Performance Benchmarks
Rust VDOM and server-rendering metrics for teams evaluating djust 1.0. Reproduce the numbers, then test them against your own Django views.
Template Rendering Speed
Template rendering, djust vs Django, measured with the harness in djust/benchmarks. The speedup depends entirely on what the template does: measured 1.0x on static markup up to ~11x on variable- and filter-heavy templates (djust's own README table).
Same workload, same machine: a 100-row filtered list. Both figures come from djust/benchmarks/benchmark.py — that is ~11x. We no longer print LiveView or Livewire numbers here: we have no harness for them, and inventing a baseline to compare against is the thing this page is meant to stop doing.
VDOM Diffing Performance
Time to diff and generate patches for a complex component (100 DOM nodes, 20% change rate). Lower is better.
Client Bundle Size
Minified and gzipped JavaScript sent to the client. Smaller bundles mean faster initial page loads.
Server Memory Usage
Memory usage per concurrent WebSocket connection. Lower is better for scaling.
Request Throughput
Requests per second handled on a single server (4 cores, 8GB RAM). Higher is better.
Methodology
What is measured today
- • djust-vs-Django template rendering (Python 3.12, Apple silicon laptop)
- • Harness: djust/benchmarks/benchmark.py — run it yourself
- • Template-suite conformance: scripts/run-django-template-suite.py
What is not (yet)
- • The LiveView / Livewire / HTMX rows above are order-of-magnitude estimates, not measurements from a published harness
- • Cross-framework WebSocket concurrency numbers have no published harness yet
Reproducibility
The djust-vs-Django numbers are reproducible from our benchmarks repository — run the harness against your own templates and workload; that is the number that matters.
Experience the Performance Yourself
See djust's speed in action. Explore examples or run benchmarks locally.