Re: Zip Chip programming info / disk image?

["mdj" <mdj.mdj@gmail.com>]

> One other omission from the Zip Chip "compatibility guide" is
> that it doesn't respect the RDY input to the processor.  (This
> nailed me when attempting to use an FP accelerator card that
> used RDY to sync with the 65(C)02.)

Interesting... What's the purpose of the syncing ?

My knowledge of the Apple II's internals isn't that deep, but it would
seem to me you could achieve DMA compatibility by invalidating the
cache when RDY was pulled, since this normally only happens during DMA
- the reason a Zip Chip ignores it

Of course, the accelerator would have to write-through, cancelling out
some of the performance benefits.

But then, its not really that big a deal; the number of DMA compatible
devices that work in 8-bit Apple II's is limted. The primary ones are
SCSI controllers, and the simplistic IO model of ProDOS 8 means that
the performance difference between DMA and, err, PIO (to borrow a PC
world term) is pretty minimal on a machine with an accelerator. This is
why my main machine has a 'lowly' Rev C. SCSI card instead of a High
Speed, which I'm saving up until the inevitable time when the Rev C
loses it's grip.

Anecdotal testing (ie run stuff, rather than use benchmarks) shows that
the accelerated machine is comparable to the unaccelerated machine
doing DMA with a High Speed.

Matt