Message-Id: <199903162141.PAA21124@mail.mankato.msus.edu>
Date: Tue, 16 Mar 1999 15:40:48 -0600 (EST)
From: Jeffrey Hundstad <jeffrey DOT hundstad AT mankato DOT msus DOT edu>
Subject: Re: Benchmarks for floating point operations
To: pgcc AT delorie DOT com
In-Reply-To: <19990316203348.A25705@physik.fu-berlin.de>
MIME-Version: 1.0
Content-Type: TEXT/plain; CHARSET=US-ASCII
Reply-To: pgcc AT delorie DOT com
X-Mailing-List: pgcc AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

try

CFLAGS=-O20 -mcpu=i686 -march=i686 -fomit-frame-pointer --fast-math
-mstack-align-double -funroll-all-loops



On 16 Mar, Axel Thimm wrote:
> We are currently trying to see what we can drain maximally from PII for a
> certain flop intensive application (QCD). Until now folks were using gcc 2.8.1
> with -O2 -fomit-frame-pointer. I thought I might surprise them with egcs or
> pgcc, but the perfomance dropped from 80 to 50 Mflop/s (?)
> 
> [This was pgcc 1.1, as I cannot compile any newer snapshot/CVS, see related
> mail in this list]
> 
> Now I know of gcc to egcs regression, but I thought that pgcc was atop of both
> of them. I tried all kind of flags, but I couldn't get the old performance
> back.
> 
> Is this a known fact? Have others made similar experiences? The program is
> memory intensive (small ratio of computations per memory accesses) and perhaps
> this is what makes the difference.
> 
> Regards, Axel.