Date: Fri, 12 Mar 1999 04:07:58 +0100
To: pgcc AT delorie DOT com
Subject: Re: AMDK6/ PPro instructions
Message-ID: <19990312040758.O16842@cerebro.laendle>
Mail-Followup-To: pgcc AT delorie DOT com
References: <36E6922D DOT 9FF4A5B7 AT mitre DOT org> <19990310232019 DOT J24287 AT cerebro DOT laendle> <14055 DOT 17068 DOT 168359 DOT 877540 AT ppp-x8-35 DOT ecn DOT purdue DOT edu> <19990311221628 DOT J11562 AT cerebro DOT laendle> <19990312003221 DOT C20937 AT math DOT fu-berlin DOT de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <19990312003221.C20937@math.fu-berlin.de>; from Felix von Leitner on Fri, Mar 12, 1999 at 12:32:21AM +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
Precedence: bulk

On Fri, Mar 12, 1999 at 12:32:21AM +0100, Felix von Leitner wrote:
> > >   Are you sure about this ?  AFAIK the AMD K6 implements the pentium
> > >   instruction set CMOVZ e.t.c.  are not implemented. It is the Cyrix M2
> > >   which implements the full pentium-pro instruction set.
> > As I said, I'm not sure. Pitty on AMD that they are unable to create a
> > reasonable cpu ;)
> 
> Why do you spread slander like this, then?

As other people have pointed out, AMD lacks these instructions, so its not
really "slander". Also, haven't you noted the smiley???

> actually create them.  I just compiled some test code with
> -march=pentiumpro on that machine and it ran OK, it was the "Request"
> program in the test subdirectory of the Hermes pixel conversion library.

This only proves 1. that pgcc still does not use them as aggressively as
it should and 2. that a compiler is not forced to use every instruction in
every program...

> > Until somebody corrects me I take it for granted that -march=pentiumpro
> > programs won't run on amd-k6 (Mark: isn't that a question for the faq ;?)
> 
> Which program can I try to use that will actually make pgcc create some
> of those instructions?

Any program might. egcs/pgcc is not yet very aggressive about generating
these, but most larger programs have enough of them in them so they won't
work on incompatible chips (glibc, pgcc, perl...).

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