From owner-freebsd-hackers Wed Mar 22 04:09:17 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id EAA03313 for hackers-outgoing; Wed, 22 Mar 1995 04:09:17 -0800 Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id EAA03283 for ; Wed, 22 Mar 1995 04:08:50 -0800 Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id WAA12708; Wed, 22 Mar 1995 22:04:50 +1000 Date: Wed, 22 Mar 1995 22:04:50 +1000 From: Bruce Evans Message-Id: <199503221204.WAA12708@godzilla.zeta.org.au> To: bde@zeta.org.au, hasty@star-gate.com Subject: Re: Why IDE is bad Cc: hackers@FreeBSD.org Sender: hackers-owner@FreeBSD.org Precedence: bulk >>`cp -pR /usr/src/sys /usr/src/sys~' takes 85 seconds to copy >>2 * 12475K (293K/sec). Any IDE drive can compete with this transfer >>rate. >Now why during a normal transfer is scsi so slow? >Is it because of the updates to the inodes? I think it's partly for synchronous updates and partly for small i/o's. The bt445c-xp34301 runs at about half its maximum speed for 8K blocks, 1/4 of its maximum speed for 4K blocks, down to 1/20 of its maximum speed (about 293K/sec :-) for 512-byte blocks. This is for sequential input. Reading the same 512-byte block over and over is no faster. The AdaptecVLB with a P2100 (?) is reported to be about the same speed. The U34F is almost twice as slow. IDE is faster for small i/o's. Bruce