SIMD

Modules: simd/simd_math, simd/simd_types

SIMD modules define vector types and math helpers used by the compiler/runtime SIMD lowering paths.

Helper functions include simd_f4_*, simd_f8_*, simd_reduce_sum, simd_prefix_sum, simd_min, simd_max, and simd_dot_product. Native vector expressions such as f32x4(...) + f32x4(...) lower as LLVM vector values and do not allocate temporary heap objects. The SimdFloat4 and SimdFloat8 wrapper types remain as handle-based compatibility APIs; runtime also exposes non-allocating *_into entrypoints for codegen or FFI paths that can provide caller-owned storage.

Runtime array reductions for Array<Float> use the double-array storage used by Seen arrays, with AVX2 or NEON min/max/sum/dot/prefix paths where available and scalar fallbacks on unsupported targets. SIMD temporary storage in the compatibility wrappers is allocated through the same budget-aware aligned allocator used by the rest of the runtime.

See SIMD and GPU for usage and CLI controls.