PERRY BENCHMARK DEMO

No Garbage Collection

Memory is never automatically freed. Perry allocates but does not reclaim — ideal for short-lived processes. No GC pauses means predictable latency.

Tradeoff

No Runtime Type Checking

Types are fully erased at compile time. Perry compiles direct field offsets — no inline caches, no type guards. Faster dispatch, but no safety net at runtime.

Speed

Single-Threaded

No Worker threads or multi-threading support. All requests are serialized — concurrent benchmark requests return 503 (busy). Simplifies the execution model.

Limitation