Hi Desi,
Yes, in the Model T laptops, RAM starts at the top and grows down as
additional RAM is added. All system variables are stored in HIGH RAM,
so this growth strategy means the ROM doesn't need a bunch of extra code
for EVERY system variable depending on the amount of RAM available.
Ken
On 9/5/20 10:39 AM, Desi Villaescusa wrote:
I know, I am talking to myself again. ;-) Saturday musings...
So, I was thinking about the error I was getting with that BASIC RAM
test program I found on the Club100 site.
(Note: I am sure this is something everyone here will already know,
but just for fun...)
So I wanted to take a look at it.
My theory was that the program was maybe expecting a 32K M100???
So, I was looking thru it and found this line:
250 FORH=32*1024TO64*1024STEP256
I decided to look at the memory map, and I think I saw 32K ROM and
then the RAM, but it looks backward (to me) in that it doesn't start
wtih the 8K standard, it ends with it.
I think I see this:
8000 = 32768 (Option 3 - 32*1024) - to - 9FFF = 40959 (32k)
A000 = 40960 (Option 2 - 40*1024) - to - BFFF = 49151 (24k)
C000 = 49152 (Option 1 - 48*1024) - to - DFFF = 57343 (16k)
E000 = 57344 (Standard - 56*1024) - to - FFFF = 65535 (8k)
So I can't just change the 32 to a 24 for my 24K M100. According to
what I think I see, a 24k RAM system starts at 40960, which would be 40.
So I changed the line in the testing program to:
250 FORH=40*1024TO64*1024STEP256
Then I ran it and got "GOOD RAM".
I took my 8K M100 to work, so I can't test it there yet, but in
theory, I should be able to change that line to:
250 FORH=56*1024TO64*1024STEP256
..and it should work.
Now, I'm not thinking this is not a very robust RAM test program. ;-)
But at least it seems to work, and I now have a better understanding
of the M100 memory map.
*Sent:* Sunday, August 30, 2020 at 11:50 AM
*From:* "Desi Villaescusa" <[email protected]>
*To:* [email protected]
*Subject:* [M100] PC file transfer fun..
.....
While I was troubleshooting, I found a RAM test program on Club100.
RAMTST.BA
Ran it, RAM error at 32768.. Uh oh...
Checked that M100, RAM soldered in. OK, no easy swaps to test... Hmmm...
Finally about that time, I found 600 baud and restarting my programs
was pretty consistant, so I decided to load a largeish program.
And it worked... Hmmm...
So I get my 8k M100 and load that RAM test program..
RAM error at 32768...
OK, so that RAM test program probably isn't working (or is designed
for a 32k fully loaded M100 or ??).
