Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 3 Aug 1995 22:12:58 -0700 (PDT)
From:      John Dyson <dyson>
To:        jiho@sierra.net
Cc:        davidg@Root.COM, freebsd-questions@freefall.cdrom.com
Subject:   Re: 2.0.5 Eager to go into swap
Message-ID:  <199508040512.WAA19121@freefall.cdrom.com>
In-Reply-To: <199508040309.AA11016@diamond.sierra.net> from "Jim Howard" at Aug 3, 95 07:12:39 pm

next in thread | previous in thread | raw e-mail | index | archive | help
These issues regarding swapping are interesting, but I have some
anecdotal evidence regarding the quality of hardware making a
big difference in system perf.  As you all know, I am a member
of the FreeBSD core team and have alot at stake regarding my
professional reputation.  I work at a very big communications
company and we had the need to put together a large (FreeBSD) Unix box
and alot of small ones too.  The small ones need to run X and
Netscape in 8MB!!!  To put it mildly, my coworkers were not
impressed with the system performance (nor was I.)  Well, I
did a bit of analysis and found that the Connor 204 IDE drive
was a bit slow :-).  I installed a Quantum Fireball(?) drive
and the system is now very snappy. :-).  Problem solved...  Note
that we also upgraded the processor from a 486/33 to a DX4/100,
and noticed only a minor difference in performance.

I suspect that part of the reason for the varied performance reports
has partially to do with disk performance among other things.
Paging performance depends very highly on your disk performance
and there is almost no way around that simple fact!!!

I will also re-iterate something that I sent to someone in private
email -- IF YOU ARE ALMOST RUNNING OUT OF SWAP SPACE, YOUR PERFORMANCE
WILL SUFFER SIGNIFICANTLY!!!  This has to do with the X server taking
up all of the anonymous backing store and paging will then occur almost
exclusively against vnode backed objects.  This will kill performance
because there are probably other swap-backed pages that could be paged
out, but can't be -- thereby clogging up memory...

John
dyson@root.com




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508040512.WAA19121>