Expand description
Safe wrappers for unaligned SIMD load and store operations.
§Overview
The goal of this crate is to remove the need for “unnecessary unsafe” code
when using vector intrinsics with no alignment requirements.
Platform-intrinsics that take raw pointers have been wrapped in functions that receive Rust reference types as arguments.
§Implemented Intrinsics
§x86, x86_64
sse,sse2,avx
Currently, there is no plan to implement gather/scatter or masked load/store intrinsics for this platform.
§Other platforms
Not yet supported.
Modules§
- x86_64
- Platform-specific intrinsics for
x86_64.