From owner-freebsd-current Fri Nov 14 16:02:20 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id QAA14280 for current-outgoing; Fri, 14 Nov 1997 16:02:20 -0800 (PST) (envelope-from owner-freebsd-current) Received: from hobbes.saturn-tech.com (drussell.internode.net [198.161.228.154]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id QAA14275 for ; Fri, 14 Nov 1997 16:02:16 -0800 (PST) (envelope-from drussell@saturn-tech.com) Received: from localhost (drussell@localhost) by hobbes.saturn-tech.com (8.8.4/8.8.2) with SMTP id RAA22690; Fri, 14 Nov 1997 17:01:09 -0700 (MST) Date: Fri, 14 Nov 1997 17:01:08 -0700 (MST) From: Doug Russell To: Burton Sampley cc: current@FreeBSD.ORG Subject: Re: Follow-up to make world time???/ In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk On Fri, 14 Nov 1997, Burton Sampley wrote: > If I did try to move /usr/src and /usr/obj to different drives, since I > don't have a second SCSI drive, which would be better on a WD EIDE 3.1g > 5400 RPM drive? Moving them to seperate drives may well make a difference. (It all depends on where the main bottleneck(s) is/are, of course.) If you have the time, I'd try it both ways. The results could be interesting. If I had to guess, I'd put the /usr/src on the WD, as in mode 4, especially if you run with DMA, the read speed is quite quick. In my last little test session with the K6-233 running at 83MHz x3=250Mhz, the fastest disk configuration I found for two WDs, (I think I used a 33200 and a 21600 if I remember right) both recent models, with the DMA driver) was to mount /usr/src on one drive and /usr/obj on the other. This was even faster than having both src and obj on a ccd between the two drives. I also should point out that on the IDE disks (at least the WD disks and ASUS TX-97E board that I have tested it with) the bus speed makes a big difference to disk read speed. Running with an 83MHz bus over a 75MHz bus adds about 15% to my read speed. 75 MHz over 66 MHz is about 20%. I only run the K6 in my build box at 75 MHz because the only possible next step for 83 MHz is 208 MHz, which is a little fast for a 166 MHz chip IMHO. However, the compile speed at 83MHzx2=166 MHz is only BARELY slower than 75MHzx2.5=187.5 MHz. (On a kernel compile, for example, running at 187.5 is about 2 seconds faster. 187.5 is about 30% faster than 66x2.5=166 Mhz (or about a minute on a test kernel compile)). As far as other things to try, I'm not really sure. What bus speed are you running with? Are your memory speed knobbies in the CMOS setup set to full blast? (Using memory to match, of course...) In my testing, the K6 surely is a very fast compiling chip, which helps my times considerably compared to all my other boxes and chips. Later.......