Monolith 1.0

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.
User avatar
Scally
Posts: 97
Joined: Thu Sep 28, 2017 7:34 pm
Location: Bermondsey, London
Full name: Alan Cooper
Contact:

Re: Monolith 1.0

Post by Scally » Thu Oct 18, 2018 8:19 pm

Thanks Jonas,

Compiled and running nicely on my Raspberry Pi.

Al.
Alan Cooper
My Chess Computers

cimarronOST
Posts: 7
Joined: Sun Feb 12, 2017 6:14 pm
Location: Italy
Contact:

Re: Monolith 1.0

Post by cimarronOST » Thu Oct 18, 2018 11:35 pm

Thanks all for the helpful replies, there was indeed a typo in the Syzygy probing code, it's fixed now and source and binaries are updated (https://github.com/cimarronOST/Monolith/releases). The bug affects also 5-men Syzygy access even though it only crashes with 6-men, so please use the fixed version when using Syzygy tablebases. I've also removed the pext version for now until I can find hardware to test it on.

Many thanks in advance to all the testers and rating lists, I really appreciate it.

tpoppins
Posts: 919
Joined: Tue Nov 24, 2015 8:11 pm
Location: upstate

Re: Monolith 1.0

Post by tpoppins » Fri Oct 19, 2018 5:15 am

I don't know about the bug, Jonas, it seems to be a good one. Looks like I'll have to restart the ongoing 40/40 gauntlet with the updated exec anyway, and not because of any issues with the original release but because it's scoring 65% over the field ranging from Toga to Coiled, with a clean score against the latter and Jumbo. Apparently basing the lineup on half your Elo increase estimate was uh, a bit too conservative of me. ;) After 150 games it does look like it may be closer to +200 than +100.
Tirsa Poppins
CCRL

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

Re: Monolith 1.0

Post by CMCanavessi » Fri Oct 19, 2018 1:18 pm

tpoppins wrote:
Fri Oct 19, 2018 5:15 am
After 150 games it does look like it may be closer to +200 than +100.
Well yes, he said +190 in the 1st post :lol:
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

tpoppins
Posts: 919
Joined: Tue Nov 24, 2015 8:11 pm
Location: upstate

Re: Monolith 1.0

Post by tpoppins » Fri Oct 19, 2018 6:54 pm

Yes, he did; the Elo difference shown in self-play is usually about 2X the difference shown in tests vs. general opposition.
Tirsa Poppins
CCRL

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

Re: Monolith 1.0

Post by CMCanavessi » Fri Oct 19, 2018 7:56 pm

tpoppins wrote:
Fri Oct 19, 2018 6:54 pm
Yes, he did; the Elo difference shown in self-play is usually about 2X the difference shown in tests vs. general opposition.
Well yes, I've noticed something like that, but usually fron engines stronger than, say, 3000-3100 elo. For weaker engines usually self-play elo translates quite well to general elo, at least in my tests.
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

Fuddur
Posts: 36
Joined: Sun Mar 18, 2018 5:35 am

Re: Monolith 1.0

Post by Fuddur » Sat Oct 20, 2018 6:16 am

I'll also add Android compiles soon. A pity I'm always a couple of days too late with my releases for Graham's ongoing Division tourneys
Hi, you made a nice progress, waiting for android release!!
Thanks

Post Reply