Category talk:Emulators

Confusion about some things

I think the discussion here about the difference between a simulator and emulator is fairly convincing: an 'emulator' is something that copies exactly the external behaviour of a thing (like an ICE), whereas a 'simulator' is a thing that models precisely the internal behaviour of a thing (like the SPICE circuit simulator).

Which leaves a question: are things like SIMH 'simulators', or 'emulators', by that definition?

I recall reading that for the simulated IMP, they had a bit of trouble getting it to run, because it had a lot of external-timing dependencies. So they had to make it much more careful about its operating speed. (Which sounds like a 'simulator' thing in a way... but maybe an 'emulator' thing too.) Not that SMH exactly matches how most CPUs operate internally, either! And most CPU emulators can't be plugged in in place of a real one - with the possible exception of one running in a UniBone/QBone. So which are SIMH and friends?

Any way, if we decide that things like SIMH really are simulators, it looks like we'll have to re-name this category? Jnc (talk) 17:55, 25 July 2025 (CEST)

I subscribe to the definition provided by the stackoverflow answer. But there is still some wiggle room with regards to simulation: to what precision do we need to model? Must a simulator be cycle accurate down to the nanosecond and keep track of all internal processor states? Does that also apply to peripheral devices? What about analog level? "To simulate a computer, you must first simulate the universe" - to paraphrase Carl Sagan. The SIMH suite is not consistent in this; individual targets may or may not model internal states.

I believe there is a broad consensus out there that SIMH would be called emulators. Most other similar projects are. Larsbrinkhoff (talk) 10:10, 26 July 2025 (CEST)