02-09-2024 02:36 AM
I have an issue with the messenger library: A query&read operation tooks about 100ms. But at the level the operation is executed, there is only a delay of about 7-10 ms. The message is routed through several layers, passing queue and event based messengers and is transmitted via TCP (direct connection PC->cRIO, so no big delay should occur). Is it possible to trace the message routing, i.e. show the timing at each point the message is forwaded only by the framework? Do I need to reprogram the redirects to have the timing included? Could I overwrite one or more classes to include timing behaviour?
02-09-2024 05:04 PM
You could try the Desktop Execution Trace Toolkit to look at the Send VIs. There isn't anything built-in. I'm doubting there is a 100ms problem in messenger library itself, though. I would first suspect a TCP effect like Nagle's algorithm, as those are in the range of 100 ms.