Mida v2.3 released

[Jack03]

Hello, I'm happy to announce the release of Mida v2.3 !
The updates in the engine itself are small, and should give a +30 ELO gain against v2.2.
The greatest update is the makefile, that now makes Mida available on all OS, and also provides optimizations according to the hardware architecture.
So, if the engine is compiled and used on a computer that supports some of these optimizations, the difference with the previous version will surely be more evident.
If there are any problems with compiling, please let me know by writing here or by opening an issue on github.
Thanks to everybody who has downloaded Mida so far, and thanks to all the CCRL testers!

https://github.com/GiacomoPorpiglia/Mida

[Gabor Szots]

Hello Giacomo,

The makefile works OK.

I have noticed something, though, which is present in previous Mida versions as well. After issuing uci from the command line, nothing happens, I have to type a lot of garbage before an analysis output is obtained. To make it easier to understand what I mean, here is a snippet from my Mida communication.

Hash table Initialized with size 64 MB

Loading NNUE : mida-nn.nnue
NNUE loaded !
id name MIDA 2.3
id author Giacomo Porpiglia
uciok
isready
position startpos
go depth 20
go infinite
e2e4
readyok
position startpos
go depth 20

go infinite
d2d4
uci
wtf!
id name MIDA 2.3
id author Giacomo Porpiglia
setoption name Hash type spin default 64 min 4 max 1024
uciok
go infinite
uci
e2e4
info score cp 26 depth 1 nodes 25 nps 0 time 0 pv e2e4
info score cp 55 depth 2 nodes 106 nps 0 time 0 pv e2e4 e7e5
info score cp 34 depth 3 nodes 321 nps 0 time 0 pv e2e4 c7c5 d2d4
info score cp 60 depth 4 nodes 598 nps 0 time 0 pv e2e4 c7c5 g1f3 b8c6
info score cp 22 depth 5 nodes 1395 nps 0 time 0 pv e2e4 c7c5 g1f3 b8c6 b1c3
info score cp 60 depth 6 nodes 2335 nps 0 time 0 pv e2e4 c7c5 g1f3 b8c6 b1c3 g8f6
info score cp 34 depth 7 nodes 6359 nps 397437 time 16 pv e2e4 e7e5 g1f3 b8c6 f1c4 f8c5 c2c3
info score cp 63 depth 8 nodes 10154 nps 317312 time 32 pv e2e4 e7e5 g1f3 b8c6 f1c4 g8f6 b1c3 f8b4
info score cp 36 depth 9 nodes 21978 nps 348857 time 63 pv e2e4 e7e5 g1f3 b8c6 d2d4 e5d4 f3d4 f8c5 d4b3
info score cp 71 depth 10 nodes 34344 nps 365361 time 94 pv e2e4 e7e5 g1f3 g8f6 d2d4 f6e4 f1d3 d7d5 d4e5 b8c6
info score cp 58 depth 11 nodes 52603 nps 420824 time 125 pv e2e4 e7e5 g1f3 b8c6 d2d4 e5d4 f3d4 f8c5 d4b3 c5b6 b1c3
info score cp 59 depth 12 nodes 76843 nps 408739 time 188 pv e2e4 e7e5 g1f3 b8c6 d2d4 e5d4 f3d4 g8f6 b1c3 f8b4 c1g5 e8g8

info score cp 60 depth 13 nodes 127255 nps 428468 time 297 pv e2e4 e7e5 g1f3 b8c6 d2d4 e5d4 f3d4 g8f6 c1g5 c6d4 d1d4 h7h6 g5h4
info score cp 75 depth 14 nodes 328302 nps 456609 time 719 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5c6 d7c6 e1g1 g8f6 f3e5 d8d4 e5f3 d4e4
info score cp 46 depth 15 nodes 747185 nps 450926 time 1657 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5c6 d7c6 d2d3 f8d6 d3d4 e5d4 d1d4 f7f6 c1e3
info score cp 77 depth 16 nodes 1091565 nps 456530 time 2391 pv e2e4 c7c5 g1f3 b8c6 f1b5 e7e6 e1g1 g8e7 c2c3 a7a6 b5c6 e7c6 d2d4 d7d5 c1g5 f7f6
info score cp 71 depth 17 nodes 1534464 nps 452510 time 3391 pv e2e4 c7c5 g1f3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 e7e5 d4b5 d7d6 c3d5 f6d5 e4d5 c6b8 d1e2
info score cp 71 depth 18 nodes 2256449 nps 454105 time 4969 pv e2e4 c7c6 d2d4 d7d5 e4e5 c6c5 d4c5 e7e6 b1c3 b8c6 f1b5 c8d7 b5c6 d7c6 c1e3 d5d4 d1d4 d8d4
info score cp 71 depth 19 nodes 2649405 nps 454521 time 5829 pv e2e4 c7c6 d2d4 d7d5 e4e5 c6c5 d4c5 e7e6 b1c3 b8c6 f1b5 c8d7 b5c6 d7c6 c1e3 d5d4 d1d4 d8d4 e3d4
info score cp 71 depth 20 nodes 3417523 nps 454699 time 7516 pv e2e4 c7c6 d2d4 d7d5 e4e5 c6c5 d4c5 e7e6 b1c3 b8c6 f1b5 c8d7 b5c6 d7c6 c1e3 d5d4 d1d4 d8d4 e3d4 c6g2
info score cp 70 depth 21 nodes 12725184 nps 454974 time 27969 pv g1f3 d7d5 d2d4 g8f6 c2c4 e7e6 c1g5 f8e7 e2e3 d5c4 f1c4 c7c5 b1c3 c5d4 f3d4 e8g8 e1g1 h7h6 g5h4 c8d7 c4e2

I have experienced exactly the same thing with another engine (I don't remember which one) and I think it was some command parsing problem.

[Jack03]

Yes, I'm aware of the problem. The issue seems to be complicated, though, because it appeared out of nothing without me changing anything in the actual CLI code.
I have experienced the same problem with Berserk 12 (already with 11.1) (maybe it's the same engine you talked about), and I was "relieved", because that probably means it's not a code-related problem. My supposition is that it's related to the C input functions, somehow...
Anyway, it doesn't seem to be giving any problem in Cute-Chess or any other testing program that I know, so that's good at least!

[Gabor Szots]

Yes, it was Berserk. And the solution has been found as well. Look here:

https://github.com/jhonnold/berserk/issues/518

[Jack03]

That’s some good news! I’ll try to fix the issue when i can.
Thanks for the useful info!

[Jack03]

Edit: the fix was surprisingly easy (thanks to the Berserk issue you provided) so I have already updated the release with the fixed code that now works on my end. Thanks for pointing this out, I've been scratching my head around this problem for a while

[Gabor Szots]

Here it works too. Also, the makefile is excellent, the exe it produces is about 20 % faster than the one at your site. Good job!

[Jack03]

Thanks! I also wanted to say I solved the issue of the NNUE file being unnecessairly created every time (that was pointed out in a previous release).
I've updated v2.3 once more with this small fix. Hopefully that's the last issue I find for now

[Gabor Szots]

Mida 2.3 exceeds time in about every 12th game. I did not notice such behaviour by previous versions.

[Jack03]

Hi Gabor,
sorry for the late response. I will look into it, I changed some code in time management to make it compatible with all OS, so it's possible.
Thanks