Take a look at this and explain to me what is going on...

And also, how is top using 0.0% CPU time? ;)

VSIZE(delta) is the total address space currently allocated (including shared).

But look at the second-to-last line:
0 kernel_tas 0.0% 50:57.74 27 0 - - - 81.1M 883M

That's rather huge....
 
PS
Code: 
The delta columns are enabled by the -w option.  The deltas for #PRTS and
     VSIZE are relative to the numbers observed when top was launched.  The
     deltas for RPRVT, RSHRD and RSIZE are relative to the previous sample.
     If the -w option was not specified, top will append either a '+' or a '-'
     to the RPRVT, RSHRD, RSIZE or VSIZE columns indicating either a gain or a
     loss relative to the previous sample.
[/code
 
actually, im a little confused by all these terms. can you explain?

what exactly is a delta?
what does #PRTS #MREGS RPRVT RSHRD RSIZE & VSIZE actually mean?

I've read the man pages, but it didn't really help me understand these terms.

ps
i ran top with only the -ln option, do you want me to run it with -w ?
 
Ok well the deltas mean, as in maths and science, 'a change in'. So the deltas represented are the change in the value since you started top (shown in parentheses).

I would suggest, only for completeness, using "top -wl3 > topout.txt" (that's WL3) so that you can best see what's going on. The first top sample never gets the CPU usage. Then you can just chop off the first 2 samples from topout.txt and put them up here.

As for the rest, I've got to run but I'll go over the other defs when I get back from work (or maybe I'll do it at work, scandalous!)...
 
What's your uptime? That last task claims to have run for 210 hours.

Does anyone else know what that task it? It looks to me like some kind of switcher for the kernel (it looks like it's called kernel_tas(k manager) or something...) If it is some kind of kernel unit, I wouldn't worry that it's using so much RAM, though it is a little strange.

Are you on Jag or 10.1.5 (or ...)?
 
the most current version of jaguar

uptime
3:53PM up 4 days, 21:46, 2 users, load averages: 0.14, 0.07, 0.02

it sucks that my system is constantly swapping out...the longer my computer is on, the worse it gets and it gets slower and slower.
 
Also, i don't have any thread called "kernel_tas" running on my (10.1.5) system.

Also, as you'll notice, kernel_tas is running as PID 0 on your box. Do you have any third-party BSD apps installed (like Fink)?

Want to enter this, it will tell you which user is running the process, and where the binary is located if it's not in your $path (is it you, or root, or something else?):
Code: 
ps -aux | grep "kernel_tas" | grep -v grep
You should get a line that looks something like this as an output (for example, but you'll have kernel_tas as the app not autodiskmount):
Code: 
root     104   0.0  0.1     2396    496  ??  Ss     0:02.11 /sbin/autodiskmount -va

Woah! Your system has been running for 4 days (96 hours) yet that process has run for 210 hours! Apple must have implemented quantum computing in Jaguar (borrowing power from paralell universes :p)!
 
here is the output from top right after a reboot:
http://darklotus.dyndns.org/temp/top2.txt look at all that free / inactive ram. too bad it wont last...

I just tried
ps -aux | grep "kernel_tas" | grep -v grep
Click to expand...
and got no results, which confuses me...it's right there in top...if I try something other than kernel_tas
IE:
ps -aux | grep "kextd" | grep -v grep
root 51 0.0 0.1 15400 776 ?? Ss 9:41PM 0:02.24 kextd
Click to expand...

it works fine....
 
kernel_task is just a handy abstraction that top uses to represent the resources consumed by the Darwin kernel itself. It's not a process that shows up in ps. Generally, you don't need to worry about it.

Also, on a multi-processor machine, a program could easily spend 3 hours on each processor and get 6 hours of CPU time during 4 hours of uptime.
 
You must log in or register to reply here.
Back
Top