[A83] Re: TI-83+ Silver Edition = Gameboy

To: <assembly-83@lists.ticalc.org>
Subject: [A83] Re: TI-83+ Silver Edition = Gameboy
From: "Henk Poley" <HPoley@DDS.nl>
Date: Sun, 25 Feb 2001 20:18:55 +0100
Reply-To: assembly-83@lists.ticalc.org

Yep, virtualization, but you first need to check if the instruction is
'real' or a 'virtual' one. This checking will destroy the registers (maybe
not all), saving them takes time, checking if they are real or virtual also
takes time, but okay, you don't need to emulate ALL instructions... (So
some less speed-decrease than a total-emu)

I think the best way to 'fix' this is to write a PC/Mac/Amiga (hey, let's
say C++) program that converts the GameBoy(-game)-ROM-image into a
Flash-App, so you don't need a program on your calc that interprets the
OP-codes, because this has already been done by the "reassembler" (or how
you would call it). This program should also 'fix' the memory mapped I/Os. 
(BTW: Don't ask me to program this...)

The only major trouble would be the LCD dimensions and the grayscale/color
display... Don't think that would be anything like easy to fix... Maybe
make an interupt-routine that displays a part the (emulated) bigger
GameBoy-LCD? Pressing something like [2nd]+"arrow key" would scroll to a
part of the screen, you need to avoid that "game-keys" like
fire,up,down,left,right,restart are the same as you scrolling keys.

Link-play-emu would be really pain-in-the-ass, Ti didn't use a very
thorough solution in it's link-design... If you would (even) want to be
able to "talk" with a GameBoy, you would have to make some
hardware-solution by connecting your linkcable with the GameBoy-cable...
(But I think the speed decrease would be too much to actually be able to
synchronise the link-port of the two machines)


	Henk Poley

-->	"Do we want a gamemachine?"

> Van: Scott Dial <wrath@calc.org>
>
> You all are condemning this as impossible but there is no need to 
> emulate ALL of the instructions, just the added ones... ever heard of 
> virtualization? Basically, you let all the instructions that ok to be 
> run, run... and all the ones that are not ok, you emulate. So in this 
> case, any memory access would need to be virtualized, and all the added 
> instructions. Simple! And this overcomes the speed problems noted in the 
> other email since the other was complete emulation on a 68k chip... this 
> is unneeded on this processor since they are plenty of native 
> instructions that can be run... Also, 24k of ram, AFIAK, is enough to do 
> this. The biggest downfall will be emulating the hardware. The graphics 
> hardware and all the memory mapped I/Os...
> 
>
> ComAsYuAre@aol.com wrote:
> 
> > There is a fundamental problem in your basic assumption -- the gameboy
is NOT 
> > a z80 chip.  It's similar, but has many instructions missing (ldir,
cpir, 
> > ix/iy, shadow registers, etc) and many new instructions as well (ldi
a,(hl), 
> > swap a, etc).  Thus emulation would be a huge pain since the cpu would 
> > actually have to be emulated as well.
> > 
> > 
> > In a message dated 2/24/01 2:56:55 PM Eastern Standard Time, 
> > gardenerbm@netzero.net writes:
> > 
> > 
> >> With the advent of the TI-83 Plus silver, could someone write a
gameboy
> >> emulator for it?  Look what the 83+ could do:
> >> 
> >> (I am assuming the gameboy has a z80 chip in it)
> >> 
> >> Advantages:
> >> 
> >> -83+ silver has room for one 8-megabit game, two 4-megabit games, or
eight
> >> 1-megab
> it games!
> >> -83+ silver will be much faster (than 83+) , allowing for easier
emulation.
> >> 
> >> Disadvantages:
> >> 
> >> -You'd have to choose between link emulation (?) and sound emulation.
> >> -Graphics would have to be distorted/crushed/reduced in order to fit
on the
> >> 83+ screen.
> >> 
> >> Anyone know if this is a possibility?  Or is the 24K of ram not
enough?
> > 
> > ----
> > Jonah Cohen
> > <ComAsYuAre@aol.com>
> > http://jonah.ticalc.org
> 
> -- 
> Scott "Wrath" Dial
> wrath@calc.org
> ICQ#3608935
> Member of TCPA - tcpa.calc.org
> PGP key available