The future of computer chess

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
mclane
Posts: 18927
Joined: Thu Mar 09, 2006 6:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub

Re: The future of computer chess

Post by mclane »

I said this many times before:

The TI99/4a has a 16 BIT CPU but a very weak chess engine „Videochess“.
It came out too learly to have a better chess engine.

But there is hope.
Someone tries to make a stronger chess engine for the TI.
His engine is written in UCSD Pascal.

https://github.com/wmaalouli/Phoenix-Chess

Maybe some others would like to help making the engine little stronger on those limited resources.

The idea is to beat “video chess” and maybe also some other 8 bit engines that run on 6502/z80 cpu.
Phoenix Chess

A chess engine for the TI 99/4A computer written in UCSD Pascal

I have long been disappointed by the performance of Video Chess on the TI 99/4A which was the only professionally released chess program on the TI. There was a user port of the original Sargon program, but it was actually inferior to Video Chess overall. Initially, the idea to write my own engine was very daunting, but as my computing skills improved over the decades, it became more approachable. In March of 2015, I launched the project with UCSD Pascal as the main programming environment. At that time my proficiency in UCSD Pascal was fairly limited and my research of chess programming was bare bones, so after nearly a couple of years my results were dismal and I never got to a truly functional version. So I gave up, but the germ of that project never truly left me and kept bugging me until October 2024 when I decided to give it another shot. By that time, my UCSD Pascal proficiency was far more advanced including the use of assembly language modules, and I deeply delved into the theory of chess engines. The end result was the current version of Phoenix Chess, which at long last consistently beats Video Chess at the latter's highest level.

This is still considered a work in progress. The primary limitation remains available memory which does not allow the program to go beyond a ply depth of 5. The secondary limitation is speed of execution which is excrutiatingly slow, making the program virtually unplayable unless massively accelerated via emulation. This unfortunately is related to the fact that UCSD Pascal compiles to intermediate p-code instead of assembly language, making it quite slow compared to the latter. So why did I pick UCSD Pascal in the first place? It's because of its extremely powerful capabilities in terms of units which allow for the creation of massive programs that can run within a very limited memory space, as well as the flexibiltiy of the data structures which is essential for a chess program. Phoenix Chess is currently hovering around 4300 lines of code, and yet can run within the 32K RAM of the TI along side the OS!

A final distinction of Phoenix Chess: it's the only chess engine for legacy early personal computers from the late 70's which uses the concept of 64-bit bitboards to represent the game, thanks to the TI's 16-bit TMS 9900 CPU which allows for the direct manipulation of 16-bit words.
You can run all this or program it from the TI Emulators, e.g. on Windows systems the emulator is called classic99,
on linux machines you can run it on the emul99 emulator which can be overclocked very good depending on your system.


Vorticon, as the nick name of the programmer is, told about his project beginning in
2016 in the following forum:

https://forums.atariage.com/topic/24842 ... nix-chess/
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
User avatar
mclane
Posts: 18927
Joined: Thu Mar 09, 2006 6:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub

Re: The future of computer chess

Post by mclane »

If anyone is interested in the sources (39 kb zipped) in a zip file, please let me know.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....