Sender: rsomerv AT mail3 DOT cadvision DOT com
Message-ID: <375690F1.6611C009@cadvision.com>
Date: Thu, 03 Jun 1999 08:28:01 -0600
From: Rob Somerville <somervil AT cadvision DOT com>
X-Mailer: Mozilla 4.6 [en] (X11; U; SunOS 5.5.1 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
To: pgcc AT delorie DOT com
Subject: Re: Performance of Krzysztof program
References: <m10pUEy-001mHUC AT Dirac DOT Chemie DOT FU-Berlin DOT DE>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: pgcc AT delorie DOT com

Bernd Melchers wrote:
> 
> I tested Krzysztofs program (gausil) at the same computer
> and different operating systems:
> 
> intel BX board, P2-400MHz:
> Win98/VC++6.0/standard-release flags:
> 1.32 sec
> Linux-2.2.9/egcs-1.1.2/glibc2/-O6 -march=pentiumpro -mcpu=pentiumpro
> 3.06 sec
> (simlar for -O2)
> 
> why is the microsoft compiler so much better?
> 
> It seems that there is much room for improvements in egcs and pgcc.
> 
> i have also other values for computers in our group:
> Amd K6-233   VC++6  2.3 sec
> Amd K6-2-350 VC++6  1.6 sec
> SGI Mips R4000-100  7.27 sec
> SGI Mips R8000-75   3.28 sec
> SGI Mips R10000-225 0.854 sec
> 
> our oldest PC with the old Amd K6-233 CPU is 25% faster with Windows/VC++
> than our newest PC with P2-400 and Linux/egcs! How could this be true???
> 
> Ciao
> Bernd
> --
> Bernd Melchers             | melchers AT FU-Berlin DOT DE
> Freie Universitaet Berlin  | "We don't write software, we compose it."
> AG Macromolecular Modelling - Prof. Dr. E.W. Knapp
> for more information see http://userpage.chemie.fu-berlin.de/~melchers/

any chance VC6.0 is linking against optimized libraries? I had same problem
with Sunsoft SPARC compilers vs. EGCS, but found the issue was Special libraries that
Sunsoft compilers used for certain optimization settings ( -fast )

-bob