On Mar 19, 4:22 am, "Michael J. Mahon" <mjma...@aol.com> wrote:
So, this is actually an emulation bug.
This is perhaps the first time that I can say in all seriousness that
this is a feature, not a bug! We are not emulating a ZIP chip or
accelerator card, though I can see this would be attractive in certain
situations.
If we assume that the emulator is supposed to emulate an Apple II
with an accelerated *processor*, then the emulated cycle count should
be divided by the acceleration ratio before being used to compute the
current "phantom" video byte.
"Maximum speed" emulation would be problematic. The alternative of
decoupling (host) video generation from emulated video generation
exposes some difficult synchronization issues for floating bus
emulation. Emulator authors have taken the easiest route that provides
value.
As it is, it emulates an Apple II with both the processor and the
video refresh accelerated--and there is no such beast!
Acceleration of the entire machine gives the benefit of peripheral
card acceleration, the most useful result being fast disk access. A
coprocessor (eg Z80) would similarly benefit. Sound is adversely
affected, of course, but this is usually not important.