this will be the merge of a lifetime : SF 80 Elo+

Discussion of anything and everything relating to chess playing software and machines.

Moderators: Harvey Williamson, bob, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Joerg Oster
Posts: 789
Joined: Fri Mar 10, 2006 3:29 pm
Location: Germany

Re: When to check NNUE box

Post by Joerg Oster » Sat Aug 08, 2020 2:20 pm

zullil wrote:
Sat Aug 08, 2020 1:11 pm
Sven wrote:
Sat Aug 08, 2020 12:56 pm
JohnS wrote:
Sat Aug 08, 2020 1:26 am
zullil wrote:
Sat Aug 08, 2020 1:10 am
JohnS wrote:
Sat Aug 08, 2020 12:37 am
cma6 wrote:
Fri Aug 07, 2020 2:08 pm
"The idea is to use NNUE only on quite balanced material positions."
This was based on the fact that SF-NNUE searches much more slowly than classic SF on most hardware. But if one has fast hardware, so that SF-NNUE runs fast enough, then why not always use SF-NNUE?
How is stockfish defining 'quite balanced material positions', thanks.
https://github.com/official-stockfish/S ... 5aac9e322f
Thanks, does NNUEThreshold = Value(500) mean 5 pawns, 0.5 pawns or something else.
Actually 2.5 pawns, since the formula also uses a fixed score of 200 for pawns (usually 256 in SF).
Yes, but things have already changed due to new patches. Wish someone would maintain a Stockfish+NNUEonly fork. :wink:
https://github.com/joergoster/Stockfish ... nue_always :D
Jörg Oster

User avatar
Ovyron
Posts: 4399
Joined: Tue Jul 03, 2007 2:30 am

Re: When to check NNUE box

Post by Ovyron » Sat Aug 08, 2020 2:31 pm

zullil wrote:
Sat Aug 08, 2020 1:58 pm
Yes, I can modify source code before compiling so that "Use NNUE" means "Use NNUE only". But that's going to get old very quickly. :wink:
So, basically you just want someone that doesn't find this getting old quickly to maintain a Stockfish fork and do this for you? :mrgreen:

zullil
Posts: 6442
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: When to check NNUE box

Post by zullil » Sat Aug 08, 2020 2:33 pm

Ovyron wrote:
Sat Aug 08, 2020 2:31 pm
zullil wrote:
Sat Aug 08, 2020 1:58 pm
Yes, I can modify source code before compiling so that "Use NNUE" means "Use NNUE only". But that's going to get old very quickly. :wink:
So, basically you just want someone that doesn't find this getting old quickly to maintain a Stockfish fork and do this for you? :mrgreen:
Yes :D Like my new hero, Joerg!

(Though eventually I should spend the time to learn how to use git fully myself. :oops: )
Last edited by zullil on Sat Aug 08, 2020 2:39 pm, edited 1 time in total.

zullil
Posts: 6442
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: When to check NNUE box

Post by zullil » Sat Aug 08, 2020 2:34 pm

Joerg Oster wrote:
Sat Aug 08, 2020 2:20 pm
zullil wrote:
Sat Aug 08, 2020 1:11 pm
Sven wrote:
Sat Aug 08, 2020 12:56 pm
JohnS wrote:
Sat Aug 08, 2020 1:26 am
zullil wrote:
Sat Aug 08, 2020 1:10 am
JohnS wrote:
Sat Aug 08, 2020 12:37 am
cma6 wrote:
Fri Aug 07, 2020 2:08 pm
"The idea is to use NNUE only on quite balanced material positions."
This was based on the fact that SF-NNUE searches much more slowly than classic SF on most hardware. But if one has fast hardware, so that SF-NNUE runs fast enough, then why not always use SF-NNUE?
How is stockfish defining 'quite balanced material positions', thanks.
https://github.com/official-stockfish/S ... 5aac9e322f
Thanks, does NNUEThreshold = Value(500) mean 5 pawns, 0.5 pawns or something else.
Actually 2.5 pawns, since the formula also uses a fixed score of 200 for pawns (usually 256 in SF).
Yes, but things have already changed due to new patches. Wish someone would maintain a Stockfish+NNUEonly fork. :wink:
https://github.com/joergoster/Stockfish ... nue_always :D
Thanks, Joerg!

User avatar
M ANSARI
Posts: 3467
Joined: Thu Mar 16, 2006 6:10 pm

Re: When to check NNUE box

Post by M ANSARI » Sat Aug 08, 2020 4:53 pm

JohnS wrote:
Sat Aug 08, 2020 1:26 am
zullil wrote:
Sat Aug 08, 2020 1:10 am
JohnS wrote:
Sat Aug 08, 2020 12:37 am
cma6 wrote:
Fri Aug 07, 2020 2:08 pm
"The idea is to use NNUE only on quite balanced material positions."
This was based on the fact that SF-NNUE searches much more slowly than classic SF on most hardware. But if one has fast hardware, so that SF-NNUE runs fast enough, then why not always use SF-NNUE?
How is stockfish defining 'quite balanced material positions', thanks.
https://github.com/official-stockfish/S ... 5aac9e322f
Thanks, does NNUEThreshold = Value(500) mean 5 pawns, 0.5 pawns or something else.

Hmmm ... ok now I understand it better. I was under the impression that a more sophisticated algo was used to decide when to turn NN off. I think there will be a lot more easy ELO points gained as this formula gets modified and tested. Maybe a 20mb is too small a network and things would be better with a larger network as per Lc0. Maybe the number of pieces remaining on the board or the number of possible good branching moves available are better identifiers of when to cut NN off. Will certainly be an interesting period as this gets developed further. It is pretty incredible what a game changer Alpha Zero and AI has been in revolutionizing chess engines. I sort of expected this to happen eventually but in 20 years or so from now ... but this is happening today!

Geonerd
Posts: 72
Joined: Fri Mar 10, 2017 12:44 am

Re: this will be the merge of a lifetime : SF 80 Elo+

Post by Geonerd » Sat Aug 08, 2020 7:16 pm

MikeB wrote:
Thu Aug 06, 2020 12:32 pm

Amen - I love it when I see an atheist quoting the Bible ... an old friend of my said this once - lets just say the atheist view is correct - well ,when we die, we die and and we all go back to dust
. But just say there is a God, then atheist gets to spend his eternal years in hell and the rest of us are in the Glory of God for eternal time - what would a risk averse person do?
<Thread Hijack!>

Pascal's Wager is an awfully thin rationale for believing. As the wise Homer Simpson observed: "But Marge, suppose we've chosen the wrong god? Every time we go to church we're just making him madder and madder!"

And, seriously, do you really think any god that would torture his own creations is worthy of worship? That's a pretty twisted world view! ;)

<We now return to the normal TC bickering and sniping! :P >

Leo
Posts: 969
Joined: Fri Sep 16, 2016 4:55 pm
Location: USA/Minnesota
Full name: Leo Anger

Re: this will be the merge of a lifetime : SF 80 Elo+

Post by Leo » Sat Aug 08, 2020 8:35 pm

Geonerd wrote:
Sat Aug 08, 2020 7:16 pm
MikeB wrote:
Thu Aug 06, 2020 12:32 pm

Amen - I love it when I see an atheist quoting the Bible ... an old friend of my said this once - lets just say the atheist view is correct - well ,when we die, we die and and we all go back to dust
. But just say there is a God, then atheist gets to spend his eternal years in hell and the rest of us are in the Glory of God for eternal time - what would a risk averse person do?
<Thread Hijack!>

Pascal's Wager is an awfully thin rationale for believing. As the wise Homer Simpson observed: "But Marge, suppose we've chosen the wrong god? Every time we go to church we're just making him madder and madder!"

And, seriously, do you really think any god that would torture his own creations is worthy of worship? That's a pretty twisted world view! ;)

<We now return to the normal TC bickering and sniping! :P >
I can see that God does not live up to your standards or expectations. Nothing like the created poking his finger in the eye of the Creator.
Advanced Micro Devices fan.

Cornfed
Posts: 99
Joined: Sun Apr 26, 2020 9:40 pm
Full name: Brian D. Smith

Re: this will be the merge of a lifetime : SF 80 Elo+

Post by Cornfed » Sat Aug 08, 2020 8:48 pm

Strikes me as ideal for researching positions that arise at the ‘end of opening’ phase.
How best to set this up for simple analysis of options in those positions?
Also, taking Chessbase as an example, is all one need to do: download the file containing

1.Eval folder
2. SF-nnue-avx2.exe
Sf-nnue-bmi2.exe
Sf-nnue-popcount.exe

Save that onto your hard drive. I save mine in: This PC ->Documents ->ChessBase-> Engines (Local).
Then activate one (lets say popcount), “add engine” to ChessBase engine options with properties options “Use NNUE” checked, EvalFile: ./eval/nn.bin; and I have chosen Analysis Contempt – “Off”

….and then analyze the position you want to delve into?

carldaman
Posts: 2004
Joined: Sat Jun 02, 2012 12:13 am

Re: this will be the merge of a lifetime : SF 80 Elo+

Post by carldaman » Sat Aug 08, 2020 8:57 pm

It defeats the purpose if the NNUE cannot always be set to active [as an option], because then you're at the mercy of the engine only using NNUE when it chooses. Hopefully the SF team will wake up to the fact that people also use the engine for analysis.

User avatar
Deberger
Posts: 70
Joined: Sat Nov 02, 2019 5:42 pm
Full name: ɹǝƃɹǝqǝᗡ ǝɔnɹꓭ

Re: this will be the merge of a lifetime : SF 80 Elo+

Post by Deberger » Sun Aug 09, 2020 12:10 am

carldaman wrote:
Sat Aug 08, 2020 8:57 pm
It defeats the purpose if the NNUE cannot always be set to active [as an option], because then you're at the mercy of the engine only using NNUE when it chooses. Hopefully the SF team will wake up to the fact that people also use the engine for analysis.
With a trivial change to the source code, you can build your own executable which always uses NNUE.

However this change can never yield stronger analysis.

Over time, using "always NNUE" will be gradually weaker and weaker as the nets are optimized for their intended use in balanced positions.
Last edited by Deberger on Sun Aug 09, 2020 12:12 am, edited 2 times in total.

Post Reply