Is there be a problem with generating an event each 1 millisecond?

No, it can be done (dependent on your OS) but you might not want to use the tm() on a state machine since it isn't very accurate especially if your event freq is close to your System Timer (tick) freq.

http://publib.boulder.ibm.com/infocenter/rhaphlp/v7r5/topic/com.ibm.rhapsody.designing.doc/topics/rhp_t_dm_using_timeout_triggers.html?resultof=%22%73%79%73%74%65%6d%22%20%22%74%69%6d%65%72%22%20

You can also use OXFTDelay().

A potential problem would be that such a frequent tick takes considerable processing overhead with frequent context switching. I have used a 1ms tick but I wasn't generating an event on every tick.
An event is queued so you won't have a lot of time left to do anything else.

Take care if you are using the tm() command to generate an event. There is a property that defines the minimal timeout for tm(). It's about 5 or 10 ms.

Roy

tm should be read as time minimum. The point of it is that if nothing happens in a state then after that time, move on.

Using tm for 'scheduling' is convenient sometimes but not the intent.

I model a 'scheduler' that uses an 'accurate tick'. AccurateTick is an extension to Rhapsody's OSAL and the Scheduler is an extension to Rhapsody's OXF.