Hi!
I'm using the atan2sp function from fastmath 2.01 atan2sp.asm.
Unfortunately this implementation has a bug. Consider the following inputs:
atan2sp(0,1) result: 0, correct
atan2sp(0,-1) result: 0, wrong, should be -PI
A workaround is for that bug is pretty easy, but wrapping each use of a optimized asm function in some more c code should be avoided.
Since i have no knowledge in the TexasInstruments Assembly Language, i'd appreciate a fix for atan2sp.
Kind Regards