Both using 8 cores Xeon E5-2670 at 3GHz all cores frequency. SF10 running on 16 threads (HT on), Lc0 v0.19 openblas implementation with 8 threads.
Private opening set, 50 openings limited to 6 moves.
TC: 1min+1s
Result:
SF10 99.5 points
Lc0 0.5 points
+99 =1 -0
Winning percentage = 99.5%
Lc0 is 920Elo weaker than SF10.
LC0 network 31740 vs Stockfish 10
Moderators: hgm, Rebel, chrisw
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
LC0 network 31740 vs Stockfish 10
Last edited by Milos on Tue Dec 04, 2018 4:16 am, edited 1 time in total.
-
- Posts: 1754
- Joined: Tue Apr 19, 2016 6:08 am
- Location: U.S.A
- Full name: Andrew Grant
Re: LC0 network 31740 vs Stockfish 10
I believe the results more than I believe you would waste your time getting a working openblas version of LeelaMilos wrote: ↑Tue Dec 04, 2018 4:10 am Both using 8 cores Xeon E5-2670 at 3GHz all cores frequency. SF10 running on 16 threads (HT on), Lc0 v0.19 openblas implementation with 8 threads.
Private opening set, 50 openings limited to 6 moves.
TC: 1min+1s
Result:
SF10 99.5 points
Lc0 0.5 points
+99 =1 -0
Winning percentage = 99.5%
Lc0 is 920Elo weaker than SF10.
#WeAreAllDraude #JusticeForDraude #RememberDraude #LeptirBigUltra
"Those who can't do, clone instead" - Eduard ( A real life friend, not this forum's Eduard )
"Those who can't do, clone instead" - Eduard ( A real life friend, not this forum's Eduard )
-
- Posts: 41423
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
Re: LC0 network 31740 vs Stockfish 10
I can't get the openblas version of Lc0 to work properly with any more than 4 cores.AndrewGrant wrote: ↑Tue Dec 04, 2018 6:30 amI believe the results more than I believe you would waste your time getting a working openblas version of LeelaMilos wrote: ↑Tue Dec 04, 2018 4:10 am Both using 8 cores Xeon E5-2670 at 3GHz all cores frequency. SF10 running on 16 threads (HT on), Lc0 v0.19 openblas implementation with 8 threads.
Private opening set, 50 openings limited to 6 moves.
TC: 1min+1s
Result:
SF10 99.5 points
Lc0 0.5 points
+99 =1 -0
Winning percentage = 99.5%
Lc0 is 920Elo weaker than SF10.
If I try 5,6,7 or 8 cores (on a 2x4CPU octal), Lc0 will lock up the computer after a few moves, meaning a reboot.
gbanksnz at gmail.com
-
- Posts: 2727
- Joined: Wed May 12, 2010 10:00 pm
Re: LC0 network 31740 vs Stockfish 10
Lc0 Network 31777 Vs Stockfish 10
Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
50 openings limited to 6 moves.
TC: 1min+1s
Result:
Lc0 100 points
Stockfish 10 0 points
+100 =0 -0
Winning percentage = 100%
Stockfish 10 is 1000 Elo weaker than Lc0.
Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
50 openings limited to 6 moves.
TC: 1min+1s
Result:
Lc0 100 points
Stockfish 10 0 points
+100 =0 -0
Winning percentage = 100%
Stockfish 10 is 1000 Elo weaker than Lc0.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
But my words like silent raindrops fell. And echoed in the wells of silence.
-
- Posts: 3232
- Joined: Mon May 31, 2010 1:29 pm
- Full name: lucasart
Re: LC0 network 31740 vs Stockfish 10
More LCO fake news ?mwyoung wrote: ↑Tue Dec 04, 2018 7:00 am Lc0 Network 31777 Vs Stockfish 10
Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
50 openings limited to 6 moves.
TC: 1min+1s
Result:
Lc0 100 points
Stockfish 10 0 points
+100 =0 -0
Winning percentage = 100%
Stockfish 10 is 1000 Elo weaker than Lc0.
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
-
- Posts: 188
- Joined: Sun Dec 25, 2016 4:59 pm
Re: LC0 network 31740 vs Stockfish 10
Actually I think that particular result is pretty clearly correct at the moment (albeit a bit of a joke, I suspect, aimed at the first result in the thread being both engines on CPU)lucasart wrote: ↑Tue Dec 04, 2018 12:49 pmMore LCO fake news ?mwyoung wrote: ↑Tue Dec 04, 2018 7:00 am Lc0 Network 31777 Vs Stockfish 10
Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
50 openings limited to 6 moves.
TC: 1min+1s
Result:
Lc0 100 points
Stockfish 10 0 points
+100 =0 -0
Winning percentage = 100%
Stockfish 10 is 1000 Elo weaker than Lc0.
Cheers!Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
Re: LC0 network 31740 vs Stockfish 10
Your reply as well as the upper attempt of "joke" shows how ppl are actually clueless and lack basic logic (and the wonderful irony in your nickname that is actually name of algorithm that exclusively requires CPU to run).MonteCarlo wrote: ↑Tue Dec 04, 2018 4:55 pmActually I think that particular result is pretty clearly correct at the moment (albeit a bit of a joke, I suspect, aimed at the first result in the thread being both engines on CPU)lucasart wrote: ↑Tue Dec 04, 2018 12:49 pmMore LCO fake news ?mwyoung wrote: ↑Tue Dec 04, 2018 7:00 am Lc0 Network 31777 Vs Stockfish 10
Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
50 openings limited to 6 moves.
TC: 1min+1s
Result:
Lc0 100 points
Stockfish 10 0 points
+100 =0 -0
Winning percentage = 100%
Stockfish 10 is 1000 Elo weaker than Lc0.
Cheers!Both engines using GTX 965 Graphics Card with over 1000 CUDA Cores.
Neither SF nor Lc0 nor any other engine can actually run only on graphics card, i.e. requires general CPU because it has a search. (The only exception to that would be Zeta engine by Srdja Matovic).
So trying to run any engine (Lc0 included) exclusively on GPU is meaningless joke or no joke in question.
-
- Posts: 188
- Joined: Sun Dec 25, 2016 4:59 pm
Re: LC0 network 31740 vs Stockfish 10
Heh, well, this will be my last reply on this point, as I don't like feeding such behavior.Milos wrote: ↑Tue Dec 04, 2018 6:24 pm Your reply as well as the upper attempt of "joke" shows how ppl are actually clueless and lack basic logic (and the wonderful irony in your nickname that is actually name of algorithm that exclusively requires CPU to run).
Neither SF nor Lc0 nor any other engine can actually run only on graphics card, i.e. requires general CPU because it has a search. (The only exception to that would be Zeta engine by Srdja Matovic).
So trying to run any engine (Lc0 included) exclusively on GPU is meaningless joke or no joke in question.
Just a couple closing points:
1) I never said the comparison exploited in the joke was completely apt or held up to technical scrutiny. Not my joke, and even if it were, if all jokes relied on every element being 100% scientifically accurate we'd not have many jokes. If you wish to get riled up over such things, that's your preference, of course.
2) My reply shows nothing about my being clueless or lacking basic logic. lucasart replied to what seemed clearly to be a joke as though it were a serious factual claim, and I tried to helpfully point out that it seemed likely a joke. Nothing about that is lacking in logic or clueless.
I'm human, and couldn't resist at least this response to your unwarranted hostility. I'm done feeding now, though.
Cheers!
-
- Posts: 3232
- Joined: Mon May 31, 2010 1:29 pm
- Full name: lucasart
Re: LC0 network 31740 vs Stockfish 10
Except your joke doesn't work, because to even run an OS and chess engines you need a CPU. A GPU alone can't even boot the computer.MonteCarlo wrote: ↑Tue Dec 04, 2018 6:46 pmHeh, well, this will be my last reply on this point, as I don't like feeding such behavior.Milos wrote: ↑Tue Dec 04, 2018 6:24 pm Your reply as well as the upper attempt of "joke" shows how ppl are actually clueless and lack basic logic (and the wonderful irony in your nickname that is actually name of algorithm that exclusively requires CPU to run).
Neither SF nor Lc0 nor any other engine can actually run only on graphics card, i.e. requires general CPU because it has a search. (The only exception to that would be Zeta engine by Srdja Matovic).
So trying to run any engine (Lc0 included) exclusively on GPU is meaningless joke or no joke in question.
Just a couple closing points:
1) I never said the comparison exploited in the joke was completely apt or held up to technical scrutiny. Not my joke, and even if it were, if all jokes relied on every element being 100% scientifically accurate we'd not have many jokes. If you wish to get riled up over such things, that's your preference, of course.
2) My reply shows nothing about my being clueless or lacking basic logic. lucasart replied to what seemed clearly to be a joke as though it were a serious factual claim, and I tried to helpfully point out that it seemed likely a joke. Nothing about that is lacking in logic or clueless.
I'm human, and couldn't resist at least this response to your unwarranted hostility. I'm done feeding now, though.
Cheers!
I can also beat Leela and Stockfish at Tetris and Pacman easily.
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
-
- Posts: 188
- Joined: Sun Dec 25, 2016 4:59 pm
Re: LC0 network 31740 vs Stockfish 10
I'll repeat what I already said to Milos and you quoted:
As I've already said, it seemed to me that the post from mwyoung was intended to be a joke but was not interpreted as such, so I was trying to be helpful. I didn't realize that would be interpreted as an invitation to provide condescending explanations that someone else's joke wouldn't make sense as a factual report (well, ok, I expect such nonsense from Milos).Not my joke, and even if it were, if all jokes relied on every element being 100% scientifically accurate we'd not have many jokes. If you wish to get riled up over such things, that's your preference, of course.
My apologies for trying to help. Won't happen again.