From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Newsgroups: comp.os.msdos.djgpp,gnu.gcc.help
Subject: Re: gprof
Date: 16 Dec 2002 13:30:56 GMT
Organization: Aachen University of Technology (RWTH)
Lines: 23
Message-ID: <atkkig$lc0$1@nets3.rz.RWTH-Aachen.DE>
References: <PXvK9.400323$MGm1 DOT 232091 AT news02 DOT bloor DOT is DOT net DOT cable DOT rogers DOT com>
NNTP-Posting-Host: acp3bf.physik.rwth-aachen.de
X-Trace: nets3.rz.RWTH-Aachen.DE 1040045456 21888 137.226.32.75 (16 Dec 2002 13:30:56 GMT)
X-Complaints-To: abuse AT rwth-aachen DOT de
NNTP-Posting-Date: 16 Dec 2002 13:30:56 GMT
Originator: broeker@
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

In comp.os.msdos.djgpp Oliver Richman <frl8 AT rogers DOT com> wrote:
> Ok, so i've read the gprof info file, but I am still in the dark about
> symspecs. I've tried numerous ways to do it, but they just don't seem to
> work.

> I want to get rid of malloc, free, std::vector, __gcxx:: and whatever else.
> All I want to see are profiles for MY c++ functions, not anyone else's
> functions.

That's going to give you nightmares --- there are just too many such
functions your code is going to be calling.  It's also not very useful
to presume that time spent in libc/libc++ routines that *your* code
called is not part of the CPU consumption of that very code.  The
resulting profile would be quite meaningless.  If, e.g., a routine
burns 100% of its CPU time in calls to strlen() or malloc(), you
really should want to *see* that, not ignore it.

It should help to avoid linking any profiling-enabled libraries.
Which, in the case of DJGPP, happens automatically --- there are no
profiling-enabled ones in the distributed .zip files, to my knowledge.
-- 
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.