Date: Sat, 19 Apr 1997 15:50:08 -0600 From: Warner Losh <imp@village.org> To: Amancio Hasty <hasty@rah.star-gate.com> Cc: current@freebsd.org Subject: Re: Speed deamons: How to build a build box? Message-ID: <E0wIi1I-0007Qw-00@rover.village.org> In-Reply-To: Your message of "Sat, 19 Apr 1997 14:11:56 PDT." <199704192111.OAA03728@rah.star-gate.com> References: <199704192111.OAA03728@rah.star-gate.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <199704192111.OAA03728@rah.star-gate.com> Amancio Hasty writes: : The disks don't seem to be running that fast. Post the disk geometry : and probably someone one the list will be able to help you set it up. : : I get this with my 5400 quantum disk drive : 16 8192 3273603 74051160 I get about 1/2 of that on write. But this is on a Quantum 4500 rpm disk. And not a very good one from others have been saying. The jaz drive benchmarks faster :-) : The read value is because of caching. I get about 86M for read caching. I'm very happy with that :-). : You may want to post your disklabel info: : setenv EDITOR emacs : disklabel -e -r /dev/rsd0 : then save the geometry to a file or if you are running X just cut : and paste. disklabel -r sd1 > xxx is faster, since emacs isn't needed :-). I know that I have the geometry setup right. I spent a great deal of time doing it by hand since sysinstall couldn't cope (it was for 2.1.6, which may be why) and disklabel sd1 auto failed :-(. All partitions are on cyl boundries. newfs has been told the correct (rather than the default) cyl sizes, etc. I'm almost positive that I set that up right. Besides, the IOZONE numbers were on the raw paritions before I newfs them.... Here's what I have, none the less: type: unknown disk: label: flags: bytes/sector: 512 sectors/track: 56 tracks/cylinder: 5 sectors/cylinder: 280 cylinders: 22435 sectors/unit: 6281856 rpm: 4500 interleave: 1 trackskew: 0 cylinderskew: 0 headswitch: 0 # milliseconds track-to-track seek: 0 # milliseconds drivedata: 0 8 partitions: # size offset fstype [fsize bsize bps/cpg] a: 131040 0 4.2BSD 0 0 0 # (Cyl. 0 - 467) b: 262080 131040 swap # (Cyl. 468 - 1403) c: 6281856 0 unused 0 0 # (Cyl. 0 - 22435*) e: 131040 393120 4.2BSD 0 0 0 # (Cyl. 1404 - 1871) f: 409360 524160 4.2BSD 0 0 0 # (Cyl. 1872 - 3333) g: 614320 933520 4.2BSD 0 0 0 # (Cyl. 3334 - 5527) h: 4733960 1547840 4.2BSD 0 0 0 # (Cyl. 5528 - 22434) [[ Again note that I set bps/cpg manually, and am using 8k/1k file system ]] bps/cpg was set to 5*56. Is that too small? Warner P.S. Overclocking from 180 to 200 (bus speed increased from 66 to 66, but still running CPU at 3x) results are in: 180 2:34 200 2:21 or 8% faster.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E0wIi1I-0007Qw-00>