TalkChess.com

Try it with time control ten times slower

There was an analysis in the programming forum a while back that showed software advance in strength is exponential, just like the hardware advance.

https://www.chessprogramming.org/Retrograde_Analysis http://home.hccnet.nl/h.g.muller/EGT7/7-men.html A word on this: It is fun and interesting to examine and code an algorithm for EGTB, but there are a lot of implementations available already. If you want to do it for your own enjoyment, there is n...

Doesn't this indicate that the reductions are flawed? Otherwise, where does the Elo increase come from? If a fatter tree is better then maybe the reductions are incorrect or too extreme. On the other hand, it seems possible that the small extra randomness introduced by the thread local storage might...

Elo is Correct. ELO and elo are ignorance.
That having been said, I've probably made both mistakes.
It's like degrees Celsius. Would you write CELCIUS or celcius?

Why not a hybrid?

Half of the threads are using Lazy SMP to update the hash table.
The other half of the threads are using DTS/YBW/ABDADA to update the hash table.
Or some other mix as a function of thread count, hash table size, and time.

Where do you imagine that DTS sits within this grid?
On top? In the middle? On the bottom?

I guess that the main problem here is that ZCT is the only open source DTS program I can think of.

Something is missing from the minimization of these: 1. Number of CPU instructions executed 2. Number of memory load/stores executed (in L1, L2, L3 cache, and DDR4 RAM) 3. The number of MESI messages passed between CPU cores. Consider a program that does eval() on the root node, and then waits until...

Am I the only one who feels repugnance for "throw it at the wall and see if it sticks" searches? Now, Lucas has pointed out, and I think quite eloquently, that Lazy SMP is beautiful in its simplicity. I think we were all shocked at Dan Homan's published results. And therein lies the rub. Lazy SMP wo...