delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/1997/06/10/04:09:46

Date: Tue, 10 Jun 1997 11:57:58 +0200 (MET DST)
From: Jan Hubicka <hubicka AT ta DOT jcu DOT cz>
To: Peter Palotas <blizzar AT hem1 DOT passagen DOT se>
cc: djgpp AT delorie DOT com
Subject: Re: Speed of DJGPP?
In-Reply-To: <3.0.16.19970609194139.3567dcf6@hem1.passagen.se>
Message-ID: <Pine.LNX.3.91.970610115532.9267A-100000@tabor.ta.jcu.cz>
MIME-Version: 1.0

On Mon, 9 Jun 1997, Peter Palotas wrote:

> 
> At 13.28 1997-06-09 +0200, you wrote:
> > On Thu, 5 Jun 1997, Peter Palotas wrote:
> >> Okay, then I misinterpreted what the guy wrote. This makes more sense.
> >> Although, I presume DJGPP makes faster code than BC!? Although, it produces
> >> about the same code as Watcom from what I have understood, so the question
> 
> >not exactly same. In more complex tests(like floating point mandelbot loop)
> >is gcc still twice faster then wc.
> 
> Well, that is not completely true. I wrote a madelbrot program for my math
> class just about two weeks ago, and a friend of mine had a friend of his
strange. Maybe it is because we use different internal loop. I may send 
you sources if you wish with my results from watcom and djgpp.
> write another mandelbrot program for him, I wrote mine in DJGPP (ofcourse)
> and he wrote his in Watcom. There was a small diffrence in speed, his was a
> little bit faster on a Pentium, and we used the exact same algorithm in the
> main loops. But as far as I know, DJGPP does not optimize code for Pentiums
> which Watcom does so that might explain it. But then I wonder how did they
> do it with quake? I mean Quake must have Pentium optimizations, did they
> write thos in inline ASM or something or what?
I think that oprimized for pentium means that they used floating point 
calculation instead of fixed point..
 > 
> >> is, why does it have to be slower?
> 
> >I personally think that it is sloweb because compiler is more unix oriented-
> >has more separated phases. As far as I know watcomm has just one program 
> >that does whole compilation and outputs in object..gcc has separaed 
> >assembler,compiler etc...so it forces some temporary files.
> >I am also not sure if watcomm uses something like RTL
> 
> What's RTL?
RTL is register transger language - language similiar to lisp used 
internaly by compiler and optimizers.
> Well, yeah, DJGPP uses more temp-files, I'm just wondering why not try and
> solve that problem by integrating the diffrent compilers, preprocessors and
thats just because dos is stupid OS. Good OS does this w/o integrating 
everything into one executable...

Honza
> assemblers a  bit more, so that they don't have to use temporary files. 
> 
> 
> 
> ------- Peter Palotas alias Blizzar -- blizzar AT hem1 DOT passagen DOT se ---------
> 
> *********************************************************************
> * A brief description of DJGPP:                                     *
> * Never before has so few people done so much for so many... DJGPP! *
> *********************************************************************
> 
> 

------------------------------------------------------------------------------
                   Have you browsed my www pages? Look at:
                       http://www.paru.cas.cz/~hubicka
      Koules-the game for Svgalib,X11 and OS/2,  Xonix-the game for X11
      czech documentation for linux index, original 2D computer art and
              funny 100 years old photos and articles are there!

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019