Date: Thu, 18 Jan 2001 16:51:28 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Mike Darrett <ez073236 AT mailbox DOT ucdavis DOT edu>
cc: djgpp AT delorie DOT com
Subject: Re: compiler efficiency
In-Reply-To: <Pine.GSO.4.21.0101172014390.7634-100000@sandman.ucdavis.edu>
Message-ID: <Pine.SUN.3.91.1010118164953.7327B-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


On Wed, 17 Jan 2001, Mike Darrett wrote:

> I was running some algorithm benchmarks on DJGPP vs Borland C++ 5.5, and
> was shocked to see that DJGPP outperformed Borland C++ on some stack
> tests. Using a linked list to simulate a stack, adding and removing 80,000
> entries took 1.8 seconds on my AMDK6-2 350, but took 2.5 seconds on
> Borland C++, compiled without the -tW option (since it is a console app).
> 
> Any ideas? Is Borland C++ using thunking to access memory? Is DJGPP simply
> more efficient?

Did you try to look at the machine code in each case?  That might give 
you some clues.

Another possibility might be code and/or data alignment.  Latest versions 
of GCC and DJGPP do a good job there; does Borland come anywhere near?