LCZero

Fri May 24, 2019 2:15 pm by · Fri May 24, 2019 2:15 pm by

Client.exe is not needed to play. It is used if you want to support the project with the computing power of your computer.

AMD FX 8350 is a pretty modern processor, but it is much better to use the GPU for the LC0. I have a NVidia GT 1030 graphics card. It's a cheap card, but it's still 10x faster than a CPU (Intel i5 mobile or AMD Ryzen 3). For example, the difference is that on the CPU the depth of analysis is 3-4 and now 8-10.
Currently a good choice for the LC0 is the RTX card (much better than GTX). Thanks to it you can increase your speed by another 10x-20x.
If you think about OpenCL (AMD card), my observations show that the speed will be 50% of that of CUDA (NVidia). I checked both versions of LC0 on my GPU.

I downloaded:

lc0-v0.21.0-windows-cuda.zip

My path:

C:\Program Files (x86)\ChessBase\Engines\LC0_CUDA

in the directory:

cublas64_100.dll
cudart64_100.dll
cudnn64_7.dll
lc0.exe
weights_run1_42161.pb

I also use fritz and install lc0.exe as a normal chess engine. And that's it.

I don't change the engine parameters. The network file is detected automatically.

Fri May 24, 2019 2:37 pm by · Fri May 24, 2019 2:37 pm by

dear jaglo, my card graphic is very weak, Nvida gtx 650...

Fri May 24, 2019 3:26 pm by · Fri May 24, 2019 3:26 pm by

It's not bad. Your card has 67% of my card speed.

https://gpu.userbenchmark.com/Compare/Nvidia-GT-1030-vs-Nvidia-GTX-650/m283726vs3154

That's enough to do some research. It should still be faster than the CPU version.

Install lc0-v0.21.1-windows-cuda.zip

Fri May 24, 2019 4:51 pm by · Fri May 24, 2019 4:51 pm by

ok... Jaglo,.. I will try... I hope it works out...hug

Fri Jun 28, 2019 3:55 am by · Fri Jun 28, 2019 3:55 am by

Massimiliano Goi wrote:Just made this mini-tournament... Lc0 was often in advantage just to drop all of it in the endings, most probably because it was short in time. Neural Network used: 44000.

Sun Jun 30, 2019 6:33 pm by · Sun Jun 30, 2019 6:33 pm by

deleted post

Last edited by jorge91 on Sat Jul 06, 2019 10:39 pm; edited 1 time in total (Reason for editing : deleted post)

Sat Jul 06, 2019 10:40 pm by · Sat Jul 06, 2019 10:40 pm by

Eman 3.81 vs Lc0 21.2, 1 hour thinking game - NOW LIVE

Fri Jul 26, 2019 8:28 pm by · Fri Jul 26, 2019 8:28 pm by

NOW LIVE: SugaR-NN 260719 vs Lc0 21.3

Fri Jul 26, 2019 10:00 pm by · Fri Jul 26, 2019 10:00 pm by

Who is this?

Sat Aug 17, 2019 11:56 pm by · Sat Aug 17, 2019 11:56 pm by

why not introduce a uci leela version

Mon Sep 23, 2019 5:02 pm by · Mon Sep 23, 2019 5:02 pm by

Hello friends

I would like to understand where is located the experience file generated by the learning process

Sun Oct 06, 2019 12:46 pm by · Sun Oct 06, 2019 12:46 pm by

The best networks I think are:

32988

42848

J13B.2-136 (see results on CCC)

T60 training is in progress and I haven't checked it thoroughly yet.

They can be downloaded from here:

https://lczero.org/networks/

https://github.com/jhorthos/lczero-training/wiki/Leela-Training

Sun Oct 06, 2019 12:57 pm by · Sun Oct 06, 2019 12:57 pm by

Depends on hardware on single RTX i find this is very good if not the best
J20-460

this:
J13B.2-136 (see results on CCC)

On CCC it play on multiple GPUs 2080Ti and its need very good hardware.

Tue Oct 08, 2019 11:24 am by · Tue Oct 08, 2019 11:24 am by

Hello

Thanks for pointing the best networks to use on Lc0. There are so many...

What are the correct nodes to set for each of these networks?

by