https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120209
--- Comment #2 from Paul Cercueil <paul at crapouillou dot net> ---
> The test case is somewhat bogus. Changing the "-1.0f" to "1.0f" will
> generate the "fldi1" instruction for the comparison as well.
Yes, I got confused while writing the bug report. The problem is loading the
-1.0f constant.
> Based on our previous discussion, I guess you'd expect this case to generate
> the sequence "fldi1; fneg" to load the constant -1.0, which would be better
> in this case.
>
>
> > The compiler also does not seem to understand that fr1 contains -1.0f which
> > it can add to fr0 directly, and instead it will load 1.0f with fldi1 this
> > time and substract it to fr0.
>
> That's because there are two constants in your case "-1" and "1".
>
> If the test case is changed to
>
> if (x < -2.0f)
> x2 *= -2.0f;
>
> ... then we will see that constants are shared in registers.
Well, but I'm doing
if (x < -1.0f)
x2 += -1.0f;
So that's only one constant, -1.0.
