This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix multiply-add regressions after expand-from-SSA
Hi,
On Thu, 30 Apr 2009, Paolo Bonzini wrote:
> > Actually the helper could look like
>
> Couldn't this lead to wrong code for
>
> a_1 = b_1 * c_1;
> b_2 = b_1 + 1.0;
> c_3 = b_2 + a_1;
>
> where b_1 and b_2 could be assigned to the same partition and so the
> final result would be "b_2 + b_2 * c_1"?
TER already has to handle this case, and as he uses only the TER results
this should be safe. IOW: if that would be broken it would have been
broken since quite some time as TER did exactly the same transformations.
(And yes, TER works with the final partitions to detect that it can move
definitions forward)
Ciao,
Michael.