OpenChess

I got some ample time last couple of hours so I decided to implement EGBBs in the world's stongest engines. I have implemented them quickly and they seem to work fine for up to 6 men. Source code is here https://github.com/dshawul/Stockfish/commits/master . Once everything is finalized I will start making binaries. This is just an announcement of the project.

OVERFLOW, stack OVERFLOW

I would be very interested to see the results running against this list of 6man test positions:

• k7/7p/8/2nKb3/8/8/8/8 w - -
  k2N2K1/8/8/8/5R2/3n4/3p4/8 w - -
  8/pKR5/8/p7/8/8/2p5/2k5 w - -
  8/k7/8/2R5/8/4q3/8/4B2K w - -
  8/8/8/Pk5P/3p1n2/8/3K4/8 w - -
  8/8/8/5bk1/8/8/4B1PP/5K2 b - -
  8/8/8/4p3/5k2/8/4K3/8 w - -
  8/7p/8/3K3P/8/8/8/5k2 b - -
  8/7P/8/5Bk1/P7/8/1b4K1/8 b - -
  8/5p2/7k/6p1/5P1P/8/8/7K w - -
  8/1p6/4p3/1K2k3/8/P7/8/8 w - -
  8/1p4P1/5K2/2k5/8/R7/6r1/8 b - -
  7K/k7/8/P7/6pP/8/8/8 w - -
  5k2/8/8/3R4/6K1/8/3b2pP/8 w - -
  3k4/3q4/8/4Q3/5P2/6K1/8/8 b - -
  1B1b4/7K/1p6/1k6/8/8/8/8 w - -

Since the positions are with in 6 men it returns the correct result immediately. Search depth goes to maximum = 100, but it will not give you PV.
Most of them are draws so it doesn't think further. But for wins it tries to choose among the winning options by searching first half of the total depth in case of 6-men
and the whole depth in case of 5-men. The former maybe slow so limiting it further maybe a good idea.
The results are all correct. I checked against the online shredder database http://www.shredderchess.com/online-che ... abase.html

Since the positions are with in 6 men it returns the correct result immediately. Search depth goes to maximum = 100, but it will not give you PV.

You can get a PV by not probing the EGTB in PV nodes. This is what I do in GNU Chess.