Koivisto v4.0 release

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
EugenioBruno
Posts: 2
Joined: Sun Oct 18, 2020 12:07 pm
Full name: Eugenio Bruno

Koivisto v4.0 release

Post by EugenioBruno » Mon Nov 09, 2020 8:36 pm

Koivisto v4.0

A while ago, we released Koivisto v3.0 after a significant improvement in strength. It's time for another release again :)

We'd like to thank Andrew Grant for the OpenBench framework (and putting up with our questions about it!), without which we wouldn't be able to test so many new ideas; it has greatly improved our experimentation process since we started using it around Koivisto v2.0.
We'd also like to thank noobpwnftw for significant help with testing.

Our latest regression test shows significant Elo gain over Koivisto v3.0:

Code: Select all

60+0.6:
ELO   | 141.20 +- 18.81 (95%)
SPRT  | 60.0+0.6s Threads=1 Hash=64MB
LLR   | 0.00 (-2.94, 2.94) [0.00, 0.00]
Games | N: 960 W: 501 L: 131 D: 328

15+0.1:
ELO   | 120.88 +- 13.38 (95%)
SPRT  | 15.0+0.15s Threads=1 Hash=16MB
LLR   | 0.00 (-2.94, 2.94) [0.00, 0.00]
Games | N: 1952 W: 1002 L: 349 D: 601
Furthermore, from cyclic time controls, one can expect significantly higher performance than before. Testing by CCRL members made issues apparent in how we handled cyclic time controls. This test shows the same version of the engine, but using the new cyclic time control manager.

Code: Select all

40/15:
ELO   | 84.12 +- 20.83 (95%)
SPRT  | 40/15s Threads=1 Hash=16MB
LLR   | 3.11 (-2.94, 2.94) [0.00, 5.00]
Games | N: 640 W: 260 L: 108 D: 272
We are always thankful for community testing. In particular we're anxious to see whether our SMP scales as well as we hope :)

Changes:

- updated tt-usage.
- does forced moves immediately.
- added bishop pawn endgame relations evaluation term.
- retuned entire evaluation with a larger dataset.
-> tuned mobility scores instead of computing them.
-> tuned king safety scores instead of computing them.
- cleaned up the code and added comments.
- asymmetric piece square tables based on king-king relation.
- changed reduction/extension policy.
- improved general time management.
- fixed cyclic time management which leads to the engine using a lot more time in games with cyclic time controls.
- fixed speed-related issues.
- added a logo and banner.

The Koivisto engine authors, Kim, Finn and Eugenio.

bastiball
Posts: 45
Joined: Tue Oct 20, 2020 2:18 am
Full name: Basti Dangca

Re: Koivisto v4.0 release

Post by bastiball » Mon Nov 09, 2020 11:37 pm

It won't start, it says "The procedure entry point pthrerad_getaffinity_np could not be located in the dynamic library *files path to the engine"

Koivisto
Posts: 19
Joined: Fri Sep 04, 2020 8:30 pm
Full name: Kim Kahre

Re: Koivisto v4.0 release

Post by Koivisto » Mon Nov 09, 2020 11:48 pm

bastiball wrote:
Mon Nov 09, 2020 11:37 pm
It won't start, it says "The procedure entry point pthrerad_getaffinity_np could not be located in the dynamic library *files path to the engine"
Seems like the windows binaries are dynamically linked. I don't have access to a windows machine so I notified Finn and Eugenio.

bastiball
Posts: 45
Joined: Tue Oct 20, 2020 2:18 am
Full name: Basti Dangca

Re: Koivisto v4.0 release

Post by bastiball » Tue Nov 10, 2020 12:08 am

Forgot to tell that I use Window Binaries and I'll wait for the update, thanks a lot.

Luecx
Posts: 25
Joined: Thu Jun 18, 2020 7:20 pm
Full name: Finn Eggers

Re: Koivisto v4.0 release

Post by Luecx » Tue Nov 10, 2020 5:53 am

bastiball wrote:
Tue Nov 10, 2020 12:08 am
Forgot to tell that I use Window Binaries and I'll wait for the update, thanks a lot.

Okay so I had a quick look and you are right, the compiles dont seem to work properly.
If you are able to compile, try compiling it yourself using:

Code: Select all

make native
which should produce multiple statically compiled binaries.

I reuploaded the static compiles. Hopefully it will work now :)

bastiball
Posts: 45
Joined: Tue Oct 20, 2020 2:18 am
Full name: Basti Dangca

Re: Koivisto v4.0 release

Post by bastiball » Tue Nov 10, 2020 6:44 am

It's working now. Thanks :D :D

Gabor Szots
Posts: 757
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

Re: Koivisto v4.0 release

Post by Gabor Szots » Tue Nov 10, 2020 11:28 am

Hi Eugenio,

I'm going to test it as soon as I have a chance but my Koivisto 3.0 test is still running. You know, that 'a while ago' was in fact only 3 weeks ago...
Gabor Szots
CCRL testing group

User avatar
Sylwy
Posts: 3572
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI (Romania) - the historical capital of MOLDOVA

Re: Koivisto v4.0 release

Post by Sylwy » Tue Nov 10, 2020 8:26 pm

My test started:

Image

EugenioBruno
Posts: 2
Joined: Sun Oct 18, 2020 12:07 pm
Full name: Eugenio Bruno

Re: Koivisto v4.0 release

Post by EugenioBruno » Wed Nov 11, 2020 5:36 pm

Gabor Szots wrote:
Tue Nov 10, 2020 11:28 am
Hi Eugenio,

I'm going to test it as soon as I have a chance but my Koivisto 3.0 test is still running. You know, that 'a while ago' was in fact only 3 weeks ago...
Haha, indeed! Thanks you for the testing - I believe it was actually some of your games that showed us issues with our time management.

I (unfortunately) don't think we'll be able to extrapolate this release frequency in the future. We've had a very, very rough guideline of making a new major release if we get +100 vs the previous version, and since the regression test was +140, we decided to release - with apologies and thanks to all the testers for the headaches caused! :)

User avatar
Sylwy
Posts: 3572
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI (Romania) - the historical capital of MOLDOVA

Re: Koivisto v4.0 release

Post by Sylwy » Wed Nov 11, 2020 7:48 pm

EugenioBruno wrote:
Wed Nov 11, 2020 5:36 pm
....if we get +100 vs the previous version, and since the regression test was +140, we decided to release - with apologies and thanks to all the testers for the headaches caused! :)
After 60 games in my test (TC=4'+2"-around 12-15 minutes per game) Koivisto 4.0 shows near +40 Elo points over Koivisto 3.0. Let's see what will be in the end.

Image

Image

Post Reply