Date:	Wed, 23 Jun 1999 15:05:31 +0200
To:	Pgcc Mailing List <pgcc AT delorie DOT com>
Subject:	Re: performance issue
Message-ID:	<19990623150531.A9441@cerebro.laendle>
Mail-Followup-To:	Pgcc Mailing List <pgcc AT delorie DOT com>
References:	<19990621182836 DOT O28893 AT io DOT txc DOT com>
Mime-Version:	1.0
In-Reply-To:	<19990621182836.O28893@io.txc.com>; from Igor Schein on Mon, Jun 21, 1999 at 06:28:36PM -0400
X-Operating-System:	Linux version 2.3.6 (root AT cerebro) (gcc driver version pgcc-2.95 19990524 (prerelease) 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 Mon, Jun 21, 1999 at 06:28:36PM -0400, Igor Schein wrote:
> 
> I'm compiling a computationally-intensive program where performance is
> an important issue.  I have AMD K6, so I use -mk6.  I do have an
> option to use either -march=i486 ( default ) or -march=k6.  With
> latter, I get slightly worse performance benchmarks.  So I was
> wondering,  is K6 optimization going to be improved before the next
> release?  I obviously expect -march=k6 to produce better benchmarks.

The principle problem with the k6 is that I don't have one, so I can't
check ;) I've ported the k6 code from egcs because I thought it was done
by people _having_ an k6, however, and now Jan claims that the egcs code
is basically bad for performance.


> P.S.  If I don't specify -mk6, gcc segfaults.  So while that's not
> a work-stopper for me now, you might want to take a look at it.  

If its not too large, yes.

> What is the default cpu target?

Depends on how it was configured. Usually its the cpu the compiler was
built on.

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

- Raw text -