Message-Id: <200208081329.PAA23478@mailproxy.de.uu.net>
Date: Thu, 08 Aug 2002 15:30:14 +0200
From: Sebastian Ude <ude AT handshake DOT de>
To: pgcc AT delorie DOT com
Subject: Re: bug? in pg77
In-Reply-To: <XFMail.020808143553.bohne@spec.inet.dkfz-heidelberg.de>
References: <XFMail DOT 020808143553 DOT bohne AT spec DOT inet DOT dkfz-heidelberg DOT de>
X-Mailer: Spruce 0.7.6 for X11 w/smtpio 0.9.0
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
Reply-To: pgcc AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: pgcc AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk



On Thu, 08 Aug 2002, a DOT bohne AT dkfz-heidelberg DOT de (Andreas Bohne) wrote:
> Date: Thu, 08 Aug 2002 14:35:53 +0200 (CEST)
> To: pgcc AT delorie DOT com, a DOT bohne AT dkfz-heidelberg DOT de
> From: a DOT bohne AT dkfz-heidelberg DOT de (Andreas Bohne)
> Reply-To: pgcc AT delorie DOT com
> Subject: bug? in pg77
>
> Hi,
>
> we are working in science I wanted to run a molecular dynamic simulation.
> It is a free software called Tinker and it is written in Fortran 77.
>
> I did run some benchmarks and now look:
>
> g77:
> real    63m41.531s
> user    63m25.240s
> sys     0m4.730s
>
>
>
> pg77:
>
> real    922m12.183s !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
> user    917m50.840s
> sys     0m54.250s
>
>
> For compiling I did use :
> CC="pg77"
> OPTI="-O3 -funroll-all-loops -ffast-math -march=i686 -mcpu=i686 "

Although you may try to play with the optimization options, just don't use
PGCC anymore - newer GCC versions optimize as good, most likely even better
as PGCC 2.95.2 on Pentium (and up) CPUs.

Nevertheless I am scared that g77 from PGCC produces THAT bad code, but I
only used the C and C++ compilers extensively and as PGCC development has
AFAIK been stopped for the reason above, just go with GCC 3.x.


- Sebastian