Time triggered networks priority

[RobDekker]

I have 2 PCs running ETS and would like to use time-triggered shared variables to send data between the 2 targets.  Does anyone know what the priority of the time-triggered shared variables are and how much processor time they require?  Both targets have high priority loops running at 2 milliseconds with %70 cpu usage (close to the maximum that we can get them to run at) and we don't want to slow that down with the shared variables.  The targets are both Pentium 4 with 2.8 GHz processors and 1 GB of ram.

thanks,

Rob

[eek]

Excellent questions, Rob.

In regards to the priority of timed-shared, they are written/read at the priority of the VI/Timed-structure they are located in.

And as far as the processing time required, most of the cost is a result of the read/write calls and mainly depends on the amount of data your handling. I'm not aware of any official benchmarking at this point.

But let's say ballbark: I would imagine that anything around 4-16 byte reads/writes would be on the order of ~10us. Just keep in mind that as the amount of data grows, so will the cost of interrupt because more data must be copied from received packets and more cache misses are generated as data received from the I/O is located in the main memory.

I hope this sheds some light onto determining whether to incorporate those timed shared variables into your application. Let us know if you have any more questions/comments regarding this discussion.

Cheers,

Emilie K | Applications Engineer | National Instruments