Effect of -fno-reorder-blocks?

[Jan Hubicka]

> > Date: Tue, 12 Nov 2002 11:46:17 -0500
> > From: Diego Novillo <dnovillo@redhat.com>
> 
> > On Tue, 12 Nov 2002, Hans-Peter Nilsson wrote:
> > 
> > > If this is simple for you to check, can I please ask you to do a
> > > comparative SPECint{95, 2000} run with and without
> > > -fno-reorder-blocks together with whatever other flags you use
> > > with -O2, for main trunk and perhaps also the 3.2 branch?
> 
> > Sure.  I've added -fno-reorder-blocks to both SPEC2000 and SPEC95
> > runs.
> 
> Thanks!
> 
> FWIW, I see only that change in the SPEC95 results.  Nothing
> below <URL:http://people.redhat.com/dnovillo/spec2000/> mentions
> use of -fno-reorder-blocks.  Though the SPEC95 result seems is
> conclusive; if -fno-reorder-blocks really was of any *general*
> positive effect (i.e. if the code conditioned by
> flag_reorder_blocks caused general pessimization), it should be
> visible in those results.
P4's trace cache should be, theoretically, completely immune about the
code layout, so it is not best testing platform.

Honza
> 
> >  They should take effect in tonight's run (5:00 UTC).
> 
> I see nothing conclusive about the effect of -fno-reorder-blocks
> when running SPEC95.
> 
> Note to others: the changes (some for better, some for worse) in
> results listed for SPEC95 at 2002-11-14 do not seem to be
> related to this change: by log inspection use of
> -fno-reorder-blocks *was* there for the 2002-11-13 run, and then
> there was no apparent change in testsuite results.
> 
> Thanks again for doing this.
> 
> brgds, H-P