Lost samples in Windows environment question

[Tom]

This is a general enquiry. A long time ago we used to have a machine
that ran Real-Time Unix and locked in memory any Fortran or C code
which you were about to run. This was 15 years ago and later versions
of the software and hardware (Concurent systems) still exist. I was
always told that Windows was unsuitable for real-time applications as
it goes and does its houskeeping at odd times and samples can be lost.
How does Labview overcome such a problem?

Tom

[Ben]

Hi Tom,

LV over comes that issue by running under a RT OS.

Pharlap is the OS that is used on the PXI platforms.

LV RT can also run on Field Point target.

I wrote a demo program that plot the actual wait time of sub-second waits for a customer and plotted the results.

As long as you stayed above 50ms or so, the actual times of the loop where close to 50ms. They would vary by a couple of ms MOST OF THE TIME.

But when they wher bad, they where real bad. Loop times sometimes would jump to 1.5 seconds.

I have done the same thing to prove to myself that LV RT was really deterministic. Using software timing the jitter was on the order of microseconds.

Using hardware timing, you can get nanosecond resolution.

I hope this helps,

Ben