[Bug math/32411] THREEp96 seems wrong

[jakub at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=32411

--- Comment #1 from Jakub Jelinek <jakub at redhat dot com> ---
It has been over 25 years, so I don't remember any details.
I agree that likely it should be what the comment says rather than what is
actually used, but I think before changing it one should check old vs. new
version on say 10 millions of random values in between lowmark and himark and
compute epsilon vs. mpfr computed values.

-- 
You are receiving this mail because:
You are on the CC list for the bug.