This is the mail archive of the gcc-patches@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]
Other format: [Raw text]

Re: [PING][PATCH][REVISED] Fix PR middle-end/PR28690, modify swap_commutative_operands_p


On Fri, Jul 13, 2007 at 12:35:42PM -0500, Peter Bergner wrote:
> On Fri, 2007-07-13 at 08:19 +0200, Paolo Bonzini wrote:
> > > If that doesn't help, then I suggest we add a target hook to
> > > the changes in commutative_operand_precedence so that it
> > > behaves the way it used to for everyone other than POWER6.
> > 
> > I think this is overkill.  The change makes sense on every architecture.
> 
> Ok, then given I already have approval from Ian to commit this, I'll
> commit this and we can work on resolving any further issues people might
> have after that.

I am waiting for SPEC CPU 2006 results with this change on top of
the proposed change:

--- gcc/rtlanal.c.foo   2007-07-11 19:53:13.000000000 -0700
+++ gcc/rtlanal.c       2007-07-11 19:54:12.000000000 -0700
@@ -2939,9 +2939,11 @@ swap_commutative_operands_p (rtx x, rtx 
   if (result)
     return result < 0;
 
+#if 0
   /* Group together equal REGs to do more simplification.  */
   if (REG_P (x) && REG_P (y))
     return REGNO (x) > REGNO (y);
+#endif
 
   return false;
 }

It will take 2 days. SPEC CPU 2K results aren't very encouraging.
One benchmark is down by 7% on IA32.  I suspect AMD processors may
have similar issues. Given the SPEC CPU 2K/2006 impact I have been
seeing on IA32 and Intel64, and the current addresing model has been
used for such a long time, I don't think it is a good idea to make
such a change which affects every architecture for one architecture.


H.J.

 


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