{% extends "base.html" %} {% block title %}Performance Benchmarks - HYDRA-24 Demo{% endblock %} {% block content %}
Real benchmarks proving HYDRA-24 is production-ready
HYDRA-24 adds <5% overhead across all dataset sizes
| Dataset Size | Baseline (ms) | With HYDRA-24 (ms) | Overhead |
|---|---|---|---|
| {{ item.dataset_size }} | {{ item.baseline_ms }} | {{ item.hydra_ms }} | {{ item.overhead_pct }}% |
24-bit fingerprints = 3 bytes per value (30x more compact)
| DAG Nodes | HYDRA-24 (MB) | Full Lineage (MB) | Savings |
|---|---|---|---|
| {{ item.dag_nodes }} | {{ item.hydra_mb }} | {{ item.full_lineage_mb }} | {{ ((1 - item.hydra_mb / item.full_lineage_mb) * 100) | round(0) | int }}% |
DAG reconstruction in milliseconds, even for deep graphs
| DAG Depth | Reconstruction Time (ms) |
|---|---|
| {{ item.dag_depth }} layers | {{ item.reconstruction_ms }} ms |
Proven at production scale
Why HYDRA-24 beats alternatives