Re: Apple //c smartport compact flash adapter

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

On Feb 4, 12:10 pm, Steven Hirsch <snhir...@gmail.com> wrote:

> IIRC, the ZipChip does not accelerate code running in expansion ROM.  That was
> the motivation behind my hack to the RamFast EPROM that copies a little loop
> into the stack page to do block I/O.  It was originally suggested by Drew
> Vogan (RamFast designer) and seemed to make quite a significant improvement
> over polled I/O.

That's a nice trick. No accelerator speeds up the $C800 space since it
could hold anything and change unpredictably.

SmartPort (in theory) suffers worse than others due to being dependant
on 1Mhz speed for IWM comms, then the fact that the "slowdown" timeout
is generalised to worst case.

In practice this would mean the validating the checksum is at least
somewhat impacted by the accelerator being slowed down, and possible
completely impacted, lowering the overall achievable "practical"
transfer rate.

I recall an issue of Nibble back in the day having reasonably
comprehensive disk benchmarks in a HD review which included the
CT-20c. That might shed some light on what can reasonably be expected.
Perhaps Robert could run some simple BLOAD time loops in an Applesoft
program....

Matt