Nemorino - new engine

[Florentino]

Hello,
may I present Nemorino, my new engine:

Features

• UCI and XBoard support (experimental - all testing is done using UCI)
  Syzygy Tablebase support
  Pondering
  Standard Chess and Chess960
  Multi Core support (Very Lazy SMP)
  Multi PV analysis mode
  64-bit Windows only

Strength is about 2900 Elo on CCRL 40/4 scale

https://bitbucket.org/christian_g_nther ... o/overview

[Damir]

Hi Christian

Unfortunately I can not create Nemorino as new engine in Fritz GUI.
The engine field is empty. Is it possible for you to fix this ?

[Florentino]

Hello Damir,
I fixed it. Can you please download and try again.
Christian

https://bitbucket.org/christian_g_nther/nemorino

[cdani]

Congratulations for your new engine!!

[thunderstruck]

When using more than 1 core, nps seems not to be correct.
Seems to be the value of only 1 core.

[Florentino]

Yes, that's true. I'm using lazy SMP and basically all threads but the first one are only helper threads used for filling up the transposition table. So the best move is still determined by the first core, it only finds more entries in the hash table. Therefore in my opinion counting only the nodes from the 1st core is more appropriate.

[Damir]

Thanks Christian. Congratulations on your new engine. I hope you keep on improving it.

[Guenther]

Hello,
may I present Nemorino, my new engine:

Features

• UCI and XBoard support (experimental - all testing is done using UCI)
  Syzygy Tablebase support
  Pondering
  Standard Chess and Chess960
  Multi Core support (Very Lazy SMP)
  Multi PV analysis mode
  64-bit Windows only

Strength is about 2900 Elo on CCRL 40/4 scale

https://bitbucket.org/christian_g_nther ... o/overview

It seems your compilation is only for modern popcnt machines?
Any chance for a non popcnt version?

(The downloaded version immediately exits on my WIN 7 64 Ultimate
Qcore from 2009, both in console mode and in GUI mode.)

[Florentino]

Hello Guenther,
I just sent you a pm. I don't provide a non-popcount version as I don't have an opportunity to test it.

[Guenther]

Hello Guenther,
I just sent you a pm. I don't provide a non-popcount version as I don't have an opportunity to test it.

I received your mail. The normal non-popcnt version works here.
To be sure I am running a test game right now.
(BTW the gcc version would not work...throws an exception)

I noticed already something which could be improved, if it is normal behaviour.
The test game is at tc 40/420 (moves/time per sec) with 2 threads, ponder off,
256MB hash and a 12 ply GUI book.
It used around 29% of all available time until the cut for its first move after book end!
While I think using more time at this point is a good idea this seems a bit out of proportion?

Guenther