Login
Search for:


Previous | Next

From: Marko_Mäkelä
Date: 1998-06-04
Subject: Re: RAM refresh
On Wed, 3 Jun 1998, Tibor Biczo wrote:

> I've measured again. :) This time the RAM refresh cycles.
> During RAM refresh the memory area $FF00-$FFFF is acessed.
> It is acessed by the CPU memory config, so if RAM is turned on,
> the read value will come from RAM, and if ROM it will come
> from the ROM address.

Interesting, but logical since the VIC-II accesses the highest addresses
(videobankbase+$3F00-$3FFF) for RAM refresh. But it does so in decreasing order, i.e. $3FFF $3FFE and so on. If the details are not in Christian Bauer's article, I can dig up my measurement results.

Does the refresh counter wrap on raster line $100? What about the last raster line? The VIC-II changes $d012 one cycle too late (or was it too early) at the end of the last raster line. Well, I could see that from my measurement data, but I don't have it here right now.

Now does anybody have the interest (and time) to write a C16/plus4 program that runs entirely in open address space? On the C64 it was rather easy to write one that requires only 8 addresses to run. I baptized it to de00-de07. But it took me 2 weeks to design a program that runs in the color memory, because there the data is fetched both by the videochip and by the processor. I filled the stack with $DA, and the program executed RTS and other opcodes ending in 0 most of the time. That program is called dadb, for the address where it spends most of its time. The program is included with Frodo 4.1; the de00-de07 program is missing, but instead there's the de00all program written by Andreas Boose.

Marko

Copyright © Plus/4 World Team, 2001-2024