P90X, a 90-day program that I started 2010-04-08, has caused an unprecedented disruption of my work for Cosmic Horizon. Other activities share the blame. I concluded my 2-year study of Japanese on 2010-05-13. And it's baseball season. Being the organizer of Austin MLB comes with some responsibilities, but I've recently managed to get the time expenditure to a reasonable level.

As I come back to work, it's a lot like emerging from a coma.

The Software Requirements Specification doesn't say much about the SPARC-V9 Standard Reference Model (see Section 1.7.2).

I think the reference model should run in another thread. The monitor idea previously discussed (i.e. a monitor that watches instructions fetched by the DUV) may not be bad. Given that we are taking a black box approach, we can still gather information about that which is necessary (although perhaps not sufficient). If we are monitoring the DUV's instruction addresses as they are read from memory, then it is necessary for the address to appear on the bus before the instruction at that address can be executed, although this is insufficient evidence to prove that the instruction has been fetched into the core, let alone executed and committed.

Monitors are self-contained components that observe e.g. inputs to the DUV for scoreboard updates.

What kind of UML diagram would be good here?

communication diagram