DON Chess Engine

[Modern Times]

He can't seriously expect people to want to use this engine when it is so unstable. Also, is this just another Stockfish derivative ? In both cases I see no reason to use it. If what I've said is correct, just use Stockfish.

[Elorejano]

Dont use it, Its your right.
But please, don't give your opinion on what the rest of us should do.
Thanks

[Sylwy]

1.-He can't seriously expect people to want to use this engine when it is so unstable. 2.-Also, is this just another Stockfish derivative ? 3.- In both cases I see no reason to use it. If what I've said is correct, just use Stockfish.

1.-Error! The versions I tested (bmi2) work perfectly!
2.-Without a doubt DON is a Stockfish derivative....but it has "something" that makes it perform better. It leads my tournament from the beginning and is constantly distancing itself from the best Stockfish so far (see UHO-Top15 rating list of Stefan Pohl).
3.-I don't want to be radical. I use both. After all, I care about the quality of the chess played. I will also add Stockfish 18 to my tournament (when it comes out), as well as Rashid's "crafted" version based on SF18.

[Chacal X]

Well, if it only works on BM12, then it shouldn't be available for download on SSE41 or SSE3, which don't work for anyone.

[Sylwy]

Well, if it only works on BM12, then it shouldn't be available for download on SSE41 or SSE3, which don't work for anyone.

+1

[Sylwy]

My (and ChatGPT's, of course) logo proposal:

[erashid]

DON – New Release Announcement

Following the successful release of DON 1.0, now excited to introduce the next evolution of the engine.

DON is a modern chess engine focused on search quality, efficiency, and practical playing strength. This new release builds upon the solid foundation of version 1.0, delivering meaningful improvements in stability, evaluation accuracy, and overall performance.

Key areas of enhancement include refined search behavior, improved decision consistency, and various internal optimizations to achieve stronger, more reliable play. As with previous versions, the primary goal remains clear: maximizing effective strength through smarter search rather than unnecessary complexity.

This update represents another step forward in DON’s development, and further improvements are already underway.

Stay tuned for future updates.

Try to enjoy.
I uploaded to 3 servers.

1. mediafire
https://www.mediafire.com/folder/06ewg9vo8mfuj/DON

2. mega
https://mega.nz/folder/Cc1zkbAR#bnycmCX38aGjK6YavnMWqw

3. pixeldrain
https://pixeldrain.com/l/cvMpTsdy

I uploaded the following builds:

Windows:
1. DON-windows-armv8
2. DON-windows-armv8-dotprod
3. DON-windows-x86-64-ssse3
4. DON-windows-x86-64-ssse3-popcnt
5. DON-windows-x86-64-sse41-popcnt
6. DON-windows-x86-64-avx2
7-. DON-windows-x86-64-bmi2
7*. DON-windows-x86-64-bmi2-comp
8-. DON-windows-x86-64-avxvnni
8*. DON-windows-x86-64-avxvnni-comp
9-. DON-windows-x86-64-avx512
9*. DON-windows-x86-64-avx512-comp

Ubuntu:
1. DON-ubuntu-x86-64-sse41-popcnt
2. DON-ubuntu-x86-64-avx2
3-. DON-ubuntu-x86-64-bmi2
3*. DON-ubuntu-x86-64-bmi2-comp
4-. DON-ubuntu-x86-64-avxvnni
4*. DON-ubuntu-x86-64-avxvnni-comp

macOS:
1. DON-macos-x86-64-sse41-popcnt
2. DON-macos-x86-64-avx2
3-. DON-macos-x86-64-bmi2
3*. DON-macos-x86-64-bmi2-comp
4. DON-macos-m1-apple-silicon

Andriod:
1. DON-android-armv7
2. DON-android-armv7-neon
3. DON-android-armv8
4. DON-android-armv8-dotprod

*NOTE: *comp = Compressed version
- ~50% lower memory usage, typically 1–2% faster.
- Recommended for long gameplay and analysis.

Thanks
Ehsan Rashid

[erashid]

1. DON dev-20260215 won a major tournament of new chess engine versions.
2. HypnoS EN 1.0 came in second, and
3. Stockfish 18 JA came in third.

Tcheran 11.0 had trouble running in the Banksia GUI.


https://scontent.fkhi2-3.fna.fbcdn.net/ ... e=69A0B609

Thanks

[Rebel]

No source code once again.

[Ras]

Moderation: Source code missing again. No further posts until provided.

And also...

DON is a modern chess engine focused on search quality, efficiency, and practical playing strength. This new release builds upon the solid foundation of version 1.0, delivering meaningful improvements in stability, evaluation accuracy, and overall performance.

Key areas of enhancement include refined search behavior, improved decision consistency, and various internal optimizations to achieve stronger, more reliable play. As with previous versions, the primary goal remains clear: maximizing effective strength through smarter search rather than unnecessary complexity.

This update represents another step forward in DON’s development, and further improvements are already underway.

You didn't write this. ChatGPT (or another LLM) did.