Raid

[IbaiBuR]

What are are talking about ?

Actually this is as joke that you say Raubfish is bad...come on..
And via great Raubfish SC engine :
You can not imagine..how thousands of successful games played..!!

And just in case..with the buggy ones.
Believe Me... I would lost my interest since a long time...

And I say again:
Raubfish SC is better than all those buggy chess engines!!
And be sure...for a experienced tester who looks for stability and fun:
Approx. +10 Elo means nothing...!!

On other hand,
When you provide/inform better engine (less DRAWİSH and more STABLE than than Raubfish SC,)
Then I will try to inform you about those buggy engines, which are worse than Raubfish SC
Btw, actually I already mentioned...you should know some of them...right ? or you forgot already ??

And to be more clear (if you will suggest about which are better than Raubfish):
Tell please the best/better engine for 'book tours' on my tournament hardware:
[Dual CPU] AMD EPYC 7B12 128 Cores /256 Threads

If you can not...please...ignore my current posting...

Dont try to deviate the discussion... You said you did not care about GPLv3, and that was what I was mainly talking about...

If some engines crash because you dont know how to use them (even if a lot of people told you that you were using them incorrectly) it is not my fault really.

Standard Stockfish Dev is probably better overall than Raubfish (because it is updated, as simple as that). Maybe not for your book tests, because to really test books sometimes it is better to use non-NNUE engines, but better overall.

I wont post anymore here if you want to continue with this discussion either open a new thread or in your thread, we are already polluting this thread too much.

Regards.

[Sedat Canbaz]

Don't try to deviate the discussion... You said you did not care about GPLv3, and that was what I was mainly talking about...

If some engines crash because you dont know how to use them (even if a lot of people told you that you were using them incorrectly) it is not my fault really.

Standard Stockfish Dev is probably better overall than Raubfish (because it is updated, as simple as that). Maybe not for your book tests, because to really test books sometimes it is better to use non-NNUE engines, but better overall.

I wont post anymore here if you want to continue with this discussion either open a new thread or in your thread, we are already polluting this thread too much.

Regards.

Agreed with one point here,
This thread is gone to different direction...
And my apologies from Raid's author

But I joined in this thread,
Simply because SolistaChess said that Raid is identical to SF...
Where no any prove that Raid is identical...at least via Simex...

Plus my target was simply that there are many engines,
Which are released as buggy (non-stable) ones...
And fun is more important than mentioned GPLv3 license...
This is my view...and I respect other opinions as well...

But you are wrong that Raubfish SC is bad...
Actually you are bad via this saying..hope helps !)
And prove that I am wrong...if you can not... please be careful...

And again and again you wrote nonsense .. please re-read again..
Actually you have seriously problems to read as before ...LOL

If not so clear, I never said that,
I do not care about GPLv3 license...I said: who cares ?

Btw, about your rest comments that I do not know how to test engines..
HEHE )) 19 years boy..will teach Me...where have you been before ??))
You have no clue what is going on my tournament machine for sure!

And as last words, remember that too,
I am more... than your age (in eng testings!) and hope helps...!

That's all

Greetings )

[Sedat Canbaz]

AMD EPYC 7B12 128 Cores /256 Threads Thats a beast

Thank you

[Sedat Canbaz]

you do 64 concurrency , your machine must be strong, i normally do 8-10 only

Actually via 64 concurrency :
I run the chess engines at half (50%) of the overall CPU cores usage!
I mean, still there are too many plenty of free seats in this BIG bus!

And the real problem is that,
Some of the passengers are not capable to seat comfortable (during journey)
They seems to be as caprice ladies...very likely, wanting in VIP seats area ))

But in SCCT, same conditions for all...sorry.. )

Sure just in case of learning eng testings...
Via concurrent matches (up to 10) is all right,
But via 64 concurrent (in gauntlet) they suffer...
E.g learning is not so much affective...

Greetings

[Sedat Canbaz]

One thing more

The most terrible...
Day by day... I see more learning engines, which are started to crash..
Sure I mean in gauntlet mode.. via 64 concurrent

That's why as usual I say,
Not always newer versions are better...!!

But I admit also that,
There are many learning engines, which are stable even via 64 concurrent!

Greetings

[Sedat Canbaz]

@to IbaiBuR member

Since you asked...
Here are the ones which crashed directly in gauntlet matches,
Sure via concurrent 64:

EMAN 9.40
ProteusSF Aureo
Tactical 220823
SpecTral 5.0

NOTE: via 32 concurrent is all right....no any crash is recorded...

------------------------------------------------------------

About the other important issue - CTG buggy code...
As far as I noticed...all, which have this CTG option/future:
They used to play all the lost games...and till to the end of line!
By the way, what does it mean.. do you know ? if still not clear:
Try ChessBase (Fritz) GUIs...and see the difference...

[IbaiBuR]

@to IbaiBuR member

Since you asked...
Here are the ones which crashed during gauntlet matches,
Sure via concurrent 64:

EMAN 9.40
ProteusSF Aureo
Tactical 220823
SpecTral 5.0

--------------------

About the other important issue - CTG buggy code...
As far as I noticed...all, which have this CTG option/future:
They used to play all the lost games...and till to the end of line!
By the way, what does it mean.. do you know ? if still not clear:
Try ChessBase (Fritz) GUIs...and see the difference...

Thanks for pointing it out (in friendly way much better). The thing is that the 4 engines you are mentioning, as far as I am aware, use the same learning code...

My guess is that, without concurrent experience option there is too much read / write operations been carried out at the same time in the experience file and at some point the engine just crashes.

I have experienced something similar, not with concurrency, but when running some of the training scripts (I normally launch 2 scripts with concurrency 1, 2 threads each script to not stress the CPU), and I experienced some crashes too but I think that on my side it was due to high RAM usage...

With the same conditions, in Linux they did not crash, only in Windows (as expected, since Windows is a RAM eater)

I was not aware of the other issue really since I normally only use the bin book option... But you have to take in account that the CTG code is not perfect because the CTG format is closed source compared to the bin book format...

So there may be some unknown information about CTG books that makes the code to not work correctly...

[Sedat Canbaz]

@to IbaiBuR member

Since you asked...
Here are the ones which crashed during gauntlet matches,
Sure via concurrent 64:

EMAN 9.40
ProteusSF Aureo
Tactical 220823
SpecTral 5.0

--------------------

About the other important issue - CTG buggy code...
As far as I noticed...all, which have this CTG option/future:
They used to play all the lost games...and till to the end of line!
By the way, what does it mean.. do you know ? if still not clear:
Try ChessBase (Fritz) GUIs...and see the difference...

Thanks for pointing it out (in friendly way much better). The thing is that the 4 engines you are mentioning, as far as I am aware, use the same learning code...

My guess is that, without concurrent experience option there is too much read / write operations been carried out at the same time in the experience file and at some point the engine just crashes.

I have experienced something similar, not with concurrency, but when running some of the training scripts (I normally launch 2 scripts with concurrency 1, 2 threads each script to not stress the CPU), and I experienced some crashes too but I think that on my side it was due to high RAM usage...

With the same conditions, in Linux they did not crash, only in Windows (as expected, since Windows is a RAM eater)

I was not aware of the other issue really since I normally only use the bin book option... But you have to take in account that the CTG code is not perfect because the CTG format is closed source compared to the bin book format...

So there may be some unknown information about CTG books that makes the code to not work correctly...

Thanks for your valuable comments!

Ah ah... these learning engines ))

Yes... for these engines or exp files..we argued a lot...
Even sometimes we used not as friendly words to each other...
And now I am thinking that we made mistakes really...
But it is time to stop...better late than never )

In other words, I think that for all chess programs..
It does not matter how is strong, stable, buggy. etc..
It is as worst idea to fight each other in non-friendly way...
Simply is pointless! because you or Me or any other friend...
Do not deserve this...right ?

Greetings )

[OrgZ]

Don't try to deviate the discussion... You said you did not care about GPLv3, and that was what I was mainly talking about...

If some engines crash because you dont know how to use them (even if a lot of people told you that you were using them incorrectly) it is not my fault really.

Standard Stockfish Dev is probably better overall than Raubfish (because it is updated, as simple as that). Maybe not for your book tests, because to really test books sometimes it is better to use non-NNUE engines, but better overall.

I wont post anymore here if you want to continue with this discussion either open a new thread or in your thread, we are already polluting this thread too much.

Regards.

Agreed with one point here,
This thread is gone to different direction...
And my apologies from Raid's author

But I joined in this thread,
Simply because SolistaChess said that Raid is identical to SF...
Where no any prove that Raid is identical...at least via Simex...

Plus my target was simply that there are many engines,
Which are released as buggy (non-stable) ones...
And fun is more important than mentioned GPLv3 license...
This is my view...and I respect other opinions as well...

But you are wrong that Raubfish SC is bad...
Actually you are bad via this saying..hope helps !)
And prove that I am wrong...if you can not... please be careful...

And again and again you wrote nonsense .. please re-read again..
Actually you have seriously problems to read as before ...LOL

If not so clear, I never said that,
I do not care about GPLv3 license...I said: who cares ?

Btw, about your rest comments that I do not know how to test engines..
HEHE )) 19 years boy..will teach Me...where have you been before ??))
You have no clue what is going on my tournament machine for sure!

And as last words, remember that too,
I am more... than your age (in eng testings!) and hope helps...!

That's all

Greetings )

All is well Mr Sedat, no worries

[Sedat Canbaz]

All is well Mr Sedat, no worries

Thanks...Mr. Tanick )

On other hand,
Many other members are hijacking other threads...
But when I start ...a real BIG problem... hehe )))

But there is a small point,
They are much better in 'hijacking' than Me !!
And other difference...they never say sorry etc..
They are always right and as perfect guys )))

On other hand,
Now I remembered...you know, I forgive..but I do not forget...))
E.g when Deeds member (via fake account) opened a thread,
viewtopic.php?f=3&t=3648&start=30

Actually opening a fake account is real scandal, it's very serious, bad thing... !!
I mean as worst than everything, and even in those days...they praised him..
E.g Deeds accused Me as LIAR... and unfortunately same/some artists were
In action again...yes...they hijacked everything on those days..even nowadays..
And now IbaiBuR member will talk that II hijack threads... just LOL..

Btw, I wonder too,
Where is Deeds now? maybe too busy with EXP sales ))

Greetings )