> It seems to me that the low-level code that emulates the CPU would > be better served by unit tests for the individual instructions > that are emulated.
Even beyond that, the ability to boot and usefully use an OS and applications on the emulated system (which combines I/O with instruction execution and probably memory mapping) is the true Acid test. I do agree that someone who was into emulator writing for the fun of software development, as opposed to the experimental industrial archaeology of running old OS's and applications, might insist that other levels of testing are relevant too. Maybe not surprisingly, what is often found is that the original peripheral or CPU or memory management had poorly documented or undocumented quirks that are necessary to emulate to do anything useful. Closing the loop on the whole process, and ending up with a better architectural description of the original CPU and peripherals as they actually worked (as opposed to how they were documented to work) is where the true joy in emulator development comes. I think Bob's papers do an excellent job of showing the process and fun involved. It becomes a sort of experimental industrial archaeology experiment that entrances me and others. Tim. _______________________________________________ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
