From owner-freebsd-hackers Sat Dec 23 21:34:32 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id VAA00185 for hackers-outgoing; Sat, 23 Dec 1995 21:34:32 -0800 (PST) Received: from lisa.rur.com (G338.257.InterLink.NET [199.202.234.53]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id VAA00175 for ; Sat, 23 Dec 1995 21:34:28 -0800 (PST) Received: (from leo@localhost) by lisa.rur.com (8.6.11/8.6.9) id AAA03527; Sun, 24 Dec 1995 00:34:24 -0500 Date: Sun, 24 Dec 1995 00:34:22 -0500 (EST) From: Leo Papandreou To: David Greenman cc: Torbjorn Granlund , freebsd-hackers@freebsd.org Subject: Re: Pentium bcopy In-Reply-To: <199512240357.TAA00460@corbin.Root.COM> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-hackers@freebsd.org Precedence: bulk On Sat, 23 Dec 1995, David Greenman wrote: > > Anyway, your optimization looks interesting and I do intend to try it out. > >Thanks for your efforts and please don't get too discouraged. > > I need to do some more testing, but a quick test shows that for copying > page-sized amounts, it's about 5% faster than bcopy on a 150Mhz P6 (Orion) > and about 25% faster on a 90Mhz Pentium (Triton, PB cache). > ...not 5 times faster, but definately an improvement. Thanks! > > -DG > > 150Mhz P6: > [corbin:davidg] ./copytest > copy 8617 > bcopy 9069 > 17.686u 0.000s 0:17.70 99.8% 35+230k 0+0io 0pf+0w > > 90Mhz P5: > [implode:davidg] ./copytest > copy 22724 > bcopy 28490 > 51.214u 0.015s 0:51.34 99.7% 35+232k 0+0io 0pf+0w > Oh great, a benchmark. You know everyone is going to try this. Here's my results on a P100 ASUS P54TP4, 2.0.5-RELEASE (still) size in ints 32 1024 2048 4096 8192 copy 75 1886 15820 35841 70672 memcpy 282 7741 15928 40631 81525 Seems like copy() is very much faster for smaller copies and merely faster for the larger ones. Nice.