Date: Mon, 16 Dec 1996 18:25:34 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: john miller <jwvm AT elvis DOT umd DOT umich DOT edu>
cc: djgpp AT delorie DOT com
Subject: Re: Long initialization with V2?
In-Reply-To: <58lf03$9qc@lastactionhero.rs.itd.umich.edu>
Message-ID: <Pine.SUN.3.91.961216181906.18715V-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On 11 Dec 1996, john miller wrote:

> I have encountered a strange problem with V2.01 in which one application
> runs very slowly.

How much (in seconds, not in percents) is ``much longer''?

> The same code compiled with version 1.12 is around 40%
> faster but profiling times are essentially identical. Apparently, the 
> initialization is taking much longer with v2 compared to v1.

It's usually the other way around: v2 programs start up and run faster 
than their v1.x brethren.

> One last item: I suspect that memory allocation might be the culprit. The 
> code is running under win95.

Did you run the v1.12 version on Windows 95 too?  If so, I cannot 
understand how did you manage to profile it (in v1.x, profiling didn't 
work in DPMI mode at all, it would produce an all-zeroes report).

If the two versions were run in two different environments, please run
them in both, and post the memory reports printed by go32-v2 (when invoked
with no arguments) in both cases.