Slow Rybka.

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Louis F

Slow Rybka.

Post by Louis F »

I have just installed the engine Rybka v2.3.2a.w32 to my ChessBase 9.0 GUI but when I fire it up it only searches at about 100 kN/s whereas Fritz 10 and Junior 10 run at 1000-1800 kN/s. Is Rybka supposed to run this slow? I have a Dell Inspirion 531 with an AMD Athlon 64 X2 Dual Core Processor 3800+ running at 2.01 GHz with 958 MB RAM. I set the hash table size at 150 MB. Should I have installed the Rybka v2.3.2a.x64 engine instead?
cooldalek

Re: Slow Rybka.

Post by cooldalek »

no - rybka counts nodes differently to other engines. Your engine is working fine
Vempele

Re: Slow Rybka.

Post by Vempele »

Should I have installed the Rybka v2.3.2a.x64 engine instead?
Yes. If it doesn't work, you'll know your Windows is 32-bit. If it does, it'll be 60% faster than the 32-bit version.
User avatar
Mike S.
Posts: 1480
Joined: Thu Mar 09, 2006 5:33 am

Re: Slow Rybka.

Post by Mike S. »

Even if engines count the nodes identically (or almost), the ratio between node rates can be up to ~10:1 for engines of similar strength. Some years ago I have compared Fritz and Junior to Hiarcs and King, and the typical node rade ratio between the "high node rate" and "low node rate" engines was 9:1, IIRC. But they were close in strength.

(At least there never was any rumour that these node counts were done much differently.)

I have a chess computer which calculates 2 nodes/sec. :mrgreen: Not 2 kN/s (2000), only TWO nodes per second! Still, it can play chess at ~1500 Elo level.
Regards, Mike
User avatar
Rubinus
Posts: 1212
Joined: Thu Jan 18, 2007 4:05 pm
Location: Prague
Full name: Pavel Háse

Re: Slow Rybka.

Post by Rubinus »

Depth, not nodes is feature. Not slowly than most opponents.
Depth+1 - Rybka is fastest. But somewhere will not change halfvariant.
User avatar
tjfroh
Posts: 8087
Joined: Thu Mar 09, 2006 9:51 pm
Location: Near the Intel Plant in the Land of Manana
Full name: Timothy Frohlick

Re: Slow Rybka.

Post by tjfroh »

Mr. Fagliano,

I have an Intel E-6850 that runs Rybka 2.3.2a LK 64bit at about 220 KNPS on a Cyberpower machine with 2 GBs of RAM under
Windows Vista Home Premium 64 bit. The 32 bit version runs about the same nodes per second.

Vas Rajlich knows that his node counter counts about 1/10th as many nodes as Junior or Shredder. He also knows that Rybka kills
those programs.

When you upgrade your computer in one or two years, Rybka 3.x will run at about 800 to 1000 KNPS because of the processor speedup.

Enjoy,

TJ
A New Way Comes Upon Earth.
God is an infinitely variable Constant.
Man marks his ground with ideologies.
Galaxies are the dreidels of God.
War is a punishment for implacability.
Peace flows from forgiveness of sins.
Louis F

Re: Slow Rybka.

Post by Louis F »

Vas Rajlich knows that his node counter counts about 1/10th as many nodes as Junior or Shredder. He also knows that Rybka kills
those programs.
I'm glad to know that my Rybka engine is working fine. But I thought a node was a node. How can one node counter count a node differently that another?
User avatar
M ANSARI
Posts: 3733
Joined: Thu Mar 16, 2006 7:10 pm

Re: Slow Rybka.

Post by M ANSARI »

If you want to improve your Knps to around 500 Knps on Rybka ... you can do it for around $500 if you already have a running system. Win XP64bit is the OS of choice for Rybka MP.

http://64.68.157.89/forum/viewtopic.php?t=15997
Terry McCracken

Re: Slow Rybka.

Post by Terry McCracken »

M ANSARI wrote:If you want to improve your Knps to around 500 Knps on Rybka ... you can do it for around $500 if you already have a running system. Win XP64bit is the OS of choice for Rybka MP.

http://64.68.157.89/forum/viewtopic.php?t=15997
How much for a whole new system? Without making serious compromises but getting the best bang for your buck?

I'm afraid I'm a bit behind the times :oops:

Terry
Gandalf

Re: Slow Rybka.

Post by Gandalf »

I'm glad to know that my Rybka engine is working fine. But I thought a node was a node. How can one node counter count a node differently that another?
This has been discussed a few times on the Rybka forum, you might want to read this topic and/or run a search for the others.