[Bug libquadmath/105101] incorrect rounding for sqrtq

[already5chosen at yahoo dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105101

--- Comment #11 from Michael_S <already5chosen at yahoo dot com> ---
(In reply to Michael_S from comment #10)
> BTW, the same ideas as in the code above could improve speed of division
> operation (on modern 64-bit HW) by factor of 3 (on Intel) or 2 (on AMD).

Did it.
On Intel it's even better than anticipated. 5x speedup on Haswell and Skylake,
4.5x on Ivy Bridge.
Unfortunately, right now I have no connection to my Zen3 test system, so can't
measure on it with final variant. But according to preliminary tests the
speedup should be slightly better than 2x.