Komodo 13 released

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Andrew
Posts: 119
Joined: Wed Mar 08, 2006 11:51 pm
Location: Australia

Re: Komodo 13 released

Post by Andrew » Tue May 07, 2019 10:06 am

JJJ wrote:
Tue May 07, 2019 8:24 am
Anyone can use any engine, I don't mind about it. If you prefer Komodo it is ok to me. Each engine adds something.
Exactly JJJ, we need to think beyond just an elo number.

Andrew

Andrew
Posts: 119
Joined: Wed Mar 08, 2006 11:51 pm
Location: Australia

Re: Komodo 13 released

Post by Andrew » Tue May 07, 2019 10:11 am

Thanks very much for the latest update! I've also been reading over the 7 piece EGTB updates you
have supported but it will be a while till I download many of these. I upgrades to a 256 Gb USB drive
recently which was enough to download all the 6 men Syzgy files but now this? :D Oh welll...

Anyway good luck in finding future improvements with Komodo and I hope you can run lots of games
against Stockfish and LCzero which I'm sure will provide a few Aha moments! (regardless of elo considerations)

Andrew

Karol Majewski
Posts: 94
Joined: Sat Jul 09, 2011 2:18 pm

Re: Komodo 13 released

Post by Karol Majewski » Tue May 07, 2019 11:42 am

Guys, just a quick question: Does Komodo MCTS make any use of Hash option or it only uses MCTS Hash option? I want to increase MCTS Hash value but then I have to decrease Hash value. So I wonder if decreasing normal Hash has any impact on analysis.

User avatar
Kirk
Posts: 5630
Joined: Sat Mar 11, 2006 2:44 am

Re: Komodo 13 released

Post by Kirk » Tue May 07, 2019 1:48 pm

lkaufman wrote:
Mon May 06, 2019 11:05 pm
Komodo 13.01 has been released at komodochess.com. Elo gains since the last release, version 12.3, are about ten elo in standard mode, 30 elo in MCTS single thread mode, and over 40 elo in MCTS multi-thread mode. Since the last annual version, Komodo 12, elo gains in MCTS mode are in the 300 to 400 elo range depending on time control and number of threads. We predict that the MCTS version will earn a rating around 3440 on CCRL blitz, and that on most lists it will be ahead of all other CPU engines except for Stockfish, normal Komodo, and Houdini.
While the default mode is still the strongest one for engine vs engine play, the MCTS mode is now stronger whenever MultiPV is set to 2 or more. Since most people use MultiPV for analysis and since the MCTS mode tends to produce more "human" moves because it does not assume that the opponent will play perfectly (by its criteria), we feel that the MCTS mode is now the more useful one for the majority of users, except perhaps for correspondence players. According to my tests, Komodo 13.01 MCTS can now defeat any CPU based engine, including recent Stockfish, if both are set to use MultiPV of 6 or more. This is because Komodo MCTS pays no performance penalty for obtaining this information, while normal alpha-beta engines pay a very heavy price to give analysis for several moves. For users who don't have a good GPU and who use high values of MultiPV, Komodo 13 MCTS should be the strongest engine available regardless of price.
Pricing remains the same (free for subscribers), and we give 20% discount for anyone who bought Komodo 12 or later version or any subscription that included Komodo 12. We have now made Komodo 10 free, and have made Komodo 11 available at a 70% discount. Anyone who bought Komodo 12.3 on or after April 15 can get a free upgrade to Komodo 13.01 on request.
We anticipate continued development of both versions, which may be separated at some future date. Subscribers will get both for a year. Our priority is the MCTS version, but we often find ideas that help both versions. The gap between them is getting rather small (see chess.com computer chess championship to illustrate) and there is no theoretical reason why Komodo MCTS can't pass normal Komodo even without MultiPV use.
An FYI. The sale price for the older Komodo 11 ends up as full price in the cart
“He knew all the tricks, dramatic irony, metaphor, pathos, puns, parody, litotes and... satire. He was vicious”

mjlef
Posts: 1427
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: Komodo 13 released

Post by mjlef » Tue May 07, 2019 2:39 pm

Amstaff wrote:
Tue May 07, 2019 9:53 am
Hello Larry, I just downloaded both version 64 and 32. Win 64 seems to work fine on my machine, but Win 32 has an issue in my older laptop running windows 7. The engine does not seem to start or show any analysis in both GUIs I loaded it in (SCID and Deep Shredder 13).
Maybe someone else has this issue, if not I will simply try another download.
Best wishes,
Gerald
My guess is an overly sensitive antivirus program is blocking it. The antivirus programs seem way more sensitive about 32 bit programs than 64 bit programs. I will follow up with you and see what we can do.

BTW, we did not include a 32 bit version in the last release. Only 1 person asked for it. My last 32 bit ancient laptop died, making it hard for me to test 32 bit versions properly. Maybe I can find something on ebay.

Mark

mjlef
Posts: 1427
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: Komodo 13 released

Post by mjlef » Tue May 07, 2019 3:39 pm

Karol Majewski wrote:
Tue May 07, 2019 11:42 am
Guys, just a quick question: Does Komodo MCTS make any use of Hash option or it only uses MCTS Hash option? I want to increase MCTS Hash value but then I have to decrease Hash value. So I wonder if decreasing normal Hash has any impact on analysis.
Both Hash tables are used. Hash is used by the short alpha-beta searches used to estimate winning probability and the policy (how likely a move is to be played) and MCTS Hash is used to store the Monte Carlo Tree. For regualr home machines the values should be good enough unless you plan to do very long analysis, in which case increase MCTS Hash.

mjlef
Posts: 1427
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: Komodo 13 released

Post by mjlef » Tue May 07, 2019 3:47 pm

Gabor Szots wrote:
Tue May 07, 2019 6:32 am
Thanks Larry.

Where do I set MultiPV under the Shredder GUI?
In the Deep Shredder GUI (I think they are all alike), f6 makes it analyze more lines and f5 less lines. If you right click over the game analysis/PV information, it gives a menu of options.

Mark

mjlef
Posts: 1427
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: Komodo 13 released

Post by mjlef » Tue May 07, 2019 3:50 pm

Andrew wrote:
Tue May 07, 2019 10:11 am
Thanks very much for the latest update! I've also been reading over the 7 piece EGTB updates you
have supported but it will be a while till I download many of these. I upgrades to a 256 Gb USB drive
recently which was enough to download all the 6 men Syzgy files but now this? :D Oh welll...

Anyway good luck in finding future improvements with Komodo and I hope you can run lots of games
against Stockfish and LCzero which I'm sure will provide a few Aha moments! (regardless of elo considerations)

Andrew
We think the KRPPvsKRP and KPPPvs KPP are the most useful 7 piece Syzygy files to use first. Note that it is quite possible to promote a pawn to a queen or other piece, and this can make the engine think it is bad if that specific table is missing, since to the engine, it looks like it has done from a win to a position where it is just vary ahead. I wanted to mention that since it can look odd when you analyze.

User avatar
CMCanavessi
Posts: 835
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: Komodo 13 released

Post by CMCanavessi » Tue May 07, 2019 6:18 pm

Mark, Larry, have you considered updating the (now very very old) free version of Komodo? Maybe something like K 10.4 or something like that would be a nice boost.

Edit: just read on another thread that K 10 is now free, great news!!
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

cma6
Posts: 138
Joined: Thu May 29, 2014 3:58 pm

Re: Komodo 13 released

Post by cma6 » Tue May 07, 2019 7:16 pm

Both Hash tables are used. Hash is used by the short alpha-beta searches used to estimate winning probability and the policy (how likely a move is to be played) and MCTS Hash is used to store the Monte Carlo Tree. For regualr home machines the values should be good enough unless you plan to do very long analysis, in which case increase MCTS Hash.
[/quote]

Hi Mark:
For long correspondence analysis with 64 GB of RAM on system (36-core dual-Xeon) where 4 GB is used for normal Komodo hash, what MCTS hash would you recommend and any change in 4 GB for Komodo overall?
Also with respect to Syzygy settings, should one stick with previously recommended Probe depth of 3 and Probe limit of 6?

Post Reply