X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-bounces using -f
Message-ID: <3C679916.8915C74F@yahoo.com>
From: CBFalconer <cbfalconer AT yahoo DOT com>
Organization: Ched Research
X-Mailer: Mozilla 4.75 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Alignment problem
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1020211080340 DOT 21907F-100000 AT is>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 66
Date: Mon, 11 Feb 2002 10:25:21 GMT
NNTP-Posting-Host: 12.90.167.59
X-Complaints-To: abuse AT worldnet DOT att DOT net
X-Trace: bgtnsc05-news.ops.worldnet.att.net 1013423121 12.90.167.59 (Mon, 11 Feb 2002 10:25:21 GMT)
NNTP-Posting-Date: Mon, 11 Feb 2002 10:25:21 GMT
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Eli Zaretskii wrote:
> 
> On Mon, 11 Feb 2002, CBFalconer wrote:
> 
> > > [1] c:\c\hashlib>*gcc -c -pg -a -O2 -W -Wall -gstabs+ malloc.c
> > >
> > > [1] c:\c\hashlib>gcc -pg -a -o hshtestp hashtest.o cokusMT.o hashlib.o malloc.o
> > >
> > > [1] c:\c\hashlib>timerun hshtestp 4 10000
> > > Timer 3 on: 22:39:40
> > >
> > > Your program caused a divide overflow error.
> > > If the problem persists, contact your program vendor.
> 
> What version of GCC is this?
> 
> > dwpmi also shows up, but this is probably an effect, not a cause.
> 
> Sorry, I don't understand this comment.  What DPMI functions appeared
> hight in the profile, and why do you think it's an effect of something?
> 
> Perhaps simply post the flat profile part (the first part of Gprof's
> output), in case there are some hints there that you overlooked.
> 
> > Now what?
> 
> Try invoking Gprof with the --line option, it might tell more about the
> parts of `free' that take most of the running time.
> 
> > > Timer 3 off: 22:42:17  Elapsed: 0:00:02.80
> 
> Is this 2.8 seconds elapsed?  If so, it might be too short for a
> meaningful profile, especially if you succeed in getting line by line
> information.  Can you modify your program so that it runs longer, by
> either enlarging the amount of allocated objects, or even by simply
> running the same code in `main' in a loop?

gcc 2.953.  Lots of time, see below.  At least until I can get a
detailed look inside free.  I suspect free is talking back to the
dpmi memory manager.  As I said, an effect, not a cause.  The -l
option just puts markers on entrance/exit from free.  That's why I
think I need the -a compile option, which results in a
zero-divide.

Flat profile:

Each sample counts as 0.0555556 seconds.
  %   cumulative   self              self     total
 time   seconds   seconds    calls  ns/call  ns/call  name
 60.00      1.17     1.17    10013 116515.20 116515.20  free
 22.86      1.61     0.44                             __dpmi_int
  2.86      1.67     0.06    10064  5520.23  5520.23  malloc
  2.86      1.72     0.06                             inserted
  2.86      1.78     0.06                             mcount
  2.86      1.83     0.06                             putintbl
  2.86      1.89     0.06                             reloadMT
  2.86      1.94     0.06                             t1hash

That 2.8 secs includes program load, screen output, etc.

-- 
Chuck F (cbfalconer AT yahoo DOT com) (cbfalconer AT XXXXworldnet DOT att DOT net)
   Available for consulting/temporary embedded and systems.
   (Remove "XXXX" from reply address. yahoo works unmodified)
   mailto:uce AT ftc DOT gov  (for spambots to harvest)