved by at least 1% on 0/4 datasets. All nine binaries verified exact B32 nonces with zero discrepancies and zero spills.

05 / Validation

Exactness and statistical controls

An initial runner-only failure was archived before measurement; the unused optional field read was removed and the whole campaign was resealed and repeated.

06 / Interpretation

What the result means

The small maximum is compatible with tuning noise. T512-N32 remains the baseline.

Limitations

  • One GPU architecture was tested.
  • The search covers only nine registered geometries.
  • A sub-gate maximum is not promoted.
07 / Reproduction

Evidence trail

Rebuild all nine binaries, preserve the upstream arithmetic, repeat the balanced four-header matrix and enforce the 1% gate.

Canonical variants

CANONICAL-EXP-133PREREGISTERED-CAMPAIGNSEALED-AUDIT

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