My test for Clover 2.3.1 x64-bmi2

[lucametehau]

I have been told that it takes CCRL some time to test an engine, so I decided to wait. I'd say it was worth waiting, because I saw that Graham wants to do a Clover gauntlet (thx btw ). As for CEGT, I have no idea (2.2 is still in testing phase but it will finish some day...).

[Wolfgang]

... As for CEGT, I have no idea (2.2 is still in testing phase but it will finish some day...).

It IS finished and the games were sent to our webmaster who makes the updates. But as I stated a number of times, he has very limited time at the moment for several reason.
So, Clover 2.2 will be in our next update, hopefully this, maybe next week. Version 2.3.1 or the then actual version will be tested too of course. But it's nearly impossible to test every version as soon as it comes out with a considerable number of games and not only 100 vs. one opponent...
I hope for your understanding...

[Wolfgang]

...Version 2.3.1 or the then actual version will be tested too of course. But it's nearly impossible to test every version as soon as it comes out with a considerable number of games and not only 100 vs. one opponent...
I hope for your understanding...

One of my Ryzen 5 machines just became free by chance, so I start a quick test for at least 1000 games
- Ryzen 5 1600 @ 3,5 GHZ
- avx2 compile
- Shredder Classic 13 GUI (5 instances)
https://cegt.forumieren.com/t1502-tourn ... sions#2948

[Sylwy]

You have the opportunity of being more selective.

......also he only plays 100 fast TC games against a single opponent not the greatest way to approximate a rating, which sums up the "quality" of such tests.......

The problem is old. I already answered last year (see image below). I don't understand why you're upset. I have not published (and I will not publish) the final result Clover-Cheng!

[mar]

You have the opportunity of being more selective.

......also he only plays 100 fast TC games against a single opponent not the greatest way to approximate a rating, which sums up the "quality" of such tests.......

The problem is old. I already answered last year (see image below). I don't understand why you're upset. I have not published (and I will not publish) the final result Clover-Cheng!

haha not really, I simply stated a fact, that's all, maestro you are free to (not) publish whatever you want (not that I would care)
also my reply was addressed to Graham

[Sylwy]

You have the opportunity of being more selective.

......also he only plays 100 fast TC games against a single opponent not the greatest way to approximate a rating, which sums up the "quality" of such tests.......

The problem is old. I already answered last year (see image below). I don't understand why you're upset. I have not published (and I will not publish) the final result Clover-Cheng!

haha not really, I simply stated a fact, that's all, maestro you are free to (not) publish whatever you want (not that I would care)

Perfect ! Thanks for understanding ! Accurate calibration requires time and computational resources. That's why my reference is always CCRL (less CEGT, not to mention SSDF). However, an individual can test a limited number of engines (Mr Pohl's case). In addition, my test time is always blitz (4 '+ 2 "-so about 12-15 minutes per game). Indicative - that's why there are my tests of 100 games.

[lucametehau]

...Version 2.3.1 or the then actual version will be tested too of course. But it's nearly impossible to test every version as soon as it comes out with a considerable number of games and not only 100 vs. one opponent...
I hope for your understanding...

One of my Ryzen 5 machines just became free by chance, so I start a quick test for at least 1000 games
- Ryzen 5 1600 @ 3,5 GHZ
- avx2 compile
- Shredder Classic 13 GUI (5 instances)
https://cegt.forumieren.com/t1502-tourn ... sions#2948

Thanks Wolfgang! You said you had a problem with your webmaster, that's why I waited.

But it's nearly impossible to test every version as soon as it comes out with a considerable number of games and not only 100 vs. one opponent...

That's what I said and it's true, I have nothing against it.

[Sylwy]

I have been told that it takes CCRL some time to test an engine, so I decided to wait. I'd say it was worth waiting, because I saw that Graham wants to do a Clover gauntlet (thx btw ). As for CEGT, I have no idea (2.2 is still in testing phase but it will finish some day...).

Well, wait for ! You are talented ..... but young. Things are a little different ..... Just an example (just one):







The situation (after 59 games) in my test (tomorrow night the final result):



==================

Let's not forget (this is for Mr. Martin Sedlak): yes, I use a single core (out of 4) and Arena 3.5.1 GUI (or sometimes Fritz_17_x64 GUI) because the other three cores have another destination (video editing, etc , etc, etc). Computer chess is just a hobby.

[lucametehau]

Well, it doesn't really matter, as long as it gets tested (at least for me ). As for ClassicAra, I think that's not the first version that got fully tested by CCRL, so I'd assume it has a some priority over me.

So I think we can put an end to this pointless "fight". I support all of you and I will show my concerns when I have some. Peace!

[Wolfgang]

...Peace!