From owner-freebsd-hackers Tue Apr 9 12:01:15 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id MAA03945 for hackers-outgoing; Tue, 9 Apr 1996 12:01:15 -0700 (PDT) Received: from Sisyphos (Sisyphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id MAA03919 for ; Tue, 9 Apr 1996 12:01:07 -0700 (PDT) Received: by Sisyphos id AA05584 (5.67b/IDA-1.5 for freebsd-hackers@FreeBSD.org); Tue, 9 Apr 1996 21:00:11 +0200 Message-Id: <199604091900.AA05584@Sisyphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Tue, 9 Apr 1996 21:00:11 +0200 In-Reply-To: Bruce Evans "Re: HDD cpu usage (IDE vs. SCSI)." (Apr 2, 21:47) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: Bruce Evans Subject: Re: HDD cpu usage (IDE vs. SCSI). Cc: dutchman@spase.nl, freebsd-hackers@freebsd.org Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk On Apr 2, 21:47, Bruce Evans wrote: } I would prefer lower latency to lower overhead in most cases. IDE disks } have natural advantages in this area (no complicated SCSI protocol to } interpreted by the slow i/o processor on the controller). Well, wasn't there some discussion about this in the mail lists some time ago ? Let's see, what I find in my mail archive ... % On Oct 8, 7:15, Bruce Evans wrote: % } That's not saying much. An IDE controller has a much lower command % } overhead than the buslogic: % } Cheap IDE on 486DX/33 ISA SAMSUNG SHD-3212A (slow disk): % } output for disklatency /dev/rwd0: % } Command overhead is 573 usec (time_4096 = 2830, time_8192 = 5087) % } transfer speed is 1.81489e+06 bytes/sec % } % } A high command overhead causes slow file system operations for % } everything except large i/o's. % } % } What are the command overheads of other popular controllers? % % NCR 53c810 (driven by a 486DX2/66): % % Command overhead is 751 usec (time_4096 = 1216, time_8192 = 1682) % transfer speed is 8.79954e+06 bytes/sec % IDE: 512/1815890 s + 573 us = 855 us % NCR: 512/8799540 s + 751 us = 809 us I've got to admit, that this only proves that a (cheap) IDE drive is slower than a reasonable SCSI drive. What I'd really like to see is "disklatency" output obtained on a fast EIDE system. Perhaps it is time to repost the sources ? Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se