This library heavily relies on unsafe and nightly features to achieve the best performance. The primary use case for this library is in games or graphics development, where speed matters more than precision
When using SIMD functions in this package, compile with LTO and opt-level=3 to ensure that auto-vectorization takes place. All SIMD functions have a feature cap at AVX2, and nothing in this library utilizes anything from AVX512. If certain functions vectorize on lower requirements, that's a bonus.
Most of the functions in here are faster than equivalent functions in sleef, at the expense of safety. (TODO: add comparison to sleef_rs)
Approximate f32 Functions:
Allows setting a variable precision level as a const generic.
sinandcos- Does include wrapping with a range reduction, but will become less accurate as the input gets larger.
log(constbase and variable base)- This is based on a log base 2 approximation, and is scaled for other bases. The fastest version of this is
constbase 2.0.
- This is based on a log base 2 approximation, and is scaled for other bases. The fastest version of this is
Exact Unsigned Integer Functions:
ilogwithconstbase- Has multiple implementations depending on the base to achieve maximum performance. The fastest impl is for log base 2.
expwithconstcoefficient- Similar to previous function, but calculates
COEFF^xinstead.
- Similar to previous function, but calculates
Any help on the library is greatly appreciated. If you'd like to contribute, just submit a PR and I'll respond to it as soon as I can. For development, I'd recommend looking at the genertated assembly often. For development of individual functions, I would recommend using a tool like Compiler Explorer, and using something like llvm-mca (available in CE under "tools") often to get an idea of performance on different platforms.