A rust translation of [DirectXMath], a SIMD linear algebra library for use in games and graphics apps.
âœ”ï¸ = Implemented, ⌠= Not yet implemented, 🚧 = In progress
| API Feature | Implemented | | --- | ------ | | Vector | âœ”ï¸ | | Vector 2D | âœ”ï¸ | | Vector 3D | âœ”ï¸ | | Vector 4D | âœ”ï¸ | | Quaternion | âœ”ï¸ | | Matrix | âœ”ï¸ | | Streaming Transforms | ⌠| | Color | ⌠| | Load/Store/Accessors | âœ”ï¸ † | | Plane | âœ”ï¸ | | Collision | âœ”ï¸ | | Documentation | âœ”ï¸ †† |
† Some alternate forms of the load/store operations are not yet implemented.
†† Most documentation has been ported and all functions have links to the official [DirectXMath documentation].
| x86 / x86-64 | ARM | No intrinsics fallback | | ------------ | --- | ---------------------- | | âœ”ï¸ | ⌠(will fallback to no-intrinsics) | âœ”ï¸ |
Intrinsics may be disabled with the no_intrinsics feature. This is the implicit behavior
if SIMD intrinsics are not available on the target architecture.
Note that only sse and sse2 are enabled by default for x86 and x86-64 targets.
The sse3, sse4.1, avx, fma, and avx2 intrinsics may be enabled via [CPU target features].
x86/64 intrinsicsRUSTFLAGS="-C target-feature=+sse3,+sse4.1,+avx,+fma,+avx2" cargo build