From owner-freebsd-hackers  Sun Jul  7 05:32:46 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id FAA28368
          for hackers-outgoing; Sun, 7 Jul 1996 05:32:46 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id FAA28262
          for <hackers@freebsd.org>; Sun, 7 Jul 1996 05:32:20 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id WAA22327; Sun, 7 Jul 1996 22:31:02 +1000
Date: Sun, 7 Jul 1996 22:31:02 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199607071231.WAA22327@godzilla.zeta.org.au>
To: grog@lemis.de, hasty@rah.star-gate.com
Subject: Re: gcc lies?
Cc: hackers@freebsd.org
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>> Dumb question , is gcc-1.42 a lot faster than gcc-2.x?

>Good question.  The answer is 'yes', at least to go by what I've just
>tried.  I compiled cccp.c (the GNU preprocessor) with both compilers
>on a P133 with BSD/OS 2.1.  cc (1.42) took about 5.4 seconds, gcc
>(2.7.2) took about 14 seconds.

gcc seems to be particularly slow at compiling cccp.c.  Here it takes 16
seconds on a P133.  It takes only 2.34 seconds with my compiler (It
doesn't actually compile with my compiler, because of non-C (long long)
in the standard headers).  The difference for compiling with -S (and fixed
headers) is much larger:

bcc -S:        0.76 real         0.67 user         0.08 sys
cc -S:         7.16 real         6.96 user         0.18 sys 
cc -O2 -S:    14.94 real        14.73 user         0.11 sys

Bruce