Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Set the SF16 reference Nps using SF11 as base #1735

Merged

Conversation

ppigazzini
Copy link
Collaborator

Fishtest with Stockfish 11 had 1.6MNps as reference Nps and 0.7MNps as threshold for the slow worker.
Set the new reference Nps according to the 26% slowdown of the SF 16 wrt SF 11 on the ARCH=x86-64-bmi2.
Set the new threshold for slow worker according to the 34% slowdown of the SF 16 wrt SF 11 on the ARCH=x86-64-modern (x86-64-sse41-popcnt).

compiler clang++ 16.0.6

  • arch=bmi2 (Intel Xeon CPU E5-2680 v3)
sf_base =  1580934 +/-  12410 (95%)
sf_test =  1163656 +/-   7679 (95%)
diff    =  -417278 +/-  11547 (95%)
speedup = -26.394% +/- 0.730% (95%
  • arch=modern (Intel core i7 3770k)
sf_base =  1865978 +/-  14061 (95%)
sf_test =  1228350 +/-  11701 (95%)
diff    =  -637627 +/-   5230 (95%)
speedup = -34.171% +/- 0.280% (95%)

Raise worker version to 206, also server side.

Fishtest with Stockfish 11 had 1.6MNps as reference Nps and 0.7MNps as
threshold for the slow worker.
Set the new reference Nps according to the 26% slowdown
of the SF 16 wrt SF 11 on the ARCH=x86-64-bmi2.
Set the new threshold for slow worker according to the 34% slowdown
of the SF 16 wrt SF 11 on the ARCH=x86-64-modern (x86-64-sse41-popcnt).

compiler clang++ 16.0.6

- arch=bmi2 (Intel Xeon CPU E5-2680 v3)
```
sf_base =  1580934 +/-  12410 (95%)
sf_test =  1163656 +/-   7679 (95%)
diff    =  -417278 +/-  11547 (95%)
speedup = -26.394% +/- 0.730% (95%
```

- arch=modern (Intel core i7 3770k)
```
sf_base =  1865978 +/-  14061 (95%)
sf_test =  1228350 +/-  11701 (95%)
diff    =  -637627 +/-   5230 (95%)
speedup = -34.171% +/- 0.280% (95%)
```

Raise worker version to 206, also server side.
@ppigazzini ppigazzini added the worker update code changes requiring a worker update label Jun 24, 2023
@ppigazzini
Copy link
Collaborator Author

By the way benches done using the SF 16 pre-release
https://github.com/official-stockfish/Stockfish/releases/tag/stockfish-dev-20230622-a49b3ba7

@ppigazzini ppigazzini added the update code change to account external change (Stockfish, book etc.) label Jun 25, 2023
@ppigazzini ppigazzini merged commit a365717 into official-stockfish:master Jun 30, 2023
@ppigazzini ppigazzini deleted the set_reference_threashold_nps branch June 30, 2023 19:29
@ppigazzini
Copy link
Collaborator Author

Triggered the workers update.

ppigazzini added a commit to ppigazzini/fishtest that referenced this pull request Feb 24, 2024
Fishtest with Stockfish 16 (official-stockfish#1735) had 1.184Mnps as reference core nps
and 0.462Mnps as threshold for the slow worker.
Set the new reference core nps to 473600nps according to the 40% slowdown
of SF_16.1 wrt SF_16 on the ARCH=x86-64-vnni256
Set the new threshold for slow worker to 226380nps according to the 51%
slowdown of SF_16.1 wrt SF_16 on the ARCH=x86-64-sse41-popcnt

- ARCH=x86-64-vnni256 (AMD Ryzen 7 PRO 7840U)
```
sf_base =  2340238 +/-  14494 (95%)
sf_test =  1406173 +/-   7804 (95%)
diff    =  -934065 +/-  11607 (95%)
speedup = -39.913% +/- 0.496% (95%
```
- ARCH=x86-84-bmi2 (Intel Xeon E5-2680 v3)
```
sf_base =  1213365 +/-   6456 (95%)
sf_test =   615113 +/-   2899 (95%)
diff    =  -598251 +/-   6130 (95%)
speedup = -49.305% +/- 0.505% (95%)
```
- ARCH=x86-64-sse41-popcnt (Intel Core i7 3770K)
```
sf_base =  1205741 +/-  21782 (95%)
sf_test =   590071 +/-   5765 (95%)
diff    =  -615670 +/-  16287 (95%)
speedup = -51.062% +/- 1.351% (95%)
```

Raise worker version to 230 (also server side)
ppigazzini added a commit to ppigazzini/fishtest that referenced this pull request Feb 24, 2024
Fishtest with Stockfish 16 (official-stockfish#1735) had 1.184Mnps as reference core nps
and 0.462Mnps as threshold for the slow worker.
Set the new reference core nps to 473600nps according to the 40% slowdown
of sf_16.1 wrt sf_16 on the ARCH=x86-64-vnni256
Set the new threshold for slow worker to 226380nps according to the 51%
slowdown of sf_16.1 wrt sf_16 on the ARCH=x86-64-sse41-popcnt

- ARCH=x86-64-vnni256 (AMD Ryzen 7 PRO 7840U)
```
sf_base =  2340238 +/-  14494 (95%)
sf_test =  1406173 +/-   7804 (95%)
diff    =  -934065 +/-  11607 (95%)
speedup = -39.913% +/- 0.496% (95%
```
- ARCH=x86-84-bmi2 (Intel Xeon E5-2680 v3)
```
sf_base =  1213365 +/-   6456 (95%)
sf_test =   615113 +/-   2899 (95%)
diff    =  -598251 +/-   6130 (95%)
speedup = -49.305% +/- 0.505% (95%)
```
- ARCH=x86-64-sse41-popcnt (Intel Core i7 3770K)
```
sf_base =  1205741 +/-  21782 (95%)
sf_test =   590071 +/-   5765 (95%)
diff    =  -615670 +/-  16287 (95%)
speedup = -51.062% +/- 1.351% (95%)
```

Raise worker version to 230 (also server side)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
update code change to account external change (Stockfish, book etc.) worker update code changes requiring a worker update
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant