From owner-freebsd-bugs Fri Jun 25 10:11:52 1999 Delivered-To: freebsd-bugs@freebsd.org Received: from uni-sb.de (uni-sb.de [134.96.252.33]) by hub.freebsd.org (Postfix) with ESMTP id 38A3914A12; Fri, 25 Jun 1999 10:11:46 -0700 (PDT) (envelope-from schuerge@wjpserver.CS.Uni-SB.DE) Received: from cs.uni-sb.de (cs.uni-sb.de [134.96.252.31]) by uni-sb.de (8.9.3/1999031900) with ESMTP id TAA12831; Fri, 25 Jun 1999 19:11:42 +0200 (CEST) Received: from wjpserver.cs.uni-sb.de (wjpserver.cs.uni-sb.de [134.96.247.42]) by cs.uni-sb.de (8.9.3/1999031900) with ESMTP id TAA06904; Fri, 25 Jun 1999 19:11:37 +0200 (CEST) Received: (from schuerge@localhost) by wjpserver.cs.uni-sb.de (8.9.3/8.9.3/wjp-SVR4/1999052600) id TAA01416; Fri, 25 Jun 1999 19:11:37 +0200 (MET DST) From: Thomas Schuerger Message-Id: <199906251711.TAA01416@wjpserver.cs.uni-sb.de> Subject: Re: kern/12381: Bad scheduling in FreeBSD In-Reply-To: <3773A8F9.61FD6998@we.lc.ehu.es> from "Jose M. Alcaide" at "Jun 25, 1999 06:06:17 pm" To: "Jose M. Alcaide" Date: Fri, 25 Jun 1999 19:11:37 +0200 (MET DST) Cc: sheldonh@FreeBSD.ORG, schuerge@cs.uni-sb.de, freebsd-bugs@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL57 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > I tried a ftp (1.7 MB): > > With system idle: > 1755377 bytes received in 1.97 seconds (872.12 KB/s) > > With setiathome running with nice 20: > 1755377 bytes received in 1.84 seconds (930.99 KB/s) > > (Obviously the difference comes from the network load). > > And also a "tar xzf ports.tar.gz": > > With system idle: > real 5m31.084s > user 0m3.413s > sys 0m21.317s > > With setiathome running with nice 20: > real 5m59.629s > user 0m4.163s > sys 0m24.460s > > I did the tests using FreeBSD 3.2-RELEASE, running on a PentiumII-350, > with an UltraDMA disk (using 0xa0ff flags and softupdates), and an > Intel Etherexpress Pro/100 network card. It's different on my machine: I'm using 4.0-CURRENT on a dual PII/400 system with 256 MB RAM and an UW-SCSI harddisk with softupdates enabled (Asus P2B-DS mainboard). Updating the ports with 2 rc5des (notice: 2 processors) processes in the background (niced to +20): time cvsup -g -h cvsup2.de.freebsd.org /usr/share/examples/cvsup/ports-supfile Connected to cvsup2.de.freebsd.org Updating collection ports-all/cvs Finished successfully 5.113u 6.662s 9:39.62 2.0% 2572+3471k 15408+48io 0pf+0w Without rc5 in the background: 5.569u 5.727s 2:45.00 6.8% 2648+3581k 14750+48io 0pf+0w So updating the ports takes 3.5 times as long when the there are the two processes in the background. This is really audible; the harddisk is accessed much more often when the processes are not running. I tested it several times with almost the same results. Transfer of 4542 KB of data over the network: Average transfer speed without rc5des in the background: 325 KB/sec Average transfer speed with rc5des in the background: 240 KB/sec I have exported a directory via NFS and NFS accesses are VERY MUCH slower from a remote machine, e.g. listing the directory with hundreds of files in it takes about half a second when nothing is in the background and 4 seconds when the rc5 processes are running. I had similar results when I only had one processor in my system. Ciao, Thomas. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message