how much overhead is required to trigger synchronized events on different cores?
Overhead as in memory use for the OS? The ESP32 has freeRTOS built in. Most of the libraries do not add more memory use overhead to load.
how do you synchronize events on separate processors, using an OS to an asynchronous event?
what is the latency between the trigger and the execution?
what is the latency between executions on the different processors?
If you'd read the materials you'd see there are semaphores and queues and events and other things to allow for cross core and cross process communications.
The ESP32 has a nano second timer which you can use to time those events. I've not done it my self.
Heck, the ESP32 has 3 ram areas one for each core and one shared by both cores.
i'm familiar with multi-processor hard real-time systems and OS mechanisms.
asynchronous, not periodic
your ignoring my questions. how long? sec, msec, usec, nsec
how long what?
I've not measured propagations times of the various components of freeRTOS.
I think that the OP asked the wrong question and everyone has been leading down the rabbit hole of RTOS and multicore processors.
The OP never answered the basic question- what do you mean by "execute in parallel"?
"at the same time" is relative -- while the bell is being rung, while the LEDs are flashing, the gate is moving