Date:	Fri, 12 Mar 1999 17:35:30 +0100
To:	pgcc AT delorie DOT com
Subject:	Re: AMDK6 optimized kernel and others
Message-ID:	<19990312173530.E18971@cerebro.laendle>
Mail-Followup-To:	pgcc AT delorie DOT com
References:	<Pine DOT BSF DOT 3 DOT 96 DOT 990311222721 DOT 17329A-100000 AT recomnet DOT recomnet DOT net> <Pine DOT GSO DOT 3 DOT 95 DOT 990312082710 DOT 13124A-100000 AT c2m2-17 DOT nada DOT kth DOT se>
Mime-Version:	1.0
In-Reply-To:	<Pine.GSO.3.95.990312082710.13124A-100000@c2m2-17.nada.kth.se>; from Michael Hanke on Fri, Mar 12, 1999 at 08:34:09AM +0100
X-Operating-System:	Linux version 2.2.3 (root AT cerebro) (gcc driver version pgcc-2.93.09 19990221 (gcc2 ss-980929 experimental) executing gcc version 2.7.2.3)
From:	Marc Lehmann <pcg AT goof DOT com>
Reply-To:	pgcc AT delorie DOT com
X-Mailing-List:	pgcc AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Fri, Mar 12, 1999 at 08:34:09AM +0100, Michael Hanke wrote:
> > without size optimization for K6. Also I noted that gzip compiled for
> > pentium is slower that gzip compiled for amdk6 on amd machine, this kinda
> > shows that amdk6 optimization actually works quite nicely. I use Stampede
> This note gives me the opportunity to ask about the real gain of pgcc
> on AMD chips. I have an old K5 processor. Since I am mainly
> interested in scientific computing, I would like to know the possible
> gain for fpu intense applications (e.g. BLAS). And the best possible

Nobody knows them. I guess

-O6 -funroll-all-loops -mstack-align-double

might do well. You can try -mk6, -mk5 (-mamdk6, -mamdk5 in the release)
or -m486 or -mpentium. It'd be nice to hear back what the fastest combination
is.

> flags (IEEE arithmetic is essential!). Recently, I am using gcc 2.7.2
> with -m486. Moreover, most programs are

You might also try -mieee-fp. Unfortunately, the x86 fpu is not ieee
compliant (not the hardware, that is, of course you can emulate it in
software), so you might need -ffloat-store and worse, but these make your
program really slow.

> written in FORTRAN. Is there a pg77 available or should I resort to
> f2c?

Unless my mind is totally boggled I even have pg77 binaries on the pgcc
server(s).

> The main exception in my applications are qt/KDE/lyx. Does anybody
> have experiences in the gain in speed for these programs?

Well, given that my perl-gtk programs run just as fast as their c-written
counterparts I don't think program speed is the limiting factor for these.
(there are exceptions, of course, an opimized imlib or gimp does make a
difference ;)

--  
      -----==-                                             |
      ----==-- _                                           |
      ---==---(_)__  __ ____  __       Marc Lehmann      +--
      --==---/ / _ \/ // /\ \/ /       pcg AT goof DOT com      |e|
      -=====/_/_//_/\_,_/ /_/\_\       XX11-RIPE         --+
    The choice of a GNU generation                       |
                                                         |

- Raw text -