Re: Fw: Error in documentation for tan(), and poor implementation of sin() and cos()

> I'll now try the same method for double-extended and quadruple precision.

for double-extended I didn't find a larger error than the ones from [1].

For quadruple, I found a larger error only for tan:

NEW tan 0 -1 0x1.67f07e02ea57707a8c1f9cp+13000 [1] [1.47] 1.4662 1.466197744254779

So it seems the argument reduction is too lazy only for binary64.

Paul

[1] https://members.loria.fr/PZimmermann/papers/accuracy.pdf