CVI 5.5 timer runs faster than system time in a multithreaded application under Win2000

[robs]

The only correlation i could detect so far was the CVI timer always running at least as fast as the system time, occasionally three times as fast.
The timer in the same application worked well (i.e. synchronously with system time) on Win NT.

[Giovanni]

Hi!
I also experienced problems with the Timer() function of LabWindows/CVI in multithreaded applications. In my case the operating system was Windows NT.
To ovecome the Timer() problem I implemented a workaround using a secondary thread for keeping an absolute time reference updated using GetTickCount() from the Win32 API.

Bye bye
Giovanni

[Jason F.]

There have been a few unconfirmed reports of this behavior with the Timer() function, but it appears that it may be caused by a certain hardware/software combination from which the Timer function relies on to obtain its data (it is rarely seen).

The workaround that Giovanni posted is certainly viable, and on this particular system of yours I would suggest trying it. However, if it doesn't work I would have to say that there may be something with this particular combination of Windows 2000 and this pc that you might need to contact either Microsoft or the vendor about since GetTickCount is an SDK function.

Now, if you would like for us to hunt down the cause of this problem with the Timer function we would like for you to contact us through the www.ni.com/ask webpage p
roviding us with information like the following:

- CPU brand/speed
- Mainboard brand/model/revision
- Mainboard chipset brand/exact model
- RAM type/amount
- FSB and cpu multiplier factor
- OS with exact revision and service pack info.