Shared Variable Engine clock is consistently inconsistent

[pltrogon]

From the annals of the weird:

I noticed a while back that the DSM displays a different time than my system clock when I manually change a NSV value. I didn't think much of this, but then I noticed that it is not just wrong but inconsistently wrong. Here's the time difference, in seconds, between the system clock and the SVE timestamp for a dozen NSV value changes [the SVE clock always lags behind]:

35, 14, 1, 18, 29, 2, 21, 24, 14, 17, 29, 20

I checked these numbers in two different ways: (1) with the "Time stamp" event property node with NSV events using the DSC module and (2) manually changing values with the DSM and comparing to the displayed Windows time.

What's the deal? The timestamp is completely useless to me with a seemingly random offset from the system time; it's by far the best way to deal with this little bit of unfortunate SV event handling behavior. Filtering event timestamps against system time is fairly standard procedure and I can't see it being useful with this going on.

This should be trivial to reproduce--does anybody else see this? If so, is there an explaination--or better, a fix?

[I should note that I have seen this but I'm running Intel core 2 duo, not an AMD multicore. I'm using LV and DSC 2010 SP1]

[pltrogon]

I just found this post, which somehow eluded me on my previous searches. No solution mentioned but it is definitely looking like a TSC issue.

[JohnCS]

You may have to set up the Time Synchronization Services on the computer with the DSM running. 

In LabVIEW Help, look for Shared Variable Engine Page (Options Dialog Box)

In DSM Help, look for Shared Variable Engine Time Server Configuration Box

I don't know if both are necessary but I made all the computers on my system point at the same time server. The time server is configured as an Authoritative Time Server under Windows. It is in turn slaved to our DSN server which follows a Public NTP server. 

JohnCS 

[pltrogon]

That's a very useful thing to know about, but it didn't work--maybe because all the SVs are deployed under the local computer.