hello thur,

i know that the 3 numbers that are outputed from the 'uptime' command are based on the load averages of the past 1, 5, and 15 minutes..

but what is a load average? lol

thanks for the help

The CPU usage. 1.00 = 100% of what your CPU can do. Anything over that means requests are stacking up and waiting for free cycles to run. It's like the boss coming in and dumping more work on your desk when you already have a full day.

How is that possible?
01:38:11 up 10 days, 9:14, 4 users, load average: 2.82, 3.09, 2.85

and gdesklets shows about ~5-25% cpu load

Regards.

The algorithm for computing load averages is quite trivial -- it's just an average of the number of processes in the CPU's run queue during the given period (1, 5, and 15 minutes) -- on SMP machines it's the sum of the number of processes in each run queue, again averaged. If you have a CPU load of over two, it means, on average, two processes were runable on the CPU during the given period. If you only have one processor, that obviously means that the CPU had to switch back and forth between multiple processes trying to run on it.

Related info:

The average number of processes running in the last 1, 5 15 minutes may seem to be fairly useless info, but check into the batch command. Batch allows you to run a program(s) when that load average goes below .8 It can be set to a different number with the atrun command. It may come in handy for you some day. If you are running an X windows terminal, try this little perl script to watch the load average in the title bar of the terminal. Run it in the background.

loadavg.pl &

#!/usr/bin/perl
use strict;
$|++;
my $host = `/bin/hostname`;
chomp $host;

while(1) {

#open /proc/loadavg file
open(LOADAVG,"/proc/loadavg") || die "Cannot open /proc/loadavg: $!\n";

my @load=split(/ /,<LOADAVG>);
close(LOADAVG);

print "\033]0;";
print "$host: $load[0] $load[1] $load[2] at ",scalar(localtime);
print "\007";

sleep 2;
}

top shows line like this

Cpu(s): 10.4% us, 52.5% sy, 0.0% ni, 0.0% id, 24.5% wa, 5.7% hi, 6.9% si

us - user
sy - system
ni - nice

and others?

It's all in the man page

WA Size of the swapped out part of the disk
SI Data + Stack size
HI (I am unsure, but I beleive it deals with the highest priority task

I suppose all these values are related to CPU load

If you take a sum of all values - you'll get 100.
There is notthing in man page

According to the man page, The total may not always total to just 100%. True, most people do not even bother with "being nice", but .....

Time spent in niced tasks will also be counted in system and user
time, so the total will be more than 100%.

I'm interested in getting cpu load to build graphs. They are 100 height.

If I take user, system, nice loads and add them, i always get less then 100%

The rest are:
0.0% id, 24.5% wa, 5.7% hi, 6.9% si

id - idle, but i can't realise what are others

http://neptune.homelinux.com:6060/ot...load-3hour.png

What is
Cpu(s): 8.6% us, 4.0% sy, 0.0% ni, 0.0% id, 87.4% wa, 0.0% hi, 0.0% si ?