This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] Optimize manual byte swap implementations v3
- From: Richard Guenther <rguenther at suse dot de>
- To: Andi Kleen <andi at firstfloor dot org>
- Cc: Andreas Krebbel <krebbel at linux dot vnet dot ibm dot com>, gcc-patches at gcc dot gnu dot org
- Date: Fri, 5 Jun 2009 13:07:00 +0200 (CEST)
- Subject: Re: [PATCH] Optimize manual byte swap implementations v3
- References: <20090209145520.GA32536@bart> <49931E43.firstname.lastname@example.org> <499414EA.email@example.com> <firstname.lastname@example.org> <email@example.com> <49971B86.firstname.lastname@example.org> <499727DB.email@example.com> <49986FE4.firstname.lastname@example.org> <email@example.com> <49F566D0.firstname.lastname@example.org> <20090605083856.GA6148@bart> <email@example.com>
On Fri, 5 Jun 2009, Andi Kleen wrote:
> "Andreas Krebbel" <firstname.lastname@example.org> writes:
> > I think the run-time of the pass is dominated by the time needed to
> > walk over the ssa statements. I doubt that tweaks in the recognition
> > of bswap will buy a lot. Merging it with sincos would help here but
> > would cost us the flexibility to move the pass to somewhere else. I
> > can certainly do that if this is considered the way to go.
> It would be nice if there was a single pass that does all these simple
> walks in parallel, serving both sincos and yours and any others that
> might come up.
That doesn't make it easily maintainable nor relocatable within
the pass pipeline. Also I don't think a stmt walk adds 0.5% compile-time.