Re: Demos

[Eric Jacobs <no@no.no>]

Mitchell Spector wrote:
> 
> >>Demo (software preview): SoftPC (fake-port)
> >
> >*double take* HUH?
> 
>     It was a hoax. Some programmer(s) wrote what was supposed to be a
> demostration copy of SoftPC for the IIgs. They had it setup so if you
> typed some commands (i.e. 'dir', 'dir/w', 'type autoexec.bat', 'time',
> 'C:', 'cd dos', etc) it would respond back with a pre-prepared output
> to make it _appear_ it was really emulating a PC. They even made the
> floppy drive seek unnecessarily and the display update slower than
> normal to further make the user believe the GS was emulating an IBM
> through software.
> 
>     Of course it didn't take much poking around the files on the disk
> to see it was a fake. Not too mention it is technically _impossible_
> to emulate an IBM at anywhere near runnable speeds on a stock 2.6 MHz
> IIgs (you could write SoftPC, but it'd probably take a day or two
> before the emulated machine would bring up a DOS prompt ;). The only
> real way to get the GS to run IBM software is with a PC-Transporter.
> (on the topic of software emulation, I'm still curious how that
> Nintendo Gameboy emulator will run, that might actually be semi-doable
> on a 15 MHz IIgs. This emulator is real, but may be kind of pokey).
> 
> Mitchell Spector
> spec@vax2.concordia.ca

Do you remember how slow a 4.77 MHz XT runs? It's pretty sluggish.
Making an XT emulator for the GS would be a lot easier than writing
a Gameboy emulator, I would expect. Using an accelerator, the direct 
page (registers) could go in the cache. Then you'd have to try some 
serious assembler tricks to decode and dispatch the instructions. The
8088 is very bad at cycle efficiency: some instructions can take a
very long time to complete. The 65816 is much better at using cycles
wisely. So it may actually be possible, though it wouldn't be very
fast, it might work.

-ej