EPICS Home

Experimental Physics and Industrial Control System


 
1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: Windows IOC time error messages
From: Mark Rivers <[email protected]>
To: "'Andrew Johnson'" <[email protected]>, "[email protected]" <[email protected]>
Date: Fri, 3 Jun 2016 00:00:52 +0000
Hi Andrew,

Thanks for the quick reply.  

The problem was only seen on one machine, and only started this afternoon.

I rebooted (and installed a pending Windows update) and the problem has gone away.  If it comes back I will let you know.

Mark


-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Andrew Johnson
Sent: Thursday, June 02, 2016 6:18 PM
To: [email protected]
Subject: Re: Windows IOC time error messages

Hi Mark,

On 06/02/2016 05:25 PM, Mark Rivers wrote:
> My Windows IOCs just started printing the following messages at a high rate:
>  
> currentTime::getCurrentTime(): 0.000003 sec time discontinuity detected
> currentTime::getCurrentTime(): 0.000003 sec time discontinuity detected
> currentTime::getCurrentTime(): 0.000004 sec time discontinuity detected
> currentTime::getCurrentTime(): 0.000004 sec time discontinuity detected
> currentTime::getCurrentTime(): 0.000002 sec time discontinuity detected
> currentTime::getCurrentTime(): 0.000004 sec time discontinuity detected

> This is base 3.14.12.5.
>  
> What can cause this?  The time on the Windows machine appears to agree
> within 1 second to that on a Linux server.

This message gets printed when the WIN32 time provider sees the current
time as being *earlier* than when it last looked, in your case by around
2-4 microseconds.

On Windows the current time is retrieved by fetching the value of a
high-frequency Performance Counter whose frequency is measured against
the OS wall-clock time (which is only provided at a lower resolution and
may take some time to be read) using a PLL. It's actually the value of
that counter which appears to be stepping backwards, so it may be that
on an SMP system the values read from the Counter can differ slightly
between CPUs, although I'm only guessing there. Microsoft's
documentation implies that the difference between threads should be no
more than ±1 tick, but I don't know how big a tick is in seconds.

I can't explain why this might have only just started though, did these
systems get a Windows update recently?

- Andrew

-- 
Arguing for surveillance because you have nothing to hide is no
different than making the claim, "I don't care about freedom of
speech because I have nothing to say." -- Edward Snowdon


References:
Windows IOC time error messages Mark Rivers
Re: Windows IOC time error messages Andrew Johnson

Navigate by Date:
Prev: Re: Windows IOC time error messages Andrew Johnson
Next: Error msg Stefen Paul
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: Windows IOC time error messages Andrew Johnson
Next: Error msg Stefen Paul
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  <20162017  2018  2019  2020  2021  2022  2023  2024