fma32

[Vutshi]

Hi,

Thank you for the great tool. I am curious whether it is difficult to add support of FMA operation for float numbers additionally to the current double implementation. It would be really helpful for testing and debugging various numerical algorithms in plenty of software using float due to speed reasons.

Cheers

[nschloe]

The new pyfma version also supports np.float32 and np.float128 arrays (float, long double in C).