From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <10301112259.AA12850@clio.rice.edu>
Subject: Re: gcc 3.2.1 optimizer degradation (strlen, -O2)
To: djgpp-workers AT delorie DOT com
Date: Sat, 11 Jan 2003 16:59:28 -0600 (CST)
Cc: eliz AT is DOT elta DOT co DOT il, gcc AT gcc DOT gnu DOT org
In-Reply-To: <2.7.9.WQJB.H8KFN8@pauzner.dnttm.ru> from "Leonid Pauzner" at Jan 11, 2003 11:01:56 PM
X-Mailer: ELM [version 2.5 PL2]
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> gcc321 -O2
> strlen.c   20 sec
> strlen2.c  17 sec
> (library?) 14 sec  

The library version was compiled with gcc 2.8.1, which seems faster
than the 16 seconds below with 2.95; but in any case your modified
code seems faster.

> gcc295 -O2
> strlen.c   16 sec  (main.c was compiled without -O)
> strlen2.c  11 sec  (main.c was compiled without -O)
> (builtin)   8 sec

I'm not sure the accuracy of the numbers, but it appears:

2.8.1  14 seconds
2.9.5  16 seconds
3.2.1  20 seconds

A dump of the assembler for each of these would be interesting.