SPCC: Testrun of Lc0 BT4-6147500pt332 finished

[pohl4711]

My UHO-Top15 Ratinglist is the world's first engine-ratinglist, using UHO-openings, and the world's first ratinglist offering additionally Gamepair-statistics.

NN-testrun of Lc0 0.31.1 BT4-6147500pt332 finished (all new dev-binaries cause disconnects in my testruns, but the official release binary 0.31.1 runs stable):
https://www.sp-cc.de/nn-vs-sf-testing.htm

https://www.sp-cc.de

Also take a look at the EAS-Ratinglist, the world's first engine-ratinglist not measuring strength of engines but engines's style of play:
https://www.sp-cc.de/eas-ratinglist.htm

(Perhaps you have to clear your browsercache with <strg>+<shift>+<delete> to reload the graphics/diagrams on my website)

[Jouni]

No progress in 1-2 years? Even if there were, we don't know if it is from net or binary!

[pohl4711]

No progress in 1-2 years? Even if there were, we don't know if it is from net or binary!

Sadly, all new dev binaries produce disconnects in my testruns. Only the release binary 0.31.1 runs stable. So, right now, new binaries can not be tested.
Because the Lc0 devs test only with fixed nodes per move (not thinking-time), they dont see this bug in their testings, so there is no bugfix at the horizon, I am afraid.

[cc2150dx]

No progress in 1-2 years? Even if there were, we don't know if it is from net or binary!

Sadly, all new dev binaries produce disconnects in my testruns. Only the release binary 0.31.1 runs stable. So, right now, new binaries can not be tested.
Because the Lc0 devs test only with fixed nodes per move (not thinking-time), they dont see this bug in their testings, so there is no bugfix at the horizon, I am afraid.

Have you looked into using the 'Dag' versions of Lc0 ? https://ci.appveyor.com/project/Ergodice/lc0/history

[pohl4711]

No progress in 1-2 years? Even if there were, we don't know if it is from net or binary!

Sadly, all new dev binaries produce disconnects in my testruns. Only the release binary 0.31.1 runs stable. So, right now, new binaries can not be tested.
Because the Lc0 devs test only with fixed nodes per move (not thinking-time), they dont see this bug in their testings, so there is no bugfix at the horizon, I am afraid.

Have you looked into using the 'Dag' versions of Lc0 ? https://ci.appveyor.com/project/Ergodice/lc0/history

Of course. These binaries from Ergocide are the binaries, which disconnect from time to time (but too often, to use them). I tried some of them over the last months. All buggy. Only "good" news here: Some of my aborted testruns showed no better Elo performance than the official released binary 0.31.1.

[cc2150dx]

No progress in 1-2 years? Even if there were, we don't know if it is from net or binary!

Sadly, all new dev binaries produce disconnects in my testruns. Only the release binary 0.31.1 runs stable. So, right now, new binaries can not be tested.
Because the Lc0 devs test only with fixed nodes per move (not thinking-time), they dont see this bug in their testings, so there is no bugfix at the horizon, I am afraid.

Have you looked into using the 'Dag' versions of Lc0 ? https://ci.appveyor.com/project/Ergodice/lc0/history

Of course. These binaries from Ergocide are the binaries, which disconnect from time to time (but too often, to use them). I tried some of them over the last months. All buggy. Only "good" news here: Some of my aborted testruns showed no better Elo performance than the official released binary 0.31.1.

When you mentioned 'dev binaries' I thought you meant https://ci.appveyor.com/project/LeelaCh ... c0/history those dev binaries.