Date:	Sat, 13 Mar 1999 21:57:29 +0100
To:	Henrik Berglund SdU <pgcc AT delorie DOT com>
Subject:	Re: 3DNow Optimizations, was Re: AMDK6 optimized kernel and others
Message-ID:	<19990313215729.G21035@cerebro.laendle>
Mail-Followup-To:	Henrik Berglund SdU <pgcc AT delorie DOT com>
References:	<Pine DOT BSF DOT 3 DOT 96 DOT 990312182057 DOT 6560A-100000 AT recomnet DOT recomnet DOT net> <Pine DOT LNX DOT 4 DOT 10 DOT 9903131213250 DOT 10630-100000 AT pm11-198 DOT gso DOT infi DOT net>
Mime-Version:	1.0
In-Reply-To:	<Pine.LNX.4.10.9903131213250.10630-100000@pm11-198.gso.infi.net>; from Jonathan Mark Brooks on Sat, Mar 13, 1999 at 12:16:20PM -0500
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 Sat, Mar 13, 1999 at 12:16:20PM -0500, Jonathan Mark Brooks wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> 
> Just out of curiosity, will we see 3DNow! optimizations in the pgcc
> compiler?

If AMD gives me a dual amd-k7 machine, yes.

> Also, my understanding is that the next revision of ecgs is supposed
> to incorporate the pgcc x86 specific optimizations.

No. the process of enhancing egcs to pgccs level is gradual. Its not a
question of just applying the patch to egcs, its a continual re-engineerig
process to adapt the pgcc changes to the egcs environment.

Egcs has two goals that currently conflict with pgcc:

- must be rock-solid on x86 AND on other architectures. I'm working on
  making pgcc run on other archs, though, so this might go away.
- the compiler mustn's use ages to compile a program. pgcc often trades
  a large increase in compile time for relatively small code improvements.
  This is not acceptable for egcs, so the algorithms have to be improved.

> I saw the thread but I wasn't sure why this is a problem with 2.2 kernels,
> i.e., why you can't use the ecgs or pgcc compilers without error.

You can use pgcc-1.1.1, actually. you can also *usually* use the
snapshots.  I just don't want to hear any bug reports, as I'm tired of
debugging the kernel.

See the egcs faq (http://egcs.cygnus.com/), what it says about the kernel
also applies to pgcc, with he added limitation that I don't want to see
ugly bug reports, and my desire for seeing "pgcc is broken, it can't
compile linux" is growing towards negative infinity.

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

- Raw text -