[Bug middle-end/53073] [4.8 Regression] 464.h264ref in SPEC CPU 2006 miscompiled

[rguenther at suse dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53073

--- Comment #10 from rguenther at suse dot de <rguenther at suse dot de> 2013-01-29 09:10:25 UTC ---
On Mon, 28 Jan 2013, sje at gcc dot gnu.org wrote:

> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53073
> 
> Steve Ellcey <sje at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |sje at gcc dot gnu.org
> 
> --- Comment #9 from Steve Ellcey <sje at gcc dot gnu.org> 2013-01-28 18:50:42 UTC ---
> FYI: I reported this issue to SPEC to make sure they were aware of it.
> their reply is at:
> 
> http://www.spec.org/cpu2006/Docs/faq.html#Run.05
> 
> They do not intend to offer alternative sources.

So be it.  At a last resort we could add a switch that disables
just number-of-iteration computations based on undefined behavior.
Note that all previous releases of GCC have the same behavior - the
knowledge is just not used.  Thus such switch would pessimize
code further than reverting to previous behavior - nevertheless
such switch would be consistent with existing switches like
-fno-strict-aliasing or -fno-strict-overflow.

Richard.