RE: ti-emu: RE: A86: Re: new emulator

To: <ti-emulator@lists.ticalc.org>
Subject: RE: ti-emu: RE: A86: Re: new emulator
From: "Liyang Hu" <liyang@liyang.freeserve.co.uk>
Date: Thu, 23 Dec 1999 20:50:28 -0000
Importance: Normal
In-Reply-To: <009801bf4d0d$dc7a9d40$6452dfd1@dad>
Reply-To: ti-emulator@lists.ticalc.org


ah... well, if cycle-accurate emulation is *that* important, I might just
implement it at some stage... however, the thing should still work, shouldnt
it?

> If you make a 256 byte table of cycle counts for opcodes, you can
> get a fast
> and accurate core emulation.  If you don't, interrupts will be _very_
> inaccurate, depending on how you do things (as you can do the
> indexed lookup
> and add in only one instruction...though the C compiler might handle it
> differently).  For example, a NOP only takes 4 t-states, but an LDIR takes
> 21 (I realize that it's a repeated LDI, which is 16, but again
> that depends
> on how it's processed).  Games and such, and probably even the ROM, that
> copy large chunks of memory, will be off when it comes to interrupts.

References:

Re: ti-emu: RE: A86: Re: new emulator

From: "David Phillips" <david@acz.org>