From owner-freebsd-current  Fri Apr  5 12:40:14 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id MAA10194
          for current-outgoing; Fri, 5 Apr 1996 12:40:14 -0800 (PST)
Received: from silvia.HIP.Berkeley.EDU (silvia.HIP.Berkeley.EDU [136.152.64.181])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id MAA10186
          for <current@freebsd.org>; Fri, 5 Apr 1996 12:40:10 -0800 (PST)
Received: (from asami@localhost) by silvia.HIP.Berkeley.EDU (8.7.5/8.6.9) id MAA25813; Fri, 5 Apr 1996 12:39:49 -0800 (PST)
Date: Fri, 5 Apr 1996 12:39:49 -0800 (PST)
Message-Id: <199604052039.MAA25813@silvia.HIP.Berkeley.EDU>
To: julian@ref.tfs.com
CC: mrami@minerva.cis.yale.edu, current@freebsd.org, nisha@cs.berkeley.edu,
        tege@matematik.su.se, hasty@rah.star-gate.com
In-reply-to: <199604052011.MAA13929@ref.tfs.com> (julian@ref.tfs.com)
Subject: Re: fast memory copy for large data sizes
From: asami@cs.berkeley.edu (Satoshi Asami)
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Marc, thanks for running the test.

 * > sh runtests
 * >     size     libc             ours
 * >       32   7.629395 MB/s   7.629395 MB/s
 * >       64  12.207031 MB/s   4.695012 MB/s
 * [...]
 * >  2097152  12.164192 MB/s   7.725020 MB/s
 * >  4194304  12.290410 MB/s   7.719504 MB/s
 * > mrami[~/bcopy]$
 * 
 * these tests SEEM to be indicating that the bcopy in libc
 * is already better!  or am I misreading something?

No, it only shows the libc version is faster on a 486.  We expected
that, as the FP unit on a 486 is much slower than a Pentium.  (Even a
straight int -> FP conversion takes a lot of time I guess.)

Any other Pentium owners? :)

Satoshi