Expand description
Vectorized feature calculation for high-performance analysis.
Re-exports§
pub use VectorizedFeatures64 as DefaultVectorizedFeatures;
Structs§
- Price
Features - Batch results for price-based ML features
- Vectorized
Features - SIMD-optimized feature calculator with cache-aligned memory buffers for HFT applications with const generic capacity
- Volume
Features - Batch results for volume-based ML features calculated via SIMD
- Weighted
Features - Batch results for weighted ML features calculated via SIMD.
Type Aliases§
- Vectorized
Features32 - Type alias for a vectorized feature calculator with 32-element capacity. Memory usage: ~768 bytes (3 buffers × 32 elements × 8 bytes), heap allocated. Best for: Simple market making, latency-critical applications. Performance note: Future optimization target for stack allocation.
- Vectorized
Features64 - Type alias for a vectorized feature calculator with 64-element capacity. Memory usage: ~1.5 KB (3 buffers × 64 elements × 8 bytes), heap allocated. Best for: Standard HFT strategies with moderate order book depth analysis.
- Vectorized
Features128 - Type alias for a vectorized feature calculator with 128-element capacity. Memory usage: ~3 KB (3 buffers × 128 elements × 8 bytes), heap allocated. Best for: Deep order book analysis, research applications.