Login
Search for:


Previous | Next

From: Crown (all posts)
Date: 1998-07-02
Subject: Re: Turbo emulating
Hi there,

> When I converted 9 sec. formatting from Plus/4 to the Amiga a
>year ago (it was an easy job), somebody asked me to write a turbo
>communicating interface for the Amiga. (Between non modified 1541
>and Amiga of course.) Because there is none.
> I tried, and I failed. I tried to convert Pigmy's turbo loading-
>saving from Bard's Tale III. (It is really nice! Even can
>formatting!) But there was some instruction timing, and I lost my
>hope, I was unable to get the real timing for the instructions, and
>emulate it on the Amiga side.
On the 68000 it's a bit harder to play with instruction timing. First there is a very simple pipeline which enables consecutive instructions to overlap. The other thing and I think this one caused you troubles,
that the 68000 can access the bus in two ways, syncronous and asyncronous.
Normal memory is acessed in syncronous ways, but slower I/O hardware is acessed in an asyncronous way. That means, if the I/O is not ready for the access the CPU will insert Wait Cycles into the execution of the instruction. The number of the Wait cycles can vary at every access,
depends on the speed of the hardware, and the phase when the access was tried. The mouse in the Amiga was acessed by this way, and I think other hardware elements were also, (mostly I/O ports I think). I guess you connected the 1541 through such a slow hardware and this is why you cannot stabilize it!

By the way I figured this out, when I tried to do stable color splits with only the CPU, but the loop contained a mouse button check, and it fucked up the timings. After removing that it became stable....

Scenest:
Ok we will be at least 4 there. Lion will you have a Plus4 by you?
If yes that could be a gathering point????

Tibor Biczo / Crown of GOTU

Copyright © Plus/4 World Team, 2001-2024