I am writing a VI that reads several waveform files and
displays them in a waveform graph for the user to check before he
continues. This is executing on an RT target just because that's where
the analog output hardware is; I don't need any deterministic timing,
really. It takes a very long time, upwards of 10 to 15 seconds, for the
waveform graph to refresh after the files have been loaded.The VI
actually finishes (in the sense that the Run button changes from
it's running state to it's stopped state), there's about a 15 second
delay, and then the waveform graph updates with the loaded values. Is
this expected behavior? I originally thought this was a disk IO
bottleneck, but loading from disk is actually quite fast, about 20
ms. If I run the VI on the host machine, without targeting the RT
hardware, it's very fast, so the delay is not inherent in redrawing the
graph.
See the attached screenshot. The waveform that is being loaded
is 1000 samples in size. I'm using Labview 7.1.1, and the RT hardware
is an 8186 controller.
To make matters weirder, I tried using a different computer
as the host computer, and it's much faster. That 10 second delay is
down to 1 s or less. The only difference in the two hosts is that the
one that shows the delay is an Semperon 2800 and the fast one is a P4
2.5 GHz or so. It seems odd to me that the host processor is the
problem, since as you can see in the screenshot, the CPU of the RT
target is maxed out for several seconds.
Can anyone shed any light on why this is so slow and/or why the two hosts behave differently?
