HARA LANGUAGE · PERFORMANCE EVIDENCE

Benchmark
Observatory

Explore the result, then inspect exactly what produced it.

Loading benchmark evidence…

THE IMPORTANT NUMBERS

Prepared steady-state matrix

Rows are equivalent workloads. Columns are runtimes. Lower is better.

COMMON WORKLOAD SET

Runtime rankings

Geometric-mean slowdown against the fastest runtime per workload, with coverage kept visible.

WHERE TIME GOES

Preparation and execution

Lifecycle phases are separate. Cold totals are not presented as pure runtime startup.

NO HIDDEN IMPLEMENTATION

Exact tested code

Workload source, preparation, runner harness and correctness checksum.

PROVENANCE

Runs, versions and machine

Results only mean something when their environment is visible.