shoneill wrote:
> Apart from this, I do believe that using refnums force the code to run in
> the UI thread (anyone care to correct me?).
Correct!
> This may or may not have implications on the execution speed of your application.
It always has and definitely in a bad way! Besides of forcing the
diagram code of your subVI to go into the UI thread (and maybe waiting
on other parts in LabVIEW using that thread at the moment) it also has
another performance problem. While updating a control terminal (or even
local) will pass the actual data to the control buffer and immediately
continue, letting the control at some time in the future be updated from
the actual buffer element by the front panel redrawing routine, updating
a property node value will ALWAYS go to the control and actually draw
the new data before returning.
This last point has no big implications if you update the control only
once every second. But if you pass a certain limit such as 50 updates
per second, or the control is very complex to redraw, the property node
value update will actually limit the speed of your subVI.
For terminal or local updates, the UI thread may miss some of the values
sometimes by not redrawing each and every one, but I want to see the guy
who can distinguish 50 updated values per second on the screen.
Rolf Kalbermatter
Rolf Kalbermatter
My Blog 
DEMO, Electronic and Mechanical Support department, room 36.LB00.390