We validate simulator accuracy against public benchmark references. Detailed raw numbers are kept as internal validation assets; this page only reports aggregate stats. Last validated 2026-05-13.
Each reference point is reproduced by configuring the simulator with the same hardware envelope (cores / clock / RAM / disk / network) and comparing measured throughput.
An internal script automates environment mapping + ramp-up traffic + peak measurement. Anything within ±25% counts as a pass. The mean absolute error level is sufficient for architectural decision support.
Limits: Queue / pool effects and dynamic backpressure are modeled in steady-state. Spike loads may have an extra ±10% error. Pair simulator results with load testing for actual production decisions.
Detailed reference sources / raw numbers / environment configs are kept private as internal validation assets. For education / research data sharing requests, email us.
Click the confidence button in the simulator toolbar to see the same aggregate stats inline.
Open Simulator →