TalkChess.com

The author fixed the memory bug. If only he can fix the 64 Mb hash maximum as well.

terminator wrote:

Code: Select all

RELEASE NOTES

Version 5.3B, 2008/06/25.
# Fixed a memory bug caused by the MP implementation.
# Fixed an evaluation bug in the endgame.

Version 5.3, 2008/06/24.
# Many evaluation improvements.
# Option for dual core CPUs, 1.5X speedup.
# Fixed a bug in UCI pondering.
# Restored "perft" command. 

The author fixed the memory bug. If only he can fix the 64 Mb hash maximum as well.

Thanks....
Use it as a UCI engine to get more Mb for the hash table....

How can I allocate more than 64 MB? The Arena GUI changes it back to 64 if I try to set it to something larger. If I set it to, say, 256 MB in Fritz, it only allocates 64 MB.

tmokonen wrote:How can I allocate more than 64 MB? The Arena GUI changes it back to 64 if I try to set it to something larger. If I set it to, say, 256 MB in Fritz, it only allocates 64 MB.

So it's limited to 64 Mb even if you use it as a UCI engine....

I believe that Delfi is hard-code limited to 64 MB of hash as I remember the author saying that increasing hash did nothing.

That's a while ago though, and since it now is MP it could benefit from more hash.

May be you all can use 45 Gigs of hash and turn Delfi in to another Rybka. For any kind of practical use 64Mb of hash should be enough.

Tony Thomas wrote:May be you all can use 45 Gigs of hash and turn Delfi in to another Rybka. For any kind of practical use 64Mb of hash should be enough.

Correct and it's not hard for the author to increase the size of the hash table if he thinks that the engine will benifit from that....

Dr.Wael Deeb wrote:

Tony Thomas wrote:May be you all can use 45 Gigs of hash and turn Delfi in to another Rybka. For any kind of practical use 64Mb of hash should be enough.

Correct and it's not hard for the author to increase the size of the hash table if he thinks that the engine will benifit from that....

Mike Sherwin says that he doesnt see any advantage from using large hashtables. I believe it was Christophe Theron who did an experiment with doubling hash tables, and he found that you gain 5 points for each doubling, or something small. I wonder if Terminator is always just a gloomy/grumpy gus.

Tony Thomas wrote:

Dr.Wael Deeb wrote:

Tony Thomas wrote:May be you all can use 45 Gigs of hash and turn Delfi in to another Rybka. For any kind of practical use 64Mb of hash should be enough.

Correct and it's not hard for the author to increase the size of the hash table if he thinks that the engine will benifit from that....

Mike Sherwin says that he doesnt see any advantage from using large hashtables. I believe it was Christophe Theron who did an experiment with doubling hash tables, and he found that you gain 5 points for each doubling, or something small. I wonder if Terminator is always just a gloomy/grumpy gus.

Aren't you afraid to be terminated Tony

Tony Thomas wrote:Mike Sherwin says that he doesnt see any advantage from using large hashtables. I believe it was Christophe Theron who did an experiment with doubling hash tables, and he found that you gain 5 points for each doubling, or something small. I wonder if Terminator is always just a gloomy/grumpy gus.

There is no need to hardcode 64 Mb unless Delfi is intended for blitz chess. I haven't tested Delfi in the playchess engineroom but it would be whacked by progs like Glaurung which are more customisable. I'll see whether I can find the author's contacts to ask him to join this debate.