Implementation of basic math data types with high level frontend and low level backend
LLML is currently supported on x86/x8664 and ARM/aarch64, with plans for _WASM support and a naive implementation, available for all targets, in the future.
As of today, LLML supports up to SSE3 for x86/x86_64, and Neon for ARM/aarch64. Support for AVX and AVX2 is expected in the near future, as an optional feature.
f32
and f64
expi
)sqrtc
)powc
) and power by imaginary (powci
)f32
and f64
sum
), norm and unit vectordot
) and cross product (cross
)sqrt
) and fast square root (sqrt_fast
, only faster in f32
x86/and x86_64)f32
and f64
tr
) and determinant (det
)inv
) and unsafe (inv_unsafe
) inversetransp
)Implements random generation for all of the crate's data types with rand
Implements serialization and deserialization for all of the crate's data types with serde