Date: Sun, 15 Feb 1998 13:54:11 +1030 From: Greg Lehey <grog@lemis.com> To: Julian Elischer <julian@whistle.com>, current@FreeBSD.ORG Subject: Re: Working (apparently) soft-update code available. Message-ID: <19980215135411.54721@freebie.lemis.com> In-Reply-To: <Pine.BSF.3.95.980213144636.23295C-100000@current1.whistle.com>; from Julian Elischer on Fri, Feb 13, 1998 at 03:07:52PM -0800 References: <Pine.BSF.3.95.980213144636.23295C-100000@current1.whistle.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 13 February 1998 at 15:07:52 -0800, Julian Elischer wrote: > > look in: > ftp://ftp.freebsd.org/pub/FreeBSD/incoming/softupdates4.tgz > > first extract the README and read it before extracting everything. > this should patch cleanly against last night's -current. > > try a make world before and after.. > particularly, > make /var/tmp and/or /usr/obj use softupdates. I tried this with a "make world". The figures are only approximate, since I didn't time the same make world under identical conditions, but they give a rough idea. I have my sources mounted on drive /src, and the objects are written to an old Conner 4 GB drive, /S. With both of these drives mounted async, a make world took 177 minutes. With soft updates, it took 156. Note that I'm still installing on /usr, which is an IDE drive with neither async nor soft updates. Here's the mount output after the build. I started the build effectively immediately after booting, so you can assume that most of these values were 0 when the make world started. I was, however, doing some other things on the system. In addition, I had /var/tmp on /home, which wasn't mounted /dev/wd0a on / (NFS exported, local, writes: sync 837 async 2326) /dev/wd0s1e on /usr (NFS exported, local, writes: sync 23567 async 10110) /dev/sd1h on /home (NFS exported, local, writes: sync 137441 async 8059) /dev/sd2e on /S (NFS exported, local, soft-updates, writes: sync 1305 async 66484) /dev/sd0e on /src (NFS exported, local, soft-updates, writes: sync 6 async 19466) > for a real laugh, try timing deletion of a large directory hierarchy.. === root@freebie (/dev/ttyp0) /usr 43 -> time rm -fr /usr/obj 2>/dev/null real 0m0.179s user 0m0.001s sys 0m0.012s Is that what I should have seen? I'm going to do some more comparisons, including another 'make world' with /var/tmp on /S, and again with neither async nor soft updates. It'll take a while :-) Greg To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980215135411.54721>