CCC has serious hardware update!

[mwyoung]

SF-NNUE scaled just fine going from 1 to 2, 4, and 8 threads (on 12 core box).
I did not test beyond 8, and the big tourney rigs have way more than 8 of course...
Results posted in SF-NNUE Discord several months ago, IIRC.

Playing what engine? As the testing sites did not, and I did not in my testing. Even in self play or plsying another NNUE engine. As posted earlier in this thread.

[brianr]

it was SF-NNUE v SF-11

Code: Select all

8 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:     118     89    21.0      33  63.6   13   16    4  48.5     100
2 SF11   :       0   ----    12.0      33  36.4    4   16   13  48.5     ---

4 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      58     44    58.0     100  58.0   27   62   11  62.0     100
2 SF11   :       0   ----    42.0     100  42.0   11   62   27  62.0     ---

1 CPU
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      45     19   281.5     500  56.3  128  307   65  61.4     100
2 SF11   :       0   ----   218.5     500  43.7   65  307  128  61.4     ---

Pretty high error bars, but still 100% CFS, FWIW

[mwyoung]

it was SF-NNUE v SF-11

Code: Select all

8 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:     118     89    21.0      33  63.6   13   16    4  48.5     100
2 SF11   :       0   ----    12.0      33  36.4    4   16   13  48.5     ---

4 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      58     44    58.0     100  58.0   27   62   11  62.0     100
2 SF11   :       0   ----    42.0     100  42.0   11   62   27  62.0     ---

1 CPU
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      45     19   281.5     500  56.3  128  307   65  61.4     100
2 SF11   :       0   ----   218.5     500  43.7   65  307  128  61.4     ---

Pretty high error bars, but still 100% CFS, FWIW

Thank you. That is ok with the error bars. Any settings I need to know.

I will test this on my system.

[mwyoung]

it was SF-NNUE v SF-11

Code: Select all

8 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:     118     89    21.0      33  63.6   13   16    4  48.5     100
2 SF11   :       0   ----    12.0      33  36.4    4   16   13  48.5     ---

4 CPUs
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      58     44    58.0     100  58.0   27   62   11  62.0     100
2 SF11   :       0   ----    42.0     100  42.0   11   62   27  62.0     ---

1 CPU
# PLAYER :  RATING  ERROR  POINTS  PLAYED   (%)    W    D    L  D(%)  CFS(%)
1 SF-NNUE:      45     19   281.5     500  56.3  128  307   65  61.4     100
2 SF11   :       0   ----   218.5     500  43.7   65  307  128  61.4     ---

Pretty high error bars, but still 100% CFS, FWIW

Thank you. That is ok with the error bars. Any settings I need to know. And what time controls.

I will test this on my system.

Interesting ignoring the error bars. That you are suggesting that SF NNUE scales better then SF11. As you add more cores.

[AndrewGrant]

Both my testing and CCRL testing between SF NNUE, and Dragon NNUE have thousands of games. All showing the scaling issue with NNUE.

Please post your data, complete with full testing conditions. I will be compiling the evidence presented in these threads when I post mine. As of now, there is a sample of 25 games you posted, a sample of SF games rejecting your belief, and conjecture.

[mwyoung]

Both my testing and CCRL testing between SF NNUE, and Dragon NNUE have thousands of games. All showing the scaling issue with NNUE.

Please post your data, complete with full testing conditions. I will be compiling the evidence presented in these threads when I post mine. As of now, there is a sample of 25 games you posted, a sample of SF games rejecting your belief, and conjecture.

I did post it.

[AndrewGrant]

The 72-game Stockfish Komodo matchup?

[mwyoung]

Both my testing and CCRL testing between SF NNUE, and Dragon NNUE have thousands of games. All showing the scaling issue with NNUE.

Please post your data, complete with full testing conditions. I will be compiling the evidence presented in these threads when I post mine. As of now, there is a sample of 25 games you posted, a sample of SF games rejecting your belief, and conjecture.

I did post it. And It is on CCRL AND CEGT.

Remember you trashing CCRL data. That test alone was over 1000 games.Both of these testing sites have Stockfish and Komodo nnue engine test with 1 and more cores.

As I said the data is there for all to see.

[AndrewGrant]

Both my testing and CCRL testing between SF NNUE, and Dragon NNUE have thousands of games. All showing the scaling issue with NNUE.

Please post your data, complete with full testing conditions. I will be compiling the evidence presented in these threads when I post mine. As of now, there is a sample of 25 games you posted, a sample of SF games rejecting your belief, and conjecture.

I did post it. And It is on CCRL AND CEGT.

Remember you trashing CCRL data. That test alone was over 1000 games.Both of these testing sites have Stockfish and Komodo nnue engine test with 1 and more cores.

As I said the data is there for all to see.

Could you please post the specific data points, and values which you are deriving your belief from? If you tell me to get it myself, you'll reject the information I choose to take. Please present to me your argument for NNUE scaling differing from Classical scaling, featuring data points from CCRL and CEGT with links.

[mwyoung]

Both my testing and CCRL testing between SF NNUE, and Dragon NNUE have thousands of games. All showing the scaling issue with NNUE.

Please post your data, complete with full testing conditions. I will be compiling the evidence presented in these threads when I post mine. As of now, there is a sample of 25 games you posted, a sample of SF games rejecting your belief, and conjecture.

I did post it. And It is on CCRL AND CEGT.

Remember you trashing CCRL data. That test alone was over 1000 games.Both of these testing sites have Stockfish and Komodo nnue engine test with 1 and more cores.

As I said the data is there for all to see.

Could you please post the specific data points, and values which you are deriving your belief from? If you tell me to get it myself, you'll reject the information I choose to take. Please present to me your argument for NNUE scaling differing from Classical scaling, featuring data points from CCRL and CEGT with links.

Let me get this right. I posted data, you chimed in and gave a B.S. formula that NNUE clearly ignored in my posted data. And saying I was being deceitful.

I post independent data from CCRL. Again showing NNUE ignores your scaling formula. And then you also trash CCRL testing.

And now you want me to gather all the data for you.....

It would have been much wiser to do that first

The data is there. CCRL, CEGT, and posted data here. More to come.