From the doc:

2.2. Idle time

Your kernel is likely to support halting the processor when a CPU is idle. As the typical hardware events like CPU_CLK_UNHALTED do not count when the CPU is halted, the kernel profile will not reflect the actual amount of time spent idle. You can change this behaviour by booting with the idle=poll option, which uses a different idle routine. This will appear as poll_idle() in your kernel profile.


2009/11/13 "Mgr. Šimon Tóth" <>
I'm trying to profile a batch system (Torque). It consists of a server
daemon, scheduler daemon and many node daemons.

Now the problem is, that when I get the result from oprofile using
CLK_UNHALTED the number of samples does not correspond to the time the
specific daemon was running (0.5 sec, vs. 10 minutes). Also the profile
looks weird, the server has only light load (5-10% cpu use) but oprofile
shows that a majority of the samples were recorded in cpu heavy functions.

I did try valgrind, but the slowdown makes the results unusable
(although the profile did look much better, actually showing that there
was a lot of waiting for system calls).

What should I do to get better results?

Mgr. Simon Toth

Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.
oprofile-list mailing list