[tcjinaz]

What Greg said. If you aren't measuring it, you can't fix it. There's a corollary: if you're measuring it too often, you're interfering (hello Dr Heisenberg). On Greg's suggestion, we ran profiling running off of our workstation, and found a number of stupid user tricks that had pegged out CPU usage, mostly relating to sending data too often to the SD. Werner slapped us good.

I am CAN ignorant, but are you calls to the CAN VI's waiting for completion (is your code waiting for them to return), or are they fire and forget, passing control as soon as the hardware involved is loaded up?

This may be unrelated, but we learned the hard way that the Compressor Start VI does not return. We had it sequentially preceding Compressor Control Loop, and had a terrible time until we parallelized them. Thanks to Doug, the CSA at the Arizona Regional, for steering our analysis where it needed to go.

I am looking forward to a higher level of CAN integration in the new controller.