Cfish

[Joachim26]

This is a quite dumb fake because Syzygy aka RdM is obviously a great programmer and would never ever write the date like "010422".
Edit: Even hobby programmers like me write 230222
plus the NNUEv4/5 net plus RdM has retired from CFish dev
Thus: 100% Fake

Ok, good to know, thank you!

To be 100.00% sure you could ask him. He is quit active in the moment on talkchess:
http://www.talkchess.com/forum3/viewtopic.php?f=2&t=81509&start=20

[Joachim26]

There is another clear hint in the date, that it's not a CFish
0104 = 1. April = April 1st a lil Joker the Faker...

[SkyNet]

So dear friend, buddy, mate .. no more grandma's slippers right ?
Night Nurse's then ..

I'm mostly interested for android engines since SF became toooo slow, if the newer engine win against Spider (which i made long time ago, from Cfish code), then i will use the newer one. As for PC versions - i don't care, SF-PB is the only i use/need.

[Joachim26]

So dear friend, buddy, mate .. no more grandma's slippers right ?
Night Nurse's then ..

I'm mostly interested for android engines since SF became toooo slow, if the newer engine win against Spider (which i made long time ago, from Cfish code), then i will use the newer one. As for PC versions - i don't care, SF-PB is the only i use/need.

My guess is: CFishNNFatBuild vs Spider LOS: 99.9%
What did you change back then in Spider?

[Joachim26]

So dear friend, buddy, mate .. no more grandma's slippers right ?
Night Nurse's then ..

I'm mostly interested for android engines since SF became toooo slow, if the newer engine win against Spider (which i made long time ago, from Cfish code), then i will use the newer one. As for PC versions - i don't care, SF-PB is the only i use/need.

If you are looking for another fast AND strong Android engine try this one:
https://github.com/Joachim26/StockfishNPS/releases/download/Win_modern_and_armv8_dev_release/SFSnpsManComp_armv8
It's SFsmallNPS manually pgo compiled and needs vondele's 12 MB net as described in the release section.

On Windows SFSnps is still a bit stronger than CFishNN. On Android it could be vice versa, because NEON sucks
Edit: Both engines have a "pure nnue" uci-option, but I never tested them against each other with 'pure' enabled.

[SkyNet]

My guess is: CFishNNFatBuild vs Spider LOS: 99.9%
-------------------------------------------------------
What did you change back then in Spider?

Don't know yet, your engine and Leo's crashes in cutechess on android.
-----------------------------------------------------
I didn't do much (together with ChessMan), i mean without support for the bigger network, the more i added new patches, the worse the engine became, in addition, some patches were impossible to add because i'm not an expert in C language (or C++ eather).
Regarding the added patches over RDM's Cfish, you can look here http://outskirts.altervista.org/forum/viewtopic.php?p=31970&view=single_post#p31970 More updated Spider (which became Cfish again) here http://outskirts.altervista.org/forum/viewtopic.php?p=33891&view=single_post#p33891 Source code is gone because of the expired link, and i didn't kept a copy.

[Joachim26]

My guess is: CFishNNFatBuild vs Spider LOS: 99.9%
-------------------------------------------------------
What did you change back then in Spider?

Don't know yet, your engine and Leo's crashes in cutechess on android.
-----------------------------------------------------
I didn't do much (together with ChessMan), i mean without support for the bigger network, the more i added new patches, the worse the engine became, in addition, some patches were impossible to add because i'm not an expert in C language (or C++ eather).
Regarding the added patches over RDM's Cfish, you can look here http://outskirts.altervista.org/forum/viewtopic.php?p=31970&view=single_post#p31970 More updated Spider (which became Cfish again) here http://outskirts.altervista.org/forum/viewtopic.php?p=33891&view=single_post#p33891 Source code is gone because of the expired link, and i didn't kept a copy.

Thanks for the two links, sources doesn't matter, since Sopel, I, and others now exactly could reproduce the code you have used for compiling. Everything's fine with GPLv3 now .
It's interesting that you and Chessman got regression soon after applying only few patches. When I started I expected that this would happen sooner or later also to my dev builds, after skipping the update to NNUEv3 net architecture.
So I did a lot of testing, after applying 1, 2, or max 3 commits, and tweaked the code a bit, if there was some obvious regression. After about 100 commits I'm still quite sure, that the released (win) builds slowly but steadily get stronger.
Concerning the Android builds: I don't make tournaments but I check from time to time the bench on Termux (and compare it with the win build) and do engine selfplay games in Droidfish. Sopel did an impressively great job concerning the NEON support in CFish but missed one(!) #define and didn't know about that till yesterday.

[SkyNet]

Thanks for the two links, sources doesn't matter, since Sopel, I, and others now exactly could reproduce the code you have used for compiling. Everything's fine with GPLv3 now .
It's interesting that you and Chessman got regression soon after applying only few patches. When I started I expected that this would happen sooner or later also to my dev builds, after skipping the update to NNUEv3 net architecture.
So I did a lot of testing, after applying 1, 2, or max 3 commits, and tweaked the code a bit, if there was some obvious regression. After about 100 commits I'm still quite sure, that the released (win) builds slowly but steadily get stronger.
Concerning the Android builds: I don't make tournaments but I check from time to time the bench on Termux (and compare it with the win build) and do engine selfplay games in Droidfish. Sopel did an impressively great job concerning the NEON support in CFish but missed one(!) #define and didn't know about that till yesterday.

Yep, i forgot to mention if you begin to examine the Sopel's patches from the beginning you'll see that one (very simple) patch is missing, it my be the one that you mentioned.
Regarding your updates, i would suggest to remove the patch with "persistent ttsearch - persistent ttdeserialize", it is useless for the current network that you use. Also remove the Robert Houdart's patch, it is useless.
Regarding your tests, i would suggest you to test your updated version vs stable version eg your update vs Cf 13 or SF 14.
You see the the thing is that SF team testing SF master vs new update using thousand of games to see the effect of the specific patch, but in general it doesn't mean that the new patch will be good in general. You can found the regression/progression tests on other sites, and sometimes the results of development versions are not good against the stable versions.
Anyway, if you need any help you can PM me and when have a free time i''l help you with everything i can.

[Joachim26]

Thanks for the two links, sources doesn't matter, since Sopel, I, and others now exactly could reproduce the code you have used for compiling. Everything's fine with GPLv3 now .
It's interesting that you and Chessman got regression soon after applying only few patches. When I started I expected that this would happen sooner or later also to my dev builds, after skipping the update to NNUEv3 net architecture.
So I did a lot of testing, after applying 1, 2, or max 3 commits, and tweaked the code a bit, if there was some obvious regression. After about 100 commits I'm still quite sure, that the released (win) builds slowly but steadily get stronger.
Concerning the Android builds: I don't make tournaments but I check from time to time the bench on Termux (and compare it with the win build) and do engine selfplay games in Droidfish. Sopel did an impressively great job concerning the NEON support in CFish but missed one(!) #define and didn't know about that till yesterday.

Yep, i forgot to mention if you begin to examine the Sopel's patches from the beginning you'll see that one (very simple) patch is missing, it my be the one that you mentioned.
Regarding your updates, i would suggest to remove the patch with "persistent ttsearch - persistent ttdeserialize", it is useless for the current network that you use. Also remove the Robert Houdart's patch, it is useless.
Regarding your tests, i would suggest you to test your updated version vs stable version eg your update vs Cf 13 or SF 14.
You see the the thing is that SF team testing SF master vs new update using thousand of games to see the effect of the specific patch, but in general it doesn't mean that the new patch will be good in general. You can found the regression/progression tests on other sites, and sometimes the results of development versions are not good against the stable versions.
Anyway, if you need any help you can PM me and when have a free time i''l help you with everything i can.

Thanks for your offer to help me. At the moment you are already doing this by performing the first Android tournaments with CFNN.armv8. As I mentioned earlier, I'm quite curious about the results, which then can be compared with the Windows tournaments.
Above there are several things which could be discussed. It's just to late to do this now. But one point I want to clearify immediately:
CFNN does show the exact nps output since I have, certainly, corrected it in position.c:
...
pos->sideToMove = !pos->sideToMove;
pos->nodes++;
set_check_info(pos);
...

I guess that Leon didn't change it (his build "seems" to be faster), but I'm to lacy now to search his sources.
Fun fact of the story is, that I prepared a '"funny" Pascal-style' PR on Sep 2, 2021 (after FTv1.0 release) to correct the nps output, but Sopel just rejected it back then.
https://github.com/Sopel97/FatTitz/commit/cc85ccff23aad1d2391781bcbecc34d6cf7731ac

Greetings from here and me to Poland, Tomasz, have a nice weekend

Edit: Here his reaction, today more funny for me than my PR
https://github.com/Sopel97/FatTitz/pull/8

[Joachim26]

i.e. in this post here:
http://outskirts.altervista.org/forum/viewtopic.php?p=53263&view=single_post#p53263

Cfish 190223, a renamed CFishNN_dev, i.e. MY CFISH, is doing quite well in a Tournament against SFdev, CorChessDev and some other engines. This is an interesting Tournament. I hope it will not be terminated soon. If the tournament is running long enough, let's make a deal.
Hey JKS, if you stop renaming my engine and just use the exact name, I will stop using PWs! Deal?

Edit: Man, 15 points and already stopped....
Anyway, please use MY exact name on your blog, and we can get rid of this f... PW.

[Joachim26]

Thanks to Anton (Dohne ) and SkyNet I now know where my name is coming from:

https://en.m.wikipedia.org/wiki/Joaqu%C3%ADn#:~:text=Joaqu%C3%ADn%20ist%20ein%20spanischer%20m%C3%A4nnlicher,Bedeutung%20des%20Namens%20siehe%20hier.

BTW: 'Dohne' is this 'guy' here:
https://banksiagui.com/forums/viewtopic.php?t=102&start=29
I think it is time again for me to leave the public part of this forum.

Edit: Neither benkrik nor I have send something to Dohne. I don't communicate with such Trolls if not absolutely necessary.

[Joachim26]

Top
Anton101
Re: Charisma NY23
Post Fri Mar 10, 2023 5:49 pm

Joaquin, don't play dumb, I'm not a Troll and I'm not what you write... Keep developing CFish NN, no one uses it anyway, perhaps some clueless... Rubbish and waste of time. And please don't be two-faced.

Well, I already gave you a lot of my time, without a doubt in a short time your developments will be forgotten, since nobody is interested.


Troll Dohne (Anton) doesn't understand that CERTAINLY CFishNN with it's old NNUEv2 net is not really competitive to SFdev clones like his ones with the newest net architecture NNUEv5. It was NEVER my intention to compete with these clones. I just want to learn a bit about GitHub and "engine development in a nutshell".
I certainly continue with this on GitHub, but I certainly will also stop today with reacting to such clueless kids.
Cheers

[EzioMagnifico]

CFish NN 230626 .

UCI chess engine, Rating CEDR=3673 . ssfgmfgsoke

Download : https://workupload.com/file/4VzXcaVaG8S

Tests vs last release:

TC 5+0.05s
Score of CFishNN_230626_modern vs CFishNN_230521_modern: 243 - 224 - 533 [0.509]
Elo difference: 6.6 +/- 14.7, LOS: 81.0 %, DrawRatio: 53.3 %
1000 of 1000 games finished.

TC 10+0.1s
Score of CFishNN_230626_modern vs CFishNN_230521_modern: 250 - 203 - 547 [0.523]
Elo difference: 16.3 +/- 14.5, LOS: 98.6 %, DrawRatio: 54.7 %
1000 of 1000 games finished.

TC 25+0.25s
Score of CFishNN_230626_modern vs CFishNN_230521_modern: 221 - 202 - 577 [0.509]
Elo difference: 6.6 +/- 14.0, LOS: 82.2 %, DrawRatio: 57.7 %
1000 of 1000 games finished.

nn-46832cfbead3.nnue

[alainprost]

There hasn't been a new engine share for a long time ???