Maybe there are some performance and precision benefits to be gained from implementing `core-math`'s MIT licensed algorithms? https://gitlab.inria.fr/core-math/core-math/-/tree/master/src/binary32
Maybe there are some performance and precision benefits to be gained from implementing
core-math's MIT licensed algorithms?https://gitlab.inria.fr/core-math/core-math/-/tree/master/src/binary32