LCZero is using my cores, not my GPU.

[Gian-Carlo Pascutto]

If I was to buy a video card just to train Leela, what would be recommend? I can get it through work; so may as well.

Hard to give specific recommendations, but at this point generally something NVIDIA GTX 1060, 1070, 1080 or 1080 Ti based, basically depending on how much you want to spend. Above the 1080 Ti and below the 1060 the price/performance curve drops off rapidly.

As for the different vendors out there, check some reviews for specific models you can get your hands on. Personally, I prefer a quiet card that's energy efficient over an overclocked one that runs 100Mhz faster, has a bunch of LEDs and is loud. (It adds up quickly with a bunch in my office!)

AMD cards will run fine if you have them lying around, and are well supported by Leela, but not necessarily by other deep learning based things. For that reason it's still hard to recommend them.

[op12no2]

Hard to give specific recommendations, but at this point generally something NVIDIA GTX 1060, 1070, 1080 or 1080 Ti based, basically depending on how much you want to spend. Above the 1080 Ti and below the 1060 the price/performance curve drops off rapidly.

As for the different vendors out there, check some reviews for specific models you can get your hands on. Personally, I prefer a quiet card that's energy efficient over an overclocked one that runs 100Mhz faster, has a bunch of LEDs and is loud. (It adds up quickly with a bunch in my office!)

AMD cards will run fine if you have them lying around, and are well supported by Leela, but not necessarily by other deep learning based things. For that reason it's still hard to recommend them.

Very useful. Thanks very much for that.

[Vinvin]

The last line of the report was this:
(1269/5310) KWG=32 KWI=2 MDIMA=8 MDIMC=8 MWG=64 NDIMB=8 NDIMC=8 NWG=16 SA=0 SB=0 STRM=1 STRN=1 VWM=8 VWN=2 0.7962 ms (10.5 GFLOPS)

How do I start up with those settings?

They're saved in the leelaz_opencl_tuning file and remembered.

All for naught, it is clear to me that my GPU is not being used.
The end-around I found on the net failed, and my back-off to an older driver is hung for an hour now.

I guess it is time to buy a nice Nvidia card.

Some numbers as references for the speed :
On my old GTX 750 Ti : 0.7 Kn/s (700 n/s)
On a GTX 1080 Ti : around 2.5 Kn/s
On the TCEC server (43 cores @2.8 GHz) : 5 Kn/s

2 live video to watch LC0 playing :
https://www.twitch.tv/ccls
and :
https://www.twitch.tv/tcec_chess_tv

Some more numbers on my CPUs i7@6*4GHz :
with 6 cores : 0.55 Kn/s
with 1 cores : 0.1 Kn/s
with 12 threads (with hyperthreading) : 0.7 Kn/s

[duncan]

I said "publicly available".

We have no way to test against Fineart. Of course, they are free to setup Leela Zero in a handicapped manner in a private match and claim superiority. This is an approach that has been proven successful in the past and is hence likely to be tried again.

is Leela-HX01 /yao yours ?
http://www.yss-aya.com/cgos/19x19/bayes.html

[Gian-Carlo Pascutto]

is Leela-HX01 /yao yours ?
http://www.yss-aya.com/cgos/19x19/bayes.html

I never run on CGOS. It's a bit like trying to measure your engine's improvement by playing in the ChessBase engine room, except that you get disconnected (and hence lose on time) more.

[jpqy]

If you want to see some Benchmarks with LCZero gpu & cpu you can check this spreadsheet:
https://docs.google.com/spreadsheets/d/ ... edit#gid=0

Or even add your own benches..

JP.

[Vinvin]

If you want to see some Benchmarks with LCZero gpu & cpu you can check this spreadsheet:
https://docs.google.com/spreadsheets/d/ ... edit#gid=0

Or even add your own benches..

JP.

Thanks, Jean-Paul !
Very useful !

[Vinvin]

Some numbers as references for the speed :
On my old GTX 750 Ti : 0.7 Kn/s (700 n/s)
On a GTX 1080 Ti : around 2.5 Kn/s
On the TCEC server (43 cores @2.8 GHz) : 5 Kn/s

2 live video to watch LC0 playing :
https://www.twitch.tv/ccls
and :
https://www.twitch.tv/tcec_chess_tv

Some more numbers on my CPUs i7@6*4GHz :
with 6 cores : 0.55 Kn/s
with 1 cores : 0.1 Kn/s
with 12 threads (with hyperthreading) : 0.7 Kn/s

Clear view of my hardware on depth=26 (except for 1 thread, only d=25) :

Code: Select all

GPU :
-----
Platform name:    NVIDIA CUDA
Platform vendor:  NVIDIA Corporation
Device ID:     0
Device name:   GeForce GTX 750 Ti
Device type:   GPU
Device vendor: NVIDIA Corporation
Device driver: 388.13
Device speed:  1110 MHz
Device cores:  5 CU
Device score:  1112

t = 6
info depth 26 nodes 130268 nps 925 score cp -21 time 140435 pv e7e5 g1f3 d7d6

t = 1
info depth 26 nodes 131065 nps 603 score cp 20 time 217360 pv e2e4 e7e5 g1f3 d7d6 b1c3 g8f6 d2d4 e5d4 f3d4 f8e7

with HT : t = 12
info depth 26 nodes 131160 nps 940 score cp 20 time 139545 pv e2e4 e7e5 g1f3 d7d6


CPU : BLAS Core: Sandybridge
-----
t = 6
info depth 26 nodes 128313 nps 637 score cp 19 time 201406 pv e2e4 c7c5 c2c3 g8f6 

t = 1
info depth 25 nodes 72693 nps 117 score cp 19 time 619450 pv e2e4 c7c5 c2c3 e7e6 d2d4 a7a6 g1f3 d7d5 e4d5 e6d5 c3c4

[duncan]

There aren't enough to be useful as a baseline, Leela Zero is far stronger than anything else that is publicly available.

may I ask which one did you test against and what was the score?