Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-Id: <200206171334.g5HDYMw36662@pilot28.cl.msu.edu>
Date: Mon, 17 Jun 2002 09:34:21 EDT
To: Marcus Elderic Koenig <elderic AT t-online DOT de>, cygwin AT cygwin DOT com
Subject: Re: gcc 3.1 slower than 2.95?
From: Harold L Hunt <huntharo AT msu DOT edu>

Marcus,

First, this is probably way off topic for the Cygwin list.  You should be
asking on the gcc lists.

However, I can make one comment.  In a scientific experiment you only change
one variable each time.  However, you changed two:

> -> gcc 2.95 
> g++ -O2 -mcpu=pentium -Wall settest.cpp -lwinmm
> 
> -> gcc 3.1
> g++ -O3 -mcpu=pentium -Wall settest.cpp -lwinmm

You changed both the compiler (from 2.95 to 3.1) and the optimization flag
(from -O2 to -O3).  Don't do that.  Make both tests at -O2 or -O3.  The -O2
flag seems to be about as high as anyone goes without doing some extensive
analysis on their own to determine if -O3 would be of any benefit.  Also, the
-O3 flag can do things that make the code size much larger which can negate
the effect of a processor cache, etc.  Stick with -O2 for both tests.


Harold

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/