Eden and my status

[Roman Hartmann]

Hi Nicolai,
sorry to hear about your loss. I lost my first move generator (which was a horrible mess btw) due a harddisc failure on my notebook. Usually I backup all my stuff but as it happened during my spare time at xmas I didn't made any backups ...

I'm sure Eden will come back stronger than before.

best regards
Roman

[hgm]

Instead, in my quest for knowledge in areas I have no clue about, I have ordered the Pico Edition of the XGamestation (http://www.xgamestation.com/view_product.php?id=34), a "retro" game console that you can build yourself. I'm planning to write a chess program for it, just to see what I can achieve with only 2k of memory, in SX28 assembly code.

It would be a really first-class job if you could pull that off. The problem doesn't seem so much the 2K program memory, but that these things only seem to have 136 bytes of data memory. That is really tight for a chess program. I vaguely seem to remember that Frans Morsch once wrote a Chess program for a similar chip, in the early '90s.

In 1980 I did write a chess program for a 6800 CPU that could run in 2KB of RAM. Program and data had to share that, but the data could take a larger fraction of that, and the program could draw on the system ROM for I/O functions. And that program (Usurpator I) did not implement castling and e.p., nor move-legality checking.

Later I developed a fully FIDE compliant Chess program for 6502, (Usurpator II), that needed 4KB to run (program+data), and I then proceeded building it into a matchbox. There I used an 8KB RAM chip, as in addition to the program I had to add all I/O software as well.

[nczempin]

Instead, in my quest for knowledge in areas I have no clue about, I have ordered the Pico Edition of the XGamestation (http://www.xgamestation.com/view_product.php?id=34), a "retro" game console that you can build yourself. I'm planning to write a chess program for it, just to see what I can achieve with only 2k of memory, in SX28 assembly code.

It would be a really first-class job if you could pull that off. The problem doesn't seem so much the 2K program memory, but that these things only seem to have 136 bytes of data memory. That is really tight for a chess program. I vaguely seem to remember that Frans Morsch once wrote a Chess program for a similar chip, in the early '90s.

In 1980 I did write a chess program for a 6800 CPU that could run in 2KB of RAM. Program and data had to share that, but the data could take a larger fraction of that, and the program could draw on the system ROM for I/O functions. And that program (Usurpator I) did not implement castling and e.p., nor move-legality checking.

Later I developed a fully FIDE compliant Chess program for 6502, (Usurpator II), that needed 4KB to run (program+data), and I then proceeded building it into a matchbox. There I used an 8KB RAM chip, as in addition to the program I had to add all I/O software as well.

Well, I'm hoping I can salvage some of that 2k Memory, perhaps that's a dead-end from the start.

In any case I'll count on you to give me some advice when the time comes; I'm much more a bloated-Java-Enterprise-applications-programmer than a low-level coder.

But I want to learn more about hardware and embedded software development, and that's where I'm starting.

If I really only have those 136 bytes only, I'll guess a legal move generator would be sufficient for me to judge the project a success.
And of course printing the move on a TV
I guess I won't have any hash tables

The Micro does have some more RAM. I'll also think about just extending the Pico step-by-step; hardware design sounds like a lot of fun, and I've never had the soldering etc. experience; I went straight to Software when I was 12.

I'm definitely salivating with expectation of the package from the States!