New Engine: ChessMatrix

Questions and comments related to CCRL testing study
Post Reply
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

New Engine: ChessMatrix

Post by KristianEkman »

Name: Kristian Ekman
Nationality: Sweden
Engine name: ChessMatrix
The engine is open-source.
Github repository: https://github.com/KristianEkman/ChessMatrix
Language: C
Engine binary (64-bit): https://github.com/KristianEkman/ChessM ... tag/v1.0.2
Estimated ELO: 2000
Protocol: UCI

Been working on this for quite som time. A big first release for me :)
If you have time, can you please run it against the others?

Kind regards
Kristian
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

Opps, found a small bug.
Can you run this instead?
https://github.com/KristianEkman/ChessM ... tag/v1.0.3

Kind regards
/Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

Hi Kristian,

The exe I downloaded tells me its version is 1.0.1.

Tried the sources with gcc and clang, I got a 10 kB error report, it starts like this (not that I understand what those messages mean but they may be useful to you):
main.c:309:38: warning: 'enum PieceType' declared inside parameter list will not be visible outside of this definition or declaration
309 | void PutFreePieceAt(int square, enum PieceType pieceType, int side01) {
| ^~~~~~~~~
main.c:309:48: error: parameter 2 ('pieceType') has incomplete type
309 | void PutFreePieceAt(int square, enum PieceType pieceType, int side01) {
| ~~~~~~~~~~~~~~~^~~~~~~~~
main.c:320:41: warning: 'enum PieceType' declared inside parameter list will not be visible outside of this definition or declaration
320 | void InitPiece(int file, int rank, enum PieceType type, Side color) {
| ^~~~~~~~~
main.c:320:51: error: parameter 3 ('type') has incomplete type
320 | void InitPiece(int file, int rank, enum PieceType type, Side color) {
| ~~~~~~~~~~~~~~~^~~~
moves.c: In function 'RemoveInvalidMoves':
moves.c:805:2: warning: implicit declaration of function 'memcpy' [-Wimplicit-function-declaration]
805 | memcpy(game->MovesBuffer, &validMoves, validMovesCount * sizeof(Move));
| ^~~~~~
moves.c:805:2: warning: incompatible implicit declaration of built-in function 'memcpy'
moves.c:7:1: note: include '<string.h>' or provide a declaration of 'memcpy'
6 | #include "sort.h"
+++ |+#include <string.h>
7 |
moves.c: In function 'ValidMoves':
moves.c:816:2: warning: incompatible implicit declaration of built-in function 'memcpy'
816 | memcpy(moves, g_mainGame.MovesBuffer, g_mainGame.MovesBufferLength * sizeof(Move));
| ^~~~~~
moves.c:816:2: note: include '<string.h>' or provide a declaration of 'memcpy'
moves.c: In function 'ValidMovesOnThread':
moves.c:827:2: warning: incompatible implicit declaration of built-in function 'memcpy'
827 | memcpy(moves, game->MovesBuffer, game->MovesBufferLength * sizeof(Move));
| ^~~~~~
moves.c:827:2: note: include '<string.h>' or provide a declaration of 'memcpy'
search.c: In function 'PrintBestLine':
search.c:419:13: warning: initialization of 'char *' from incompatible pointer type 'char (*)[1800]' [-Wincompatible-pointer-types]
419 | char* pv = &buffer;
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

Hi!
I just forgot to edit the version number in the uci output.
https://github.com/KristianEkman/ChessM ... tag/v1.0.4

The program is not compatible for gcc compillation yet.
You should be able to compile the code using Visual Studio 2019 on Windows 10.

Cheers
/Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

I have started a gauntlet with 1.0.4 under Fritz. Unfortunately it displays its thinking only when making the move.
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

Hi Gabor!
I made a new release here:
https://github.com/KristianEkman/ChessM ... tag/v1.0.5
It should write out the thinking after every searched depth.

I hope it will work better this time.
Let me know if there are more errors.

Kind Regards
/Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

KristianEkman wrote: Sun Jul 19, 2020 11:13 am Hi Gabor!
I made a new release here:
https://github.com/KristianEkman/ChessM ... tag/v1.0.5
It should write out the thinking after every searched depth.

I hope it will work better this time.
Let me know if there are more errors.

Kind Regards
/Kristian
Hi Kristian,

The opponents did not complain, only I.

I think I am going to stick with 1.0.4 for the moment or I may use 1.0.5 if the only change was the output. Can you confirm that?
I have submitted some of its games already, you will be able to see its results at the next update (which is normaly on Saturdays but this week maybe Wednesday too).

BTW, you wrote Relsease 1.0.5.
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

There is a change in evaluation calculation from 1.0.4 to 1.0.5.
Probably there is no major improvment but it would be fun to know if the results differ.
Is it possible?

/Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

KristianEkman wrote: Sun Jul 19, 2020 12:18 pm There is a change in evaluation calculation from 1.0.4 to 1.0.5.
Probably there is no major improvment but it would be fun to know if the results differ.
Is it possible?

/Kristian
I myself have many engines to test so I think you'll have to be patient. Other testers may come to the rescue though.
User avatar
Sergio Martinez
Posts: 3396
Joined: Tue Mar 12, 2013 11:38 pm
Sign-up code: 10159
Location: Murcia, Spain

Re: New Engine: ChessMatrix

Post by Sergio Martinez »

Hi Kristian

I am going to start a gauntlet with version 1.0.5
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

Sergio Martinez wrote: Sun Jul 19, 2020 9:41 pm Hi Kristian

I am going to start a gauntlet with version 1.0.5
Opponents?
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

Hi, Ill have a much better version soon.
You can skip version 1.0.5 for now.

Thanks.
/Kristian
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

This is the release I´d like to have tested when you can.
https://github.com/KristianEkman/ChessM ... tag/v1.0.6
Estimted ELO: 1950

Cheers!
Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

KristianEkman wrote: Tue Aug 11, 2020 8:46 am This is the release I´d like to have tested when you can.
https://github.com/KristianEkman/ChessM ... tag/v1.0.6
Estimted ELO: 1950

Cheers!
Kristian
Hi Kristian,

I'm going to test it sooner or later. However, 1.0.4 also has a 1950 rating so where is the improvement?
KristianEkman
Posts: 8
Joined: Sun Jul 12, 2020 5:51 am
Sign-up code: 10159

Re: New Engine: ChessMatrix

Post by KristianEkman »

Hi, it is just estimates from my part.
Did you measure 1.0.4 to be 1950?
In that case 1.0.6 should be better.
I couldn't find your results for 1.0.4 anywhere.

Kind Regards
/Kristian
User avatar
Gabor Szots
Posts: 12890
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: New Engine: ChessMatrix

Post by Gabor Szots »

KristianEkman wrote: Tue Aug 11, 2020 11:03 am Hi, it is just estimates from my part.
Did you measure 1.0.4 to be 1950?
In that case 1.0.6 should be better.
I couldn't find your results for 1.0.4 anywhere.

Kind Regards
/Kristian
CCRL has ranking lists, you know, and your engines's results are there.

http://www.computerchess.org.uk/ccrl/40 ... 0_4_64-bit
Post Reply