Wall-clock runtime
Wall-clock results describe the end-to-end runtime of the present workflow. They can remain long because computational correctness, reproducibility, and traceability still have priority over aggressive optimization.
.mayan-hero {
display: grid;
grid-template-columns: minmax(0, 1.06fr) minmax(280px, 0.94fr);
align-items: stretch;
gap: 26px;
min-height: min(720px, calc(100vh - 132px));
margin: 8px 0 14px;
padding: 64px 40px;
position: relative;
overflow: hidden;
isolation: isolate;
border: 1px solid rgba(255, 255, 255, 0.08);
border-radius: 22px;
background:
linear-gradient(135deg, rgba(47, 111, 159, 0.16), transparent 38%),
linear-gradient(315deg, rgba(31, 138, 112, 0.14), transparent 42%),
linear-gradient(135deg, #020817 0%, #06162d 54%, #020817 100%);
box-shadow:
0 24px 80px rgba(2, 8, 23, 0.38),
inset 0 0 0 1px rgba(255, 255, 255, 0.05);
}
.mayan-hero::before {
position: absolute;
inset: 18px;
z-index: -1;
content: "";
border: 1px solid rgba(255, 255, 255, 0.08);
background-image:
linear-gradient(rgba(255, 255, 255, 0.05) 1px, transparent 1px),
linear-gradient(90deg, rgba(255, 255, 255, 0.05) 1px, transparent 1px);
background-size: 42px 42px;
mask-image: linear-gradient(90deg, black, transparent 84%);
}
.mayan-hero::after {
position: absolute;
inset: 0;
z-index: -1;
content: "";
background: linear-gradient(
115deg,
transparent 0,
transparent 36%,
rgba(255, 255, 255, 0.05) 50%,
transparent 64%,
transparent 100%
);
opacity: 0.6;
pointer-events: none;
}
.mayan-hero-copy {
display: flex;
flex-direction: column;
justify-content: center;
min-width: 0;
}
.mayan-hero .eyebrow {
color: rgba(32, 214, 199, 0.92);
}
.mayan-hero h1 {
max-width: 10ch;
margin-bottom: 20px;
color: var(--surface-strong);
font-size: clamp(3.8rem, 8vw, 7rem);
line-height: 0.93;
letter-spacing: 0;
}
.mayan-hero .lead {
max-width: 780px;
color: rgba(255, 255, 255, 0.76);
font-size: clamp(1.02rem, 1.6vw, 1.2rem);
line-height: 1.62;
}
.mayan-hero .lead + .lead {
margin-top: 12px;
}
.mayan-hero .action-row {
margin-top: 26px;
}
.mayan-hero .button.primary {
border-color: transparent;
background: linear-gradient(135deg, #2f6f9f, #20d6c7);
color: var(--surface-strong);
box-shadow: 0 16px 32px rgba(32, 214, 199, 0.2);
}
.mayan-hero .button.secondary {
border-color: rgba(255, 255, 255, 0.16);
background: rgba(255, 255, 255, 0.06);
color: var(--surface-strong);
}
.mayan-hero .button:hover,
.mayan-hero .button:focus-visible {
transform: translateY(-2px);
}
.mayan-hero-aside {
display: grid;
align-content: center;
gap: 14px;
min-width: 0;
}
.summary-card {
min-width: 0;
padding: 18px 18px 20px;
border: 1px solid rgba(255, 255, 255, 0.12);
border-radius: var(--radius);
background: rgba(255, 255, 255, 0.06);
box-shadow: 0 12px 30px rgba(2, 8, 23, 0.22);
color: var(--surface-strong);
}
.summary-label {
margin: 0;
color: rgba(255, 255, 255, 0.66);
font-size: 0.74rem;
font-weight: 800;
letter-spacing: 0.06em;
text-transform: uppercase;
}
.summary-card strong {
display: block;
margin: 18px 0 10px;
font-size: clamp(1.7rem, 4vw, 3rem);
line-height: 0.98;
}
.summary-text {
margin: 0;
color: rgba(255, 255, 255, 0.72);
line-height: 1.55;
}
.signal-panel {
.status-strip p,
.measurement-card p,
.measurement-interpretation p,
.data-panel p,
.measurement-grid {
display: grid;
grid-template-columns: repeat(2, minmax(0, 1fr));
gap: 12px;
margin-bottom: 16px;
}
.measurement-card {
min-width: 0;
min-height: 188px;
padding: 20px;
border: 1px solid rgba(29, 34, 40, 0.16);
border-radius: var(--radius);
background: rgba(255, 254, 250, 0.9);
box-shadow: 0 10px 30px rgba(29, 34, 40, 0.05);
}
.measurement-card h3 {
margin-bottom: 14px;
font-size: 1.08rem;
}
.measurement-card-accent {
background: #172028;
color: var(--surface-strong);
}
.measurement-card-accent p {
color: rgba(255, 255, 255, 0.72);
}
.measurement-interpretation {
margin-top: 4px;
}
.section-heading {
@media (max-width: 920px) {
.intro-section,
.mayan-hero,
.split-layout,
.intro-section {
.mayan-hero {
min-height: 0;
padding: 54px 30px;
}
.mayan-hero-aside {
grid-template-columns: 1fr;
}
.measurement-grid {
grid-template-columns: 1fr;
}
.metric-grid {
.mayan-hero {
padding: 44px 22px;
border-radius: 14px;
}
.mayan-hero::before {
inset: 14px;
background-size: 34px 34px;
}
.mayan-hero h1 {
font-size: clamp(3.1rem, 16vw, 5.4rem);
}
.mayan-hero .lead {
font-size: 1rem;
}
.status-strip,
.mayan-hero .action-row,
.mayan-hero .button {
width: 100%;
}
.measurement-card,
.summary-card {
min-height: 0;
}
.rule-row {
background-size: 42px 42px;
-webkit-mask-image: linear-gradient(90deg, black, transparent 84%);
mask-image: linear-gradient(90deg, black, transparent 84%);
display: grid;
grid-template-columns: 0.75fr 1fr 1fr 1.1fr;
grid-template-columns: 0.75fr 1fr 1.08fr 1.08fr;
gap: 1px;
.data-panel code {
overflow-wrap: anywhere;
word-break: break-word;
font-family: "SFMono-Regular", Consolas, "Liberation Mono", monospace;
font-size: 0.86rem;
line-height: 1.55;
}
.panel-heading {
Measurement-first research layer
MAYAN ALFA
Independent ARM64 benchmark and validation program focused on reproducible prime-counting observations and controlled archive interpretation.
The current release separates wall-clock runtime of the intentionally unoptimized workflow from clean-binary timing with detailed instrumentation disabled. Optimization and deeper instrumentation will follow in a later phase.
How we measure
Validation first. Performance claims only after complete canonical comparison.
Measurement protocol
Why the same project can report different runtimes
Clean-binary timing
Clean-binary timing is reported as an indicative throughput signal with detailed instrumentation disabled. It is methodologically different from wall-clock runtime and should not be collapsed into the same speed claim.
Instrumentation horizon
More detailed instrumentation will be introduced after the current correctness-focused phase is stabilized and optimization work becomes the active priority.
Interpretive rule
Limit, step, execution mode, and archive tier are always reported together. The numeric result is therefore read as a bounded observation, not as an isolated slogan.
Interpretive rule
Limit, step, execution mode, and archive tier are always reported together. The numeric result is therefore read as a bounded observation, not as an isolated slogan.
Correctness-first stage
The current release keeps computational accuracy and reproducible output ahead of optimization speed.
Binary timing still indicative
Clean-binary timing remains indicative until later instrumentation turns it into a fully auditable performance layer.
Publication boundary
The public presentation layer stops at 10B, while 100B, 200B, and 500B remain controlled archive evidence.
Benchmark board
P0.7 Canonical Binary Benchmark
Measurement families
Interpretation map
Limit
Tier
MAYAN
PRIMESIEVE
10M
public
measured
measured
100M
public
measured
measured
1B
public
measured
measured
10B
public ceiling
measured
measured
100B
controlled
archive evidence
archive evidence
200B
controlled
archive evidence
archive evidence
500B
controlled reference
wall-clock reference
wall-clock reference
Interpretation
Current public wordingWall-clock runtime, clean-binary timing, and controlled archive evidence are separate methodological objects. They must be read in their own context before any speed claim is made.
The site is written to keep the scientific reading explicit: correctness-first workflow, long wall-clock runtime, and indicative binary timing are not merged into a single number.
Archive integrity
Reference package and source boundary
TESTY.zip SHA-256
This bundle contains the current performance evidence used for the measurement analysis.
Source directory
The current source directory is the primary working boundary for measured data, validation outputs, and program sources.
Publication tiers
Public pages stop at 10B. 100B, 200B, and 500B remain controlled archive evidence and are included in the GOLD package of detailed overviews and measurements.
Project memory
Master guides
Základní strategie
Vize, identita a publikační základ
- 01_VIZE_A_IDENTITA_PROJEKTU_EXPANDED.docx
- 02_PUBLIC_PRIVATE_ARCHITEKTURA_EXPANDED.docx
- 03_GITHUB_PUBLIKACE_EXPANDED.docx
- 04_ZENODO_DOI_PUBLIKACE_EXPANDED.docx
Release discipline
Public release rules
Keep the audit trail
Historical runs and raw logs remain part of the traceable evidence.
Bound the public range
10M-10B is the public presentation layer; 100B+ stays controlled.
Protect the mathematical core
The benchmark layer records and packages results, but it does not change the computation itself.
Stage optimization later
Detailed instrumentation and speed tuning come after the correctness-first phase is stabilized.
