Expand description
Vector type with multi-backend support
This module provides the core Vector<T> type with SIMD-optimized operations
across multiple backends (Scalar, SSE2, AVX2, AVX-512, NEON, WASM SIMD).
GPU thresholds intentionally set to usize::MAX to disable GPU for element-wise ops. See docs/performance-analysis.md - GPU is 2-65,000x SLOWER than scalar for these ops.
Modules§
- dispatch
- Backend dispatch macros for vector operations
Structs§
- Vector
- High-performance vector with multi-backend support