Jimmy Myhrman
Jimmy Myhrman
Also chiming in to suggest an f32 option. While on most platforms performance of f32 and f64 math is similar (or even f64 being faster), it is also the case...
There is also incentive around making it generic over float type such that you can also go for [fastfloat](https://github.com/robsmith11/fastfloat) type that implements fast-math float operations (as often done in C/C++...
Curious if there has been any progress on mitigating this issue? It currently poses a risk to using basedrop in production code without a patch.
Excellent, thanks for the update!