Efficient Execution
Algovex uses an array-based execution model to process large datasets efficiently. The engine grows a single rolling window per series to keep per-bar work linear.
Array-Based Calculation
Indicators operate over arrays of bars rather than per-bar loops, which keeps the hot path predictable and easy to reason about.
Batch Calculation • Predictable Hot PathIncremental Windows
Scan-all calculators persist their state and advance one bar per step, turning repeated full re-scans into linear work.
O(n) Scan • Cached StateData Ingest Pipeline
Loading large historical datasets (1-minute bars, up to 25 years) requires an efficient pipeline. Algovex caches candle data so repeat loads are fast.
Historical candles are cached so repeat loads of large datasets are instant.
Up to 25 years of 1-minute bars (Pro) load from the cache instead of being fetched again.
Engine Core
HIGH-PERFORMANCE ENGINE
O(n)
Linear Scan
Per-Bar Work Stays Linear
By caching scan state and advancing one bar per step, repeated full re-scans collapse into linear work, so long histories load quickly.
Factor Matrix Determinism
Deterministic
Reproducible Output
O(n)
Per-Bar Cost
High-Performance
Engine