This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Third Draft "Unsafe fp optimizations" project description.



> <<My point is that gcc should feel free to apply transformations that
> do not change results in IEEE arithmetic, even if they
> might change the results in some other arithmetic (old Cray
> arithmetic, say, where x / 2.0 may not have been equivalent to x *
> 0.5).
> >>
>
> I am not so sure that's acceptable, especially when we have quite a
> few pseudo-IEEE machines around.

Currently this particular transformation 1.0/2.0 -> 0.5 is
checked for exactness in gcc's idea of the target's arithmetic and
it is further filtered by a machine-dependent function check_float_value.

We do not in general try to be bug compatible with all target
arithmetics, but I'm sure no one would mind if you wanted to put
in the effort to implement that.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]