Perft Positions and Node Counts

[ChessRenewal]

In the video series Programming A Chess Engine in C, I think I recall a file being displayed with numerous perft positions and node counts - very useful for testing an engine's move generation. Is that file or a similar one available?

[smatovic]

Idk about the video series, here two .epd files with perft positions from chess programming wiki:

https://gitlab.com/smatovic/ZetaDva/-/b ... type=heads
https://gitlab.com/smatovic/ZetaDva/-/b ... type=heads

--
Srdja

[Steve Maughan]

Etherial has two perft files that I've used to prove a move generator is working perfectly:

https://github.com/AndyGrant/Ethereal/t ... /src/perft

Steve

[Ajedrecista]

Hello Jay:

In the video series Programming A Chess Engine in C, I think I recall a file being displayed with numerous perft positions and node counts - very useful for testing an engine's move generation. Is that file or a similar one available?

Videos 43 and 44 of the excellent series Programming A Chess Engine in C show an EPD file called perftsuite.epd. A comment by the author on the Part 44 says the following:

[...]

Plenty of engines do divide, I used to use and engine called 'Roce'.

In my experience, if you have very few differences, try positions with castling, promotions and enpassant. I think I provide a load of test positions in this series.

Here's where I got it:
http://www.rocechess.ch/perft.html

[...]

Going to the web and downloading the file does the job. Once unzipped, you will see that the file is the same as in the series... or almost. I saw that perft(6) of the second position of the file (the famous Kiwipete position) is missing in the file from the Roce site, while is present in the EPD of the video.

Regards from Spain.

Ajedrecista.

[ydebilloez]

2 testfiles:
https://sourceforge.net/p/belofte/gitre ... tsuite.epd (162 positions)
https://sourceforge.net/p/belofte/gitre ... e-full.epd (6800 positions)