Skip to main content

Benchmark

v0.3 against v0.2

Same corpus, same machine, five runs each. The median of each run set is reported.

Results

v0.3 vs v0.23 improved, 1 regressed, 1 unchanged

v0.3 compared with v0.2
Metricv0.2v0.3ChangeVerdict
p95 compile latencylower is better420 ms310 ms−110 ms−26.2%improved
Throughputhigher is better1,200 pages/s1,480 pages/s+280 pages/s+23.3%improved
Output sizelower is better70.2 kB71.9 kB+1.7 kB+2.4%regressed
Peak memorylower is better48.2 MB48.5 MB+0.3 MB+0.6%unchanged
Cache hit ratehigher is better81%87%+6%+7.4%improved
Method
Median of 5 runs over the 11 gallery fixtures, warm process, Node 22.
Limitations
One machine and one corpus; pages with adapter diagrams were excluded.

Where the time goes

Compile time per phase (v0.3)

Total41 ms

  1. Parse and validate9 ms22%
  2. Normalize6 ms15%
  3. Render blocks14 ms34%
  4. Assemble CSS7 ms17%
  5. Verify document5 ms12%

Output budget

Budget80 kB

  1. Stylesheet31 kB39%
  2. Embedded font24.5 kB31%
  3. Markup12.4 kB15%
  4. Runtime4 kB5%

Annotated capture

The compiler health dashboard with four stats and a bar chart
Dashboard fixture rendered by v0.3 at 1200×750.
  1. Compile time The stat counts up on arrival and settles on the authored text.

  2. Legacy bars Each bar carries a hover value label.

  3. AK Render bars

  4. Theme toggle Present on every page by default.

References

  1. Benchmark notes for v0.3

    docs/artifacts · 2026

    Raw timings and the machine profile.

  2. Render benchmark script

    AK Render maintainers

  3. Web Vitals

    Google · web.dev · 2024

    Why p95 latency is reported rather than the mean.