1K Real Chess for ZX80

Any discussions related to the creation of new hardware or software for the ZX80 or ZX81
Moggy
Posts: 1606
Joined: Wed Jun 18, 2008 2:00 pm

Re: ZX80 porting of Super Micro Chess 1K

Post by Moggy » Sat Oct 14, 2017 5:34 pm

Hi Sefano.

Yes I would be happy to debug test versions on real ZX80 so please do post them here as I'm sure others will be glad to help too. :D

Best Regards.

Moggy.

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: ZX80 porting of Super Micro Chess 1K

Post by marste » Sun Oct 15, 2017 11:18 pm

I already tested the running on real hardware time ago with a standard cassette load, both the standard and the flicker free versions (that infact I had to correct a bit for timing), and they were starting and appearing correctly on the screen! Hope they still work since I didn't change much that should interfere.

The debugging is now more on "strange moves" that are sometime computed as best one (especially the "suicidal" ones) and/or king checkings. I've to understand when it is happening and then why...

Attached two of the version for ZX80. They are both for "flexible play" (meaning you can choose each move if it is to compute automatically or define yourself the move). They differ for the display routine used (standard or flicker-free) and the deep of calculation / reaction time. Just load and then "RUN".
Attachments
1kRealChess80ff-F25-a.o
(947 Bytes) Downloaded 23 times
1kRealChess80-F37-a1.o
(916 Bytes) Downloaded 19 times

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: ZX80 porting of Super Micro Chess 1K

Post by marste » Thu Oct 19, 2017 10:18 am

PS: yesterday evening I run a deep debugging session and I identified that the problem is in the alfa beta implementation

I think this weekend I'll get a new (hopefully working) version.

Cheeers!
_Stefano

Moggy
Posts: 1606
Joined: Wed Jun 18, 2008 2:00 pm

Re: ZX80 porting of Super Micro Chess 1K

Post by Moggy » Thu Oct 19, 2017 11:45 pm

Excellent work Stefano. :D

The black square with white dot is a good choice and the flicker cursor version is the best one visually I think.


Regards.

Moggy.

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: ZX80 porting of Super Micro Chess 1K

Post by marste » Sun Mar 18, 2018 11:50 pm

It took me more than I originally tought but here there the first version of the 1K Real Chess for ZX80!!

Just the flexible verison, still not optimized and without openiing logic, but hope it's working.

A big thanks to Paul Veltjens for the real HW support and Paul Farrow for the recursive step-over debugging functionality! :)

Happy hacking!
_Stefano
Attachments
1kRealChess80-DEBUG.o
(881 Bytes) Downloaded 8 times

Moggy
Posts: 1606
Joined: Wed Jun 18, 2008 2:00 pm

Re: ZX80 porting of Super Micro Chess 1K

Post by Moggy » Mon Mar 19, 2018 9:17 pm

Works OK in emulator but when used on real hardware which has the ZXpand ROM the cursor movement is all over the place. sometimes it jumps 1 square then more than 1 square more so on horizontal movement which I think is a bug in the ROM used in the 80 with the ZXpand which also shows itself when you try and list more than a screenful of files, when the word "MORE?" appears a normal speed keypress gives a blank screen whereas hitting the key at light speed lets you list the remaining files so a keypress timing problem I would have thought.

Also it's hard to tell where the cursor is at any time, the previous method you used with the flashing cursor was far superior.


Having said all that it is still a great addition to list of ZX80 software. :D

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: ZX80 porting of Super Micro Chess 1K

Post by marste » Tue Mar 20, 2018 7:03 pm

So, the good news seems that the engine logic is working! ;)

Regarding the move input I didn't change, and last time yourself preferred the "flicker" version! If the ROM is bugged can it be corrected there? This weekend I'll try again on my ZX80. I'll try to ask also the opinion of Paul.

Moggy
Posts: 1606
Joined: Wed Jun 18, 2008 2:00 pm

Re: ZX80 porting of Super Micro Chess 1K

Post by Moggy » Tue Mar 20, 2018 10:05 pm

marste wrote:
Tue Mar 20, 2018 7:03 pm
So, the good news seems that the engine logic is working! ;)

Regarding the move input I didn't change, and last time yourself preferred the "flicker" version! If the ROM is bugged can it be corrected there? This weekend I'll try again on my ZX80. I'll try to ask also the opinion of Paul.
Hi Stefano.

The attached file is the version I prefer.

The version you posted does not flicker or flash the cursor it is always steady when I say flicker version I mean the one with the "flashing" cursor not screen flicker, my apologies for not making this clear.
As for the ROM bug it is something Paul is aware of but with the flicker version it does not seem to present itself maybe because of the time taken to execute the flicker makes the cursor movement more steady as it takes longer to show both states of the cursor.

The logic seems to be ok after the opening move and plays the weak or less strong game.

These are not criticisms Stefano only guidance and I know the effort put into this project is immense and much appreciated by myself. :D .
Attachments
1kRealChess80ff-F25-a.o
(947 Bytes) Downloaded 4 times

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: 1K Real Chess for ZX80

Post by marste » Fri Mar 23, 2018 6:05 pm

Then will be easily done! ;)

This weekend I'll post new version!

User avatar
marste
Posts: 123
Joined: Sun Aug 10, 2014 9:58 pm
Location: Italy
Contact:

Re: 1K Real Chess for ZX80

Post by marste » Sat Mar 24, 2018 2:43 pm

So, here it is: the flicker free release of 1k real chess for ZX80. Hope is working! 8-)

Attached there are 2 versions; one using alfa/beta pruning on 4 ply, the other quiescence search on 3-5 ply (I've still to merge of the two capabilities).

Both are still without initial move logic and full optimization is to be verified, but with the given (I would say quick) reaction time the stenght should be decent.

@Moggy let me know if they're working ok and in case which of the two version you think is stronger! :)
Attachments
1kRealChessZX80ff-35-DEBUG.o
(918 Bytes) Downloaded 7 times
1kRealChessZX80ff-4a-DEBUG.o
(945 Bytes) Downloaded 7 times

Post Reply