[ntp:questions] still not able to get NTP to sync on windows 7 even w/ more updated binaries

Dave Hart davehart at gmail.com
Tue Mar 1 00:13:05 UTC 2011


On Mon, Feb 28, 2011 at 21:50 UTC, Anonymous wrote:
> On 2/26/2011 2:44 AM, Dave Hart wrote:
>> http://davehart.net/moo/ntpd-QPC-20090614-0900.zip
>>
>> That represents the last in a series of 4.2.4p6-based
>>  private releases of mine.  I originally developed the
>>  rewritten ntpd interpolation for Windows on this version
>>  of ntpd, but those changes were not integrated into the
>>  reference implementation until 4.2.5p162.  At that point,
>>  my interest in maintaining the private 4.2.4-based
>>  releases went way down, as 4.2.4 code had forked from
>>  ntp-dev at the end of 2006, and there had been a major
>>  cleanup pass on the ntpd source in the summer of 2007
>>  which made porting changes between 4.2.4 and 4.2.6
>>  nontrivial.
>>  Aside from David J Taylor, a few friends of his,
>>   and myself, everyone using the new interpolation code
>>   on Windows has used 4.2.5p162 or later.
>
> It reads like, versions later than 4.2.5p162
>  e.g. the current ntp-dev-4.2.7p134.tar.gz
>  should have similar interpolation functionality,
>  to 4.2.4p6-DLH-QPC-o;  However David J Taylor still sees
>  otherwise on Win 7 in the wild?

David has seen that my forked 4.2.4p6-based binary does better on some
systems than official releases which incorporate nearly identical
interpolation changes.  My working theory has been there may have been
a regression between 4.2.4 and 4.2.6 in ntpd performance on
low-precision clocks, such as the precision=-10 Vista/Win7
millisecond-ticking clock.  Using official 4.2.4 versions, this is not
possible to test as the interpolation code was never disabled with
older Windows ntpd, even when it was not working well at all, so the
system will believe it has a -18 or better precision but in fact
bounce around 10s of milliseconds.

I suspect the easiest way to track down such a regression is on a
POSIXy system with a low-resolution clock.  I pursued this a bit using
OpenWRT but the hardware I was using had too precise a clock to
reproduce the difference.

> Does the source for that version
>  (QPC-20090614-0900 / 4.2.4p6-DLH-QPC-o / 2009.03.15.239)
>  still exist?
>
>  I'm guessing,
> <davehart.net/ntp/refclock/ntp-4.2.4p6-DLH-QPC-20090315-src.zip>
>
>
> Is it forked off of
>  <archive.ntp.org/ntp4/ntp-4.2/ntp-4.2.4p6.tar.gz> ?

Correct, and that version is very close to the source for the later
June release, which as I recall, was a one-off for David Taylor
backporting some 4.2.5 feature he wanted onto the 4.2.4p6-based fork
that was working better for him.  I can likely dig up the source if
anyone wants it.

Cheers,
Dave Hart



More information about the questions mailing list