From: hansoft AT geocities DOT com (Hans Bezemer)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: How fast is DJGPP?
Date: Fri, 15 Aug 1997 07:45:08 GMT
Organization: HanSoft & Partners
Lines: 27
Message-ID: <33f40704.779715@news.nl.net>
References: <33F32C35 DOT 7C8B AT voyageur DOT ca> <33F34CB0 DOT 6521 AT voyageur DOT ca>
Reply-To: hansoft AT geocities DOT com
NNTP-Posting-Host: utr97-2.utrecht.nl.net
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

On Thu, 14 Aug 1997 13:21:36 -0500, "J.E." <cellis AT voyageur DOT ca> wrote:

>I forgot to mention that, when I talk about the speed of DJGPP I am
>refering to the speed of the exe's it produces.
Well, I never got to TC++ 3.0, but remained with TC2. I've made the
4tH compiler, which is a Forth compiler with signed 32 bit numbers (16
bit is "standard" in these environments). I used 'long's for these
numbers.

4tH is highly portable, so it is not tweeked to any specific compiler.
It compiles under K&R or ANSI-C compilers and I have confirmed ports
to different environments like FreeBSD, Linux, RS/6000, DPX/2, DPX/20,
Solaris, TC++ 3.0 for Windows, TC2, Coherent, and (of course) DJGPP.

4tH works perfectly in a tiny memory model (DOS com-file, 64K), which
is about the fastest memory model you can have. That port was done
with TC2. It does the bench in 21 secs.

I compiled the very same code to DPGPP V2.72 using -O3 and the whole
thing used only 9 secs for the very same bench. As a little extra one
can compile very, very large 4tH programs. ;)

You can test this for yourself, since 4tH is free and comes with full
source. http://www.geocities.com/SiliconValley/Bay/2334

Hans