home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!caen!zaphod.mps.ohio-state.edu!cs.utexas.edu!sun-barr!news2me.EBay.Sun.COM!exodus.Eng.Sun.COM!appserv.Eng.Sun.COM!appserv!limes
- From: limes@ouroborous.eng.sun.com (Greg Limes)
- Newsgroups: comp.sys.sun.admin
- Subject: Re: Four processor vs two processor Galaxy systems
- Date: 9 Nov 92 13:28:59
- Organization: Sun Microsystems Computer Corporation
- Lines: 28
- Message-ID: <LIMES.92Nov9132859@ouroborous.eng.sun.com>
- References: <ivan.721286441@durras>
- NNTP-Posting-Host: ouroborous
- In-reply-to: ivan@durras.anu.edu.au's message of 9 Nov 92 05:20:41 GMT
-
- In article <ivan.721286441@durras> ivan@durras.anu.edu.au (Ivan Angus - ne Dean) writes:
- | Has anyone managed to produce a summary of what sort of applications
- | cause a four processor system, e.g. 4/690 to run slower than a two
- | processor version of the same system ?
-
- Easy measurement: run "vmstat" for a while. Adding two more CPUs will
- be a good idea if you tend to have more than two processes runnable
- ("r" column under "procs"), and are spending most of your time in user
- mode ("us" column under "cpu").
-
- | ... The sort of job mix we would subject it to includes
- | running up to sixty Xterminals and a number of telnet style logins, with most
- | users running compilations, plus a small amount of database work. The
- | users are primarily first year University students, so the compilations
- | tend to be small but frequent.
-
- I've had good luck with running lots of compiles in parallel on
- a four-processor 690, but it may depend on how much you bottleneck
- on your disk subsystem.
-
- | Along the same lines, is upgrading to 4.1.3 a good idea ? Presumably
- | perfomance patches for Galaxy systems have been rolled into 4.1.3 ?
-
- #insert standard 4.1.3 rant here: "lots of bugs fixed, do it", but
- feel free to discount my opinion as I am probably biased and totally
- insulated from any costs to upgrade ...
-
- -- Greg Limes [limes@eng.sun.com]
-