[Bug tree-optimization/77644] missed optimization with sqrt in comparison

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77644

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
Err, abs(A) < sqrt (B) isn't handled yet, no?  Though I'm not sure if it's wise
to emit A * A < B for that given A * A is going to drop of quite some bits in
precision and likeliness to overflow / underflow to +Inf / denormal is high.