Re: TomsFastMath v0.10 released

Wolfgang Ehrhardt wrote:
I was not talking about invmod but gcd. To repeat: TFM uses a (usually
much slower) Euclidian gcd and not a binary like LTM. So you had to
code another algorithm and not rewrite the LTM gcd in terms of TFM,
and my question was if there was a specific reason - there are some
situations (cf eg GMP doc) where a gcd using mod is faster.

IIRC it was a hack. I've put it on my TODO list to revise for v0.11