linux-rt-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Understanding cyclictest result
@ 2017-02-03  8:13 Ran Shalit
  2017-02-03 15:27 ` Sebastian Andrzej Siewior
  0 siblings, 1 reply; 2+ messages in thread
From: Ran Shalit @ 2017-02-03  8:13 UTC (permalink / raw)
  To: linux-rt-users

Hello,

I am beginner with RT patch, and cyclictest.

I have installed ubuntu 12.04 lts, and used the binary of 3.18 kernel from:
https://bitbucket.org/thismaechler/ubuntustudio-14.04-realtimeaudio/src/65ac2d9fce70f63479ef4a8f4764743cc7dc9b11/rt-kernel/HOWTO?at=master&fileviewer=file-view-default
I have verified with uname -a that it is linux-rt version.

I then tested with cyclictest , for interval of 1msec (1000usec)
 cyclictest -t1 -p 80 -n -i 1000 -l 10000
and got max value of 180usec.

On comparing my results to resuls examples given in
https://rt.wiki.kernel.org/index.php/Cyclictest
it seems like a very large number.
I know it depends on hardware (I am using Atom, one core only.), yet
becuase I'm not too experienced with it, I am not sure if my real
problem is HW or doing wrong testing/RTpatch.
I also see that it is much better then results before applying RT
patch (3000usec)


Another thing is that I also noticed that cpu load of 100% gives same
max value as if I just interrupt the console with several enter
clicks, from this I assume that load of 50% will probably won't change
this max value to a lower one.

Best Regards,
Ran

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Understanding cyclictest result
  2017-02-03  8:13 Understanding cyclictest result Ran Shalit
@ 2017-02-03 15:27 ` Sebastian Andrzej Siewior
  0 siblings, 0 replies; 2+ messages in thread
From: Sebastian Andrzej Siewior @ 2017-02-03 15:27 UTC (permalink / raw)
  To: Ran Shalit; +Cc: linux-rt-users

On 2017-02-03 10:13:15 [+0200], Ran Shalit wrote:
> Hello,
Hello,

…
> and got max value of 180usec.
> 
> On comparing my results to resuls examples given in
> https://rt.wiki.kernel.org/index.php/Cyclictest
> it seems like a very large number.
> I know it depends on hardware (I am using Atom, one core only.), yet
> becuase I'm not too experienced with it, I am not sure if my real
> problem is HW or doing wrong testing/RTpatch.
> I also see that it is much better then results before applying RT
> patch (3000usec)

So Carsten has
- Atom Z530 @1600 MHz with 382us [0]
- Atom N2800 @1860 MHz with 43us [1]

[0] https://www.osadl.org/Profile-of-system-in-rack-1-slot-8.qa-profile-r1s8.0.html#hwlats
[1] https://www.osadl.org/Profile-of-system-in-rack-c-slot-4.qa-profile-rcs4.0.html#hwlats

> Best Regards,
> Ran

Sebastian

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-02-03 15:27 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-02-03  8:13 Understanding cyclictest result Ran Shalit
2017-02-03 15:27 ` Sebastian Andrzej Siewior

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).