This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Fixes tree-loop-distribute-patterns issues
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: Carlos O'Donell <carlos at redhat dot com>
- Cc: Roland McGrath <roland at hack dot frob dot com>, Adhemerval Zanella <azanella at linux dot vnet dot ibm dot com>, "GNU C. Library" <libc-alpha at sourceware dot org>, Siddhesh Poyarekar <siddhesh at redhat dot com>
- Date: Fri, 21 Jun 2013 10:15:04 +0200
- Subject: Re: [PATCH] Fixes tree-loop-distribute-patterns issues
- References: <51C1BFE9 dot 4070805 at linux dot vnet dot ibm dot com> <51C1CEFC dot 9000100 at redhat dot com> <51C1FE4C dot 3020400 at linux dot vnet dot ibm dot com> <20130619221130 dot 7B91A2C10E at topped-with-meat dot com> <51C31177 dot 90303 at linux dot vnet dot ibm dot com> <20130620175832 dot 0E6FA2C133 at topped-with-meat dot com> <20130620213141 dot GA4833 at domone dot kolej dot mff dot cuni dot cz> <20130620205919 dot 9156B2C135 at topped-with-meat dot com> <20130621020055 dot GA4729 at domone dot kolej dot mff dot cuni dot cz> <51C3A75D dot 8070606 at redhat dot com>
On Thu, Jun 20, 2013 at 09:07:41PM -0400, Carlos O'Donell wrote:
> On 06/20/2013 10:00 PM, OndÅej BÃlka wrote:
> > One solution is mandate to run benchmarks with fixed version of gcc and
> > fixed flags.
>
> That is not a bad idea...
>
> > Second variant could be have assemblies and regeneration script that would
> > be ran with specific gcc.
>
> ... and neither is that.
>
> I see these details as things that we as a community will
> need to work through as part of the refinement of the
> benchmarks we use as our criteria for performance patch
> acceptance.
>
Now when I think about that once gcc version and flags are fixed we can use -O3
and we will also see if there is regression againist vectorized.
This goes both ways as it will also show if gcc vectorizer consistently
makes some mistake that can be fixed.