Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

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
Nordlandia
Posts: 2634
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by Nordlandia » Thu Nov 14, 2019 5:14 pm

Any recommended tweaks for slightly enhanced performance?

CPUCT or similar tweaks.

https://github.com/LeelaChessZero/lc0/wiki/Lc0-options

Based of this thread:
viewtopic.php?f=2&t=70035

corres
Posts: 3083
Joined: Wed Nov 18, 2015 10:41 am
Location: hungary

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by corres » Thu Nov 14, 2019 6:22 pm

Nordlandia wrote:
Thu Nov 14, 2019 5:14 pm
Any recommended tweaks for slightly enhanced performance?
...
As the power of "Super" is higher only about 10% those parameters what is good for normal 2080 Ti also good for Super.
There are no "ideal" parameters because they depends on the used NET and TC, but in the most cases are good for lc0 ver.0.22.0
Threads = 2
Backend = cudnn-fp16
NNCacheSize = 10000000
any other are Default.

For LTC may be better
Threads = 2
Backend = cudnn-fp16
NNCachesize = 20000000
MiniBatchSize = 512
CPuct = 3.4
CPuctBase = 10000

User avatar
Nordlandia
Posts: 2634
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by Nordlandia » Fri Nov 15, 2019 2:58 am

I usually test casual engine games at 30m+30s time control. And recently two computer match via Fritz engine cloud with ponder enabled.

So 30 minute + 30 second increment in addition to ponder enabled. i7-5960X 4.5GHz and i7-5930K 4.5GHz. So the main hardware is fairly even.

Stockfish on 5960X and Leela on 5930K.

corres
Posts: 3083
Joined: Wed Nov 18, 2015 10:41 am
Location: hungary

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by corres » Fri Nov 15, 2019 2:28 pm

Nordlandia wrote:
Fri Nov 15, 2019 2:58 am
I usually test casual engine games at 30m+30s time control. And recently two computer match via Fritz engine cloud with ponder enabled.
So 30 minute + 30 second increment in addition to ponder enabled. i7-5960X 4.5GHz and i7-5930K 4.5GHz. So the main hardware is fairly even.
Stockfish on 5960X and Leela on 5930K.
It depends on the used NET and other factors (Stockfish params, opening book, etc.)

User avatar
Ponti
Posts: 483
Joined: Wed Mar 15, 2006 5:13 am
Location: Curitiba - PR - BRAZIL
Contact:

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by Ponti » Fri Nov 15, 2019 4:21 pm

corres wrote:
Fri Nov 15, 2019 2:28 pm
Nordlandia wrote:
Fri Nov 15, 2019 2:58 am
I usually test casual engine games at 30m+30s time control. And recently two computer match via Fritz engine cloud with ponder enabled.
So 30 minute + 30 second increment in addition to ponder enabled. i7-5960X 4.5GHz and i7-5930K 4.5GHz. So the main hardware is fairly even.
Stockfish on 5960X and Leela on 5930K.
It depends on the used NET and other factors (Stockfish params, opening book, etc.)
Which one is the current best version of the Leela neural network ?
The binary my computer uses (Ryzen 1800X) is lc0-v0.19.1.1-windows-opencl. File NN is 32141.
I don't own a great video-card (using a Radeon Sapphire HD 7790), though I'm planning to buy a new one.

Thanks in advance,
A. Ponti
A. Ponti
AMD Ryzen 1800x, Windows 10.
FIDE current ratings: standard 1913, rapid 1931

corres
Posts: 3083
Joined: Wed Nov 18, 2015 10:41 am
Location: hungary

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by corres » Fri Nov 15, 2019 8:13 pm

Ponti wrote:
Fri Nov 15, 2019 4:21 pm
corres wrote:
Fri Nov 15, 2019 2:28 pm
Nordlandia wrote:
Fri Nov 15, 2019 2:58 am
I usually test casual engine games at 30m+30s time control. And recently two computer match via Fritz engine cloud with ponder enabled.
So 30 minute + 30 second increment in addition to ponder enabled. i7-5960X 4.5GHz and i7-5930K 4.5GHz. So the main hardware is fairly even.
Stockfish on 5960X and Leela on 5930K.
It depends on the used NET and other factors (Stockfish params, opening book, etc.)
Which one is the current best version of the Leela neural network ?
The binary my computer uses (Ryzen 1800X) is lc0-v0.19.1.1-windows-opencl. File NN is 32141.
I don't own a great video-card (using a Radeon Sapphire HD 7790), though I'm planning to buy a new one.
Thanks in advance,
A. Ponti
It is your task to find it or ask peoples who use opencl.

shrapnel
Posts: 1338
Joined: Fri Nov 02, 2012 8:43 am
Location: New Delhi, India

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by shrapnel » Sat Nov 16, 2019 4:09 am

Ponti wrote:
Fri Nov 15, 2019 4:21 pm
I don't own a great video-card (using a Radeon Sapphire HD 7790), though I'm planning to buy a new one.
Well, don't make the mistake of buying an AMD Radeon Card again, especially if you want to use AI Chess Engines, like lc0 and Fat Fritz.
i7 5960X @ 4.1 Ghz, 64 GB G.Skill RipJaws RAM, Twin Asus ROG Strix OC 11 GB Geforce 2080 Tis

User avatar
Nordlandia
Posts: 2634
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by Nordlandia » Sat Nov 16, 2019 1:45 pm


corres
Posts: 3083
Joined: Wed Nov 18, 2015 10:41 am
Location: hungary

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by corres » Sat Nov 16, 2019 2:44 pm

Nordlandia wrote:
Sat Nov 16, 2019 1:45 pm
I'm using T40B.pb from http://data.lczero.org/files/networks-contrib/
T40B.4-160 or -260 are strong nets for a single Ti.
Maybe 42872 is stronger in some situations.
For 2-4 x RTX 2080 Ti J13B.2-188 is even better.

User avatar
Nordlandia
Posts: 2634
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Lc0 settings for Leela on GPU- Nvidia RTX 2080 Super

Post by Nordlandia » Sat Nov 16, 2019 3:02 pm

corres wrote:
Sat Nov 16, 2019 2:44 pm
Nordlandia wrote:
Sat Nov 16, 2019 1:45 pm
I'm using T40B.pb from http://data.lczero.org/files/networks-contrib/
T40B.4-160 or -260 are strong nets for a single Ti.
Maybe 42872 is stronger in some situations.
For 2-4 x RTX 2080 Ti J13B.2-188 is even better.
Which of 160 or 260 do you recommend

I usually test setup positions. Often imbalances ones.

Post Reply