[Asterisk-Users] Cheap Time sources which is best?

[Tzafrir Cohen]

On Sun, Sep 25, 2005 at 11:26:19AM -0400, Steve Gladden wrote:
> >>> What does zttest tell you?
> 
> Well on the 2.4 box it tells me (see below output).
> I've not really found any documentation on zttest on what is good or
> what is bad so at this point the numbers don't mean much.
> Also curiously enough (zttest) is not available on the 2.6 kernel system
> Both systems are latest CVS-HEAD just freshly compiled/installed and running.
> 
> For whatever reason zttest is not included in the build when done on our
> 2.6 kernel box.

You could easily build it in the zaptel source directory ('make zttest')
and run it from there. 

> It's in the source code but no binary was compiled and installed.
> So I don't even have the other box to compare to right now. :-( sadness!
> Actual in-use performance on both boxes seems to be great...

As it only uses a bunch of ioctls to zaptel, and the zaptel
kernel/userspace interface rarely changes, you would probably be lucky
even with the old binary.

> Typically have 2 or 3 simultaneous G711u calls going on and a couple
> gsm calls average of 5 calls going on....
> at times add several voicemail sessions and a meetme.
> 
> zztest output from 2.4 kernel box to follow:
> 
> Steve
> -----------------------------------------------------------------
> Opened pseudo zap interface, measuring accuracy...
> 99.987793% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586%
> 99.975586% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586%
> 99.975586% 99.975586%
> 99.975586% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586%
> 99.975586% 99.975586%
> 99.975586% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586%
> 99.975586% 99.975586%
> 99.975586% 99.975586% 99.975586% 99.975586% 99.975586% 99.975586%
> 99.975586% 99.975586%
> 99.975586% 99.975586% 99.975586% 99.987793% 99.975586% 99.975586%
> 99.975586% 99.975586%
> 99.975586% 99.975586%
> --- Results after 49 passes ---
> Best: 99.987793 -- Worst: 99.975586

So it seems to miss almost two of every 8192 (8000?) interrupts.

-- 
Tzafrir Cohen         | tzafrir at jbr.cohens.org.il | VIM is
http://tzafrir.org.il |                           | a Mutt's  
tzafrir at cohens.org.il |                           |  best
ICQ# 16849755         |                           | friend