Date: Sun, 21 Jul 1996 10:46:03 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Thomas Demmer <demmer AT LSTM DOT Ruhr-UNI-Bochum DOT De>
Cc: djgpp AT delorie DOT com
Subject: Re: gprof timing
In-Reply-To: <31EF4333.2781@LSTM.Ruhr-UNI-Bochum.De>
Message-Id: <Pine.SUN.3.91.960721104300.5526H-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Fri, 19 Jul 1996, Thomas Demmer wrote:

> I am right now optimizing (well, who is not) a grid generator
> code with gprof.
> Now gprof tells me the total execution time is 44 seconds. 
> The wall clock in fact indicates 87 seconds.
> Recompiling without -pg and running the test case gives
> 82 seconds. 

I think `gprof' can only report the cumulative time that it measured, but
it doesn't tell you how much elapsed time did the program run.  The 
relative times (in percents) are computed relative to the total time 
reported by `gprof' (44 sec in your case), so the profile is correct 
anyhow.