Full workflow runtime
Full workflow runtime describes the end-to-end execution path of the current correctness-first implementation. It preserves reproducibility, traceability, and audit value before aggressive optimization is introduced.
orrectness-first computational framework
MAYAN ALFA
MAYAN ALFA is an independent computational observation framework focused on correctness-first measurement, reproducible validation, and structured archive continuity.
The current implementation focuses on ARM64 benchmarking and prime-counting observation workflows. Wall-clock runtime, clean-binary timing, and controlled archive evidence remain separate methodological layers.
Measurement Layers
Why MAYAN ALFA separates runtime, timing, and evidence
Clean-binary timing
Clean-binary timing is reported as a directional throughput signal with detailed instrumentation disabled. It is methodologically separate from full workflow runtime and must not be merged into a single speed claim.
Instrumentation horizon
Deeper instrumentation will be introduced after the correctness-first phase is stabilized and optimization becomes the active development priority.
Interpretation rule
Limit, step, execution mode, and archive tier are always reported together. Each number is therefore read as a bounded observation within the framework, not as an isolated performance slogan.
Correctness-first release
The current release preserves computational accuracy, reproducible output, and traceable evidence before optimization speed.
Directional timing layer
Clean-binary timing remains a directional throughput signal until deeper instrumentation makes the performance layer fully auditable.
Public release boundary
The public presentation layer is bounded at 10B. 100B, 200B, and 500B remain preserved as controlled archive evidence.
Measurement board
P0.7 Canonical Measurement Layer
Measurement families
Evidence map
Limit
Layer
MAYAN
PRIMESIEVE
10M
public
measured
measured
100M
public
measured
measured
1B
public
measured
measured
10B
public boundary
measured
measured
100B
controlled
archive evidence
archive evidence
200B
controlled
archive evidence
archive evidence
500B
controlled reference
workflow reference
workflow reference
Measurement reading
Public interpretation ruleFull workflow runtime, clean-binary timing, and controlled archive evidence are separate methodological layers. Each must be read in its own context before any performance claim is made.
The framework keeps the scientific reading explicit: correctness-first execution, directional binary timing, and controlled archive evidence are not merged into a single performance number.
Evidence integrity
Evidence archive, reference packages, and release boundary
MAYAN ALFA GITHUB PUBLIC RELEASE.zip SHA-256
sha256:264b797b04e01695772f090881edb1cf2bab30aa73ab74d1575c8d4985926e64
MAYAN ALFA PAPER WORKPACK.zip SHA-256
sha256:b301c0f03e8a31349cf2fa06ce0382dc7509929647f24cdb77990425e0600a03
MAYAN ALFA PUBLIC QA.zip SHA-256
sha256:3695cd2ffc046dcd32ac8e2fe3e68df84972002c6e60804e06f86d5d6cbfa178
MAYAN ALFA PUBLIC RELEASE.zip SHA-256
sha256:18518792898ecceb13f2935660e4756efe41169d48c08b3f7c61774709e77889
These packages preserve the current public evidence set used for measurement review and release validation.
Release boundary
The public archive is derived from the validated MAYAN_ALFA release boundary.
Internal development paths, local machine identifiers, and protected infrastructure references are intentionally excluded from public distribution.
Zenodo DOI
Version DOI — v0.7:
https://doi.org/
Concept DOI — all versions:
https://doi.org/
Public release boundary sanitized for GitHub and Zenodo preservation.
Archive architecture
The archive separates public presentation, release hosting, long-term preservation, and validation depth into distinct evidence layers.
Public anchors
- PUBLIC Open presentation layer for public-facing framework material.
- GitHub Versioned repository for public code, release packages, and outputs.
- Zenodo DOI-based preservation layer for versioned public releases.
Evidence tiers
- BRONZE Initial validation layer for early evidence staging.
- SILVER Validated public evidence layer up to 10B.
- GOLD Controlled evidence layer from 100B to 500B.
- PLATINUM Future scaling layer reserved for 1T+ to 10T+ observations.
Framework memory
Research guide archive
Foundation
Vision, Identity, and Release Foundation
Release discipline
Public release rules
01
Preserve the audit trail
Historical runs, raw logs, and release evidence remain part of the traceable record.
02
Bound the public layer
10M-10B remains the public measurement range; 100B+ remains controlled archive evidence.
03
Protect the core
Public releases preserve measurement evidence while protected core logic stays outside distribution.
04
Stage optimization later
Deeper instrumentation and speed tuning follow after the correctness-first phase is stabilized.
