Hello, been a while, dual CPU mobos
Ralph Doncaster
ralph-Zd07PnzKK1IAvxtiuMwx3w at public.gmane.org
Tue Sep 7 03:15:59 UTC 2004
On Mon, 6 Sep 2004, Christopher Browne wrote:
> The economy of things normally is that if you have a system with a mere
> 512MB of RAM, it is almost certain that spending a couple hundred
> dollars to upgrade to 2GB of RAM will provide the cheapest and most
> effective performance upgrade.
There is no such thing as a "normal" system and nothing is "almost
certain". Bumping the RAM in my POP server (>5000 entries in /etc/passwd)
from 1GB to 2GB isn't going to make a material difference:
ralph at ns ~$ free
total used free shared buffers cached
Mem: 905240 804992 100248 0 35316 478100
-/+ buffers/cache: 291576 613664
Swap: 2024168 60688 1963480
ralph at ns ~$ uptime
10:45pm up 96 days, 1:40, 5 users, load average: 0.69, 0.87, 0.84
vmstat will give you a good idea of where (if any) system bottlenecks are.
swap io system cpu
si so bi bo in cs us sy id
0 0 0 2 4 4 2 2 3
0 0 0 0 159 47 1 1 98
0 0 0 52 254 183 1 9 90
0 0 128 0 192 89 4 2 94
0 0 0 0 224 98 5 6 89
0 0 129 0 234 72 5 4 91
0 0 0 0 222 117 17 18 66
0 0 0 32 252 110 9 4 87
0 0 0 0 190 89 4 2 94
0 0 0 0 240 131 15 6 79
0 0 128 0 163 60 0 3 97
I can see that my CPU is mostly idle, so upgrading my 1Ghz duron CPU isn't
going to make a difference. There's not much disk I/O, so my software
RAID-1 setup is fine. Performance-wise this server is running fine.
IMHO, the cheapest way to improve performance is through software. I used
to run the pop server that came with Courier IMAP. I now get more than
double the performance from nupop. hdparm tuning and using noatime on
your filesystem will often give a big improvement to disk performance.
Also consider tmpfs for /tmp. I backed out of using it on 2.4.24 due to
crashes, but will probably give it a shot on my next kernel upgrade.
Ralph Doncaster, IStop.com president
6042147 Canada Inc.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
More information about the Legacy
mailing list